Star formation from the interstellar medium of galactic disks is a basic process controlling the evolution of galaxies. Understanding the star formation rate in a local patch of a disk with a given gas mass is thus an important challenge for theoretical models. Here we simulate a kiloparsec region of a disk, following the evolution of self-gravitating molecular clouds down to subparsec scales, as they form stars that then inject feedback energy by dissociating and ionizing UV photons and supernova explosions. We assess the relative importance of each feedback mechanism. We find that H2-dissociating feedback results in the largest absolute reduction in star formation compared to the run with no feedback. Subsequently adding photoionization feedback produces a more modest reduction. Our fiducial models that combine all three feedback mechanisms yield, without fine-tuning, star formation rates that are in excellent agreement with observations, with H2-dissociating photons playing a crucial role. Models that only include supernova feedback—a common method in galaxy evolution simulations—settle to similar star formation rates, but with very different temperature and chemical states of the gas, and with very different spatial distributions of young stars.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)