Spectroscopy of Galaxies: Evolution of Escape Fractions, Metallicity Gradients and Stellar Metallicity

We use spectroscopic observations to investigate galaxies from the age of reionization to the peak of star formation, and the local universe. This thesis presents three projects to better understand characteristics of galactic feedback, i.e., how it regulates galactic gas flows. We used deep absorption line spectroscopy to estimate the escape fractions fesc of star-forming gravitationally-lensed galaxies at z ≃ 5. The approach is to measure the covering fraction of neutral hydrogen in a galaxy from the amount of non-ionizing UV radiation absorbed by low-ionization metal species. With the boost of signal by gravitational lensing, we observed four 4 < z < 5 star-forming galaxies with DEIMOS, doubling the sample size of existing observations at the redshift range. We found that the escape fractions across our sample varies from galaxy to galaxy and appear to have no significant evolution over time. We inferred a median absolute escape fraction of Lyman continuum photons of 19 ± 6%. Accounting for possible biases and uncertainties, the absolute escape fraction could be reduced to no less than fesc,abs ~ 11%. Moreover, we spatially resolved and detected variation of escape fraction within a galaxy for the first time. The significant variations within the galaxy suggest that the escape fraction is governed by small-scale structure. We found a tentative anti-correlation between the star-formation rate (local or integrated) and the inferred escape fraction. This supports that the escape fraction is associated with the delay time after an episode of star formation. Feedback seems to be more effective in governing both a low SFR and a smaller HI covering fraction. Spatially resolved spectroscopic observations of galaxies at the peak era of star-formation activities are effective in providing insight into the primitive disks. Specifically, we used metallicity gradients to provide constraints on the amount and extent of feedback produced in star-forming galaxies. We observed 15 star-forming galaxies at z ~ 2 with OSIRIS to obtain their kinematic properties and gas-phase metallicity gradients. With helps from AO correction and gravitational lensing, the typical spatial resolution in our study is less than a half-light radius of a typical L* galaxies at z ≃ 2. Combining with the sample in Jones et al., 2013, we approximately tripled the existing metallicity gradient measurements. We found a lower fraction of rotationally-supported systems than reported from larger kinematic surveys with coarser spatial resolution, which might be partially due to a our improved spatial resolution. We demonstrated that a high spatial resolution is crucial for an accurate diagnosis of the kinematic properties and dynamical maturity of z ≃ 2 galaxies. As for metallicity gradients, we found a much higher fraction of z ≃ 2 galaxies having weak or flat metallicity gradients than in previous studies. We correlated the metallicity gradient with the total metallicity and found that all galaxies with low total metallicities have flat gradients (< 0.1 dex per kpc-1). For galaxies with high metallicities ([N II]/Hα > 0.1), there is a divergence between isolated or rotationally-supported and dynamically-immature systems with the latter showing zero gradients irrespective of the integrated metallicity. The results indicate that relatively strong feedback (e.g. high mass loading factors or high SN energy output) is required in order to explain the majority of the observed flat gradients. In the second part of the thesis, we observed quiescent galaxies at z < 1 and archaeologically constrained galactic feedback via its cumulative influence on stellar metallicities. First, we present the stellar mass-[Fe/H] relationship in the Cl0024+17 galaxy cluster at z ~ 0.4. We derived the metalliticies via full spectrum stellar population synthesis modeling of individual quiescent galaxies. Our results provide the metallicities of the lowest galaxy mass (M* = 109.7M⊙) at which individual stellar metallicity has been measured beyond the local universe. We found that the mass-[Fe/H] relationship evolves with the redshift at which the galaxy is observed today. Furthermore, we found an even stronger evolution of the mass-[Fe/H] relation with the time at which the galaxy formed (rather than the time at which it is observed). Galaxies that formed earlier have lower Fe abundance than galaxies that formed later. Lastly, we measured magnesium (Mg) abundances and extended the observed redshift to z ~ 0.55. We found that while the mass-[Fe/H] relation evolves significantly over the observed redshift range, the mass-[Mg/H] relation does not. This is due to the shorter star formation histories of quiescent galaxies at higher redshifts. Fe is mainly produced in Type Ia SN. It has a longer recycling time than Mg, which is mainly produced in core-collapse SN. Using core-collapse SN elements as a metal indicator lessens the complication of delayed recycling time and allows us to effectively use galactic chemical models with instantaneous recycling to quantify average outflows that these galaxies experience over their lifetime. We found that the average mass-loading factor η is a power-law function of galaxy stellar mass, η ∝ M*-0.21±0.09, consistent with the results of other observational methods and with the predictions where outflow is caused by star formation feedback in turbulent disks.