ISOCAM and DENIS survey of 0.8 square degrees in the bar of the LMC. Detection of the whole T-AGB star population

C. Loup ¹, M-R. Cioni ², E. Josselin ¹, J. Blommaert ³, G. Simon ⁴, M. Groenewegen ⁵, C. Allard ⁴, E. Bertin ⁶, P-A. Duc ⁶, N. Epchtein ⁷, P. Fouqueinst ^4,8, H. Habing ², F. Kerschbaum ⁹, A. Omont ¹, N. Trams ³, J. van Loon ¹⁰, L. Waters ¹⁰, P. Whitelock ¹¹, & A. Zijlstra ¹²

¹ Institut d'Astrophysique de Paris, CNRS, France
² Sterrewacht Leiden, the Netherlands
³ ISO Data Centre, VILSPA, Madrid, Spain
⁴ Observatoire de Paris-Meudon, France
⁵ Max Planck Institut für Astrophysik, Garching, Germany
⁶ European Southern Observatory, Garching, Germany
⁷ Observatoire de Nice, France
⁸ European Southern Observatory, Chile
⁹ Institut für Astronomy, Wien, Austria
¹⁰ Astronomish Instituut, University of Amsterdam, the Netherlands
¹¹ SAAO, South Africa
¹² UMIST, Manchester, United Kingdom

Searches for AGB stars in the LMC with ``old'' observational techniques started in the sixties and were still producing results in the nineties. They were oriented towards two different concepts, searches for stars with a determined spectral type M or C with prism-pbjective surveys, and searches for Long Period Variables (LPVs) mostly with sets of I photographic plates (lately analysed with modern techniques). About 4400 AGB stars were thus identified in the whole LMC. The main results of these works were : first the need to add ``convective overshoot'' in theoretical evolutionary models to explain the luminosity function of carbon stars, second the well established period- luminosity relation for mira variables with periods smaller than 400 days. All the AGB stars identified in these surveys are ``optical'' stars in the sense that they experience (very) faint mass-loss. Even limited to such stars with very faint mass-loss, our knowledge of the thermal-pulsing-AGB (T-AGB) population is complete in only three regions of the Bar of the LMC.

First searches for so-called ``obscured'' AGB stars, i.e. T-AGB stars with high mass-loss rates, faint or even invisible in the optical range, started only in 1986, using the IRAS data. About 50 obscured AGB have thus been discovered in the whole LMC, with mass-loss rates of the order of a few to several 1D-5 Mo/yr. However, due to the sensitivity limits of IRAS, this sample is strongly biased towards the reddest and the most luminous sources.

This let us with an obvious lack in our knowledge of the T-AGB population : all the stars with intermediate mass-loss rates, as well as stars with luminosities lower than 1D4 Lo and high mass-loss rates, were totally unknown. The major goal of our ISOCAM minisurvey was to specifically discover this missing population in order to test evolutionary models with a complete sample of T-AGB stars. We have surveyed 0.8 square degrees in the Bar with ISOCAM at 4.5 and 12 microns, and with DENIS in the IJKs bands. Full-filling our expectations, the combination of both surveys actually allow us to easily discover all the T-AGB stars, with all the range of mass-loss rates. Typically, in an area of 16'x16' (1 raster), we find 20 stars with significant mass-loss rates (to be compared with 1 obscured source previously identified with IRAS), and 1000 stars with no or faint mass-loss. Based on accurate dust radiative transfer models, we developed tools to accurately estimate bolometric luminosities and mass-loss rates on huge samples of stars. We will present the luminosity function thus obtained for a complete sample of T-AGB stars, as well as the mass-loss rates distribution. One of the main result is the bump formed by thermal pulsing AGB stars in the global luminosity function, bump which was not planned by theory and could be a tracer of the old star formation history.