Stellar Population Models and Chemical Enrichment in Early-Type Galaxies

The stellar populations within galaxies contain information about how those galaxies formed. Hidden in its integrated light, the abundances of different chemical elements in the atmospheres of its constituent stars give clues about a galaxy's past. Through modelling of stellar populations, these clues may be revealed. This thesis is focused on developing such models, with particular interest given to abundance patterns, to address a well known limitation of SSPs built from purely empirical stellar spectra that reside in the local solar neighbourhood. Such stars have the imprint of the formation history of the Milky Way and would not fully represent a system that has formed differently. The work is focused on the generation of a new semi-empirical stellar spectra library. Using predictions of how atmospheric abundances affect theoretical stellar spectra, empirical spectra are differentially corrected to create semi-empirical spectra with different [alpha/Fe] ratios. I then used these semi-empirical star spectra in the computation of new semi-empirical single-age, single-metallicity stellar populations models (SSPs) with variable [alpha/Fe]. The empirical MILES stellar library is used to test three different, state-of-the-art libraries of theoretical stellar spectra that are commonly used in stellar population analysis. The aim is to find where models best represent real star spectra, in a differential way and hence identify good choices of models to use in the creation of semi-empirical stars. We find that most spectral line strengths are well reproduced by the models, particularly those indices that are sensitive to iron and sodium. Exceptions include the higher order Balmer lines (Hdelta, Hgamma), in which the models predict more variation than in data, particularly at low effective temperatures. We also investigate the impact of microturbulence on line-strengths, and find that although the absolute effect can be large (up to ~1-2A), the differential effect is minimal. Corruptions with C_2 line lists for the Coelho set of models are identified and corrected for, improving them for future applications. Using cool giant models, we show that using differential predictions of models produce less or similar agreements with observations than the models' absolute predictions. This result justifies using the theoretical models in a differential way only, when predicting the effect of abundance patterns on SSPs. Next, I generate a new, high resolution, theoretical stellar spectral library using existing methodologies (ATLAS9 model atmospheres and ASSET radiative transfer). These models are fully consistent, in that the abundances have been varied the same way in both model atmosphere and spectral synthesis components. The final grid spans a large range of effective temperature, surface gravity, metallicity, $\alpha$-element and carbon abundances. All theoretical spectra have a large wavelength coverage (1680-9000A), with a fine linear sampling of delta lambda=0.05A. The choice of microturbulence for each model is based on a modification of a previously derived formula for effective temperatures below 6000K, which relates the effective temperature and surface gravity to the microturbulence. The library is tested in both high and low (MILES) resolution regimes, through index measurements and spectral plots. The new theoretical library can reproduce the effects caused by variations in parameters compared to observations of individual stars and other theoretical libraries. Using interpolations within the theoretical library, I create families of theoretical MILES stars with different [alpha/Fe] abundances. Ratios of theoretical spectra are then used to predict how star spectra change with atmospheric abundances, referred to as differential corrections. The application of these corrections to empirical MILES stars produces families of semi-empirical (sMILES) stars with different [alpha/Fe] abundances than found in the local solar neighbourhood. The final result is 5 families of 801 sMILES stars with [alpha/Fe] abundances ranging from -0.2 to 0.6 dex at MILES resolution (FWHM=2.5A) and wavelength coverage (3540.5A - 7409.6A). I then use sMILES stars to create semi-empirical SSPs, with varying [alpha/Fe] abundances for a large range of ages and metallicities. The range and sampling of [alpha/Fe] abundance represent an improvement over previously calculated models (e.g. Vazdekis et al. 2015) The intention is to make the high resolution library and sMILES stars publicly available through the online UCLan database. As new isochrones become available, based on a wider range of abundance patterns, sMILES stars may be used to create SSPs with different abundance ratios. sMILES SSP predictions are then compared to previously computed SSPs. I find reasonable agreements between model predictions for line-strength changes with age and metallicity. For intermediate and old ages, sMILES and Vazdekis et al. model predictions of [MgFe] agree well, for a range of metallicities. I identify differences for total metallicity indicators between SSP models for young, metal-rich populations. Finally, I demonstrate of a potential application of sMILES SSPs, focusing on abundance patterns in a set of stacked SDSS galaxy spectra (Ferreras et al. 2019). Based on a comparison between sMILES model predictions and observations of Mgb and Fe5270 line-strengths, I demonstrate that sMILES SSPs can identify [alpha/Fe] variations of stellar populations in these satellite galaxy spectra for different velocity dispersions and satellite-to-primary mass ratios. Using Lick indices, I find that satellite galaxies are [alpha/Fe]-enhanced to varying degrees. Future fitting of indices and full spectra, of satellites compared to primary galaxies will allow us to measure underlying stellar population parameters to investigate galactic conformity