Feedback from quasars in star-forming galaxies and the triggering of massive galactic winds

The shining of quasars is a likely trigger of massive galactic winds, able to remove most interstellar medium (ISM) from a star-forming spheroid. However, the mechanism responsible for the deposition of energy into the ISM is still unclear. Starting from a model for feedback in galaxy formation with a two-phase medium (Monaco), we propose that the perturbation induced by radiative heating from a quasar on the ISM triggers a critical change of feedback regime. In the feedback model, supernova remnants (SNRs) expanding in the hot and pressurized phase of a star-forming spheroid typically become pressure confined before the hot interior gas is able to cool. In the presence of runaway radiative heating by a quasar, a mass flow from the cold to the hot phase develops; whenever this evaporation flow is significant with respect to the star formation rate, owing to the increased density of the hot phase the SNRs reach the point where their interior gas cools before being confined, forming a thick cold shell. We show that in this case the consequent drop in pressure leads quickly to the percolation of all the shells and to the formation of a super shell of cold gas that sweeps the whole galaxy. Radiation pressure is then very effective in removing such a shell from the galaxy. This self-limiting mechanism leads to a correlation between black hole and bulge masses for more massive bulges than 10(10) M-.. The insertion of a motivated wind trigger criterion in a hierarchical galaxy formation model shows, however, that winds are not necessary to obtain a good black hole-bulge correlation. In the absence of winds, good results are obtained if the mechanism responsible for the creation of a reservoir of low-angular momentum gas (able to accrete on to the black hole) deposits mass at a rate proportional to the star formation rate. Using a novel galaxy formation model, we show under which conditions black hole masses are self-limited by the wind mechanism described above, and outline the possible observational consequences of this self limitation.