The present thesis deals with the mathematical treatment of kinetic theory and focuses more precisely on the Boltzmann equation. We investigate several properties of the solutions to the latter equation: their positivity and their hydrodynamical limits for instance. We also study the local Cauchy problem for a quantic version of the Boltzmann equation.

Une galaxie est un système complexe au sens où, autant des phénomènes se produisant à l'échelle du milieu interstellaire, comme des explosions de supernovæ ou l'activité d'un trou noir supermassif, que des interactions entre galaxies au sein de groupes ou d'amas, comme l'effeuillage par effet de marée ou par effet de bélier, influencent et conditionnent l'évolution de la galaxie dans son ensemble. Comme les processus œuvrant dans de tels systèmes font intervenir une gamme d'échelle de temps et de distance considérable, allant de l'étoile individuelle aux amas de galaxies tout entier, leur modélisation constitue un immense défi qui ne peut être relevé ni par une approche purement analytique ni par l'entremise de techniques exclusivement numériques.Cette thèse, à l'interface entre modèles semi-analytique et analyse de simulations numériques, se concentre sur l'étude de l'effeuillage des étoiles des galaxies satellites par effet de marée et sur les rétro-actions induites par les supernovæ.Ce manuscrit présente, d'une part, un modèle d'occupation des halos permettant de contraindre la masse d'étoiles perdue par les galaxies satellites depuis leur entrée dans leur groupe ou leur amas ainsi qu'un modèle d'effeuillage impulsif prédisant la masse stellaire arrachée aux satellites. Ce dernier est confronté, par le truchement du modèle d'occupation des halos, aux observations des fonctions de masses des groupes et des amas.Il expose, d'autre part, l'étude des rétro-actions des supernovæ implémenté dans les simulations numériques du projet NIHAO, conduite en séparant en différentes composantes le gaz des simulations et en comptabilisant les échanges entre ces dernières, laquelle a permis de mettre en évidence trois processus distincts par le biais desquels les supernovæ réduisent ou suppriment la formation stellaire.Enfin, il détaille les améliorations techniques et scientifiques apportées au modèle semi-analytique GalICS. / A galaxy is a complex system since as many phenomena take place at the scale of the interstellar medium, such as supernovae explosions or the activity of supermassive black holes, as interactions between galaxies within groups or clusters, such as tidal or ram pressure stripping, affect and condition the evolution of the galaxy itself as a whole. Because the processes acting in such systems involve a considerable range of times and distances, going from individual stars to entire clusters of galaxies, they modelling constitutes an immense challenge that cannot be met neither by a purely analytical approach nor by solely numerical technics.This thesis, being at the interface between semi-analytical models and the analysis of numerical simulations, focuses on the study of star stripping in satellite galaxies by tidal effects and on the supernovae induced feedback.This manuscript presents, on one hand, an halo occupation model that allows to constrain the stellar mass lost by satellite galaxies since they entered their group or their cluster, and a model of impulsive stripping that predicts the stellar mass ripped out of satellites. The latter is compared, through the halo occupation model, to the observed mass functions of groups and clusters.It exposes, on the other hand, the study of the supernovae feedback implemented in the numerical simulations of the NIHAO project, performed separating the simulated gas into different components and counting the exchanges that take place between them. This allowed for the highlighting of three distinct processes through which supernovae reduce or suppress their star formation.

La digitation visqueuse est une instabilité hydrodynamique apparaissant lorsque, dans un milieu poreux, un fluide moins visqueux déplace un fluide plus visqueux. L’objectif de notre thèse est l’étude expérimentale des propriétés des motifs de digitation lorsque l'échantillon de fluide visqueux est de taille finie et lorsqu'une réaction chimique modifie la viscosité dans un milieu poreux modèle, en l’occurrence une cellule de Hele-Shaw. En particulier, notre étude a permis de quantifier la contribution de dispersion et de la digitation visqueuse à l’étalement dans l’espace d’échantillons de taille finie en fonction des paramètres expérimentaux (contraste de viscosité, vitesse de déplacement et taille de l’échantillon). Pour les fluides réactifs, nous analysons la digitation induite par une réaction A + B → C dont le produit C est plus visqueux que les réactifs A et B, ceux-ci ayant la même viscosité. Nous mettons en évidence l’effet des concentrations en réactifs, du choix du fluide vecteur et du débit d’injection sur le motif de digitation.

En milieu aquatique, 90% des microorganismes se présentent sous forme de biofilm plutôt que dans un état planctonique. Les biofilms peuvent se former sur la plupart des surfaces humides, en particulier, les milieux poreux en raison de leur grande surface spécifique. La formation du biofilm dans les milieux poreux représente un domaine précieux pour la recherche scientifique en raison de sa pertinence pour de nombreux processus industriels, telles que le traitement des eaux, la bio-médiation des sols, la récupération du pétrole et le stockage du CO2. Cependant, le développement du biofilm n’est pas simplement une agrégation passive de cellules, il implique des interactions biologiques, physiques et chimiques avec le microenvironnement. Les études macroscopiques ont démontré que les conditions hydrodynamiques dans les milieux poreux jouent un rôle décisif sur la dynamique d'accumulation des biofilms, ce qui influence à son tour les propriétés hydrodynamiques comme la porosité, la perméabilité et la chute de pression. Dans cette thèse nous avons mis au point une méthodologie et un dispositif expérimental permettant la caractérisation de la structure d’un biofilm.A partir de cette procédure, une étude expérimentale sur l’influence de l’écoulement sur la formation et la structure des biofilms a été effectuée sur une souche bactérienne Pseudomonas putida. Les biofilms sont développés dans des micros cellules d’écoulement de type Hèle-Shaw (en PDMS ou PMMA) et alimentés en continue avec un milieu nutritif. La caractérisation de la colonisation avant croissance du biofilm a été également réalisée afin de pouvoir caractériser la variabilité statistique et la reproductibilité des expériences. La formation du biofilm sur un support solide dans un écoulement cisaillé a été évaluée après 24h, 48h et 72h de développement pour deux conditions hydrodynamiques, Re=0.04 (0.0021 Pa) et Re=2 (0.094 Pa). Les observations ont été effectuées sous microscope confocal à l’aide de marqueurs fluorescents. Des images 2D sont prises en différentes positions puis sont utilisées pour effectuer une reconstruction 3D du biofilm avec l’évaluation la distribution spatiale sur une zone de 12*12mm². Nous avons ensuite mis en évidence que les biofilms formés sont peu sensibles aux conditions de colonisation initiales. Nous avons également observé une stratification du biofilm selon la hauteur. La couche interne présente une faible épaisseur (5~10 µm) mais avec une structure dense, tant dis que la couche externe présente plutôt une structure filamenteuse. Le rapport des fractions volumiques entre ces deux couches peut varier de 3 jusqu’à 12, selon le temps de formation. Cet écart est autant plus important pour le cas de faible cisaillement que celui de fort cisaillement. Ceci montre que la partie supérieure du biofilm semble être contrôlée par les conditions hydrodynamiques. En analysant la distribution spatiale du biofilm, nous avons constaté une forte hétérogénéité après 48h de développement présente dans la structure, ainsi qu’une diminution de la fraction volumique de la biomasse après 72h, pour les deux conditions hydrodynamiques imposées. Ceci évoque de probables détachements ou des érosions du biofilm. A propos de la cinétique de croissance, on constate un taux de croissance apparent différents pour chaque temps d’observation. Ces valeurs sont largement inférieures aux taux de croissance observé en culture libre. Ce résultat indique également un possible effet de l’hydrodynamique sur la croissance du biofilm. Cette étude nous permet, à partir des mesures à l’échelle microscopique, d’obtenir des informations sur la structure et le taux de croissance apparent du biofilm, ainsi que l’effet de l’hydrodynamique sur ses propriétés à l’échelle de quelques pores. Ce changement d’échelle, permettra à terme de développer des outils pour simuler et/ou modéliser l’évolution de la morphologie et la distribution spatiale d’un biofilm dans un milieu poreux. / In the aquatic environment, 90% of microorganisms are present as a biofilm rather than free-swimming cells. Biofilms may develop on most of humid surfaces, in particular, in porous media for their high specific surface area. Biofilm formation in porous media is very interesting subject for many scientific researchers, because of its relevance to many industrial processes such as water treatment, soil bio- mediation, oil recovery and CO2 storage. However, the development of the biofilm is not just a passive aggregation of bacteria cells. It involves biological, physical and chemical interactions with the bacteria’s micro-environment. Several studies in macroscopic scale have shown that hydrodynamic conditions in porous media play an essential role on the dynamics of biofilm growth, which in turn affects hydrodynamic properties of porous media such as porosity, permeability and pressure drop. In this thesis we have developed an experimental device and an appropriate methodology for the characterization of biofilm’s structure. An experimental study on the influence of fluid flow on the formation and structure of biofilms was performed using a bacterial strain Pseudomonas putida. Biofilms were grown in micro Hele-Shaw flow cell (in PDMS or PMMA) under laminar flows (Re=0.04~2) and fed continuously with a nutrient medium. Characterization of initial colonization was also carried out in order to examine the statistical variability and reproducibility of experiments. Biofilm formation on a solid support under a sheared flow (Re=0.04 (0.0021 Pa) and Re = 2 (0.094 Pa)) was evaluated after 24, 48 and 72h of development. Observations were made under a confocal laser scanning microscopes using fluorescent tag. 2D images were taken at different positions in the flow cell and used to perform a 3D reconstruction of biofilm’s structure and an evaluation of its spatial distribution for an observation area of 12 *12mm². The results show that biofilms formation is not sensitive to initial colonization. A stratification of biofilm was also observed. The inner layer has a thin thickness (5~10 µm), but with a dense structure, while the outer layer show rather a filamentous structure. The ratio of volume fractions between these two layers varies from 3 to 12, depending on the formation time. This difference is more important in the case of low shear stress than that of high shear stress, which means that the upper part of the biofilm seems to be controlled by the hydrodynamic conditions. By analyzing the spatial distribution of the biomass, we found that after 48h, the biofilm present a significant heterogeneity and the volume fraction of biomass decreases after 72h for both two hydrodynamic conditions, which suggests probable detachments or erosions of biofilm. Concerning the growth kinetics, different apparent growth rates were observed for each observation time. These values are significantly below the growth rates observed in free culture medium. This result also indicates a possible effect of hydrodynamics on the growth of biofilm. This experimental study of biofilm formation in micro-scale allowed us to obtain the information on the biofilm structural and its apparent growth rate, as well as the hydrodynamic effect on its properties across several pores of the porous media. This scaling up makes it’s possible to develop eventually mathematical models to simulate the evolution biofilm’s morphology and its spatial distribution in the porous medium.

Ce travail de recherche s'intéresse à la dynamique de bulles en ascension à grand nombre de Reynolds dans un liquide fortement confiné entre deux plaques (cellule de Hele-Shaw). Dans le régime étudié, les trajectoires des bulles et leurs déformations sont contenues dans le plan de la cellule. La dynamique essentiellement bidimensionnelle favorise en particulier l'observation d'interactions entre bulles. Cette étude expérimentale comprend donc deux volets : l'analyse de la dynamique d'une bulle isolée et de son sillage et celle des interactions hydrodynamiques entre deux bulles. La cinématique des bulles (forme, trajectoire et vitesse) est mesurée à partir de visualisations par ombroscopie sur une large gamme de tailles caractérisées par un diamètre équivalent dans le plan, noté d. La dynamique des sillages est quant à elle étudiée par Vélocimétrie par Image de Particules (PIV) à Haute Fréquence. Concernant l'étude de la bulle isolée, nous avons exploré la situation où les bulles montent dans un liquide au repos et celle où elles sont soumises à un écoulement descendant à contre-courant. En liquide au repos, pour des bulles de taille suffisante qui ne sont pas mobiles dans l'interstice d'épaisseur e nous avons montré que la vitesse moyenne d'ascension Vb est proportionnelle à (e/d)⅙ √gd, et que le nombre de Reynolds défini par Re=Vbd/v fixe la déformation des bulles. De plus des lois d'échelle simples sont obtenues dans la gamme 1200≤Re≤3000 et e/d ≤ 0,4 pour les grandeurs décrivant les oscillations de trajectoire dans le repère de la bulle. Par ailleurs, des mesures de vitesse nous ont permis de caractériser la structure du sillage associé aux oscillations de trajectoire de la bulle. Nous avons tout d'abord étudié en détail les caractéristiques du détachement tourbillonnaire. Ces mesures de vitesse dans les sillages ont également mis en évidence l'existence de deux dynamiques distinctes sur deux échelles de temps nettement séparées : la période d'oscillation de la bulle et le temps visqueux défini à partir de e. En écoulement à contre-courant, un résultat intéressant consiste en la disparition de la phase intermédiaire d'appariement tourbillonnaire dans l'allée de von Karman de bulles oscillantes pour la plus importante des vitesses du contre-écoulement. La caractérisation de la cinématique des bulles isolées et des perturbations de vitesse qu'elles induisent dans le liquide a permis d'aboutir à des lois d'échelle suffisamment robustes pour pouvoir prédire leur comportement instationnaire simplement à partir de leur taille. Cette connaissance s'avère cruciale dans l'analyse des interactions entre deux bulles pour explorer les écarts de leur comportement cinématique par rapport au cas isolé. Les expériences d'interaction entre deux bulles consistent à injecter deux bulles successives et à observer leur mouvement ainsi que celui qu'elles induisent dans la phase liquide. Le suivi des bulles par ombroscopie permet de distinguer plusieurs modes d'interaction entre les bulles : attraction horizontale, entrainement vertical, éjection du sillage ou rebond, contournement, positionnement préférentiel et coalescence. Certains mécanismes d'interaction ont été plus spécifiquement étudiés à l'aide de mesures par vélocimétrie. Nous avons ainsi pu quantifier l'effet du sillage de la première bulle sur la deuxième, et notamment caractériser l'interaction bulle - tourbillon. / We study the dynamics of bubbles rising in a liquid confined in a thin-gap cell (Hele-Shaw cell of thickness e). In the regime investigated corresponding to high Reynolds numbers, bubble paths and deformations occur in the plane of the cell. This two-dimensional dynamics facilitates the observation of bubbles interaction. The aim of the investigation is twofold: the analysis of the coupling between the motion of an isolated oscillating bubble and its wake, and the analysis of the hydrodynamical interactions between two bubbles. Bubble motions (shape, trajectory and velocity) are measured from visualizations using shadowgraphy for a large range of bubble sizes characterized by their in-plane equivalent diameter d. The behaviour of the wake is explored using High Frequency Particle Image Velocimetry (HF PIV). We investigated the kinematics of an isolated bubble when its size d increases. We showed that the mean vertical velocity of the bubble Vb is proportional to (e/d)⅙ √gd, and that the Reynolds number Re=Vbd/v determines its mean deformation. Simple scaling laws were then obtained in the range 1200≤Re≤3000 and e/d ≤ 0,4 for all the quantities describing the path oscillations of the bubble in its reference frame. Moreover, measurements of the liquid velocity allowed us to characterize the structure of the wake associated to the oscillating bubbles. We first investigated in detail the characteristics of vortex shedding. We then showed that the time evolution of the bubble wake depends on two contrasted time scales. The first corresponds to short times on the order of the period of oscillation and the second to the effect of wall friction becoming predominant for times comparable to the viscous time scale based on the gap thickness e. In the presence of a sufficiently strong counterflow, we observed the disappearance of the intermediate phase of vortex pairing in the wake of an isolated oscillating bubble. The characterization of the bubble kinematics and of the bubble-induced velocity perturbation in the liquid phase for the isolated bubble provided scaling laws robust enough to predict their periodic motion. This knowledge is fundamental for the discussion of hydrodynamical interactions, allowing us to discuss the kinematics of interacting bubbles as compared to their kinematics as isolated bubbles. Experiments consisted in the injection of two successive bubbles in the cell, the observation of their motions and the measurement of the perturbations induced in the liquid phase. Visualizations of the bubbles motions allowed us to observe several types of interactions: horizontal attraction, vertical entrainment, ejection or bouncing, preferential positioning, and coalescence. Some mechanisms occurring during interaction have been more precisely studied using HF PIV, in particular the effect of the wake of the leading bubble on the trailing bubble, and the associated bubble-vortex interaction.

We investigate the rotational and thermal properties of star-forming molecular clouds using hydrodynamic simulations. Stars form from molecular cloud cores by gravoturbulent fragmentation. Understanding the angular momentum and the thermal evolution of cloud cores thus plays a fundamental role in completing the theoretical picture of star formation. This is true not only for current star formation as observed in regions like the Orion nebula or the ρ-Ophiuchi molecular cloud but also for the formation of stars of the first or second generation in the universe. <br><br> In this thesis we show how the angular momentum of prestellar and protostellar cores evolves and compare our results with observed quantities. The specific angular momentum of prestellar cores in our models agree remarkably well with observations of cloud cores. Some prestellar cores go into collapse to build up stars and stellar systems. The resulting protostellar objects have specific angular momenta that fall into the range of observed binaries. We find that collapse induced by gravoturbulent fragmentation is accompanied by a substantial loss of specific angular momentum. This eases the "angular momentum problem" in star formation even in the absence of magnetic fields. <br><br> The distribution of stellar masses at birth (the initial mass function, IMF) is another aspect that any theory of star formation must explain. We focus on the influence of the thermodynamic properties of star-forming gas and address this issue by studying the effects of a piecewise polytropic equation of state on the formation of stellar clusters. We increase the polytropic exponent γ from a value below unity to a value above unity at a certain critical density. The change of the thermodynamic state at the critical density selects a characteristic mass scale for fragmentation, which we relate to the peak of the IMF observed in the solar neighborhood. Our investigation generally supports the idea that the distribution of stellar masses depends mainly on the thermodynamic state of the gas. <br><br> A common assumption is that the chemical evolution of the star-forming gas can be decoupled from its dynamical evolution, with the former never affecting the latter. Although justified in some circumstances, this assumption is not true in every case. In particular, in low-metallicity gas the timescales for reaching the chemical equilibrium are comparable or larger than the dynamical timescales. <br><br> In this thesis we take a first approach to combine a chemical network with a hydrodynamical code in order to study the influence of low levels of metal enrichment on the cooling and collapse of ionized gas in small protogalactic halos. Our initial conditions represent protogalaxies forming within a fossil HII region -- a previously ionized HII region which has not yet had time to cool and recombine. We show that in these regions, H₂ is the dominant and most effective coolant, and that it is the amount of H₂ formed that controls whether or not the gas can collapse and form stars. For metallicities Z <= 10^-3 Z_sun, metal line cooling alters the density and temperature evolution of the gas by less than 1% compared to the metal-free case at densities below 1 cm^-3 and temperatures above 2000 K. We also find that an external ultraviolet background delays or suppresses the cooling and collapse of the gas regardless of whether it is metal-enriched or not. Finally, we study the dependence of this process on redshift and mass of the dark matter halo. / Sterne sind fundamentale Bestandteile des Kosmos. Sie entstehen im Inneren von turbulenten Molekülwolken, die aus molekularem Wasserstoffgas und Staub bestehen. Durch konvergente Strömungen in der turbulenten Wolke bilden sich lokale Dichtemaxima, die kollabieren, falls die zum Zentrum der Wolke gerichtete Schwerkraft über die nach außen gerichteten Druckkräfte dominiert. Dies ist der Fall, wenn die Masse des Gases einen kritischen Wert überschreitet, der Jeansmasse genannt wird. Die Jeansmasse hängt von der Dichte und der Temperatur des Gases ab und fällt im isothermen Fall mit steigender Dichte stetig ab, so dass während des Kontraktionsprozesses immer kleinere Teilmassen instabil werden. Es kommt zur Fragmentierung der Molekülwolke zu protostellaren Kernen, den direkten Vorläufern der Sterne. <br><br> In der vorliegenden Arbeit werden die zeitliche Entwicklung des Drehimpulses der protostellaren Kerne und der Einfluss der thermischen Eigenschaften des Gases mit Hilfe von dreidimensionalen hydrodynamischen Simulationen untersucht. Hierbei konzentrieren wir uns auf zwei fundamentale Probleme, die jede Theorie der Sternentstehung lösen muss: das "Drehimpulsproblem" und die Massenverteilung der Sterne (IMF). Die thermischen Eigenschaften des Gases sind nicht nur von Bedeutung für die derzeitige Sternentstehung in beobachtbaren Regionen wie z.B. der Orionnebel oder die ρ-Ophiuchi Molekülwolke, sondern auch für die Entstehung von Sternen der ersten und zweiten Generation im frühen Universum. <br><br> Wir betrachten die Entwicklung des spezifischen Drehimpulses von protostellaren Kernen und vergleichen unsere Resultate mit beobachteten Werten. Wir finden eine gute Übereinstimmung zwischen den spezifischen Drehimpulsen der protostellaren Kerne in unserem Model und denen der beobachteten Kerne in Molekülwolken. In unseren Simulationen geht der gravitative Kollaps mit einem Verlust an spezifischem Drehimpuls einher. Somit kann das Drehimpulsproblem der Sternentstehung auch ohne Betrachtung der Magnetfelder entschärft werden. <br><br> Ein weiterer Schwerpunkt der Arbeit ist die Untersuchung des Einflusses der thermodynamischen Eigenschaften des Gases auf die Massenverteilung der Sterne, die aus diesem Gas entstehen. Wir verwenden eine stückweise polytrope Zustandgleichung, die die Temperatur-Dichte-Beziehung genauer beschreibt. Wir zeigen, dass Veränderungen in der Zustandgleichung bei einer bestimmten Dichte einen direkten Einfluss auf die charakteristische Massenskala der Fragmentierung haben und somit den Scheitelpunkt der Sternmassenverteilung in der solaren Umgebung bestimmen. <br><br> Des Weiteren sind die thermodynamischen Eigenschaften des Gases auch für die Sternentstehung im frühen Universum von Bedeutung. Das primordiale Gas, aus dem die ersten Sterne gebildet wurden, enthält keine Metalle (Elemente schwerer als H oder He), da diese erst durch Kernreaktionen in Sternen gebildet werden. In dieser Arbeit untersuchen wir den Einfluss einer geringen Metallizität auf das Kühlungs- und Kollapsverhalten von Gas, aus welchem die zweite Generation von Sternen entstanden ist. Dieses Gas ist anfänglich heiß und ionisiert und befindet sich in kleinen protogalaktischen Halos aus dunkler Materie. Unsere hydrodynamischen Simulationen, die auch ein adäquates chemisches Netzwerk beinhalten, zeigen, dass die Temperatur- und Dichteentwicklung des Gases während der Anfangsphase des Kollapses durch eine geringe Metallizität im Gas kaum beeinflusst wird. Wir stellen weiterhin fest, dass externe ultraviolette Strahlung den Kühlprozess des Gases ohne Metallizität und des Gases mit geringer Metallizität gleichermaßen verzögert oder sogar verhindert. Außerdem untersuchen wir den Einfluss der Rotverschiebung und der Masse des Halos aus dunkler Materie auf die Kühlung und den Kollaps des Gases.

L'objectif de cette thèse est de comprendre la structure des binaires gamma, binaires à collision de vents composées d'une étoile massive et d'un pulsar jeune. Ces binaires possèdent probablement une structure similaire aux binaires à collision de vents composées de deux étoiles massives, avec des particularités liées à la nature relativiste du vent de pulsar. L'interaction de deux vents supersoniques d'étoiles massives crée une structure choquée qui présente des signatures observationnelles du domaine radio aux rayons X. Plusieurs instabilités ainsi que le mouvement orbital des étoiles influent sur la structure choquée. Afin de comprendre leur impact, j'ai effectué des simulations à haute résolution de binaires à collision de vents à l'aide du code hydrodynamique RAMSES. Ces simulations sont numériquement coûteuses à réaliser, surtout lorsque un des vents domine fortement l'autre. A petite échelle, les simulations soulignent l'importance de l'instabilité de couche mince non-linéaire dans les collisions isothermes alors que l'instabilité de Kelvin-Helmholtz peut fortement modifier la structure choquée dans une collision adiabatique. A plus grande échelle, cette instabilité peut parfois détruire la structure spirale à laquelle on s'attend si la différence de vitesse entre les vents est trop importante. WR 104 est une binaire dont on observe la structure spirale grâce à l'émission de poussières. Les simulations de ce système montrent un bon accord avec la structure observée et indiquent que des processus de refroidissement du gaz sont nécessaires à la formation de poussières. Pour modéliser les vents de pulsar dans les binaires gamma, RAMSES a été étendu à l'hydrodynamique relativiste. J'utilise ce nouveau code pour réaliser des simulations préliminaires de binaires gamma. Elles montrent effectivement une structure similaire aux binaires stellaires, avec de légères corrections relativistes . Ce code est adapté à l'étude de divers systèmes astrophysiques tels que les jets relativistes, les sursauts gamma ou les nébuleuses de pulsar et fera partie de la prochaine version de RAMSES qui sera rendue publique. / The aim of this thesis is to understand the structure of colliding wind binaries composed of a massive star and a young pulsar, called gamma-ray binaries. They are expected to display a similar structure to colliding wind binaries composed of massive stars, with some particularities due to the relativistic nature of the pulsar wind. The interaction of the supersonic winds from massive stars creates a shocked structure with observational signatures from the radio domain to the X-rays. The structure is affected by various instabilities and by the orbital motion of the stars. To understand their impact, I carried out high resolution simulations of colliding wind binaries with the hydrodynamical code RAMSES. They are computationally demanding, especially when one of the winds strongly dominates the other one. Small scale simulations highlight the importance of the Non-linear Thin Shell Instability in isothermal collisions while the Kelvin-Helmholtz instability may strongly impact the dynamics of adiabatic collisions. I found that, at larger scales, this instability can destroy the expected large scale spiral structure when there is an important velocity gradient between the winds. WR 104 is a system that displays a spiral structure with important dust emission. The simulation of this system shows a good agreement with the observed structure and indicates cooling processes are necessary to enable dust formation. To model the pulsar wind in gamma-ray binaries, an extension of RAMSES has been developed, that incorporates relativistic hydrodynamics. I used this new relativistic code to perform preliminary simulations of gamma-ray binaries. They display a similar structure to colliding wind binaries with small relativistic corrections. We expect to use this code to perform large scale simulations of gamma-ray binaries. It will be part of the next public release of RAMSES and is suited for the study of many astrophysical problems such as relativistic jets, pulsar wind nebulae or gamma-ray bursts.

Orientador: Wei-Liang Qian / Resumo: O sucesso da descrição hidrodinâmica das colisões de íons pesados relativísticos desempenha um papel vital para entender as propriedades da matéria QCD. A essência da evolução hidrodinâmica, em geral, foi atribuída à resposta dinâmica às condições iniciais flutuantes. Em particular, as características observadas nas correlações de duas partículas, referidas como ``cume'' e ``ombro'', mostraram ser reproduzidas com sucesso por simulações hidrodinâmicas com condições iniciais flutuantes evento a evento, mas não por condições iniciais médias. Posteriormente, leva ao entendimento atual, através de extensos estudos de análise hidrodinâmica/transporte baseada em eventos por eventos, que as correlações de duas partículas para o momento transversal inferior podem ser interpretadas principalmente em termos de harmônicos de fluxo $ v_n $. Notavelmente, o fluxo triangular, $ v_3 $, é atribuído principalmente à aparência da estrutura do ``ombro'' no lado externo da partícula acionadora. Além disso, entende-se que esses coeficientes harmônicos estão intimamente associados aos correspondentes $ \varepsilon_n $, as anisotropias da distribuição inicial de energia. No entanto, a linearidade entre $ v_n $ e $ \varepsilon_n $ se torna menos evidente para harmônicos maiores que $ n = 2 $. Isso sugere que as próprias flutuações de evento a evento carregam informações importantes, além da linearidade observada. Se alguém se restringe apenas à análise das relações/correlações médias de eventos entr... (Resumo completo, clicar acesso eletrônico abaixo) / Doutor

Ventos (em inglês outflows) de ampla abertura e larga escala sâo uma característica comum em galáxias ativas, como as galáxias Seyfert. Em sistemas como este, onde buracos negros supermassivos (em inglês super massive black holes, SMBHs) de núcleos galácticos ativos de galáxias (em inglês active galactic nuclei, AGN) coexistem com regiões de formação estelar (em inglês star forming, SF), nâo está claro das observações se o AGN SMBH ou o SF (ou ambos) são responsaveis pela indução desses ventos. Neste trabalho, estudamos como ambos podem influenciar a evolução da galáxia hospedeira e seus outflows, considerando galáxias tipo Seyfert nas escalas de kilo-parsec (kpc). Para este objetivo, estendemos o trabalho anterior desenvolvido por Melioli & de Gouveia Dal Pino (2015), que considerou ventos puramente hidrodinâmicos impulsionados tanto pela SF quanto pelo AGN, mas levando em conta para este último apenas ventos bem estreitos (colimados). A fim de obter uma melhor compreensão da influencia (feedback) desses mecanismos sobre a evolução da galáxia e seus outflows, incluímos também os efeitos de ventos de AGN com maior ângulo de abertura, já que ventos em forma de cone podem melhorar a interação com o meio interestelar da galáxia e assim, empurrar mais gás nos outflows. Além disso, incluímos também os efeitos dos campos magnéticos no vento, já que estes podem, potencialmente, ajudar a preservar as estruturas e acelerar os outflows. Realizamos simulações tridimensionais magneto-hidrodinâmicas (MHD) considerando o resfriamento radiativo em equilíbrio de ionização e os efeitos dos ventos do AGN com dois diferentes ângulos de abertura (0º e 10º) e razões entre a pressão térmica e a pressão magnética beta=infinito, = 300 e 30, correspondentes a campos magnéticos 0, 0,76 micro-Gauss e 2,4 micro-Gauss respectivamente. Os resultados de nossas simulações mostram que os ventos impulsionados pelos produtos de SF (isto é, pelas explosões de supernovas, SNe) podem direcionar ventos com velocidades 100-1000 km s¹, taxas de perda de massa da ordem de 50 Massas solares/ano, densidades de ~1-10 cm-3 e temperaturas entre 10 e 10 K, que se assemelham às propriedades dos denominados absorvedores de calor (em inglês warm absorbers, WAs) e também são compatíveis com as velocidades dos outflows moleculares observadas. No entanto, as densidades obtidas nas simulações são muito pequenas e as temperaturas são muito grandes para explicar os valores observados nos outflows moleculares (que têm n ~150-300 cm³ e T<1000 K). Ventos colimados de AGN (sem a presença de ventos SF) também são incapazes de conduzir outflows, mas podem acelerar estruturas a velocidades muito altas, da ordem de ~10.000 km s¹ e temperaturas T> 10 K, tal como observado em ventos ultra rapidos (em inglês, ultra-fast outflows, UFOs). A introdução do vento de AGN, particularmente com um grande ângulo de abertura, causa a formação de estruturas semelhantes a fontes galácticas. Isso faz com que parte do gás em expansão (que está sendo empurrado pelo vento de SF) retorne para a galáxia, produzindo um feedback \'positivo\' na evolução da galáxia hospedeira. Descobrimos que esses efeitos são mais pronunciados na presença de campos magnéticos, devido à ação de forças magnéticas extras pelo vento AGN, o qual intensifica o efeito de retorno do gás (fallback), e ao mesmo tempo reduz a taxa de perda de massa nos outflows por fatores de até 10. Além disso, a presença de um vento de AGN colimado (0º) causa uma remoção significativa da massa do núcleo da galáxia em poucos 100.000 anos, mas este é logo reabastecido pelo de gás acretante proveniente do meio interestelar (ISM) à medida que as explosões de SNe se sucedem. Por outro lado, um vento de AGN com um grande ângulo de abertura, em presença de campos magnéticos, remove o gás nuclear inteiramente em alguns 100.000 anos e não permite o reabastecimento posterior pelo ISM. Portanto, extingue a acreção de combustível e de massa no SMBH. Isso indica que o ciclo de trabalho desses outflows é de cerca de alguns 100.000 anos, compatível com as escalas de tempo inferidas para os UFOs e outflows moleculares observados. Em resumo, os modelos que incluem ventos de AGN com um ângulo de abertura maior e campos magnéticos, levam a velocidades médias muito maiores que os modelos sem vento de AGN, e também permitem que mais gás seja acelerado para velocidades máximas em torno de ~10 km s¹, com densidades e temperaturas compatíveis com aquelas observadas em UFOs. No entanto, as estruturas com velocidades intermediárias de vários ~100 km s¹ e densidades até uns poucos 100 cm³, que de fato poderiam reproduzir os outflows moleculares observados, têm temperaturas que são muito grandes para explicar as características observadas nos outflows moleculares, que tem temperaturas T< 1000 K. Além disso, estes ventos de AGN não colimados em presença de campos magnéticos entre T< 1000 K. Alem disso, estes grandes ventos AGN de angulo de abertura em fluxos magnetizados reduzem as taxas de perda de massa dos outflows para valores menores que aqueles observados tanto em outflows moleculares quanto em UFOs. Em trabalhos futuros, pretendemos estender o espaço paramétrico aqui investigado e também incluir novos ingredientes em nossos modelos, como o resfriamento radioativo fora do equilíbrio, a fim de tentar reproduzir as características acima que não foram explicadas pelo modelo atual. / Large-scale broad outflows are a common feature in active galaxies, like Seyfert galaxies. In systems like this, where supermassive black hole (SMBH) active galactic nuclei (AGN) coexist with star-forming (SF) regions it is unclear from the observations if the SMBH AGN or the SF (or both) are driving these outflows. In this work, we have studied how both may influence the evolution of the host galaxy and its outflows, considering Seyfert-like galaxies at kilo-parsec (kpc) scales. For this aim, we have extended previous work developed by Melioli & de Gouveia Dal Pino (2015), who considered purely hydrodynamical outflows driven by both SF and AGN, but considering for the latter only very narrow (collimated) winds. In order to achieve a better understanding of the feedback of these mechanisms on the galaxy evolution and its outflows, here we have included the effects of AGN winds with a larger opening angle too, since conic-shaped winds can improve the interaction with the interstellar medium of the galaxy and thus push more gas into the outflows. Besides, we have also included the effects of magnetic fields in the flow, since these can potentially help to preserve the structures and speed up the outflows. We have performed three-dimensional magneto-hydrodynamical (MHD) simulations considering equilibrium radiative cooling and the effects of AGN-winds with two different opening angles (0º and 10º), and thermal pressure to magnetic pressure ratios of beta=infinite, 300 and 30 corresponding to magnetic fields 0, 0.76 micro-Gauss and 2.4 micro-Gauss, respectively. The results of our simulations show that the winds driven by the products of SF (i.e., by explosions of supernovae, SNe) alone can drive outflows with velocities ~100-1000 km s¹, mass outflow rates of the order of 50 Solar Masses yr¹, densities of ~1-10 cm³, and temperatures between 10 and 10 K, which resemble the properties of warm absorbers (WAs) and are also compatible with the velocities of the observed molecular outflows. However, the obtained densities from the simulations are too small and the temperatures too large to explain the observed values in molecular outflows (which have n ~ 150-300 cm³ and T<1000 K). Collimated AGN winds alone (without the presence of SF-winds) are also unable to drive hese outflows, but they can accelerate structures to very high speeds, of the order of ~ 10.000 km s¹, and temperatures T> 10 K as observed in ultra-fast outflows (UFOs). The introduction of an AGN wind, particularly with a large opening angle, causes the formation of fountain-like structures. This makes part of the expanding gas (pushed by the SF-wind) to fallback into the galaxy producing a \'positive\' feedback on the host galaxy evolution. We have found that these effects are more pronounced in presence of magnetic fields, due to the action of extra magnetic forces by the AGN wind producing enhanced fallback that reduces the mass loss rate in the outflows by factors up to 10. Furthermore, the presence of a collimated AGN wind (0º) causes a significant removal of mass from the core region in a few 100.000 yr, but this is soon replenished by gas inflow from the interstellar medium (ISM) when the SNe explosions fully develop. On the other hand, an AGN wind with a large opening angle in presence of magnetic fields is able to remove the nuclear gas entirely within a few 100.000 yr and does not allow for later replenishment. Therefore, it quenches the fueling and mass accretion onto the SMBH. This indicates that the duty cycle of these outflows is around a few 100.000 yr, compatible with the time-scales inferred for the observed UFOs and molecular outflows. In summary, models that include AGN winds with a larger opening angle and magnetic fields, lead to to be accelerated to maximum velocities around 10 km s¹ (than models with collimated AGN winds), with densities and temperatures which are compatible with those observed in UFOs. However, the structures with intermediate velocities of several ~100 km s¹ and densities up to a few 100 cm3, that in fact could reproduce the observed molecular outflows, have temperatures which are too large to explain the observed molecular features, which have temperatures T<1000 K. Besides, these large opening angle AGN winds in magnetized flows reduce the mass loss rates of the outflows to values smaller than those observed both in molecular outflows and UFOs. In future work, we intend to extend the parametric space here investigated and also include new ingredients in our models, such as non-equilibrium radiative cooling, in order to try to reproduce the features above that were not explained by the current model.

Ventos (em inglês outflows) de ampla abertura e larga escala sâo uma característica comum em galáxias ativas, como as galáxias Seyfert. Em sistemas como este, onde buracos negros supermassivos (em inglês super massive black holes, SMBHs) de núcleos galácticos ativos de galáxias (em inglês active galactic nuclei, AGN) coexistem com regiões de formação estelar (em inglês star forming, SF), nâo está claro das observações se o AGN SMBH ou o SF (ou ambos) são responsaveis pela indução desses ventos. Neste trabalho, estudamos como ambos podem influenciar a evolução da galáxia hospedeira e seus outflows, considerando galáxias tipo Seyfert nas escalas de kilo-parsec (kpc). Para este objetivo, estendemos o trabalho anterior desenvolvido por Melioli & de Gouveia Dal Pino (2015), que considerou ventos puramente hidrodinâmicos impulsionados tanto pela SF quanto pelo AGN, mas levando em conta para este último apenas ventos bem estreitos (colimados). A fim de obter uma melhor compreensão da influencia (feedback) desses mecanismos sobre a evolução da galáxia e seus outflows, incluímos também os efeitos de ventos de AGN com maior ângulo de abertura, já que ventos em forma de cone podem melhorar a interação com o meio interestelar da galáxia e assim, empurrar mais gás nos outflows. Além disso, incluímos também os efeitos dos campos magnéticos no vento, já que estes podem, potencialmente, ajudar a preservar as estruturas e acelerar os outflows. Realizamos simulações tridimensionais magneto-hidrodinâmicas (MHD) considerando o resfriamento radiativo em equilíbrio de ionização e os efeitos dos ventos do AGN com dois diferentes ângulos de abertura (0º e 10º) e razões entre a pressão térmica e a pressão magnética beta=infinito, = 300 e 30, correspondentes a campos magnéticos 0, 0,76 micro-Gauss e 2,4 micro-Gauss respectivamente. Os resultados de nossas simulações mostram que os ventos impulsionados pelos produtos de SF (isto é, pelas explosões de supernovas, SNe) podem direcionar ventos com velocidades 100-1000 km s¹, taxas de perda de massa da ordem de 50 Massas solares/ano, densidades de ~1-10 cm-3 e temperaturas entre 10 e 10 K, que se assemelham às propriedades dos denominados absorvedores de calor (em inglês warm absorbers, WAs) e também são compatíveis com as velocidades dos outflows moleculares observadas. No entanto, as densidades obtidas nas simulações são muito pequenas e as temperaturas são muito grandes para explicar os valores observados nos outflows moleculares (que têm n ~150-300 cm³ e T<1000 K). Ventos colimados de AGN (sem a presença de ventos SF) também são incapazes de conduzir outflows, mas podem acelerar estruturas a velocidades muito altas, da ordem de ~10.000 km s¹ e temperaturas T> 10 K, tal como observado em ventos ultra rapidos (em inglês, ultra-fast outflows, UFOs). A introdução do vento de AGN, particularmente com um grande ângulo de abertura, causa a formação de estruturas semelhantes a fontes galácticas. Isso faz com que parte do gás em expansão (que está sendo empurrado pelo vento de SF) retorne para a galáxia, produzindo um feedback \'positivo\' na evolução da galáxia hospedeira. Descobrimos que esses efeitos são mais pronunciados na presença de campos magnéticos, devido à ação de forças magnéticas extras pelo vento AGN, o qual intensifica o efeito de retorno do gás (fallback), e ao mesmo tempo reduz a taxa de perda de massa nos outflows por fatores de até 10. Além disso, a presença de um vento de AGN colimado (0º) causa uma remoção significativa da massa do núcleo da galáxia em poucos 100.000 anos, mas este é logo reabastecido pelo de gás acretante proveniente do meio interestelar (ISM) à medida que as explosões de SNe se sucedem. Por outro lado, um vento de AGN com um grande ângulo de abertura, em presença de campos magnéticos, remove o gás nuclear inteiramente em alguns 100.000 anos e não permite o reabastecimento posterior pelo ISM. Portanto, extingue a acreção de combustível e de massa no SMBH. Isso indica que o ciclo de trabalho desses outflows é de cerca de alguns 100.000 anos, compatível com as escalas de tempo inferidas para os UFOs e outflows moleculares observados. Em resumo, os modelos que incluem ventos de AGN com um ângulo de abertura maior e campos magnéticos, levam a velocidades médias muito maiores que os modelos sem vento de AGN, e também permitem que mais gás seja acelerado para velocidades máximas em torno de ~10 km s¹, com densidades e temperaturas compatíveis com aquelas observadas em UFOs. No entanto, as estruturas com velocidades intermediárias de vários ~100 km s¹ e densidades até uns poucos 100 cm³, que de fato poderiam reproduzir os outflows moleculares observados, têm temperaturas que são muito grandes para explicar as características observadas nos outflows moleculares, que tem temperaturas T< 1000 K. Além disso, estes ventos de AGN não colimados em presença de campos magnéticos entre T< 1000 K. Alem disso, estes grandes ventos AGN de angulo de abertura em fluxos magnetizados reduzem as taxas de perda de massa dos outflows para valores menores que aqueles observados tanto em outflows moleculares quanto em UFOs. Em trabalhos futuros, pretendemos estender o espaço paramétrico aqui investigado e também incluir novos ingredientes em nossos modelos, como o resfriamento radioativo fora do equilíbrio, a fim de tentar reproduzir as características acima que não foram explicadas pelo modelo atual. / Large-scale broad outflows are a common feature in active galaxies, like Seyfert galaxies. In systems like this, where supermassive black hole (SMBH) active galactic nuclei (AGN) coexist with star-forming (SF) regions it is unclear from the observations if the SMBH AGN or the SF (or both) are driving these outflows. In this work, we have studied how both may influence the evolution of the host galaxy and its outflows, considering Seyfert-like galaxies at kilo-parsec (kpc) scales. For this aim, we have extended previous work developed by Melioli & de Gouveia Dal Pino (2015), who considered purely hydrodynamical outflows driven by both SF and AGN, but considering for the latter only very narrow (collimated) winds. In order to achieve a better understanding of the feedback of these mechanisms on the galaxy evolution and its outflows, here we have included the effects of AGN winds with a larger opening angle too, since conic-shaped winds can improve the interaction with the interstellar medium of the galaxy and thus push more gas into the outflows. Besides, we have also included the effects of magnetic fields in the flow, since these can potentially help to preserve the structures and speed up the outflows. We have performed three-dimensional magneto-hydrodynamical (MHD) simulations considering equilibrium radiative cooling and the effects of AGN-winds with two different opening angles (0º and 10º), and thermal pressure to magnetic pressure ratios of beta=infinite, 300 and 30 corresponding to magnetic fields 0, 0.76 micro-Gauss and 2.4 micro-Gauss, respectively. The results of our simulations show that the winds driven by the products of SF (i.e., by explosions of supernovae, SNe) alone can drive outflows with velocities ~100-1000 km s¹, mass outflow rates of the order of 50 Solar Masses yr¹, densities of ~1-10 cm³, and temperatures between 10 and 10 K, which resemble the properties of warm absorbers (WAs) and are also compatible with the velocities of the observed molecular outflows. However, the obtained densities from the simulations are too small and the temperatures too large to explain the observed values in molecular outflows (which have n ~ 150-300 cm³ and T<1000 K). Collimated AGN winds alone (without the presence of SF-winds) are also unable to drive hese outflows, but they can accelerate structures to very high speeds, of the order of ~ 10.000 km s¹, and temperatures T> 10 K as observed in ultra-fast outflows (UFOs). The introduction of an AGN wind, particularly with a large opening angle, causes the formation of fountain-like structures. This makes part of the expanding gas (pushed by the SF-wind) to fallback into the galaxy producing a \'positive\' feedback on the host galaxy evolution. We have found that these effects are more pronounced in presence of magnetic fields, due to the action of extra magnetic forces by the AGN wind producing enhanced fallback that reduces the mass loss rate in the outflows by factors up to 10. Furthermore, the presence of a collimated AGN wind (0º) causes a significant removal of mass from the core region in a few 100.000 yr, but this is soon replenished by gas inflow from the interstellar medium (ISM) when the SNe explosions fully develop. On the other hand, an AGN wind with a large opening angle in presence of magnetic fields is able to remove the nuclear gas entirely within a few 100.000 yr and does not allow for later replenishment. Therefore, it quenches the fueling and mass accretion onto the SMBH. This indicates that the duty cycle of these outflows is around a few 100.000 yr, compatible with the time-scales inferred for the observed UFOs and molecular outflows. In summary, models that include AGN winds with a larger opening angle and magnetic fields, lead to to be accelerated to maximum velocities around 10 km s¹ (than models with collimated AGN winds), with densities and temperatures which are compatible with those observed in UFOs. However, the structures with intermediate velocities of several ~100 km s¹ and densities up to a few 100 cm3, that in fact could reproduce the observed molecular outflows, have temperatures which are too large to explain the observed molecular features, which have temperatures T<1000 K. Besides, these large opening angle AGN winds in magnetized flows reduce the mass loss rates of the outflows to values smaller than those observed both in molecular outflows and UFOs. In future work, we intend to extend the parametric space here investigated and also include new ingredients in our models, such as non-equilibrium radiative cooling, in order to try to reproduce the features above that were not explained by the current model.