2007A&A...476..151B

In the last couple of years a population of very massive (M_*>10¹¹M_☉), high-redshift (z≥2) galaxies has been identified, but its role in galaxy evolution has not yet been fully understood. It is necessary to perform a systematic study of high-redshift massive galaxies, in order to determine the shape of the very massive tail of the stellar mass function and determine the epoch of their assembly. We selected high-z massive galaxies at 5.8µm, in the SWIRE ELAIS-S1 field (1deg²). Galaxies with the 1.6µm stellar peak redshifted into the IRAC bands (z≃1-3, called ``IR-peakers'') were identified. Stellar masses were derived by means of spectro-photometric fitting and used to compute the stellar mass function (MF) at z=1-2 and 2-3. A parametric fit to the MF was performed, based on a Bayesian formalism, and the stellar mass density of massive galaxies above z=2 determined. We present the first systematic study of the very-massive tail of the galaxy stellar mass function at high redshift. A total of 326 sources were selected. The majority of these galaxies have stellar masses in excess of 10¹¹M_☉ and lie at z>1.5. The availability of mid-IR data turned out to be a valuable tool to constrain the contribution of young stars to galaxy SEDs, and thus their M_*/L ratio. The influence of near-IR data and of the chosen stellar library on the SED fitting are also discussed. The z=2-3 stellar mass function between 10¹¹ and ∼10¹²M_☉ is probed with unprecedented detail. A significant evolution is found not only for galaxies with M∼10¹¹M_☉, but also in the highest mass bins considered. The comoving number density of these galaxies was lower by more than a factor of 10 at z=2-3, with respect to the local estimate. SWIRE 5.8µm peakers more massive than 1.6x10¹¹M_☉ provide 30-50% of the total stellar mass density in galaxies at z=2-3.