Particle dispersions provide a promising tool for the engineering of functional materials that exploit self-assembly of complex structures. Dispersion made from magnetic colloidal particles is a great choice; they are biocompatible and remotely controllable among many other advantages. However, their dominating dipolar interaction typically limits structural complexity to linear arrangements. This paper shows how a magnetostatic equilibrium state with noncollinear arrangement of the magnetic moments, as reported for ferromagnetic Janus particles, enables the controlled self-organization of diverse structures in two dimensions via constant and low-frequency external magnetic fields. Branched clusters of staggered chains, compact clusters, linear chains, and dispersed single particles can be formed and interconverted reversibly in a controlled way. The structural diversity is a consequence of both the inhomogeneity and the spatial extension of the magnetization distribution inside the particles. We draw this conclusion from calculations based on a model of spheres with multiple shifted dipoles. The results demonstrate that fundamentally new possibilities for responsive magnetic materials can arise from interactions between particles with a spatially extended, anisotropic magnetization distribution.

L’exposition des personnes aux champs électromagnétiques est une problématique majeure de société qui touche le domaine de l’automobile. Le développement de la technologie des véhicules électriques et hybrides nécessite la prise en compte de la problématique d’exposition aux ondes dès la phase même de la conception.Le travail présenté dans ce document s’intéresse principalement à l’exposition aux champs magnétiques basse fréquence (BF) marqués par les effets non thermiques de stimulation électriques. Dans cette gamme fréquentielle, les sources principales des risques d’exposition sont les modules de la chaîne de traction qui sont souvent situés à proximité des passagers. L’objectif principal de cette thèse est de virtualiser la mesure évaluant l’exposition aux champs magnétiques BF, jusqu’ici réalisée par un testeur de niveau d’exposition commercial, afin d’être capable de faire une levée de risque d’exposition, en amont, sans entraîner de surcoût au constructeur. Ceci est réalisé en virtualisant l’instrument de mesure.Nous consacrons d’abord notre étude à la compréhension de la mesure d’évaluation d’exposition et du fonctionnement du champmètre Narda ELT-400 utilisé chez notre constructeur automobile pour mesurer l’indice d’exposition. Ainsi, nous détaillons les différentes définitions et méthodes de calcul de l’indice d’exposition du public général aux champs magnétiques BF. Enfin, nous présentons des applications de l’utilisation du champmètre sur des cas de mesure en statique et en dynamique.La deuxième partie de l’étude traite de la construction du prototype virtuel permettant de modéliser la mesure d’indice d’exposition aux champs magnétiques BF réalisée par le champmètre Narda ELT-400. Pour ce faire, nous développons et optimisons dans un premier temps le modèle de la sonde triaxiale du champmètre sous FEKO permettant de positionner et d’orienter la sonde selon le choix de l’utilisateur. Ensuite le champ magnétique rayonné par la sonde virtuelle est traité pour obtenir une mesure virtuelle de l’induction magnétique détectée par la sonde et de l’indice d’exposition.La troisième partie est consacrée à la validation du modèle complet sur des cas réels de mesures effectuées sur des pinces manuelles de soudage par points en comparant les résultats du prototype virtuel à ceux mesurés. La validation du modèle est réalisée dans trois cas d’étude différents correspondant à des profils de champ différents, des fréquences différentes dans la gamme BF et à des formes de signal différentes.Dans la dernière partie de notre étude, une approche stochastique de simulation est proposée en appliquant la méthode de Monte-Carlo, basée sur un grand nombre de tirages aléatoires. Cette méthode permet d’étudier la dispersion des résultats due à l’imprécision plausible de positionnement et d’orientation commise par l’opérateur de la sonde. Cette méthode est appliquée à différentes distances de sources rayonnantes élémentaires. Les résultats principaux de cette partie portent sur le lien entre l’homogénéité du champ au point investigué et la variabilité de la mesure virtuelle autour de ce point suite à l’imprécision commise. / Exposure to electromagnetic fields is a major issue in society that affects the automotive domain. The development of electric and hybrid vehicles technology requires the consideration of the issue of exposure since the conception stage.The work presented in this paper is mainly concerned with exposure to low-frequency (LF) magnetic fields distinguished with non-thermal electric stimulation effects. In this frequency range, the main sources of exposure risks are the powertrain elements often located close to the passengers.The main objective of this thesis is to virtualize the measurement assessing exposure to LF magnetic fields, hitherto performed by a commercial exposure level tester, to be able to minimize exposure risks, a priori, without incurring additional costs to the manufacturer. This is realized by virtualizing the measuring instrument.We focus first on understanding the exposure evaluation measurement and on the operation of the ELT-400 Narda field meter used at our automobile manufacturer to measure the exposure index. Thus, we show the different definitions and methods of calculating exposure index of the general public to LF magnetic fields. Finally, we present applications of the use of the field meter on measuring cases in statistic and dynamic states.The second part of the study is consecrated to the construction of the virtual prototype to model the measurement of the exposure index to LF magnetic fields performed by Narda field meter ELT-400. To do this, first we develop and optimize the model of the field meter’s triaxial probe in FEKO to position and orient the probe as selected by the user. Then the magnetic field radiated by the virtual probe is processed to obtain a virtual measurement of the magnetic induction detected by the probe and of the exposure index.The third part of the study deals with the validation of the whole model on real cases of measurements done over manual spot welding clamps by comparing the virtual prototype results with those measured. The model validation is performed in three different case studies corresponding to different field profiles, different frequencies in LF domain and different forms of signal.In the last part of our study, a stochastic simulation approach is proposed by applying the Monte-Carlo Method based on a large number of random draws. This method allows studying the dispersion of results due to the plausible imprecise positioning and orientation committed by the operator of the probe. This method is applied at different distances from elementary radiating sources. The main results of this part focus on the link between the field homogeneity at the investigated point and the variability of the virtual measurement around this point due to the inaccuracy committed.

“La vie ne produit pas de corps symétrique”L. Pasteur. Briser la symétrie-miroir est un sujet fascinant qui revêt un rôle crucial en physique, en chimie et en biologie.Le but ultime de cette étude est de démontrer expérimentalement la conjecture de Curie de-Gennes. Cette hypothèse a été proposée pour la première fois par P. Curie 1894 et plus tard développée par De-Gennes. Elle indique qu’il est possible de générer une dissymétrie (excès énantiomérique (e.e.)) en soumettant un mélange racémique à l'influence combinée de champs magnétique et électrique colinéaires. Une telle influence modifie la cinétique du système : selon l'orientation relative des champs magnétique et électrique, l'un des énantiomères est formé plus rapidement que l'autre en raison d’une énergie d'activation inférieure (Ea). Une telle influence a été ultérieurement appelée « faussement chirale » par Barron. Pour démontrer expérimentalement cette conjecture, deux expériences différentes ont été développées :1) La cristallisation énantiosélective d'un réseau de coordination chiral sous des influences physiques externes (champ magnétique ou force de rotation) par diffusion lente. Un petit e.e. est supposé être créé puis amplifié par la croissance des cristaux à l'interface solvant/vapeur d'éthanol. Dans cette expérience, nous avons constaté que la démonstration de conjecture de Curie de Gennes est difficile à atteindre en raison de la forte dispersion de la moyenne des valeurs du e.e. obtenues et de la faiblesse de l'effet attendu.2) La création d'un e.e. dans un atropisomère racémique d’un cristal liquide à base de biphényle en appliquant des champs magnétique et électrique (anti) parallèles. Un arrangement supramoléculaire hélicoïdal des molécules de biphényles dans la phase nématique augmentera le signal lié à la présence d’un e.e. à un niveau mesurable. Au cours de cette seconde expérience, nous avons observé expérimentalement un e.e des dérivés à base de biphényle induit par une combinaison (anti)parallèle de champs magnétique et électrique. Cet excès dépend linéairement du produit des champs magnétique et électrique et l’ordre de grandeur des résultats observés expérimentalement correspond aux estimations théoriques. Compte tenu des résultats obtenus, sous réserve d’ultimes vérifications, cette expérience constitue une démonstration expérimentale convaincante de la conjecture Curie de-Gennes. / “La vie ne produit pas de corps symétrique” L. Pasteur Mirror symmetry breaking is an ever-fascinating topic that plays a crucial role in physics, chemistry, and biology. The ultimate goal of this study is to demonstrate experimentally the so-called Curie de-Gennes conjecture. It was proposed for the first time by P. Curie (1894) and later developed by de-Gennes. It stipulates that it is possible to generate dissymmetry (enantiomeric excess (e.e.)) by submitting a racemic mixture to the combined influence of collinear magnetic and electric fields. Such influence modifies the kinetics of the system: depending on the relative orientation of the magnetic and electric fields, one of the enantiomers is formed faster than the other because of a lower activation energy (Ea),. This influence was later named by Barron as « falsely chiral ». To demonstrate this conjecture, two different experiments were developed:1) The enantioselective crystallization of a chiral metal-organic framework under external physical influence (magnetic field or rotation force) by slow diffusion. The small e.e. is expected to be created and amplified by crystal growth at the interface aqueous solution / ethanol-saturated vapors. In this experiment, we have found that the Curie de-Gennes conjecture demonstration is difficult to achieve due to high dispersion of the average of the observed e.e. and the weakness of the expected effect. 2) Creating an e.e. in a racemic atropisomeric biphenyl based liquid crystal by applying (anti)parallel magnetic and electric fields. The helical twisting power of biphenyls molecules in the nematic phase will enhance the dichroic signal related to the expected e.e. to a measurable level. During this second experiment we have experimentally observed an e.e. in biphenyl-based nematic liquid crystals induced by (anti)parallel combination of magnetic and electric fields. This excess was linearly dependent on the product of magnetic and electric fields and, within one order of magnitude, its absolute value corresponds to the theoretical estimation. Ultimate checks shall confirm that these results constitute a convincing experimental demonstration of the Curie de Gennes conjecture.

No description available.

There is an increasing interest in nanostructured and nanocomposite surface finishings for automotive and aerospace applications. The widespread applicability of these novel materials is based on their unique mechanical, physical, and chemical properties. An advantageous production method is the electrocodeposition (ECD) process from metal plating baths containing dispersed nanoparticles. By using this technique, a broad range of substrate sizes and shapes can be coated cost-effectively. However, the prediction of the amount as well the distribution of nanoparticles within the metal film fails frequently. There is no complete understanding of the particle incorporation mechanism. The goal of this research was to improve the fundamental understanding of the ECD mechanism. In order to identify the forces affecting the codeposition behavior of nanoparticles in a metal matrix, the effects of a variety of interrelated process parameters on the composite film formation have been investigated systematically. Nanocomposites containing metal and metal oxide nanoparticles in a nickel matrix have been prepared by means of ECD from two different types of nickel plating baths, an acidic sulfamate (pH 4.3) and an alkaline pyrophosphate bath (pH 9.5). The effect of deposition conditions on the ECD process was investigated utilizing two electrode configurations, viz. a parallel plate electrode (PPE) and impinging jet electrode (IJE) and different deposition techniques, viz. direct current (DC) deposition, both pulse plating (PP) and pulse-reverse plating (PRP). The surface charge and sedimentation behavior of the nanoparticles in these electrolytes were characterized by zeta potential and stability measurements. The surface charge, hydrodynamic diameter and colloidal stability of the nanoparticles in the nickel electrolytes were mainly affected by the composition and pH of the bath. The particles tend to form agglomerates in both nickel baths. Smaller agglomerates and an improved colloidal stability occurred in the case of the alkaline bath. Composites with a maximum particle content of either ~3.6 vol-% of 13 nm Al2O3 or ~10.4 vol-% of 21 nm TiO2 were obtained using a parallel plate electrode and DC deposition conditions. Both jet plating as well as pulse plating resulted in a distinct increase of the particle codeposition. A maximum incorporation of ~12 vol-% of 50 nm Al2O3 particles in a nickel matrix was achieved using an unsubmerged IJE system, while PP and PRP resulted in composites with particle contents up to 11 vol-% of 13 nm Al2O3. The particle incorporation increased with the particle content of the electrolyte for all deposition conditions studied. A beneficial effect on the amount of codeposited particles was found with decreasing average current density. The Al2O3 and TiO2 particles were found to be negatively charged in the alkaline pyrophosphate bath, and positively charged in the acidic sulfamate bath. It could be shown that negatively charged particles codeposited preferentially within the nickel matrix. The effect of PP and PRP conditions, e.g. pulse frequency, duty cycle and value of the peak current density, on the ECD of Ni-Al2O3 composites was studied using rectangular current pulses in the order of milliseconds. In general, low duty cycles and high pulse frequencies resulted in an enhanced particle codeposition. Using the unsubmerged IJE system, the effects of jet flow rate, particle loading and current density on the particle incorporation were studied. Referring to the experimental results from the ECD of 50 nm alumina with nickel using an IJE system, a kinetic model was developed. Therefore, the particle flux to the electrode was derived from an analysis of the total force acting on the particle in front of the electrode. The model took into account the convective diffusion of particles to the electrode surface, and the effect of gravitational and buoyancy forces on the particle flux. The gravitational force was found to be important for the ECD of 300 nm particles, but not for 50 nm particles. The effect of an external magnetic field on the ECD of Co or Fe3O4 nanoparticles in a nickel matrix has been studied for different current densities, particle contents of the electrolyte and magnetic flux density. The particle incorporation showed a distinct dependency on the orientation of an externally applied magnetic field. While the particle incorporation increased in a perpendicular field (perpendicular with regard to the electrode surface), it decreased in a parallel orientation. The influence of the magnetic field on the ECD of magnetic nanoparticles with nickel was explained by the interplay of Lorentz force and magnetophoretic force. The structure and the properties of the nickel matrix were significantly altered due to the codeposition of nanoparticles. The pure nickel deposits from the sulfamate bath exhibited a strong <100> texture, and those from the pyrophosphate bath a strong <110> preferred orientation. With increasing plating current density and particle incorporation, a variation in the crystallite size and a loss of texture was observed. High resolution TEM imaging proved a complete embedding of nanoparticles by the nickel matrix without any voids. In the case of both nickel baths, the Vickers microhardness showed a tendency to increase with the amount of particle incorporation. The enhanced hardness of the composite films was associated with modifications in the microstructure of the nickel matrix as well as with the nanoparticle incorporation. The wear resistance as examined by linear abrasion test increased with decreasing current density and due to the particle incorporation. Furthermore, the incorporation of magnetic nanoparticles resulted in a distinct increase of the magnetic hardness of the nickel matrix.

The aim of this work was to investigate the effects that a magnetic field can induce during the electrodeposition of Ni and Ni-Fe alloys. Special regard was given to mass transport controlled effects. Magnetic field effects on the nucleation and growth of ferromagnetic layers and on the properties of electrodeposited layers (like grain size, texture, morphology or roughness) were investigated. The influence of a magnetic field on the magnetic properties of Ni layers and on the composition of Ni-Fe alloys was also studied. Nucleation and growth of thin Ni layers on gold electrodes under a superimposed magnetic field were analysed in-situ with the Electrochemical Quartz Crystal Microbalance technique. Three theoretical models were chosen for characterizing the Ni nucleation: Scharifker-Hills (SH), Scharifker-Mostany (SM) and Heerman-Tarallo (HT). The AFM images proved that more nuclei appear in a magnetic field in the case that the Lorentz force and the natural convection act in the same direction. From all the models, the HT model gave the best agreement with the AFM results. When the Lorentz force and the natural convection act in the same direction, an increase of the Ni partial current with the magnetic field was obtained. When they act in opposite directions, the Ni current was influenced just at the beginning of deposition (first 10 seconds). At longer times, the magnetic field has no effect on the Ni current. However, the total current (jNi+jHER) decreases with the magnetic field. In the absence of a macroscopic MHD convection, the Ni current decreases with the magnetic field the first 10-15 seconds of deposition. On longer time scales no influence of the magnetic field could be noticed for this configuration. When the magnetic field was applied perpendicular to the electric current, an increase of the hydrogen evolution reaction (HER) with the magnetic flux density was noticed. Hydrogen reduction is mass transport controlled. Therefore, the magnetic field will increase the limiting current of the HER. Optical microscopy images showed that the hydrogen bubbles were circular in the absence of the MHD convection and that they presented a tail when a Lorentz force was present. The direction of the tail depends on the net force induced by the natural and MHD convections. The interplay between the natural and MHD convections proved to be important during Ni-Fe alloy deposition, too. When the Lorentz force and the natural convection act in the same direction, an increase of the Fe content of the alloys with the magnetic field was observed. When the Lorentz force was perpendicular to the natural convection, no significant changes were observed in the composition of the layers. The alloy composition did not change with the magnetic field when the electric current was parallel to the magnetic field lines. Two surfactants were used in the case that Ni was electrodeposited from a sulfamate bath: SDS and sulfirol 8. The Ni layers obtained from a sulfamate bath with sulfirol 8 presented larger grains compared to the layers deposited from a bath free of surfactants. This increase of the grain size was attributed to the incorporation of the surfactant in the deposit. Coarser layers were obtained in a magnetic field (applied perpendicular to the electric current) when the electrodeposition was done from an electrolyte with surfactants. The number of grains increased with the magnetic field for the Ni layers electrodeposited from a bath free of surfactants and for a bath with SDS. As a consequence, the grain size decreased. In the case of the electrolyte with sulfirol 8, the number of grains decreased with the magnetic field, and their size increased. For the Ni-Fe alloys, which contained less than 10 at% Fe, the preferred crystalline orientation changes from (220), in the absence of a magnetic field, to (111), (when the magnetic field was applied perpendicular to the electric current). When the magnetic field lines were parallel to the electric current, both the (111) and (220) textures were preferred in almost the same proportion. As a general conclusion of this work it can be said that by choosing the right experimental condition, one can improve the morphology and the properties of the deposited layers by applying a magnetic field. At the same time, the mass transport processes can be influenced by a magnetic field.

Superconducting Nanowire Single Photon Detectors (SNSPDs) are a promising technology for detecting single photon emissions with high efficiency and low noise. This detector class has numerous applications in quantum optics, communication, and sensing. One typical design for these devices is the hairpin structure, in which a superconducting nanowire is patterned into a meandering shape. The combination of academic research and interest from the industry is boosting the development of hairpin SNSPD devices to achieve high detection efficiency while maintaining fast response time and low jitter, requiring optimization of the device geometry, materials properties, and sophisticated readout electronics. This thesis qualitatively enquires about different device geometries, varying turn designs and features. Moreover, proposing a promising experimental setup with the potential of being scaled up to simultaneously test numerous devices with a varying magnetic field, driving the hairpin SNSPDs to their detection limit, and enabling further quantitative studies to deepen the understanding of the underlying mechanisms currently hindering the SNSPDs. Analyzing the acquired data draws results regarding the critical current and dark counts trends. Furthermore, at low magnetic field strength, the enquired devices are found to have their critical current enhanced. Moreover, comparisons are drawn among similar design structures. Furthermore, a discussion on manufacturing defects detrimental to the SNSPD performance is initiated. Finally, further studies on this topic adopting the presented method are encouraged to acquire additional quantitative results to be compared with theoretical models describing the thin superconducting structures. / Superconducting Nanowire Single Photon Detectors (SNSPD) är en lovande teknik för att detektera utsläpp av enstaka fotoner med hög effektivitet och lågt brus. Denna detektorkategori har många tillämpningar inom kvantoptik, kommunikation och avkänning. En typisk konstruktion för dessa anordningar är hairpin-strukturen, där en supraledande nanotråd är mönstrad i en slingrande form. Kombinationen av akademisk forskning och intresse från industrin ökar utvecklingen av hairpin-SNSPD-enheter för att uppnå hög detektionseffektivitet med bibehållen snabb responstid och låg jitter, vilket kräver optimering av enhetens geometri, materialegenskaper och sofistikerad avläsningselektronik. I denna avhandling undersöks kvalitativt olika anordningsgeometrier, varierande vridningsdesign och funktioner. Dessutom föreslås en lovande experimentell uppställning med potential att skalas upp för att samtidigt testa många anordningar med ett varierande magnetfält, vilket driver hairpin-SNSPD:erna till sin detektionsgräns och möjliggör ytterligare kvantitativa studier för att fördjupa förståelsen av de underliggande mekanismer som för närvarande hindrar SNSPD:erna. Analysen av de insamlade uppgifterna ger resultat när det gäller den kritiska strömmen och trenderna för mörkertalet. Dessutom visar sig de undersökta enheterna ha en ökad kritisk strömstyrka vid låg magnetfältsstyrka. Dessutom görs jämförelser mellan liknande konstruktionsstrukturer. Dessutom inleds en diskussion om tillverkningsfel som är skadliga för SNSPD:s prestanda. Slutligen uppmuntras ytterligare studier i detta ämne med den presenterade metoden för att få ytterligare kvantitativa resultat som kan jämföras med teoretiska modeller som beskriver tunna supraledande strukturer.

本論文では，鉄損算出精度向上を目的としたプレイモデルによるヒステリシス特性を考慮した磁界解析の実用化に向けて，ヒステリシスモデルのモデリング精度向上，計算機の記憶容量低減，有限要素磁界解析の計算時間削減に関する手法を提案した．さらに，ヒステリシス特性の考慮の有無が解析結果に与える影響を検討し，提案したヒステリシス特性を考慮した有限要素磁界解析の実用性を明らかにした． / In order to achieve the highly accurate analysis of an iron loss for electric machines, this paper proposes finite element magnetic field analyses taking account of the hysteretic property by using the play model. As a consequence, it is verified that the proposed method can accurately estimate an iron loss of a magnetic materials and decrease computational costs. Furthermore, this paper investigates the influence of the hysteretic property in finite element magnetic field analyses to demonstrate its effectiveness. / 博士(工学) / Doctor of Philosophy in Engineering / 同志社大学 / Doshisha University

Current controls the motion of a manipulator. The manipulators at ABB are powered by a three-phase alternating current system where shunt resistors are utilised to measure the current to the motors. Magnetic field sensors are instead investigated to eliminate issues with power losses, the number of components and the cost of the shunt resistors. Since current produces a magnetic field, it can be measured without contact using a magnetic field sensor. However, employing non-contact magnetic field sensors in three-phase implementations introduces problems with stray magnetic fields due to the three traces being in close proximity to each other. This magnetic crosstalk will influence the sensors, hence the current measurement for each trace. In this thesis separating this influence of the magnetic fields is done through a software approach. Initially, two magnetic field sensors, a tunnel magnetoresistance sensor and a Fluxgate sensor, were tested and evaluated to gain knowledge and understanding. From the different tests, it was decided to continue with the Fluxgate sensor. Further, a partial least-squares regression was constructed to separate the magnetic field and predict the current in each trace from the Fluxgate sensor output. From a simulation created, the current could be predicted with an error of approximately 1 nA, meaning less than 0.1% when considering a simulated linear model of the Fluxgate sensor.

Les décharges Sparks sont des décharges électriques transitoires avec une courte durée de vie. Par rapport à son initiation en milieu gazeux, l’initiation de ce type de décharges dans un liquide diélectrique induit de nouveaux phénomènes physico-chimiques dans le plasma et aux interfaces plasma-liquide et plasma-électrodes. Depuis une vingtaine d’années, la recherche scientifique exploite les propriétés de ces décharges pour des applications diverses : dépollution de liquide, synthèse de nanoparticules, usinage par électro-érosion, etc. Dans ce contexte, ce mémoire a pour objectif d’apporter une meilleure compréhension de la physique des décharges Sparks dans les liquides diélectriques. Les décharges dans les liquides se caractérisent par un comportement stochastique fort. Des études statistiques d’un nombre important de décharges sur les caractéristiques électriques ont été effectuées en fonction de différents paramètres. Ces paramètres sont la distance inter-électrodes, la nature des électrodes ainsi que la polarité de la tension appliquée. L’acquisition des courbes courant-tension de chaque décharge permet de déterminer ses propriétés électriques, soient la tension de claquage, le courant de la décharge, le délai de claquage, la charge injectée, la probabilité de claquage, etc. L’influence d’un champ magnétique externe, en particulier son orientation par rapport à l’axe des électrodes, sur les caractéristiques de la décharge a ensuite été explorée. L’étude des interactions plasma-électrode en fonction de l’orientation du champ magnétique a été réalisée en analysant des images de la dispersion des impacts créés par les décharges sur la contre-électrode et de l’érosion de la pointe. De plus, nous avons démontré que la nature du matériau des électrodes, en particulier ses propriétés magnétiques, influe grandement le taux d’érosion de celles-ci. Les résultats rapportés dans ce mémoire contribueront non seulement à l’avancement de la physique des décharges dans les liquides, mais aussi au développement / optimisation des applications dans des différents domaines technologiques. / Spark discharges are transient electric discharges with a short lifetime. Compared to its initiation in a gaseous medium, the initiation of this type of discharges in a dielectric liquid induces new physico-chemical phenomena in the plasma and at the plasma-liquid and plasma-electrode interfaces. For about twenty years, scientific research has been exploiting the properties of these discharges for various applications: liquid depollution, nanoparticle synthesis, electro-erosion machining, etc. In this context, this thesis aims to provide a better understanding of the physics of Sparks discharges in dielectric liquids. Discharges in liquids are characterized by a strong stochastic behavior. Statistical studies of a large number of discharges on the electrical characteristics have been performed as a function of different parameters. These parameters are the inter-electrode distance, the nature of the electrodes and the polarity of the applied voltage. The acquisition of the current-voltage curves of each discharge allows to determine its electrical properties, i.e. the breakdown voltage, the discharge current, the breakdown delay, the injected charge, the breakdown probability, etc. The influence of an external magnetic field, in particular its orientation relative to the axis of the electrodes, on the characteristics of the discharge was then explored. The study of the plasma-electrode interactions as a function of the magnetic field orientation was performed by analyzing images of the dispersion of the impacts created by the discharges on the counter-electrode and the erosion of the tip. Furthermore, we have shown that the nature of the electrode material, in particular its magnetic properties, greatly influences the rate of electrode erosion. The results reported in this thesis will contribute not only to the advancement of the physics of discharges in liquids, but also to the development / optimization of applications in different technological fields.