Evolution of the ISM of starburst galaxies: The SN heating efficiency

The interstellar medium heated by supernova explosions (SN) may acquire an expansion velocity larger than the escape velocity and leave the galaxy through a supersonic wind. Galactic winds are effectively observed in many local starburst galaxies. SN ejecta are transported out of the galaxies by Such winds which thus affect the chemical evolution of the galaxies. The effectiveness of the processes mentioned above depends on the heating efficiency (HE) of the SNe, i.e. on the fraction of SN energy which is not radiated away. The value of HE, in particular in starburst (SB) galaxies, is a matter of debate. We have constructed a simple semi-analytic model, considering the essential ingredients of a SB environment which is able to qualitatively trace the thermalisation history of the ISM in a SB region and determine the HE evolution. Our study has been also accompanied by fully 3-D radiative cooling, hydrodynamical simulations of SNR-SNR and SNR-clouds interactions. We find that, as long as the typical time scale of mass-loss of the clouds to the ambient medium, which is often dominated by photoevaporation, remains shorter than the time scale at which the SNRs interact to form a superbubble, the SN heating efficiency remains very small, as radiative cooling of the gas dominates. If there is a continuous production of clouds by the gas swept by the SNR shells, this occurs during the first less than or equal to16 Myr of the SB activity (of similar to30 Myr), after which the efficiency rapidly increases to one, leading to a possible galactic wind formation. Under an extreme condition in which no clouds are allowed to form, other than those that were already initially present in the SB environment, then in this case HE increases to one in only few Myr. We conclude that the HE value has a time-dependent trend that is sensitive to the initial conditions of the system and cannot be simply assumed to be similar to1, as it is commonly done in most SB galactic wind models.