2015A&A...580A..43G

We study the atomic neutral hydrogen (HI) content of ∼1600 galaxies up to z∼0.1 using stacking techniques. The observations were carried out with the Westerbork Synthesis Radio Telescope (WSRT) in the area of the SDSS South Galactic Cap (SSGC), where we selected a galaxy sample from the SDSS spectroscopic catalog. Multi-wavelength information is provided by SDSS, NVSS, GALEX, and WISE. We use the collected information to study HI trends with color, star-forming, and active galactic nuclei (AGN) properties. Using near-UV (NUV)-r colors, galaxies are divided into blue cloud, green valley and red sequence galaxies. As expected based on previous observations, we detect HI in green valley objects with lower amounts of HI than blue galaxies, while stacking only produces a 3σ upper limit for red galaxies with M_HI<(5±1.5)x10⁸M_☉ and M_HI/L_r<0.02±0.006M_☉/L_☉ (averaged over four redshift bins up to z∼0.1). We find that the HI content is more dependent on NUV-r color, and less on ionization properties, in the sense that regardless of the presence of an optical AGN (based on optical ionization line diagnostics), green-valley galaxies always show HI, whereas red galaxies only produce an upper limit. This suggests that feedback from optical AGN is not the (main) reason for depleting large-scale gas reservoirs. Low-level radio continuum emission in our galaxies can stem either from star formation, or from AGN. We use the WISE color-color plot to separate these phenomena by dividing the sample into IR late-type and IR early-type galaxies. We find that the radio emission in IR late-type galaxies stems from enhanced star formation, and this group is detected in HI. However, IR early-type galaxies lack any sign of HI gas and star formation activity, suggesting that radio AGN are likely to be the source of radio emission in this population. Future HI surveys will allow for extending our studies to higher redshift, and for testing any possible evolution of the HI content in relation to star-forming and AGN properties up to cosmologically significant distances. Such surveys will provide enough data to test the effect of radio/optical AGN feedback on the HI content at lower, currently rather unexplored HI detection limit (M_HI<10⁷M_☉).