Neste trabalho primeiramente estudamos os efeitos na propagação de ondas de Alfvén em plasmas empoeirados encontrados em inúmeros ambientes astrofísicos. A relação de dispersão da onda é modificada apresentando novos mecanismos de amortecimento. Há na literatura dados observacionais que indicam a existência de condições para crescimento de partículas de poeira na base da atmosfera de estrelas gigantes e supergigantes frias. Construímos um modelo de perda de massa para essas estrelas no qual mostramos que um fluxo de ondas de Alfvén, amortecido pela presença de poeira, pode contribuir para a geração de um vento de baixa velocidade e alta taxa de perda de massa, de acordo com as observações. Já no caso de estrelas quentes, mostramos como é possível obter as condições ideais para formação e crescimento destas partículas em um modelo de colisão de ventos em sistemas binários. A partir deste modelo, é possível explicar as altas emissões em raios-X observadas, além do crescimento de grãos no pós-choque. Aplicando a idéia ao sistema binário de eta Carinae, o modelo permite a determinação dos parâmetros orbitais do sistema. A poeira contida nos ventos estelares é então ejetada para o meio interestelar. Através de um cálculo semi-empírico determinamos a importância de cada intervalo de massa estelar, em cada etapa evolutiva, no retorno de material sólido ao MI. Em regiões de formação estelar investigamos como as ondas de Alfvén, amortecidas pela presença de poeira, influenciam a estabilidade de nuvens moleculares. Em oposição às teorias encontradas na literatura, mostramos que uma nuvem molecular anã, suportada apenas por pressão magnética, não pode ser dinamicamente estável. / In this work, we firstly discuss the propagation of Alfvén waves in dusty plasmas found in several astrophysical environments. The wave dispersion relation is modified giving rise to new damping mechanisms. There are in the literature observational data indicating the presence of dust near the surface of cool giant and supergiant stars. We developed a stellar mass loss model where we show that a flux of Alfvén waves, damped by the dust presence,can generate a low velocity and high mass loss rate wind, in agreement with the observations. In the case of hot stars we show how it is possible to obtain the special conditions for dust growth in a wind collision model of massive binary systems. For this model it is possible to explain both, the high X-rays emissions and dust growth at the post-shock phase. Applying the idea to the $eta$ Carinae binary system, the model allows the determination of the system orbital and the stellar wind parameters. The stellar wind´s dust is then ejected to the interstellar medium. In a semi-empirical calculation we showed the importance of each stellar mass, at each evolutionary phase, on the dust feedback of the ISM. For star formation regions, we investigate the role of Alfvén waves, damped by the dust presence, on the stability of the molecular clouds. In spite of the works found in the literature, we showed that a dwarf molecular cloud, magnetically supported only, is unstable.

The transit method, employed by Microvariability and Oscillation of Stars (MOST), Kepler, and various ground-based surveys has enabled the characterization of extrasolar planets to unprecedented precision. These results are precise enough to begin to measure planet atmosphere composition, planetary oblateness, starspots, and other phenomena at the level of a few hundred parts per million. However, these results depend on our understanding of stellar limb darkening, that is, the intensity distribution across the stellar disk that is sequentially blocked as the planet transits. Typically, stellar limb darkening is assumed to be a simple parameterization with two coefficients that are derived from stellar atmosphere models or fit directly. In this work, we revisit this assumption and compute synthetic planetary-transit light curves directly from model stellar atmosphere center-to-limb intensity variations (CLIVs) using the plane-parallel Atlas and spherically symmetric SAtlas codes. We compare these light curves to those constructed using best-fit limb-darkening parameterizations. We find that adopting parametric stellar limb-darkening laws leads to systematic differences from the more geometrically realistic model stellar atmosphere CLIV of about 50–100 ppm at the transit center and up to 300 ppm at ingress/egress. While these errors are small, they are systematic, and they appear to limit the precision necessary to measure secondary effects. Our results may also have a significant impact on transit spectra.

We have obtained a Chandra/HETGS spectrum of one of the most massive and luminous stars in the Galaxy: the blue hypergiant Cyg OB2-12 (HIP 101364, spectral type B3 Ia+). This is the first measurement at high resolution of X-ray spectral lines in a blue hypergiant and allows comparison of X-ray properties between massive stars at different but related evolutionary stages: O-type supergiants, luminous blue variables, Wolf-Rayet stars, and blue hypergiants stars. The new data provide a look at how the most massive stars shed mass during their pre-supernova evolution. We find that In Cyg OB2-12 the resolved Si and Mg lines are broadened by about 1000 km/s (FWHM). The lines, however, do not show appreciable centroid shifts (/s), which would be much larger for canonical moderately thick winds (~500 km/s). The He-like Mg XI lines show evidence of photo-excitation, implying a wind origin close to the UV-bright photosphere. The spectrum also indicates relatively high temperature plasma, up to 22 MK (1.9 keV), showing significant continuum and emission lines below 5A (above 2.5 keV). Hence, at first glance, the spectrum resembles neither an O-star thick wind, nor a magnetically confined (narrow-line) plasma. We will present more detailed wind models using both X-ray and UV spectra to constrain fundamental physical parameters of this star.

As WR 6 is a putatively single WN4 star, and is relatively bright (V = 6.9), it is an ideal case for studying the wind mechanisms in these extremely luminous stars. To obtain higher resolution spectra at higher energy (above 1 keV) than previously obtained with the XMM/Newton RGS, we have observed WR 6 with the Chandra High Energy Transmission Grating Spectrometer for 450 ks. We have resolved emission lines of S, Si, Mg, Ne, and Fe, which all show a “fin”-shaped profile, characteristic of a self-absorbed uniformly expanding shell. Steep blue edges gives robust maximal expansion velocities of about 2000 km/s, somewhat larger than the 1700 km/s derived from UV lines. The He-like lines all indicate that X-ray emitting plasmas are far from the photosphere – even at the higher energies where opacity is lowest – as was also the case for the longer wavelength lines observed with XMM-Newton/RGS. Abundances determined from X-ray spectral modeling indicate enhancements consistent with nucleosynthesis. The star was also variable in X-rays and in simultaneous optical photometry obtained with Chandra aspect camera, but not coherently with the optically known period of 3.765 days.

The nature of the edge-on eclipsing binary Epsilon Aurigae remains perplexing, despite notable progress since the recent 2009-2011 eclipse. The binary involves an early F supergiant with a still unknown companion enshrouded in a disk. Although the eclipse geometry produces a significant broad band polarization signature, semiregular pulsations of the F supergiant are also a source of variable polarization, with an amplitude that is commensurate with the effect of the eclipse. This fact makes use of the polarization for studying the disk of the companion far more challenging. In an effort to better understand the pulsation nature of the supergiant, we explore a simple model for the stellar contribution to the polarization signal. The model does reasonably well in characterizing the gross properties of the time-variable polarization.

Co-rotating Interaction Regions (CIRs) are a well-known phenomenon in the solar wind, and is a favored culprit for certain cyclical behavior observed in the spectra of some massive stars. A prime example are the discrete absorption components (DACs) seen in the UV wind lines of many O stars. Here we report on modeling for the variable continuum polarization that could arise from the presence of CIR structures. Considerations are limited to optically thin scattering. Using a core-halo approach for winds that are thick to electron scattering, an application to observed variable polarization of WR6 (EZ CMa; HD 50896) is presented.

Among the different types of massive stars in advanced evolutionary stages is the enigmatic WN8h type. There are only a few Wolf–Rayet (WR) stars with this spectral type in our Galaxy. It has long been suggested that WN8h-type stars are the products of binary evolution that may harbor neutron stars (NS). One of the most intriguing WN8h stars is the runaway WR 124 surrounded by its magnificent nebula M1-67. We test the presence of an accreting NS companion in WR 124 using ~100 ks long observations by the Chandra X-ray observatory. The hard X-ray emission from WR 124 with a luminosity of L X ~ 1031 erg s−1 is marginally detected. We use the non-local thermodynamic equilibrium stellar atmosphere code PoWR to estimate the WR wind opacity to the X-rays. The wind of a WN8-type star is effectively opaque for X-rays, hence the low X-ray luminosity of WR 124 does not rule out the presence of an embedded compact object. We suggest that, in general, high-opacity WR winds could prevent X-ray detections of embedded NS, and be an explanation for the apparent lack of WR+NS systems.

We utilize a Poisson likelihood in a maximum likelihood statistical analysis to analyze X-ray spectragraphic data. Specifically, we examine four extragalactic supernova remnants (SNR). IKT 5 (SNR 0047-73.5), IKT 25 (SNR 0104-72.3), and DEM S 128 (SNR 0103-72.4) which are designated as Type Ia in the literature due to their spectra and morphology. This is troublesome because of their asymmetry, a trait not usually associated with young Type Ia remnants. We present \emph{Chandra X-ray Observatory} data on these three remnants, and perform a maximum likelihood analysis on their spectra. We find that the X-ray emission is dominated by interactions with the interstellar medium. In spite of this, we find a significant Fe overabundance in all three remnants. Through examination of radio, optical, and infrared data, we conclude that these three remnants are likely not "classical" Type Ia SNR, but may be examples of so-called "prompt" Type Ia SNR. We detect potential point sources that may be members of the progenitor systems of both DEM S 128 and IKT 5, which could suggest a new subclass of prompt Type Ia SNR, Fe-rich CC remnants. In addition, we examine IKT 18. This remnant is positionally coincident with the X-ray point source HD 5980. Due to an outburst in 1994, in which its brightness changed by 3 magnitudes (corrsponding to an increase in luminosity by a factor of 16) HD 5980 was classified as a luminous blue variable star. We examine this point source and the remnant IKT 18 in the X-ray, and find that its non-thermal photon index has decreased from 2002 to 2013, corresponding to a larger proportion of more energetic X-rays, which is unexpected.

Interstellar dust plays an important role in the physics of the interstellar medium, as well as the formation and evolution of stars. The presence of dust is often indicated in optical images by dark lanes which bisect spiral galaxies, or seen directly as reflection nebulosity around stars or emission nebulosity if sufficient heating is present. Of interest in this thesis is the dust that is associated with bipolar outflow sources. Bipolar outflows can occur in either evolved stars or in young stellar objects, and are so named because they consist of two lobes which are thought to be due to out-flowing dust and gas, with a dark lane between them due to thick dust in a circumstellar disk or shell which often blocks the light from the central star. The spatial distribution of the properties of dust around bipolar outflow sources has been examined using a combination of theoretical and observational techniques. To aid the interpretation of observations of bipolar outflow sources, we have modelled the wavelength dependence of light from 0.36 to 22\um, scattered by dust particles with varying characteristics. The results were then presented in the form of colour excess ratios. These model ratios can be applied to observations if the contribution due to the central star is able to be removed, such that all that remains in the image is the effect of the dust particles. The scattering of light by dust particles was modelled by varying six different characteristics: grain material, size (particle radius from 0.002 to 0.75\um), mantle temperatures and thicknesses, shape, and orientation. Of those characteristics, the largest variation in the colour excess ratios resulted from varying grain composition and size. Different scattering angles also produce a noticeable variation in the colour excess ratios, however the effect is difficult to distinguish from the general extinction due to dust around the source. Water ice mantles were also found to significantly change the colour excess ratios. Grain shape and orientation produced only small variations in the colour excess ratios. Three bipolar outflow sources were studied as part of this thesis, two evolved objects, OH~231.8+4.2 and Mz\,3, and the young T-Tauri object Rno\,91. The observations involved multi-wavelength imaging in the infrared, from which colours and colour excess ratios were obtained at various points of the bipolar outflows and then compared to the predictions made in the modelling. The most extensive data set analysed was seven images of the object OH~231.8+4.2 which were used in a multi-wavelength study in the infrared H to N bands (1.25--12\um). The central source position of the object has been confined to less than an arcsec using the longer wavelength images and an L--M colour image. The two peaks which dominate the lobes in the shorter wavelength images were found to be scattering peaks where the light from the central source is scattered from the walls of the lobes. The spatial distribution of water ice in the nebula has also been constrained to the circumstellar disk which has a torus or disk shape rather than being a spherical shell. The colour excess ratios derived for the nebula from the images also suggest slightly different dust properties between the circumstellar disk, lobe walls and within the lobe cavities. The young T-Tauri star Rno\,91 also contains ice, and was observed between J and L. The central star which illuminates the nebula was shown to be coincident with the brightest point in these images. Using colour excess ratio results for various parts of the nebula, it was shown that the dust close to the central star is likely to contain larger grains than the diffuse ISM, but with a similar composition. Moving away from the central star, the dust becomes more like that observed in the diffuse ISM. The presence of water ice on dust close to the central star was confirmed using images centred in the ice band. The protoplanetary bipolar outflow source Mz\,3 is slightly more evolved than OH 231.8+4.2, and does not have evidence of any water ice in the circumstellar disk. Images of this object were obtained between J and 10\um. The presence of warm dust throughout the inner bipolar lobes of this object is noticeable by the brightness of the lobes in the image at 10\um. Line profiles through the position of the central source of the 10\um\ image demonstrate that there is a circumstellar shell close to the central source which has an inner radius of $\lta\,375$\,AU. Colour excess ratio results for the bipolar lobes suggest that the dust associated with Mz\,3 is generally smaller than that found in the diffuse ISM. The properties of the dust in the bipolar lobes were also observed to be different to the dust closer to the central source and lying in the circumstellar disk. The small sizes for dust in Mz\,3 is consistent with the high velocity outflows that have been associated with the object.

La formation des étoiles massives reste, à ce jour, encore mal connue à cause de l'extrême quantité d'énergie que ces étoiles dégagent, limitant en conséquence leurs masses théoriques et contredisant les observations de ce type d'étoile. Les observatoires du futur (en particulier l'observatoire spatial Herschel) vont tenter de répondre à cette problématique grâce notamment aux émissions moléculaires de l'eau. L'analyse précise et correcte de ces données, dans l'avenir, nécessite donc dès aujourd'hui un travail associant des observations et des modélisations des objets concernés. C'est dans ce but que cette thèse a consisté en l'élaboration d'une méthode de modélisation dite « globale » d'objets proto-stellaires massifs (proto-amas ou cœurs denses massifs). Celle-ci a permis une description physique et une étude chimique des multiples cœurs denses massifs étudiées, et a ouvert de nombreuses voies vers des aspects évolutifs. Elle a également donné des indices pour affiner le programme d'observation en temps garanti WISH des raies moléculaires de l'eau et confirmé le rôle clef de cette molécule pour la compréhension de la formation des étoiles massives.