BeAtlas: A grid of synthetic spectra for Be stars

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 physical structure and solving the radiative transfer problem. At this point, the calculation of synthetic observables arises as a tool to investigate the physical quantities of these systems. Among the proposed models to explain Be star disks, the Viscous Decretion Disk model is the only one correctly explaining a large set of evidence that implies that viscous shear is the mechanism driving the disk outflow.

Modeling of observations requires elaborate procedures to achieve reliable results. In the case of Be stars, the system under study is very complex. The central star is affected by fast rotation and stellar pulsation. The disk structure varies in time, and depends in a complex way on several factors, e.g., disk feeding rate, viscosity, the presence of a binary companion, etc. Another source of complication is the complex interplay between stellar and disk parameters.

We present a new tool for modeling Be star observations that takes into consideration most of the relevant physical processes. Named BeAtlas, it consists of two parts. The first part is a grid of Be star models that cover all the relevant physical parameters of the star (mass, age, rotation rate, etc.) and the disk (size, density structure, density scale, etc.). BeAtlas was computed with the radiative transfer code Hdust. The stellar parameters were obtained from the Geneva stellar evolution models.

The second part is a set of computer tools to the fit of the observations with BeAtlas using Bayesian statistics to infer the posterior probabilities of each stellar and disk parameter. BeAtlas allows for using prior information, such as distance, v sin i and inclination, when available. Our main goal with BeAtlas is to provide means for studying B and Be stars observations into account all the relevant parameters and their cross correlations.

BeAtlas was first applied to study of 111 OBA stars. Using IUE data and prior information (v sin i) and Hipparcos parallax) about these stars, we were able to derive their stellar and geometrical parameters, as well as the corresponding interstellar extinction. A comparison of our results of those from the literature yielded in general good results. Deviant cases must still be analyzed in more detail, but one possibility under investigation is that they due to fundamental differences between our approach and some of the ones used in the literature (e.g., rotating vs. non-rotating models) which limits the viability of a direct comparison.

A second application was on the Be Star beta CMi, which was modelled by Klement et al. 2015 also using the Viscous Decretion Disk model. The results show the ability of BeAtlas to quickly recover the disk and stellar parameters. Furthermore, a detailed analysis of the results show that our procedure does indeed provide for more robust and realistic error estimates by properly considering the coupling between the parameters.

Finally, a third application was on the Be star alpha Arae. We present new observations (polarization, spectroscopy and sub-mm photometry) that, together with data obtained from the literature, show that alpha Arae had a stable disk for the last 50 years. We found a disk density slope (n = 2.44^{+0.27}_{-0.16}) that is inconsistent with the standard theory that predicts that isothermal steady-state disks should have n=3.5. Two effects may be responsible for this: non-isothermal effects in the disk, or the effects of a binary companion. Our results provide evidence that alpha Arae is a binary system. In our case, evidence came for the change of the SED slope in the sub-mm domain, here interpreted as being caused by an unseen binary companion that truncates the disk of the primary.