Many young, massive stars are found in close binaries. Using population synthesis simulations. we predict the likelihood of a companion star being present when these massive stars end their lives as core-collapse supernovae (SNe). We focus on stripped-envelope SNe, whose progenitors have lost their outer hydrogen and possibly helium layers before explosion. We use these results to interpret new Hubble Space Telescope observations of the site of the broad-lined Type. Ic SN 2002ap, 14 years post-explosion. For a subsolar metallicity consistent with SN 2002ap, we expect a main-sequence (MS) companion present in about two thirds of all stripped-envelope SNe and a compact companion (likely a stripped helium star or a white dwarf/neutron star/black hole) in about 5% of cases. About a quarter of progenitors are single at explosion (originating from initially single stars, mergers, or disrupted systems). All of the latter scenarios require a massive progenitor, inconsistent with earlier studies of SN 2002ap. Our new, deeper upper limits exclude the presence of an MS companion star > 8-10 M., ruling out about 40% of all stripped-envelope SN channels. The most likely scenario for SN 2002ap includes nonconservative binary interaction of a primary star initially. 23 M.. Although unlikely (< 1% of the scenarios), we also discuss the possibility of an exotic reverse merger channel for broadlined Type. Ic events. Finally, we explore how our results depend on the metallicity and the model assumptions and discuss how additional searches for companions can constrain the physics that govern. the evolution of SN progenitors.

La formation des étoiles, processus fondamental en astrophysique, résiste toujours à la compréhension. En effet, de nombreux phénomènes interagissent durant les différentes étapes, et ce sur une large gamme d’échelles. Il est donc primordial de comprendre la dynamique du milieu interstellaire, dans lequel les étoiles se forment. En particulier, il est maintenant bien établi que la structure du milieu interstellaire est fortement impactée par des processus de rétroaction de la part des étoiles qui s’y forment. D’une part cette rétroaction limite le taux de formation de nouvelles étoiles, et d’autre part elle est l’un des contributeurs à la morphologie et la dynamique des galaxies : taille du disque, éjection de matière, etc. Ce travail de thèse propose d’étudier numériquement la dynamique du milieu interstellaire, de manière à mettre en évidence l’impact des processus de rétroaction. Le processus principal qui sera étudié est les supernovae, figurant parmi les évènements les plus énergétiques dans le milieu interstellaire. Après l’étude et la modélisation en détail de l’explosion d’une unique supernova dans nuage moléculaire, un modèle numérique incluant formation d’étoiles et rétroaction par supernovae sera présenté et mis en œuvre dans des simulations d’un disque galactique stratifié à l’échelle du kiloparsec. Une extension de ce modèle pour tenir compte du rayonnement ionisant sera proposée. Il est effectivement possible de réguler la formation d’étoiles à l’aide de modèles de rétroaction par les supernovae, mais les résultats précis dépendent de manière significative des détails du schéma mis en œuvre. En utilisant la variante apparaissant comme la plus réaliste, des simulations à haute résolution du milieu interstellaire sont présentées et étudiées. En particulier, il est possible de mettre en évidence des propriétés de la turbulence compressible et magnétisée à l’échelle galactique : variation des spectres de puissance en fonction de l’altitude, alignement spontané de la vitesse et du champ magnétique, effet antagoniste de la rétroaction sur cet alignement et formation de structures. / Star formation, a fundamental process in astrophysics, remains only partially understood. Several processes are known to interact during all the steps over a large range of scales. It is therefore of highest importance to understand the dynamics of the interstellar medium, in which stars form. In particular, it is now well-known that the structure of the interstellar medium is strongly affected by feedback processes emanating from the stars that form in it. On the one hand this feedback limits the rate of formation of new stars, and on the other hand it is one of the main contributors to the shape and dynamics of galaxies: thickness of the disk, matter outflows, etc. This work aims to study numerically the dynamics of the interstellar medium, in order to highlight the impact of stellar feedback processes. The main process that will be studied is supernovae, being among the most energetic events in the interstellar medium. After the study and detailed modeling of the explosion of a single supernova inside a molecular cloud, a numerical model including star formation and supernova feedback will be presented and used in kiloparsec-scale simulations of a stratified galactic disk. An extension of this model will be suggested in order to take into account the ionizing radiation. It is indeed possible to regulate star formation with supernova feedback models, although the precise results strongly depend on the detailed scheme that is implemented. Using the most realistic-looking variant, high-resolution simulations are presented and studied. In particular, it is possible to extract properties of compressible and magnetized turbulence at the galactic scale: variation of the power spectra as a function of altitude, spontaneous alignment between velocity and magnetic field, antagonistic effect of stellar feedback onto this alignment, and structure formation.

Les recherches de SNe ont commencé il y a plus de 100 ans. Depuis, il a été possible de collecter beaucoup de données d'observations astronomiques. Les astronomes ont développé une classification détaillée et ont abouti un relatif consensus sur la nature physique de ces événements très différents. Néanmoins, beaucoup de questions restent sans réponse. En résumé, les supernovæ de type II (riche en hydrogène) résultent de l'éjection l'enveloppe des supergéantes rouges (SGR). Les principales sources de connaissance sur ces objets sont l'évolution de leur luminosité en fonction du temps (`courbes de lumière') et leurs spectres observés à différentes époques. La méthode la plus utilisée pour extraire les informations des données d'observation est la modélisation des courbes de lumières et des spectres des supernovæ. Dans le Chapitre 1 (Introduction), nous présentons successivement l’évolution stellaire, la physique des explosions et l’évolution des éjectas. Nous décrivons aussi les différents types de supernova ; l’état actuel des connaissances sur les CCSNe ainsi que ces limitations. Nous discutons de la théorie de l'évolution stellaire. Nous décrivons notre approche numérique au Chapitre 2 (Supernova modelling). Elle consiste en trois étapes principales : la modélisation de l'évolution stellaire, l'explosion de l'étoile SGR résultante, et la modélisation de l'évolution des éjectas. Nous présentons la structure modélisée des étoiles SGR ; ces modèles et techniques de calcul sont similaires aux modèles utilisés dans les chapitres suivants. Nous discutons notre méthode d'explosion d'un progéniteur quand son noyau dégénéré commence à s'effondrer. Dans le Chapitre 3 (Observational properties), nous discutons les propriétés observées en photométrie et spectroscopie des CCSNe. Nous extrayons les propriétés statistiques de l'échantillon existant. En utilisant la technique présentée, nous avons effectué une étude détaillée de SN 2008bk, une supernova bien observée (Chapitre 4). Nous pouvons contraindre les propriétés du progéniteur et des paramètres d'explosion. Notre modélisation nous permet de comparer non seulement les propriétés de base telles que la luminosité, mais aussi à analyser en détail les caractéristiques spectrales, telles que la présence de certaines raies spectrales ainsi que leur morphologie. Nous montrons qu'une étoile de 12M⊙ sur la séquence principale est un bon candidat au progéniteur de SN 2008bk. Aussi, nous discutons de la forme asymétrique de la raie Hα et concluons qu'elle provient du chevauchement avec la raie forte du Ba II 6596.9 Å. SN 2008bk, avec quelques dizaines d'autres objets, forme une sous-classe importante de CCSNe — supernovae à faible luminosité. Nous avons consacré une attention particulière à cette classe d'objets, dont l'uniformité et les données d'observation nous permet de tirer des conclusions importantes. Au Chapitre 5, nous étudions l'échantillon de 17 SNe de faible luminosité et analysons la possibilité que ces événements résultent d'explosions de progéniteurs de petite ou de grande masse. Il n'y a pas d'accord solide dans la communauté astronomique sur les progéniteurs possibles des explosions de SNe à faible luminosité. Notre analyse montre que les étoiles massives de masse inférieure (~12 M⊙) sont de bons candidats pour les progéniteurs de cette sous-classe de SNe. De plus, nos simulations d'étoiles de masse élevée (25 et 27 M⊙) montrent qu'une explosion ayant une luminosité aussi faible aurait des propriétés d’observation remarquables qui ne sont pas présentes dans les données. Dans le Chapitre 6, nous étendons notre étude sur toutes les CCSNe, en utilisant des modèles plus énergétiques que dans les Chapitres 4 et 5. Nous fournissons des preuves que ce qui différencie la diversité de SNe II est l'énergie d'explosion plutôt que la masse des éjectas (ou plus précisément la masse de l'enveloppe riche en H de progéniteur). / Dedicated SN searches started over 100 years ago. Over that time, astronomers have collected large sets of observational data. They have developed detailed classification and achieved general agreement on the nature of these events. Nevertheless, a lot of questions remain unanswered. In short, most Type II SNe (hydrogen-rich SNe) are terminal explosions of red supergiant (RSG) stars. The main source of knowledge about these objects are the way their luminosity changes with time (`light curves') and how their radiation is distributed in wavelength. One of the widely used methods to extract the information from the observational data is computer modelling. The largest part of our work lays in the numerical simulations. In Chapter 1 (Introduction), we present succinctly the necessary theory which includes stellar evolution, explosion physics and ejecta evolution. We discuss different types of SNe; the modern knowledge on CCSNe and its problems. We discuss stellar evolution theory. We describe the nucleosynthesis that takes place in the cores of massive stars and gives rise to their final chemical stratification. We describe our numerical approach in Chapter 2 (Supernova modelling). It includes three major steps: stellar evolution modelling, explosion of the resulting RSG star, and ejecta evolution modelling. We present modelled structure of RSG stars; these models and computational techniques are similar to models used in subsequent chapters. We then discuss our numerical methods of exploding a SN once its degenerate core starts collapsing. We discuss explosive nucleosynthesis and its impact on the progenitor composition, production of unstable isotopes and the basic physics of radioactive decay. In Chapter 3 (Observational properties), we discuss the photometric and spectral observational properties of core-collapse SNe. We extract statistical properties of the existing sample. Using the presented technique, we performed a detailed study of the well observed object SN 2008bk (Chapter 4). We are able to constrain its progenitor and explosion properties. Our modelling allows us to compare not only the basic properties such as luminosity, but also to analyze in detail the spectral features, such as line identification and morphology. We show that a 12 M⊙ star on the main sequence is a good candidate for the progenitor of SN 2008bk. Also we discuss the asymmetric shape of the Hα line and conclude that it stems from the overlap with the strong Ba II 6596.9 Å line. SN 2008bk, together with about 20 objects, form a subclass of low-luminosity CCSNe Type II. We devoted a particular attention to this class of objects, whose uniformity and observational data allows us to draw important conclusions. In Chapter 5 (Low-luminosity Type II-P SNe), we study the sample of 17 low-luminosity SNe and analyze the possibility that these events are the result of explosions of low- and high-mass progenitors. There is no solid agreement in the astronomical community on the possible progenitors of the low-luminosity explosions of Type II SNe. Our analysis shows that lower-mass massive stars (~12 M⊙) are good candidates for the progenitors of this subclass of SNe. Moreover, our simulations of high-mass stars (25 and 27 M⊙) show that such low brightness of the explosion of such a massive object would have notable observational properties which are not present in the data. In Chapter 6 (Kinetic energy variation), we extend our study further on the whole class of hydrogen-rich core-collapse SN, using more energetic models than in Chapters 4 and 5. We provide evidences that what differentiates the diversity of SNe II is the explosion energy rather than the ejecta mass (or H-rich envelope mass of the progenitor).

An explanation of the emission line behavior in cataclysmic variables has been among the most important and elusive problems in eruptive star research. This work expands accretion disk chromosphere models of line emission to predict line behavior qualitatively. A search for UX UMa-like thick disk cataclysmics in the Palomar Green survey sample gave space densities consistent only with luminous high accretion rate disks: Ṁ ≥ 10⁻⁷·⁰Mₒ/yr. Instead, 20 composites whose combined energy distributions were "flat", Fᵥ α ν⁰, were discovered. These typically were early K dwarfs paired with 30,000K subdwarf O stars. The study also showed that a substantial fraction of subdwarf O and B stars are binary. The detached eclipsing binary BE UMa showed a reflection effect emission line spectrum due to the close 7Rₒ separation between an EUV-emitting subdwarf O star and an M1-5 dwarf. Analysis gave the hot star physical parameters as: 80,000K ≤ T(p)(K) ≤ 100,000, 7.0 ≤ log g(p) ≤ 8.0, and log (He/H) = 1.0 ± 1.0. The BE UMa optical emission line spectrum was modeled using a quantitative photoionization-recombination stellar atmosphere-like code. A rich high excitation continuum fluorescence and recombinational spectrum including HeII λ4686 and CIII λ4650 was formed at lower optical depths corresponding to nₑ ≤ 10¹³·⁵ cm⁻³ and Tₑ = 20,000K. The model suggests that T(p) = 100,000K. Cataclysmic variables too have a central source due to loss of half the accretion energy at the white dwarf surface. This temperature is no higher than the innermost disk regions; hence, mass accretion rates determine the character of the EUV radiation. Observations of 13 cataclysmics representing most types were obtained. From these data, the H, HeI, HeII, CaII, and high excitation metal emission line behavior in cataclysmics were analyzed. Cataclysmic variable accretion rates were shown to follow a sequence; from the low excitation dwarf novae [10⁻⁹·⁵ ≤ Ṁ (Mₒ/yr) ≤ 10⁻⁸·⁵] to the high excitation novae and nova-like systems [10⁻⁸·⁵ < Ṁ(Mₒ/yr) ≤ 10⁻⁶·⁵]. Predicted line profiles were consistent with observations. Thus, the model accounts well for the considered observations.

Modern wide-field imaging transient search projects led to the discovery of a new class of rare, over-luminous stellar explosions. Events like SN 2005ap (Quimby et al. 2007a), SN 2006gy (Quimby 2006; Smith et al. 2007), SN 2006tf (Quimby, Castro \& Mondol 2007; Quimby et al. 2007b; Smith et al. 2008), SN 2008am (Chatzopoulos et al. 2010), SN 2008es (Yuan et al. 2008; Gezari et al. 2008; Miller et al. 2008) SN 2008fz (Drake et al. 2009) and SN 2008iy (Miller et al. 2010) introduced us new evidence about stellar death, since traditional ideas about the mechanisms that can power these phenomena are found to be inadequate to explain the observed properties. The Texas Supernova Search Project (TSS; Quimby et al. 2005) and its successor, the ROTSE-Supernova Verification Project (RSVP; Yuan et al. 2007) discovered most of the above mentioned over-luminous supernovae (OLSNe) over the past five years of their operation. The advantage of this project is that it is essentially free of selection bias or the limits of a targeted search; the automated wide field (~3 square degrees) ROTSE-III telescopes (Akerlof et al. 2003), scan the whole sky blindly, looking for transients down to unfiltered magnitude ~19 mag and they do not focus on pre-selected galaxies. An estimated rate for these exceptionally luminous supernovae is ~ 2.6 10^{-7} events/Mpc^{3}/yr (Quimby et al. 2009a). Current and future transient search projects such as the SDSS-II Supernova Survey (Frieman et al. 2008),the Palomar Transient Factory (PTF; Law et al. 2009), SkyMapper (Schmidt et al. 2005), PanSTARRS (Chambers 2006) and Large Synoptic Survey Telescope (Tyson \& LSST collaboration 2002) are expected to increase the number of rare over-luminous (or, under-luminous) supernove and to discover new, unusual transients. / text

Using a set of high-resolution hydrodynamical simulations run with the Cholla. code, we investigate how mass and momentum couple to the multiphase components of galactic winds. The simulations model the interaction between a hot wind driven by supernova explosions and a cooler, denser cloud of interstellar or circumgalactic media. By resolving scales of Delta x < 0.1 pc over > 100 pc distances, our calculations capture how the cloud disruption leads to a distribution of densities and temperatures in the resulting multiphase outflow and quantify the mass and momentum associated with each phase. We find that the multiphase wind contains comparable mass and momenta in phases over a wide range of densities and temperatures extending from the hot wind (n approximate to 10(-2.5) cm(-3), T approximate to 10(6.5) K) to the coldest components (n approximate to 10(2) cm(-3), T approximate to 10(2) K). We further find that the momentum distributes roughly in proportion to the mass in each phase, and the mass loading of the hot phase by the destruction of cold, dense material is an efficient process. These results provide new insight into the physical origin of observed multiphase galactic outflows and inform galaxy formation models that include coarser treatments of galactic winds. Our results confirm that cool gas observed in outflows at large distances from the galaxy (greater than or similar to 1 kpc) likely does not originate through the entrainment of cold material near the central starburst.

The collapse of a stellar core is expected to produce gravitational waves (GWs), neutrinos, and in most cases a luminous supernova. Sometimes, however, the optical event could be significantly less luminous than a supernova and a direct collapse to a black hole, where the star just disappears, is possible. The GW event GW150914 was detected by the LIGO Virgo Collaboration via a burst analysis that gave localization contours enclosing the Large Magellanic Cloud (LMC). Shortly thereafter, we used DECam to observe 102 deg(2) of the localization area, including 38 deg(2) on the LMC for a missing supergiant search. We construct a complete catalog of LMC luminous red supergiants, the best candidates to undergo invisible core collapse, and collected catalogs of other candidates: less luminous red supergiants, yellow supergiants, blue supergiants, luminous blue variable stars, and Wolf-Rayet stars. Of the objects in the imaging region, all are recovered in the images. The timescale for stellar disappearance is set by the free-fall time, which is a function of the stellar radius. Our observations at 4 and 13 days after the event result in a search sensitive to objects of up to about 200 solar radii. We conclude that it is unlikely that GW150914 was caused by the core collapse of a relatively compact supergiant in the LMC, consistent with the LIGO Collaboration analyses of the gravitational waveform as best interpreted as a high mass binary black hole merger. We discuss how to generalize this search for future very nearby core-collapse candidates.

While the progenitors of Type Ia supernovae (SNe Ia) have long been thought to be thermonuclear explosions of white dwarf stars, what triggers the explosion are still a topic of debate. This thesis considers constraints on single-degenerate progenitors of SNe Ia based on the presence of a Roche-lobe filling companion. The ejecta strips material from the companion, that maybe detectable via Hα emission during the nebular phase. Using the full structure from simulations produces line widths are larger than those produced in simple models. The structure formed by the ejecta-companion interaction produce a broken reverse shock that may be visible in X-rays via the Fe K&alpha; line at the age of Tycho's supernova remnant (SNR). If the similar structures in Tycho’s SNR are produced this way then the companion star must have been massive, with M ~ 2 M_&odot;. Detections of circumstellar material within the supernova provides another way to indirectly probe the companion star. Mass loss through winds or novae are expected to shape the circumsteller medium for single-degenerate progenitors and the velocities, v ~ 100 km s^-1 appear to be consistent with recurrent nova shells, a model that is tested by analysing simulations of RS Ophiuchi. Models of RS Ophiuchi can explain the absorption lines seen around the 2006 outburst if the mass loss is 10⁻⁶ M_&odot; yr<sup>-1,/sup>. The circumsteller medium is shown to produce in the velocity and relative strengths of the features seen in SN 2006X. However, whether density in the shells is high enough to produce the required recombination timescale and to overcome ionization by &gamma;-rays for shells at 5 × 10¹⁶ cm remains uncertain.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

To constrain the era when the first galaxies and stars appeared upcoming instruments will rely on the brightest events in the universe: supernovae and brilliant emission from massive black holes. In this dissertation, we investigate the observability of certain types of supernovae of the very first stars (Population III stars) and find that while these events are sufficiently luminous to be observed with deep-sky instruments such as the James Webb Space Telescope (JWST), they may not observe these particular types of events in their lifetimes. We next explore the origins of massive black holes and introduce the direct collapse hypothesis of supermassive black hole formation. We model CR7, an apparently metal-free, luminous, Lyman-alpha emitting galaxy, as if it were powered by a massive direct collapse black hole and find that such a black hole can account for CR7's impressive Lyman-alpha flux. We finally investigate the nature of the connection between water megamasers, very bright radio sources originating from population inversion in dense, shocked gas around massive black holes and hydroxyl megamasers which generally accompany star formation. We carry out a ~ 60 hour radio survey for water emission among galaxies hosting OH megamaser hosts to assess the connection between the two types of emission. We find marginally statistically significant evidence that OH megamasers exclude water kilomasers and confirm with high levels of significance (> 8 sigma) the presence of a water megamaser in II Zw 96, establishing this object as the second galaxy known to cohost simultaneous water and hydroxyl megamasers.