Radiative Feedback from Massive Stars: Ionizing Radiation and Its Fate in Starburst Galaxies.

Abstract

The ionizing light emitted by massive stars represents a critical feedback process in the universe, from potentially providing the ionizing photons for cosmic reionization to forming the basis for diagnostics of fundamental galactic properties. However, the ultimate fate of ionizing photons in a galaxy and the energy distribution of this radiation remain uncertain. The radiation is challenging to study directly; instead, we use both narrowband images and long-slit spectra of photoionized gas to investigate the galactic properties that control the fate of ionizing radiation and to evaluate the spectral energy distributions (SED) predicted by stellar atmosphere codes. First, we generate [SIII]/[SII] maps to trace the passage of ionizing radiation through seven starburst galaxies. With these maps, we discover optically thin ionization cones in NGC 3125 and NGC 5253. Both ionization cones are narrow, which indicates that the low-density passageways carved by wind and bubble activity are narrow as well. These results suggest that an orientation bias limits the ability to directly detect escaping Lyman continuum in starburst galaxies. Next, we turn our attention to the shape of the ionizing SED; we run photoionization simulations drawing from four different grids of atmosphere models for the ionizing source. We compare the emission-line spectra predicted by these model HII regions to long-slit observations of single-star HII regions. All families of atmosphere codes reproduce the observed emission-lines for lines with ionization potential $<35$ eV, assuming the gas distribution is clumpy. However, the predictions exhibit significant scatter about the observations for ions with higher ionization potential. Overall, we find that atmospheres generated with the WM-basic code best represent the ionizing stars. Comparing the H-alpha derived ionizing photon rates to those from the simulations reveals an offset between the different atmosphere models that is systematic with the hardness of the SEDs. Finally, we show that comparing the effective temperatures derived from nebular modeling to those expected from literature calibrations is a potential diagnostic test for the upper layers of the atmosphere models. With this thesis, we demonstrate the diagnostic power of two under-utilized methods: ionization parameter mapping and photoionization simulations of single-star HII regions.