BeAtlas: A grid of synthetic spectra for Be stars - Bayesian inference of the physical properties of B and Be Stars

• Ano de defesa: 2019
• Autor(a) principal: Mota, Bruno Correia
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Astrofísica Estelar; Astronomia; Astronomy; astrophysics; bayesian method; Be stars; Estrelas Be; Método Bayesiano; stellar physics
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/14/14131/tde-26052019-143801/

Classical Be stars are the only Main Sequence stars that possess Keplerian decretion circumstellar disks ruled by viscous processes. The physical properties of the disk can be studied by modeling its structure and solving the radiative transfer problem. In this point, the production of synthetic observables arises as a tool to investigate the physical quantities of the system. Among the proposed models to the disk formation, the Viscous Decretion Disk model is the current paradigm, correctly describing a large set of evidences that implies viscous shear is the mechanism driving the disk outflow. The modeling of stars is a challenger that demands, sometimes, very elaborate tricks to achieve confident results. Despite this, when we include other elements to the model (e.g. magnetic fields and rotation), the modeling can become very complex. In our study, we aim to study classical Be stars which are known by its rapid rotation rate and non-radial pulsations. These elements are, in some way, at the origin of circumstellar decretion disks. Our main goal is to infer stellar, disk, geometrical parameters, and the interstellar extinction, without neglecting their uncertainties and correlations, for two specific cases (alpha Arae and beta CMi) and for a set of 111 rapid rotating main sequence stars. For such a purpose, we make use of different observables, basically photometric and polarimetric data. We use a modern statistical procedure employed by the code Emcee, stellar parameters based on evolutionary models of the astrophysics group of the University of Geneva, and the code \\textsc used to model the physical processes that occur into Be disks. To constrain the solutions, domain knowledge of $v \\sin i$ and parallax were adopted. We provide a inference tool for the Monte Carlo Markov Chain modeling of Be stars and hot main-sequence O, B and A stars. Our tool includes important effects due to the rapid rotation. The potential of the observables in constraining the stellar parameters is discussed. A catalog of the inferred parameters and a discussion about their correlations are shown. We also present the first multi-technique treatment of a late- and early-type Be star, beta CMi and alpha Arae. The correlations found show that the parameters are coupled. This corroborates the hypothesis that an independent inference of any parameter could propagate errors. For beta CMi, the analysis reveals a circumstellar disk truncated by a companion or a dissipating disk. Moreover, our results suggest the sub-spectral type B8e for this star. For alpha Arae, the analysis also indicates a truncated disk, a thick circumstellar disk and B2e as spectral-type. The thesis is organized as follows. First, we introduce essential topics in Chapter 1. The tools and the grid used are shown in Chapters 2 and 3. In Chapter 4, we describe the adopted Bayesian approach used. After, we list observational dataset gathered into groups based on their features (Chap. 5). We discuss and analyze the results in Chapters 6 and 7. We summarize our conclusions in Chapter 8. The published articles, tutorials, and tables are in the appendices.