Unveiling the oldest and most massive galaxies at very high redshift

Context. The identification and characterisation of massive galaxies over a wide redshift range allow us to place stringent constraints on the cosmic history of galaxy mass assembly and on current models of galaxy formation and evolution. Aims. This work explores the existence of high-redshift massive galaxies unveiled with Spitzer +IRAC, but missed by conventional selection techniques based on optical and near-infrared observations. Methods. To this end, we used multi-wavelength imaging data available for the GOODS-South field (130 arcmin 2 ) and selected a flux-limited sample from the IRAC 3.6  μ m image to $S_{3.6}\geq 1.8 ~\mu$Jy ( m (AB)  K -band image ($K>23.5$ AB). Our selection unveiled 20 galaxies on which we performed a detailed analysis.
For each galaxy, we built a spectral energy distribution (SED) based on optical-to-8 μ m photometry. The SEDs were then used to estimate the photometric redshifts and to derive the main galaxies' physical properties. Further constraints were also obtained from the available X-ray and 24  μ m data. Results. The majority of the sample (14 out of 20) sources show degenerate/bimodal solutions for the photometric redshifts. These can either be heavily dust-enshrouded ($A_{\rm V}\sim2{-}4$) starbursts at $2 10^{12} L_{\odot}$, or massive post-starburst galaxies in the redshift interval $4 z dusty source, these latter objects favour a low-extinction solution, with four of them showing best-fit photo- z solutions at $z\sim4$. One galaxy, ID-6, the only source in our sample with both an X-ray and a 24  μ m detection, might be an extremely massive object at $z\sim 8$ detected during a post-starburst phase with concomitant QSO activity responsible for the 24  μ m and X-ray emissions (although a lower- z solution is not excluded). Conclusions. Our investigation of Spitzer -selected galaxies that have very red SEDs and completely undetected in the optical reveals a potential population of massive galaxies at $z\geq4$, which appear to include significant AGN emissions. These sources may be the oldest stellar systems at $z\sim4$, given that the estimated ages are close to the age of the Universe at that redshift. We found that these, previously unrecognised, optically obscured objects might provide an important contribution to the massive end ($M>10^{11}~M_{\odot}$) of the high- z stellar mass function, and they would almost double it. Our evidence in these mature high- z galaxies of the widespread presence of hidden AGNs may have important implications for galaxy formation, due to their feedback effects on the surrounding ISM.