Predicted properties of eclipsing binaries observable by Gaia

Using a population synthesis model we estimate the fraction of stars of different types that are likely to show eclipses, as a function of the period. The population model is based on the bse-code from Hurley et al. (2002) which is a rapid binary evolution code that include all common effects in close binaries such as mass transfer, tidal locking, wind etc. We use this to evolve millions of systems from original distributions in separation, mass and eccentricity, with ages between zero and 12 Gyr, to get a population of systems representative of our Galaxy. The binary data from our model are then analysed in order to, statistically, see how many eclipsing systems we should have. This is done with a simple model where we neglect limb-darkening and other complicating effects. Assuming a random distribution of the inclination angle the probability of an eclipse of a given depth, Am, can easily be calculated. Adding a reasonable fraction of true single stars, we can finally estimate the fraction of eclipsing binaries in limited areas of the HR-diagram, as a function of the period. A first comparison with observational data from the Hipparcos mission shows quite satisfying agreement, and extrapolation to Gaia should thus be a natural application. We find that Gaia will observe about 500 000 eclipsing binaries, this (surprisingly) small number arises from the fact that many eclipsing systems will not be detected by Gaia.