Brightest cluster galaxies in cosmological simulations: achievements and limitations of active galactic nuclei feedback models

We analyse the basic properties of brightest cluster galaxies (BCGs) produced by state of the art cosmological zoom-in hydrodynamical simulations. These simulations have been run with different subgrid physics included. Here we focus on the results obtained with and without the inclusion of the prescriptions for supermassive black hole growth and of the ensuing active galactic nuclei (AGN) feedback. The latter process goes in the right direction of decreasing significantly the overall formation of stars. However, BCGs end up still containing too much stellar mass, a problem that increases with halo mass, and having an unsatisfactory structure. This is in the sense that their effective radii are too large, and that their density profiles feature a flattening on scales much larger than observed. We also find that our model of thermal AGN feedback has very little effect on the stellar velocity dispersions, which turn out to be very large. Taken together, these problems, which to some extent can be recognized also in other numerical studies typically dealing with smaller halo masses, indicate that on one hand present day subresolution models of AGN feedback are not effective enough in diminishing the global formation of stars in the most massive galaxies, but on the other hand they are relatively too effective in their centres. It is likely that a form of feedback generating large-scale gas outflows from BCGs precursors, and a more widespread effect over the galaxy volume, can alleviate these difficulties.