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.

L’objectif de la thèse est de développer un système de type jonction tunnel tout oxyde à base de La2/3Sr1/3MnO3 (LSMO) où il serait possible de contrôler l’aimantation des électrodes magnétiques par des processus à faible consommation d’énergie. Des jonctions tunnel épitaxiées de LSMO/SrTiO3/LSMO ont été obtenues montrant un double renversement de l’aimantation à température ambiante et un taux de magnétorésistance de 71 % à 10 K. En exerçant une contrainte sur le LSMO par le substrat il a été possible de moduler l’anisotropie des couches magnétiques. Des anisotropies perpendiculaire ou dans le plan ont pu être obtenues. Afin de contrôler le renversement des moments magnétiques dans une des électrodes ferromagnétiques trois options ont été envisagées : l’utilisation de l’injection de spin à partir d’un métal à fort couplage spin-orbite (Pt, Ir) ou d’un oxyde contenant de tels ions (ici Ru dans SrRuO3), et l’utilisation du Bi2FeCrO6, un oxyde multiféroïque pouvant présenter un couplage magnétoélectrique. Malgré des résultats prometteurs, aucune solution n’a permis des tests sur des jonctions afin d’estimer leur efficacité. L’objectif final n’est pas encore atteint mais des avancées intéressantes ont été faites afin d’envisager des dispositifs permettant le stockage et la manipulation de l’information. / The objective of this work is to develop La2/3Sr1/3MnO3 (LSMO) based all-oxide magnetic tunnel junction systems where it would be possible to control the magnetization of magnetic electrodes by low energy consumption processes. Epitaxial tunnel junctions of LSMO/SrTiO3/LSMO were obtained showing a double magnetization switching at room temperature and a magnetoresistance ratio of 71 % at 10 K. Using strain engineering, it was possible to modulate the anisotropy of the LSMO magnetic layers. Perpendicular or in plane anisotropies could be thus obtained. In order to control the reversal of the magnetic moments in one of the ferromagnetic electrodes three options were considered : the use of spin injection from a metal with a strong spin-orbit coupling (Pt, Ir) or an oxide containing this type of ions (here Ru in SrRuO3), and the use of Bi2FeCrO6 multiferroic oxide that may exhibit a magnetoelectric coupling. Despite promising results, no solution has allowed tests on junctions to be carried out to estimate their effectiveness. Although the final objective is not yet achieved, interesting progress has been made on the way to information storage and manipulation devices.

京都大学 / 新制・課程博士 / 博士(工学) / 甲第23512号 / 工博第4924号 / 新制||工||1769(附属図書館) / 京都大学大学院工学研究科電子工学専攻 / (主査)教授 白石 誠司, 教授 山田 啓文, 教授 引原 隆士 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DGAM

Although ferromagnets are found in all kinds of technological applications, their natural occurrence is rather unusual because only few substances are known to be intrinsically ferromagnetic at room temperature. In the past twenty years, a plethora of new artificial ferromagnetic materials has been found by introducing defects into non-magnetic host materials. In contrast to the intrinsic ferromagnetic materials, they offer an outstanding degree of material engineering freedom, provided one finds a type of defect to functionalize every possible host material to add magnetism to its intrinsic properties. Still, some controversial questions remain: What are the mechanisms behind these ferromagnetic materials? Why are their magnetization values reported in the literature so low? Are these materials really technologically relevant ferromagnets? In this work, we aim to provide a systematic investigation of the phenomenon. We propose a universal scheme for the computational discovery of new artificial functional magnetic materials, which is guided by experimental constraints and based on first principles. The obtained predictions explain very well the experimental data found in the literature. The potential of the method is further demonstrated by the experimental realization of a truly 2D ferromagnetic phase at room temperature, created in nominally non-magnetic TiO$_2$ films by ion irradiation, which follows a characteristic 2D magnetic percolation transition and exhibits a tunable magnetic anisotropy. Furthermore, the technological relevance of these artificial ferromagnetic materials, which comes to shine when one combines the engineered magnetic with some of the intrinsic properties of the host material, is demonstrated by creating a spin filter device in a ZnO host that generates highly spin-polarized currents even at room temperature.:1 Introduction 2 Computational discovery of artificial ferromagnets 2.1 Ferromagnetism in solids 2.1.1 Exchange interaction and magnetic order 2.1.2 Artificial magnetism due to defects 2.2 Predicting defect structures from collision cascades 2.3 Finding magnetic defect candidates 2.4 Magnetic percolation 2.5 Magnetic phase diagram of anatase TiO 2 artificial ferromagnet 2.5.1 Defect creation in anatase TiO 2 2.5.2 Magnetic properties of dFP defects in anatase TiO 2 2.5.3 Constructing a magnetic phase diagram 2.6 Revisiting prior experimental results 3 Artificial ferromagnetism in TiO 2 hosts 3.1 Low energy ion irradiation 3.2 SQUID magnetometry 3.3 Experimental realization of an artificial ferromagnet in TiO2 4 Artificial magnetic monolayers and surface effects 4.1 Critical behavior and 2D magnetism 4.2 Magnetic anisotropy 4.2.1 Demagnetizing field and magnetic shape anisotropy 4.2.2 Magnetocrystalline anisotropy 4.3 Artificial ferromagnetic monolayer at TiO 2 surface with perpendicular magnetic anisotropy 4.4 DFT calculations of the defective anatase TiO 2 [001] surface 5 Spin transport through artificial ferromagnet interfaces 5.1 Artificial ferromagnetism in ZnO hosts 5.2 Spin filter effect at magnetic/non-magnetic interfaces in ZnO 5.2.1 The spin filter effect 5.2.2 Lithium and hydrogen doping in ZnO 5.2.3 Magneto-transport in artificial ferromagnetic Li:ZnO microwires 5.2.4 Spin transport through magnetic/non-magnetic interfaces 5.2.5 Minority spin filter effect 6 Conclusions and Outlook Bibliography Appendix: A List of publications B Computation inputs and codes B.1 DFT electronic structure calculations - Fleur input files B.2 Magnetic Percolation simulations B.3 SQUID raw data analysis code B.4 SRIM Monte Carlo binary collision code automation / Obwohl Ferromagnete in allen möglichen technischen Anwendungen zu finden sind, ist ihr natürliches Vorkommen eher ungewöhnlich, da nur wenige Stoffe bekannt sind, die bei Raumtemperatur intrinsisch ferromagnetisch sind. In den letzten zwanzig Jahren wurde eine Fülle neuer künstlicher ferromagnetischer Materialien durch das Einbringen von Defekten in nichtmagnetische Wirtsmaterialien entdeckt. Im Gegensatz zu den intrinsischen ferromagnetischen Materialien bieten sie einen herausragenden Grad an materialtechnischer Freiheit, vorausgesetzt man findet zu jedem möglichen Wirtsmaterial einen passenden Typus von Defekten, um dessen intrinsische Eigenschaften um Magnetismus zu ergänzen. Dennoch bleiben einige kontroverse Fragen bislang unbeantwortet: Welche Mechanismen stehen hinter diesen ferromagnetischen Materialien? Warum werden ihre Magnetisierungswerte in der Literatur meist so niedrig angegeben? Sind diese Materialien wirklich technologisch relevante Ferromagneten? In dieser Arbeit wollen wir eine systematische Untersuchung des Phänomens durchführen. Wir schlagen ein universelles ab-initio Protokoll für die computergestützte Entdeckung von neuen künstlichen funktionalen magnetischen Materialien vor, das sich an experimentellen Bedingungen orientiert. Die erhaltenen Vorhersagen erklären die in der Literatur gefundenen experimentellen Daten sehr gut. Wir demonstrieren die Wirksamkeit der Methode durch die experimentelle Realisierung einer echten 2D-ferromagnetischen Phase bei Raumtemperatur, die in nominell nicht-ma'-gne'-tischen TiO$_2$-Filmen durch Ionenbestrahlung erzeugt wird. Die so entstehende ferromagnetische Phase folgt einem charakteristischen zweidimensionalen magnetischen Perkolationsprozess und weist eine steuerbare magnetische Anisotropie auf. Weiterhin wird die technologische Relevanz dieser künstlichen ferromagnetischen Materialien gezeigt, welche besonders zum Vorschein kommt, wenn man die künstlichen magnetischen mit einigen der intrinsischen Eigenschaften des Wirtsmaterials kombiniert, und zwar indem ein Spin-Filter Element auf Basis eines ZnO-Wirts gebaut wird, das selbst bei Raumtemperatur hoch spin-polarisierte Ströme erzeugt.:1 Introduction 2 Computational discovery of artificial ferromagnets 2.1 Ferromagnetism in solids 2.1.1 Exchange interaction and magnetic order 2.1.2 Artificial magnetism due to defects 2.2 Predicting defect structures from collision cascades 2.3 Finding magnetic defect candidates 2.4 Magnetic percolation 2.5 Magnetic phase diagram of anatase TiO 2 artificial ferromagnet 2.5.1 Defect creation in anatase TiO 2 2.5.2 Magnetic properties of dFP defects in anatase TiO 2 2.5.3 Constructing a magnetic phase diagram 2.6 Revisiting prior experimental results 3 Artificial ferromagnetism in TiO 2 hosts 3.1 Low energy ion irradiation 3.2 SQUID magnetometry 3.3 Experimental realization of an artificial ferromagnet in TiO2 4 Artificial magnetic monolayers and surface effects 4.1 Critical behavior and 2D magnetism 4.2 Magnetic anisotropy 4.2.1 Demagnetizing field and magnetic shape anisotropy 4.2.2 Magnetocrystalline anisotropy 4.3 Artificial ferromagnetic monolayer at TiO 2 surface with perpendicular magnetic anisotropy 4.4 DFT calculations of the defective anatase TiO 2 [001] surface 5 Spin transport through artificial ferromagnet interfaces 5.1 Artificial ferromagnetism in ZnO hosts 5.2 Spin filter effect at magnetic/non-magnetic interfaces in ZnO 5.2.1 The spin filter effect 5.2.2 Lithium and hydrogen doping in ZnO 5.2.3 Magneto-transport in artificial ferromagnetic Li:ZnO microwires 5.2.4 Spin transport through magnetic/non-magnetic interfaces 5.2.5 Minority spin filter effect 6 Conclusions and Outlook Bibliography Appendix: A List of publications B Computation inputs and codes B.1 DFT electronic structure calculations - Fleur input files B.2 Magnetic Percolation simulations B.3 SQUID raw data analysis code B.4 SRIM Monte Carlo binary collision code automation

The focus of the work is on the epitaxial growth of Mn5Ge3 layers on Ge (001) via ultra-fast solid-state reaction between Mn and Ge using millisecond range FLA at the ambient pressure in continuous N2 flow. Epitaxial Mn5Ge3 layers were obtained both on Ge (001) and Ge (111) substrates by optimizing the fabrication parameters, Mn thickness (30 nm), FLA energy density (100-110 Jcm-2) and FLA duration time. The epitaxial relationship between the alloy film and substrate is the (100) plane of Mn5Ge3 along [001] direction parallel with the [100] direction of Ge (001) plane. It is notable that the hexagonal c axis of Mn5Ge3 on Ge (001) is parallel to the film surface plane, while the reported Mn5Ge3’s c axis on Ge (111) tends to be perpendicular to the film plane. In fact, using ultrafast-SPE the c-axis of Mn5Ge3 is always parallel to the sample surface. Mn5Ge3 films exhibit ferromagnetism which is demonstrated by the anomalous Hall effect up to the TC = 283±5 K. The films exhibit their in-plane magnetic easy axis along the hexagonal c-axis independent of the Mn5Ge3 film thickness. This provides a new avenue for the fabrication of Ge-based spin-injectors fully compatible with industrial CMOS technology. The deeper understanding of the magnetic, structural and electrical properties of (100) epitaxial Mn5Ge3 grown on Ge (001) are presented by utilizing DFT calculation (by our collaborator M. Birowska) and various experimental methods. The Mn atoms in Mn5Ge3 occupy two distinct Wyckoff positions with fourfold (Mn1) and sixfold (Mn2) multiplicity. During cooling down to 100 K the Mn5Ge3 unit-cell shows remarkable structural deformation. The nearest distance d3 between Mn2-Mn2 atoms in the hexagonal a-b plane is shortened much faster than the nearest distance d1 between Mn1-Mn1 atoms along hexagonal c axis. The DFT calculations show that below critical distance d3 < 3.002 Å, the Mn2 atoms are AFM coupled while for d3 > 3.002 Å the coupling is FM. The FM coupling between Mn1 atoms weakly depends on the atomic distance d1. Moreover, there is a transition from collinear to noncollinear spin configuration at about 70±5 K. Simultaneously, at low temperature, the angular dependent magnetoresistance shows a switching from multi-fold component to twofold symmetry. The combination of different experimental techniques with theoretical calculations enabled us to conclude that the switching between non-collinear and collinear spin configurations and the variation of anisotropic magnetoresistance in Mn5Ge3 is due to the strain induced change of the magnetic coupling between Mn2-Mn2 atoms. Finally, the effects of strain on the structural and magnetic properties of epitaxial Mn5Ge3 on Ge (111) substrate by applying ms-range FLA are investigated. The X-ray diffraction results demonstrate that during FLA process the formation of nonmagnetic secondary phases of MnxGey is fully suppressed and the in-plane tensile strain is enhanced. The temperature dependent magnetization indicates that after FLA the Curie temperature of Mn5Ge3 increases from 283±5 K to above 400 K. Further Monte Carlo simulations manifest that the change of the strain in Mn5Ge3 during ms-range FLA modifies the distance between adjacent Mn atoms in the hexagonal basal plane, which provokes the different ferromagnetic interaction between them. Consequently, the significant increase of Curie temperature is observed. This provides a good way to improve the Curie temperature of Mn5Ge3 which is promising to realize room-temperature operated Ge based spin-injectors.

Aujourd'hui se pose une question fondamentale sur le futur de l'électronique actuelle. De plus en plus, des circuits hybrides intégrant de nouvelles fonctionnalités sont fabriqués. On envisage même, à plus long terme, des circuits basés sur une technologie différente de l'approche CMOS utilisée actuellement. Une de ces technologies est la spintronique qui tire profit du spin, degré de liberté supplémentaire de l'électron. Elle a rapidement fait ses preuves par le passé dans le stockage non volatile binaire (disques durs) et s'oriente aujourd'hui vers de nouvelles mémoires magnétiques ultra-performantes et basse consommation les MRAMs (Magnetic Random Access Memories). En parallèle, une nouvelle catégorie de matériaux à fort potentiel a émergé : les matériaux bidimensionnels (2D). Ces matériaux, dont le fer de lance est le graphène (une couche d'un atome d'épaisseur de graphite), offrent de nouvelles propriétés inégalées. Leur combinaison via la fabrication d'hétérostructures et la capacité d'avoir un contrôle de leur épaisseur à l'échelle atomique pourrait devenir un atout majeur en électronique et plus particulièrement en spintronique. L'objectif de cette thèse a été l'étude de l'intégration et la démonstration du potentiel en termes de fonctionnalités et de performances de ces nouveaux matériaux 2D au sein de jonctions tunnel magnétiques (MTJs), le dispositif prototype de la spintronique. Au cours de cette thèse, nous avons poursuivi les travaux initiés au laboratoire sur l'intégration dans des MTJs du graphène obtenu via une méthode de dépôt CVD (dépôt chimique en phase vapeur) directe sur l’électrode ferromagnétique inférieure. Nous avons démontré que les propriétés de filtrage en spin et de membrane protectrice contre l'oxydation de l'électrode ferromagnétique (FM) sous-jacente s'étendaient à une unique couche de graphène. Par ailleurs, nous avons aussi pu étudier et améliorer significativement l'amplitude du filtrage en spin et du signal de magnétorésistance observé via l'optimisation des procédés de croissance et d'intégration et le choix de différentes configurations de matériaux ferromagnétiques (Ni(111), Co...). De forts effets de filtrage de spin ont ainsi pu être observés avec des magnétorésistances allant de -15% à plus de +80%, soit presque trois fois l'état de l'art. En parallèle, nous nous sommes aussi intéressés à un autre matériau 2D, le nitrure de bore hexagonal (h-BN), isolant isomorphe du graphène qui s'apparenterait à une barrière tunnel d'un seul atome d'épaisseur. Afin d’étudier le h-BN dans une MTJ, nous avons décidé d’exploiter à nouveau le principe d’une croissance directe par CVD du matériau 2D sur le matériau FM. Des mesures CT-AFM (Conductive Tip Atomic Force Microscopy) nous ont permis de démontrer les propriétés de barrière tunnel homogène du h-BN ainsi que le contrôle possible de la hauteur de barrière avec le nombre de couches de h-BN. De plus, des mesures électriques et de magnétotransport nous ont permis de confirmer l’intégration réussie de la barrière tunnel h-BN dans notre MTJ. Nous avons pu obtenir les premiers résultats de forte magnétorésistance pour du h-BN avec une amplitude de la magnétorésistance de +50%, plus d'un ordre de grandeur au-dessus de l'état de l'art, révélant le potentiel du h-BN. Nous avons enfin aussi pu démontrer l'importance du couplage entre le h-BN et l'électrode FM offrant un potentiel de contrôle inédit sur les effets de filtrage en spin et allant jusqu'à rendre le h-BN métallique. Lors de cette thèse, nous avons pu montrer que l’intégration du graphène et du h-BN dans des MTJs via la croissance directe par CVD est un procédé privilégié pour tirer pleinement profit de leurs propriétés. Les résultats obtenus de forte magnétorésistance et de filtrage en spin laissent entrevoir le fort potentiel du graphène, du h-BN mais aussi des autres nouveaux matériaux 2D à venir pour les MTJs. Ces études ouvrent une nouvelle voie d’exploration pour les MTJs : les 2D-MTJs. / Nowadays a critical issue is raised concerning the future of current electronics. Increasingly, hybrid circuits with new functionalities are manufactured. A longer term approach is even contemplated with circuits based on a technology different from the one currently used (CMOS technology). One of these envisioned technologies is spintronics, which benefits from the spin properties, the electron additional degree of freedom. Spintronics has quickly proven its worth in the past in the field of non volatile data storing (hard drives) and is today moving towards new fast and ultra-low-power magnetic random access memories the MRAMs. Meanwhile, these last few years, a new category of materials with high potential has emerged : the bidimensional materials (2D). These materials, with graphene (one atomically thick layer of graphite) as the forerunner, provide new unrivaled properties. Their combination in the form of heterostructures and the ability to obtain a control of their thickness at the atomic scale could be a major asset for electronics and more specifically spintronics. The purpose of this thesis has been the study of the integration and the demonstration of the potential in terms of functionalities and performances of these new 2D materials inside the prototypical spintronic device: the magnetic tunnel junction (MTJ). During this thesis, we have pursued the work initiated by the laboratory on the integration of graphene in MTJs with direct CVD deposition method (chemical vapor deposition) on the underlying ferromagnetic electrode. We demonstrated that the spin filtering and protective membrane properties (preventing the oxidation of the underlying ferromagnetic electrode (FM)) observed earlier expand to a graphene monolayer. Furthermore, we have also studied and improved significantly the amplitude of the spin filtering and the magnetoresistance signal observed. This was done thanks to the optimization of the growth process, integration, and choice of the different configurations of ferromagnetic materials in our structures (Ni(111), Co...). High spin filtering effects have been observed as a function of the configurations with magnetoristances ranging from -15% to beyond +80%, which is almost three times the state of the art. Meanwhile, we looked at another 2D material, the hexagonal boron nitride (h-BN), an insulating isomorph of graphene which could be considered as an atomically thin tunnel barrier. In order to study h-BN into a MTJ, we took again advantage of direct CVD growth of the 2D material on a ferromagnet. CT-AFM (Conductive Tip Atomic Force Microscopy) measurements allowed us to demonstrate the homogeneous tunnel barrier properties of h-BN and the possible control of the barrier height with the number of h-BN layers. Simultaneously, electrical and magnetotransport measurement in the complete junction allowed us to confirm the achieved integration of the h-BN tunnel barrier into our MTJ. We have been able to obtain the first results of high magnetoresistance for h-BN with values one order of magnitude beyond the state of the art. A magnetoresistance of +50% has been reached, thanks to the optimization of the growth process revealing the potential of h-BN. We have also been able to show the important role of the coupling between h-BN and the FM electrode offering an unprecedented potential of control on the spin filtering effects, ranging up to making the h-BN metallic. During this thesis, we have been able to demonstrate that the integration of graphene and h-BN in MTJs through direct CVD growth is a promising process in order to fully exploit their properties. The results obtained of high magnetoresistance and spin filtering point to the high potential for MTJs of graphene and h-BN but also to all the new 2D materials to come. These studies pave the way for exploring a new path for MTJs : the 2D-MTJs.

<div> <div> <div> <p>Process variations and increasing leakage current are major challenges toward memory realization in deeply-scaled CMOS devices. Spintronic devices recently emerged as one of the leading candidates for future information storage due to its potential for non-volatility, high speed, low power and good endurance. In this thesis, we start with the basic concepts and applications of three spintronic devices, namely spin or- bit torque (SOT) based spin-valves, SOT-based magnetic tunnel junctions and the magnetic skyrmion (MS) for both logic and machine learning hardware. </p> <p>We propose a new Spin-Orbit Torque based Domino-style Spin Logic (SOT-DSL) that operates in a sequence of Preset and Evaluation modes of operations. During the preset mode, the output magnet is clocked to its hard-axis using spin Hall effect. In the evaluation mode, the clocked output magnet is switched by a spin current from the preceding stage. The nano-magnets in SOT-DSL are always driven by orthogonal spins rather than collinear spins, which in turn eliminates the incubation delay and allows fast magnetization switching. Based on our simulation results, SOT-DSL shows up to 50% improvement in energy consumption compared to All-Spin Logic. Moreover, SOT-DSL relaxes the requirement for buffer insertion between long spin channels, and significantly lowers the design complexity. This dissertation also covers two applications using MS as information carriers. MS has been shown to possess several advantages in terms of unprecedented stability, ultra-low depinning current density, and compact size. </p><p><br></p><p>We propose a multi-bit MS cell with appropriate peripheral circuits. A systematic device-circuit-architecture co-design is performed to evaluate the feasibility of using MS-based memory as last-level caches for general purpose processors. To further establish the viability of skyrmions for other applications, a deep spiking neural network (SNN) architecture where computation units are realized by MS-based devices is also proposed. We develop device architectures and models suitable for neurons and synapses, provide device-to-system level analysis for the design of an All-Spin Spiking Neural Network based on skyrmionic devices, and demonstrate its efficiency over a corresponding CMOS implementation.</p> <div> <div> <div> <p><br></p><p>Apart from the aforementioned applications such as memory storage elements or logic operation, this research also focuses on the implementation of spin-based device to solve combinatorial optimization problems. Finding an efficient computing method to solve these problems has been researched extensively. The computational cost for such optimization problems exponentially increases with the number of variables using traditional von-Neumann architecture. Ising model, on the other hand, has been proposed as a more suitable computation paradigm for its simple architecture and inherent ability to efficiently solve combinatorial optimization problems. In this work, SHE-MTJs are used as a stochastic switching bit to solve these problems based on the Ising model. We also design an unique approach to map bi-prime factorization problem to our proposed device-circuit configuration. By solving coupled Landau- Lifshitz-Gilbert equations, we demonstrate that our coupling network can factorize up to 16-bit binary numbers. </p> </div> </div> </div> </div> </div> </div>

<p>Novel computational paradigms based on non-von Neumann architectures are being extensively explored for modern data-intensive applications and big-data problems. One direction in this context is to harness the intrinsic physics of spintronics devices for the implementation of nanoscale and low-power building blocks of such emerging computational systems. For example, a Probabilistic Spin Logic (PSL) that consists of networks of p-bits has been proposed for neuromorphic computing, Bayesian networks, and for solving optimization problems. In my work, I will discuss two types of device-components required for PSL: (i) p-bits mimicking binary stochastic neurons (BSN) and (ii) compound synapses for implementing weighted interconnects between p-bits. Furthermore, I will also show how the integration of recently discovered van der Waals ferromagnets in spintronics devices can reduce the current densities required by orders of magnitude, paving the way for future low-power spintronics devices.</p> <p>First, a spin-device with input-output isolation and stable magnets capable of generating tunable random numbers, similar to a BSN, was demonstrated. In this device, spin-orbit torque pulses are used to initialize a nano-magnet with perpendicular magnetic anisotropy (PMA) along its hard axis. After removal of each pulse, the nano-magnet can relax back to either of its two stable states, generating a stream of binary random numbers. By applying a small Oersted field using the input terminal of the device, the probability of obtaining 0 or 1 in binary random numbers (P) can be tuned electrically. Furthermore, our work shows that in the case when two stochastic devices are connected in series, “P” of the second device is a function of “P” of the first p-bit and the weight of the interconnection between them. Such control over correlated probabilities of stochastic devices using interconnecting weights is the working principle of PSL.</p> <p>Next my work focused on compact and energy efficient implementations of p-bits and interconnecting weights using modified spin-devices. It was shown that unstable in-plane magnetic tunneling junctions (MTJs), i.e. MTJs with a low energy barrier, naturally fluctuate between two states (parallel and anti-parallel) without any external excitation, in this way generating binary random numbers. Furthermore, spin-orbit torque of tantalum is used to control the time spent by the in-plane MTJ in either of its two states i.e. “P” of the device. In this device, the READ and WRITE paths are separated since the MTJ state is read by passing a current through the MTJ (READ path) while “P” is controlled by passing a current through the tantalum bar (WRITE path). Hence, a BSN/p-bit is implemented without energy-consuming hard axis initialization of the magnet and Oersted fields. Next, probabilistic switching of stable magnets was utilized to implement a novel compound synapse, which can be used for weighted interconnects between p-bits. In this experiment, an ensemble of nano-magnets was subjected to spin-orbit torque pulses such that each nano-magnet has a finite probability of switching. Hence, when a series of pulses are applied, the total magnetization of the ensemble gradually increases with the number of pulses</p> <p>applied similar to the potentiation and depression curves of synapses. Furthermore, it was shown that a modified pulse scheme can improve the linearity of the synaptic behavior, which is desired for neuromorphic computing. By implementing both neuronal and synaptic devices using simple nano-magnets, we have shown that PSL can be realized using a modified Magnetic Random Access Memory (MRAM) technology. Note that MRAM technology exists in many current foundries.</p> <p>To further reduce the current densities required for spin-torque devices, we have fabricated heterostructures consisting of a 2-dimensional semiconducting ferromagnet (Cr₂Ge₂Te₆) and a metal with spin-orbit coupling metal (tantalum). Because of properties such as clean interfaces, perfect crystalline nanomagnet structure and sustained magnetic moments down to the mono-layer limit and low current shunting, 2D ferromagnets require orders of magnitude lower current densities for spin-orbit torque switching than conventional metallic ferromagnets such as CoFeB.</p>

Spintronics is a spin-electronic field where the electron spinangular momentum, in conjunction with charge, is used to read andwrite information in magnetic sensors and logic circuits, e.g. hard disk drive (HDD), magnetic random access memory (MRAM) and broadband TeraHertz (THz) emitters. To realize the THz operations of the spin logic circuits THz manipulation of the magnetic state is pivotal. This THz manipulation of the magnetic state in anti-ferromagnetic magnetic materials can be realized by coupling the materials with THz antennas. On the other hand, these antennas enhance the THz amplitude of spin-electronic THz emitters when coupled with its output. Therefore, these THz antennas can not only be coupled with the input of magnetic logics to improve the efficiency of magnetic sate manipulation in logic devices but also with the output of the spintronic THz emitters to enhance the generated THz signal amplitude. In this project, we have examined four types of antennas: h-dipole, spiral, bow-tie, and a sub-THz antenna. All the antennas are placed on top of a MgO substrate material for simplicity. However, a bow-tie antenna is also fabricated on an antiferromagnetic substrate of TmFeO3 to check this antenna’s reliability to manipulate its magnetic state. We have studied the impact of antenna geometries on the generated electric field amplitude. We have optimized each antenna for maximum electric field norm profile, with an increase of 30% for the h-dipole and spiral antennas, and an increase of 100% for the bow-tie antenna. However, in this project we were not able to find any general conclusions about what geometrical parameters can further amplify the generated electric field. None of the antennas generated a large enough peak-to-peak electric field amplitude to manipulate the magnetic state of anti-ferromagnetic materials. However, they did successfully amplify the spintronic THz emitter output and could certainly be useful in that regard.

Résumé : Au cours de cette thèse nous nous sommes intéressés à différentes propriétés de transport électrique, thermique et thermoélectrique. En particulier, nous avons mis en évidence les effets Righi-Leduc anomal et planaire qui sont les équivalents thermiques des effets Hall anomal et planaire. Ces effets doivent impérativement être considérés dans l’interprétation des mesures d’effet Seebeck de Spin.Nous avons mis à profit les techniques développées dans le cadre de cette étude pour étudier l’effet Nernst dans InSb. Nous avons utilsé un modèle de distribution de porteur pour expliquer les non-linéarités de celui-ci à des champs magnétiques proches de 1T.Nous avons construit une expérience de résonance ferromagnétique dont le but sera d’étudier les implications des effets thermique et thermoélectrique dans les expérience de pompage de spin. Enfin, nous nous sommes intéréssés au transport électrique en géométrie Corbino. La géométrie Corbino est celle d’un disque dans laquelle il n’existe aucun bord libre ou des charges peuvent s’accumuler. Cela se traduit par l’apparition d’un courant de Hall ortho-radial dont la conséquence est l’augmentation de la résistance du disque. Nous avons mis en évidence une augmentation de résistance en géométrie Corbino dans CoGd et CoTb dont l’origine est l’effet Hall anomale. Bien que cet effet ne soit pas dissipatif, il a donné naissance à un courant dissipatif transverse. Nous avons également vu que cet effet entre en compétition avec la magnétorésistance anisotrope dans le permalloy.Du fait de la similitude entre l’effet Hall anomal et l’effet Hall de spin, on s’attend dans le platine à l’existence d’un fort courant de spin orthoradial sans possibilités d’accumulations dont la mise en évidence expérimentale fera l’objet de travaux futurs. / Abstract: During this PhD we studied different electric, thermal and thermoelectric properties. For instance, we characterized the anomalous and planar Righi-Leduc effects which are the thermal equivalent of the anomalous and planar Hall effects. These effect have to be taken into account when interpreting spin Seebeck measurements.We used the technics that we developped during this study to look at the Nernst effect in InSb. We developped, a carrier mobility distribution model to explain its non-linearity at fields around 1T.We built a ferromagnetic resonance experiment in order to study the impact of thermal and thermoelectric properties in spin pumping effect.Lastly, we studied electric transport in the Corbino geometry. Corbino geometry is the one of a disc where there are no free boundaries where electric charges can accumulate. This causes the apparition of an orthoradial Hall current which consequence is the increase of resistance of the disc. We showed an increase of resistance in the Corbino geometry in CoGd and CoTb originating from anomalous Hall effect. Despite the anomalous Hall effect does not dissipate, it produces an orthoradial current which dissipates. We also found in permalloy that this increase of resistance is counterbalanced by a decrease of resistance due to the anisotropic magnetoresistance.The similarity between anomalous Hall effect and spin Hall effect which share a common microscopic origin implies that we expect in platinum the apparition of an orthoradial spin current without possibility for the charges to accumulateition. This current should dissipate just the way it does for the anomalous Hall effect.The study of this spin current will be the topic of a futur study.