The growth and entrainment of cold gas in a hot wind

Both absorption and emission-line studies show that cold gas around galaxies is commonly outflowing at speeds of several hundred km s(-1). This observational fact poses a severe challenge to our theoretical models of galaxy evolution since most feedback mechanisms (e.g. supernovae feedback) accelerate hot gas, and the time-scale it takes to accelerate a blob of cold gas via a hot wind is much larger than the time it takes to destroy the blob. We revisit this long-standing problem using three-dimensional hydrodynamical simulations with radiative cooling. Our results confirm previous findings that cooling is often not efficient enough to prevent the destruction of cold gas. However, we also identify regions of parameter space where the cooling efficiency of the mixed, 'warm' gas is sufficiently large to contribute new comoving cold gas, which can significantly exceed the original cold gas mass. This happens whenever, t(cool, mix)/t(cc) < 1, where t(cool, mix) is the cooling time of the mixed warm gas and t(cc) is the cloud-crushing time. This criterion is always satisfied for a large enough cloud. Cooling 'focuses' stripped material on to the tail where mixing takes place and new cold gas forms. A sufficiently large simulation domain is crucial to capturing this behaviour.