Physical regimes for feedback in galaxy formation

We present a new (semi-) analytical model for feedback in galaxy formation. The interstellar medium (ISM) is modelled as a two-phase medium in pressure equilibrium, where the cold phase is fragmented into clouds with a given mass spectrum. Cold gas infalls from an external halo. Large clouds are continually formed by coagulation and destroyed by gravitational collapse. Stars form in the collapsing clouds; the remnants of exploding Type II supernovae (SNe) percolate into a single super-bubble (SB) that sweeps the ISM, heating the hot phase (if the SB is adiabatic) or cooling it (in the snowplough stage, when the interior gas of the SB has cooled). Different feedback regimes are obtained whenever SBs are stopped either in the adiabatic or in the snowplough stage, either by pressure confinement or by blowout. The resulting feedback regimes occur in well-defined regions of the space defined by vertical scalelength and surface density of the structure. In the adiabatic blowout regime, the efficiency of SNe in heating the ISM is rather low (similar to5 per cent, with similar to80 per cent of the energy budget injected into the external halo), and the outcoming ISM is self-regulated to a state that, in conditions typical of our galaxy, is similar to that found in the Milky Way. Feedback is most efficient in the adiabatic confinement regime, where star formation is hampered by the very high thermal pressure and the resulting inefficient coagulation. In some significant regions of the parameter space, confinement takes place in the snowplough stage; in this case, the hot phase has a lower temperature and star formation is quicker. In some critical cases, found at different densities in several regions of the parameter space, the hot phase is strongly depleted and the cold phase percolates the whole volume, giving rise to a burst of star formation. While the hot phase is allowed to leak out of the star-forming region, and may give rise to a tenuous wind that escapes the potential well of a small galactic halo, strong galactic winds are predicted to occur only in critical cases or in the snowplough confinement regime whenever the SBs are able to percolate the volume. This model provides a starting point for constructing a realistic grid of feedback solutions to be used in galaxy formation codes, either semi- analytical or numerical. The predictive power of this model extends to many properties of the ISM, so that most parameters can be constrained by reproducing the main properties of the Milky Way.