Improving the Precision of Type I A Supernova Cosmology

Type Ia supernovae (SNe Ia) mark the beginning and the end of stellar evolution. They are one of the most powerful cosmological probes in our universe thanks to their high intrinsic luminosities and standardizable properties. In the 1990s, the observation of SNe Ia led to the discovery of the accelerating cosmic expansion that was awarded the Nobel Prize in 2011. Despite decades of advancements, the exact details of progenitor systems (i.e. what exploded), explosion mechanisms (i.e., how they exploded), and evolution effect (i.e. are the nearby population the same as those in the early universe?) are still not fully understood yet. With more advanced observation surveys forthcoming in the near future, such as those on board the Vera C. Rubin Observatory and the Nancy Grace Roman Space Telescope, the systematic uncertainties of SNe Ia observations will continue to dominate the error budget of their distance measurements. This dissertation is motivated to improve the precision of SN~Ia cosmology by understanding the physics and diversity and SNe Ia, as well as improving the astronomical tools needed for cosmological purposes. Photometric and spectroscopic observational studies of the first homogenous sample of the peculiar 03fg-like SNe Ia obtained by the Carnegie Supernovae Project (CSP) are conducted to investigate the physics of SNe Ia. 03fg-like events are usually more luminous than the normal SNe Ia and exhibit peculiar light-curve shapes in the redder filters, such as the weak or missing secondary maxima and the delayed peak time in iYJH bands. Spectroscopically, 03fg-like SN Ia show peculiar features in both optical and near-infrared (NIR) regions, such as the stronger C, slower Si, and the lack of the H-band break. One of the most extreme objects within this subgroup, ASASSN-15hy, is used for the case study of an envelope model that could potentially explain all 03fg-like SNe~Ia. It is found that a core degenerate scenario, an explosion of a degenerate white dwarf core inside a nondegenerate envelope, with a transition from deflagration to detonation can explain the observed peculiar properties. In the case of ASASSN-15hy, the low metallicity of the progenitor is a key aspect of the model explanation. In fact, a low-metallicity host environment is a shared preference among these 03fg-like SNe~Ia. Such host environment is more common in the early universe, which brings up the concern that 03fg-like SNe~Ia may be a problem for high-redshift SN~Ia cosmology due to detection bias and potential evolution preference. Therefore, more detailed observational and theoretical studies of these events are strongly recommended. Another main focus of this dissertation is on the development of a new NIR spectral template of SNe Ia that captures the feature variations that are correlated with the light-curve shapes. Compared to optical observations, NIR observations of SN Ia are less sensitive to dust and have more uniform peak luminosities, which are beneficial for cosmological purposes. A spectral template is usually needed to fit the light curves of SNe Ia accurately in order to estimate the distance. However, the NIR part of the existing spectral templates lacks an accurate description of the intrinsic spectral variations. Using the largest and most homogeneous collection of NIR spectra of SNe Ia to date collected by CSP-II, we are able to explore the NIR spectral diversity of SNe Ia and build a new NIR spectra template. Principal component analysis and Gaussian process regression are used for the template construction, which reduces data dimensionality and models the parameter dependence, respectively. Using the new template reduces the systematic uncertainties in K-corrections by ~90% compared to those from the Hsiao template. Furthermore, this template can serve as the baseline spectral energy distribution for various light-curve fitters and can identify peculiar spectral features that might point to compelling physics. The NIR spectra data and template presented in this work will substantially improve future SN Ia cosmological experiments, for both nearby and distant samples.