Low-Metallicity Stars and High-Redshift Galaxies Through the Lens of Local Metal-Poor Star-Forming Regions

Abstract

Over the last decade, unprecedented imaging campaigns and deep spectroscopy have delivered our first glimpse of galaxies in the reionization era, and the results have stymied interpretation. The prominent nebular emission in CIII], CIV, and other high-ionization lines detected in the first rest-UV spectra at these redshifts suggests that hard ionizing radiation fields are common at z>6, in striking contrast to typical star-forming systems at lower redshift. The difficulty in interpreting this emission is fundamentally tied to the lack of empirical constraints on massive stars at the very low metallicities we expect to encounter in the first billion years of the Universe. Nearby star-forming regions hosting stars and gas extending to metallicities below that of the SMC (<20% solar) represent a critical laboratory for study of such young stellar populations. In this dissertation, I first present HST/COS UV spectroscopy of HeII-emitters which reveal a clear transition from UV spectra dominated by stellar features above 20% solar metallicity to high-ionization nebular line emission at lower metallicities, tracked by a significant hardening in the inferred ionizing spectrum. I demonstrate that nebular HeII and CIV are ubiquitous among local extremely metal-poor galaxies (XMPs, <10% solar), and that CIV in particular may be an effective signpost of rapidly-assembling systems at these low metallicities in the distant Universe. I then present a new technique for locating XMPs dominated by very young stars in broadband photometry, and apply it to SDSS imaging to uncover 32 such systems at typical effective ages of tens of megayears. I demonstrate that the nebular HeII commonly encountered in such systems is inconsistent with an origin in high-mass X-ray binaries or the most massive short-lived stars, and suggest instead that this line may provide insight onto stripped stars and other uncertain products of binary evolution at very low metallicity. I also present evidence from stellar population synthesis modeling that an overabundance of massive stars or spun-up binary products are necessary to explain the strongest stellar wind features in very young systems at higher metallicities where these winds are prominent. Finally, I conclude by outlining a path forward from testing to direct calibration of models for massive stars below the metallicity of the SMC.