Understanding the origin of strong galactic outflows

Understanding the origin of strong galactic outflows

Understanding the origin of strong galactic outflows is one of the key issues in galaxy formation theory. However, due to our incomplete picture of stellar feedback processes and the lack of numerical resolution, much remains uncertain about how gas is blown away from star-forming regions in galaxies. Using high-resolution radiation-hydrodynamic simulations of a gas-rich isolated disk, we investigated the impact of radiation feedback and supernova explosions on the star formation histories and gas outflow rates. In particular, we have introduced radiation pressure from resonantly scattered Lyman α photons for the first time, and found that it efficiently destroys giant molecular clouds in the early stage of bursty star formation episodes in a metal-poor dwarf galaxy. As a result, supernovae explode in low-density environments, carrying a large amount of gas that is comparable to the star formation rate. We also demonstrated that this early feedback self-controls the gas outflow rate by regulating a bursty star formation history, as opposed to some recent claims that multi-scattering photons can drive extremely massive outflows. As Lyman α photons scatter very efficiently in a dust-poor medium, it may also have played a crucial role in suppressing the formation of extremely metal-poor globular clusters in the early Universe.

 Figure 1: radiation-hydrodynamic simulations of a metal-poor dwarf galaxy. The left panels display the projected distribution of gas (top) and stellar density (bottom). The right panels show that the majority of Lyman α radiation is generated (top) and scattering (bottom) inside star-forming regions. Lyman α feedback suppresses the formation of a massive stellar bulge by blowing gas out from the galaxy centre. In order to cover a wide dynamic range from several hundred kpc to a few pc, the simulations were carried out on the high-performance computing facility, DiRAC, employing ~10 million adaptively refined grids.