Plasmas involving strong electron drift in crossed electric and magnetic fields are of great interest for a number of applications such as space propulsion and material processing plasma sources. Specific applications include Hall thrusters, which are high efficiency, low thrust propulsion systems used on many missions for satellite orbit corrections and for future planned interplanetary missions, as well as magnetrons of various configurations used in plasma deposition devices. Similar conditions also exist in the E-layer of the ionosphere and on the Sun. Despite many successful applications of Hall thrusters and other Hall plasma sources, some aspects of their operation are still poorly understood. A particularly important problem is the anomalous electron transport, which greatly exceeds classical collisional values. Hall plasma devices exhibit numerous turbulent fluctuations in a wide frequency range and it is believed that fluctuations resulting from plasma instabilities are likely one of the main causes of the observed anomalous transport. Plasma turbulence also affects many other important processes such as electron injection, location of the ionization region and wall erosion among others that influence the operation and efficiency of Hall thrusters. In Hall thrusters, the E0xB0 flow is made unstable due to gradients in the plasma density, temperature and magnetic field. The gradient drift instabilities are long wavelength instabilities that propagate in the azimuthal direction. A fluid theory of these unstable modes is proposed. It is shown that a full account of the compressibility of the electron flow in inhomogeneous magnetic field leads to quantitative modifications of the previously obtained instability criteria and characteristics of the unstable modes. The ExB drift also drives ion sound type instabilities in Hall thrusters. The reactive/dissipative response of the closure current to the thruster walls drives these negative energy modes. A model for this type of instabilities is proposed and analyzed for typical Hall thruster conditions. It is shown how wall parameters modify the characteristic growth rate and frequency of the unstable modes and the related anomalous transport. Nonlinear phenomena are important to understand different aspects of the Hall thruster plasma dynamics. A nonlinear fluid model for the typical Hall thruster plasma is proposed. The model takes into account electron inertia, electron collisions with neutrals, density gradients as well as various nonlinear terms that arise from the electron drift and nonlinear polarization that were included via the gyroviscous cancellation. The proposed model includes the long wavelength and the low hybrid modes destabilized by density gradients and collisions. This system of fluid equations was implemented using the computational framework BOUT++ from which a set of nonlinear simulations of plasma turbulence was performed. It is shown from these first principles nonlinear simulations that small scale low hybrid oscillations result in an anomalous electron current significantly exceeding the classical collisional current.

Dans un propulseur à courant de Hall, la création des ions et leur accélération sont régis par le même phénomène physique. L'idée du propulseur de Hall double étage (DSHT) est de découpler l'ionisation du gaz (poussée) et l'accélération des ions (ISP), de sorte à rendre le système davantage versatile. Les travaux menés durant cette thèse visent à démontrer, grâce à des essais expérimentaux, la pertinence et la faisabilité d'un tel concept. Dans un premier temps, un prototype de DSHT, baptisé ID-HALL, a été conçu et assemblé. Il est constitué d'une source inductive magnétisée insérée dans un tube en céramique et d'un étage d'accélération identique à une barrière magnétique de propulseur simple étage. La source inductive a été optimisée de sorte à réduire le couplage capacitif et à maximiser l'efficacité du transfert de puissance par ajout de pièces en ferrite et diminution de la fréquence RF d'excitation. Dans un deuxième temps, la source inductive du propulseur a été caractérisée indépendamment du propulseur en argon et xénon pour différentes pressions. Le dispositif expérimental a permis notamment de tracer une cartographie 2D de la densité et de la température. Enfin, le propulseur a été monté dans son caisson et des mesures préliminaires (caractéristiques courant-tension, mesures par sonde RPA) ont été menées. En parallèle, des simulations utilisant un modèle hybride 2D ont été effectuées en mode simple et double étage. Elles mettent en évidence un fonctionnement versatile du moteur pour des tensions inférieures à 150 V. A terme, on visera à démontrer que la densité de courant et l'énergie des ions peuvent être, dans certaines conditions, significativement découplées. / In Hall thrusters, the same physical phenomenon is used both to generate the plasma and to accelerate ions. Furthermore, only a single operating point is experimentally observed. The double stage Hall thruster (DSHT) design could allow a separate control of ionization (thrust) and ions acceleration (ISP) to make the system more versatile. The work carried out during this PhD aims to experimentally demonstrate the relevance and the feasibility of this concept. Firstly, a new design of DSHT, called ID-HALL, was proposed and a new prototype was built. It combines the concentric cylinder configuration of a single stage Hall thruster with a magnetized inductively coupled RF plasma source (ICP) whose coil is placed inside the inner cylinder. The ICP source was improved in terms of power coupling efficiency by adding ferrite parts and by decreasing the heating RF frequency. The ICP source used in the ID-HALL thruster was then characterized independently of the thruster using argon and xenon and varying pressure. The experimental setup has allowed to measure the spatial variations of the electron density and temperature. Finally, the thruster was mounted in its vacuum chamber and preliminary measures (voltage-current characteristics, RPA measurements) were led. At the same time, simulations using a two-dimensional hybrid model were performed in single and double stage. A versatile operation for voltages lower than 150 V was highlighted. An emphasis will be given to demonstrate that the current density (given by the ion flux probe) and the ions energy (given by the RPA) might be significantly decoupled.

La question de la propulsion spatiale a été un enjeu politique au coeur de la guerre froide et reste un enjeu stratégique de nos jours. La technologie chimique déjà en place sur les moteurs fusées s'avère être limitée par la vitesse d'éjection et la durée de vie des appareils. La propulsion électrique et plus particulièrement le moteur à effet Hall apparait ainsi comme la technologie la plus performante et la plus utilisée pour diriger un satellite dans l'espace. Cependant, la physique à l'intérieur d'un propulseur étant complexe, de par les champs électromagnétiques ou les processus de collisions importants, toutes les particularités de fonctionnement du moteur ne sont pas parfaitement expliquées. Au bout de centaines d'heures d'essais, certains prototypes voient leur paroi s'éroder de façon anormale et des instabilités électromagnétiques se développent au sein de la chambre d'ionisation. La mobilité des électrons mesurée est en contradiction avec les modèles analytiques et soulèvent des problématiques sur la physique du plasma à l'intérieur de ces moteurs. Par conséquent, le code AVIP a été développé afin de proposer un code 3D massivement parallèle et non-structuré à Safran Aircraft Engines modélisant le plasma instationnaire à l'intérieur du propulseur. Des méthodes lagrangiennes et eulériennes sont utilisées et intégrées dans le code et mon travail s'est concentré sur le développement d'un modèle fluide, étant plus rapide et donc mieux adapté à la conception et au design industriel. Le modèle fluide est basé sur un modèle aux moments avec une expression rigoureuse des termes de collisions et une description précise des conditions limites pour les gaines. Ce modèle a été implémenté numériquement dans un formalisme non structuré et optimisé de façon à être performant sur les nouvelles architectures de calcul. La modélisation retenue et les efforts d'optimisation ont permis de réaliser un calcul réel de moteur à effet Hall afin de retrouver les propriétés globales de fonctionnement telles que l'accélération des ions ou encore la localisation de la zone d'ionisation. Un second cas d'application a finalement reproduit avec succès les instabilités azimutales dans le propulseur avec un modèle fluide et a justifié le rôle de ces instabilités dans le transport anormal des électrons et l'érosion des parois / The space propulsion has been a political issue in the midst of the Cold War and remains nowadays a strategic and industrial issue. The chemical propulsion on rocket engines is limited by its ejection velocity and its lifetime. Electric propulsion and more particularly Hall effect thrusters appear then as the most powerful and used technology for space satellite operation. The physic inside a thruster is complex because of the electromagnetic fields and important collision processes. Therefore, all specificities of the engine operation are not perfectly understood. After hundreds of hours of tests, thruster walls are curiously eroded and electromagnetic instabilities are developping within the ionization chamber. The measured electron mobility is in contradiction with the analytical models and raises issues on the plasma behavior inside the discharg chamber. As a result, the AVIP code was developed to provide a massively parallel and unstructured 3D code to Safran Aircraft Engines modeling unsteady plasma inside the thruster. Lagrangian and Eulerian methods are used and integrated in the solver and my work has focused on the development of a fluid model which is faster and therefore better suited to industrial conception. The model is based on a set of equations for neutrals, ions and electrons without drift-diffusion hypothesis, combined with a Poisson equation to describe the electric potential. A rigorous expression of collision terms and a precise description of the boundary conditions for sheaths have been established. This model has been implemented numerically in an unstructured formalism and optimized to obtain good performances on new computing architectures. The model and the numerical implementation allow us to perform a real Hall effect thruster simulation. Overall operating properties such as the acceleration of the ions or the location of the ionization zone are captured. Finally, a second application has successfully reproduced azimuthal instabilities in the Hall thruster with the fluid model and justified the role of these instabilities in the anomalous electron transport and in theerosion of the walls

Um propulsor do tipo Hall é um mecanismo que utiliza predominantemente uma configuração de campos eletromagnéticos Hall, um campo elétrico perpendicular a um campo magnético, para confinar elétrons e acelerar íons. Os elétrons são confinados dentro de um canal de aceleração onde os campos eletromagnéticos estão presentes. Um gás neutro é lançado dentro desse canal de aceleração de forma que os elétrons confinados podem colidir com os átomos do gás e os ionizar. Os íons gerados dessas colisões, elétrons-gás, são fortemente repelidos para fora do canal de aceleração pelo campo elétrico. A expulsão desses íons é o fator responsável pela propulsão. Nesses propulsores é importante que os elétrons estejam confinados dentro do canal de aceleração e que sejam capazes de produzir o maior número possível de íons. Visando determinar quais são os parâmetros de controle– intensidade dos campos eletromagnéticos– que propiciam uma dinâmica eletrônica com essas características, derivamos, via formalismo Hamiltoniano, as equações de movimento de um elétron e as analisamos. Dessas equações de movimento encontramos funções analíticas que indicam os limites geométricos atingidos pelo elétron dentro do sistema propulsor para cada conjunto de parâmetros de controle. Essas funções constituem o critério de confinamento eletrônico utilizado nesse trabalho. Além disso, a partir das equações de movimento, mostramos quais as configurações de campos eletromagnéticos que teoricamente incrementam o desempenho dos propulsores Hall. Verificamos que nas configurações de maior desempenho a dinâmica eletrônica é caótica. Neste trabalho, o caos é determinado com o auxílio dos mapas de Poincaré e dos expoentes de Lyapunov. / A Hall thruster is a system that utilizes an electromagnetic fields configuration predominantly like Hall, an electric field which lies perpendicular to a magnetic field, to confine electrons and to accelerate ions. The electrons are confined within an acceleration chamber where the electromagnetic fields are present. A neutral gas is released within this acceleration chamber so that the confined electrons can collide with the gas and ionize it. The ions generated from these collisions, the electron-gas, are strongly repelled by the electric field system. The expulsion of these ions generate the propulsion. In these thrusters it is very important that the electrons are confined within the acceleration chamber and are able to produce the largest possible number of ions. In order to determine the control parameters, that is, the electromagnetic fields intensity which provides an electronic dynamic with these characteristics; we derived, via Hamiltonian formalism, the motion equations for an electron and we analyzed them. From these motion equations, we found functions that indicate the electron geometric boundaries within these thrusters, for each set of control parameters. In this work, these functions indicate the electronic confinement. Moreover, from the motion equations, we showed the electromagnetic fields settings which theoretically improve the Hall thruster’s performance. We found that, in these higher performance settings, the electron dynamics is chaotic. In this work, the chaos is determined by Poincaré maps and by Lyapunov exponents.

Um propulsor do tipo Hall é um mecanismo que utiliza predominantemente uma configuração de campos eletromagnéticos Hall, um campo elétrico perpendicular a um campo magnético, para confinar elétrons e acelerar íons. Os elétrons são confinados dentro de um canal de aceleração onde os campos eletromagnéticos estão presentes. Um gás neutro é lançado dentro desse canal de aceleração de forma que os elétrons confinados podem colidir com os átomos do gás e os ionizar. Os íons gerados dessas colisões, elétrons-gás, são fortemente repelidos para fora do canal de aceleração pelo campo elétrico. A expulsão desses íons é o fator responsável pela propulsão. Nesses propulsores é importante que os elétrons estejam confinados dentro do canal de aceleração e que sejam capazes de produzir o maior número possível de íons. Visando determinar quais são os parâmetros de controle– intensidade dos campos eletromagnéticos– que propiciam uma dinâmica eletrônica com essas características, derivamos, via formalismo Hamiltoniano, as equações de movimento de um elétron e as analisamos. Dessas equações de movimento encontramos funções analíticas que indicam os limites geométricos atingidos pelo elétron dentro do sistema propulsor para cada conjunto de parâmetros de controle. Essas funções constituem o critério de confinamento eletrônico utilizado nesse trabalho. Além disso, a partir das equações de movimento, mostramos quais as configurações de campos eletromagnéticos que teoricamente incrementam o desempenho dos propulsores Hall. Verificamos que nas configurações de maior desempenho a dinâmica eletrônica é caótica. Neste trabalho, o caos é determinado com o auxílio dos mapas de Poincaré e dos expoentes de Lyapunov. / A Hall thruster is a system that utilizes an electromagnetic fields configuration predominantly like Hall, an electric field which lies perpendicular to a magnetic field, to confine electrons and to accelerate ions. The electrons are confined within an acceleration chamber where the electromagnetic fields are present. A neutral gas is released within this acceleration chamber so that the confined electrons can collide with the gas and ionize it. The ions generated from these collisions, the electron-gas, are strongly repelled by the electric field system. The expulsion of these ions generate the propulsion. In these thrusters it is very important that the electrons are confined within the acceleration chamber and are able to produce the largest possible number of ions. In order to determine the control parameters, that is, the electromagnetic fields intensity which provides an electronic dynamic with these characteristics; we derived, via Hamiltonian formalism, the motion equations for an electron and we analyzed them. From these motion equations, we found functions that indicate the electron geometric boundaries within these thrusters, for each set of control parameters. In this work, these functions indicate the electronic confinement. Moreover, from the motion equations, we showed the electromagnetic fields settings which theoretically improve the Hall thruster’s performance. We found that, in these higher performance settings, the electron dynamics is chaotic. In this work, the chaos is determined by Poincaré maps and by Lyapunov exponents.

La cathode creuse est un élément clef des propulseurs à plasma. Dans un propulseur à plasma, un gaz propulsif est ionisé dans un canal de décharge puis accéléré hors de celui-ci afin de créer la poussée. Dans le propulseur de Hall en particulier, l'ionisation du gaz est provoquée par l'injection dans le canal de décharge d'un intense courant électronique (de quelques ampères à plus d'une centaine d'ampères). L'élément chargé de fournir le courant électronique de la décharge, la cathode creuse, est crucial dans le fonctionnement du propulseur. Or, celle-ci est souvent idéalisée dans les modèles de propulseur et n'est que rarement étudiée pour sa physique propre. Pourtant, le développement de propulseurs de Hall de haute puissance, destinés à terme à équiper l'ensemble des missions spatiales, requiert la mise au point de cathodes capable de délivrer un fort courant (jusqu'à plus de 100 A) sur des durées de l'ordre de la dizaine de milliers d'heures. Or, la mise au point de nouvelles cathodes s'est révélée difficile en raison de l'absence de modèle susceptible de prédire a priori les performances d'une cathode en fonction de sa conception. On se propose ici de mettre en place un modèle prédictif de cathode creuse capable de retranscrire la physique du fonctionnement de la cathode. L'objectif in fine est bien sûr d'utiliser ce modèle afin de faire le lien entre la conception de la cathode et son fonctionnement dans le but de guider le développement de futures cathodes. On présentera tout d'abord brièvement le contexte d'application des cathodes creuses, et on donnera un rapide aperçu du principe de fonctionnement global de la cathode. Ensuite, après avoir effectué un tour d'horizon des différents modèles numériques de cathode creuse préexistants dans la littérature, on détaillera le modèle de la cathode développé ici, qui incorpore une description fluide du plasma, ainsi que des transferts thermiques aux parois, qui conditionnent en grande partie le bon fonctionnement de la cathode. Un soin particulier sera apporté à la validation des résultats de simulation vis-à-vis des mesures expérimentales disponibles dans la littérature, ce qui nous permettra de perfectionner certains points du modèle afin de mieux traduire la réalité physique. En particulier, une modélisation spécifique de la région de transition entre la décharge interne de la cathode et la plume du propulseur sera réalisée. Ce modèle permettra de mettre en évidence certains phénomènes d'instabilité du plasma spécifiques de cette décharge, qui ont été jusqu'ici observés expérimentalement mais jamais pleinement intégrés aux modèles de cathode creuse. A l'aide du modèle validé, on procèdera à l'analyse physique de l'ensemble des phénomènes qui gouvernent le fonctionnement d'une cathode particulière, la cathode NSTAR développée par la NASA au Jet Propulsion Laboratory. Ensuite, on s'appuiera sur le modèle numérique pour comprendre l'impact sur le fonctionnement de la cathode des choix de conception au travers d'une étude paramétrique autour de la cathode NSTAR. Les tendances dégagées nous permettront de formuler des recommandations quant au développement de cathodes de haute puissance. Enfin, dans le but d'illustrer la versatilité du modèle développé, le comportement d'une cathode creuse employant une géométrie alternative à la cathode NSTAR sera également présenté. / A hollow cathode is a critical component of plasma thrusters. In a plasma thruster, a propellant gas is ionized in a discharge chamber and accelerated out of it so as to generate thrust. In Hall thrusters in particular, the ionization of the gas is caused by an intense electron current (from a few to hundred amps) which flows through the discharge chamber. The hollow cathode is the device which is responsible for providing the discharge current. This key element is often idealized in thruster numerical models and its physical behavior is rarely studied for its own sake. Yet, developing high power Hall thrusters, designed to propel in the long run every type of space mission, requires new hollow cathodes able to supply an intense electron current (over 100 A) over a duration on the order of ten thousand hours. So far, designing new cathodes proved difficult because of the lack of model capable of predicting the performance of a cathode based on its design. In this work, we build up a predictive model of a hollow cathode capable of simulating the physics relevant to the operation of the cathode. In the end, we aim at using this model to associate design characteristics of the cathode to key aspects of the cathode performance during operation. Our goal with this model is to guide the development of future high power hollow cathodes. We will first briefly describe the range of application of hollow cathodes related to space propulsion. Then we will give a brief account of the working principles of the cathode and we will set the numerical models available in the literature prior to this one out. The numerical model developed in this work will then be described. It includes a fluid treatment of the plasma as well as an account of the heat fluxes to the walls which largely control the performance of the cathode. Simulation results will be thoroughly compared to experimental measurements available in the literature and specific aspects of the model will be refined to match up simulation results with the physical reality. For instance, a model that specifically represents the transition region between the internal plasma of the cathode and the plume of the cathode will be described. This model will enable us to highlight plasma instability phenomena which were so far observed experimentally, yet never properly included in hollow cathode models. Using the model just developed, we will analyze the physics of a particular hollow cathode which has been developed by NASA at the Jet Propulsion Laboratory, the NSTAR hollow cathode. Then, thanks to the numerical model, we will be able to carry out a parametric study revolving around the design of the NSTAR cathode. This will allow us to bring out the influence of the design on the cathode performance and we will eventually express recommendations regarding the design of future high power cathodes. To conclude, the versatility of the numerical model built up here will also be displayed through simulations of the behavior of a hollow cathode based on an alternate geometry.

Um propulsor do tipo Hall é um mecanismo que utiliza predominantemente uma configuração de campos eletromagnéticos Hall, um campo elétrico perpendicular a um campo magnético, para confinar elétrons e acelerar íons. Os elétrons são confinados dentro de um canal de aceleração onde os campos eletromagnéticos estão presentes. Um gás neutro é lançado dentro desse canal de aceleração de forma que os elétrons confinados podem colidir com os átomos do gás e os ionizar. Os íons gerados dessas colisões, elétrons-gás, são fortemente repelidos para fora do canal de aceleração pelo campo elétrico. A expulsão desses íons é o fator responsável pela propulsão. Nesses propulsores é importante que os elétrons estejam confinados dentro do canal de aceleração e que sejam capazes de produzir o maior número possível de íons. Visando determinar quais são os parâmetros de controle– intensidade dos campos eletromagnéticos– que propiciam uma dinâmica eletrônica com essas características, derivamos, via formalismo Hamiltoniano, as equações de movimento de um elétron e as analisamos. Dessas equações de movimento encontramos funções analíticas que indicam os limites geométricos atingidos pelo elétron dentro do sistema propulsor para cada conjunto de parâmetros de controle. Essas funções constituem o critério de confinamento eletrônico utilizado nesse trabalho. Além disso, a partir das equações de movimento, mostramos quais as configurações de campos eletromagnéticos que teoricamente incrementam o desempenho dos propulsores Hall. Verificamos que nas configurações de maior desempenho a dinâmica eletrônica é caótica. Neste trabalho, o caos é determinado com o auxílio dos mapas de Poincaré e dos expoentes de Lyapunov. / A Hall thruster is a system that utilizes an electromagnetic fields configuration predominantly like Hall, an electric field which lies perpendicular to a magnetic field, to confine electrons and to accelerate ions. The electrons are confined within an acceleration chamber where the electromagnetic fields are present. A neutral gas is released within this acceleration chamber so that the confined electrons can collide with the gas and ionize it. The ions generated from these collisions, the electron-gas, are strongly repelled by the electric field system. The expulsion of these ions generate the propulsion. In these thrusters it is very important that the electrons are confined within the acceleration chamber and are able to produce the largest possible number of ions. In order to determine the control parameters, that is, the electromagnetic fields intensity which provides an electronic dynamic with these characteristics; we derived, via Hamiltonian formalism, the motion equations for an electron and we analyzed them. From these motion equations, we found functions that indicate the electron geometric boundaries within these thrusters, for each set of control parameters. In this work, these functions indicate the electronic confinement. Moreover, from the motion equations, we showed the electromagnetic fields settings which theoretically improve the Hall thruster’s performance. We found that, in these higher performance settings, the electron dynamics is chaotic. In this work, the chaos is determined by Poincaré maps and by Lyapunov exponents.