Revealing the chemical evolution history of the Milky Way Halo with RR Lyrae stars

The Galactic Halo is the component of the Milky Way that can preserve the clearest signatures of the accretion of satellites required for the built-up of the Galaxy according to the Λ cold dark matter paradigm. The nature of these primordial satellites, their relationship to the current dwarf galaxies distributed around the Milky Way, and the fraction of the current Halo stellar population that originated from them are all open questions. The abundance of different chemical elements carries information about the enrichment timescale and physical parameters (e.g. mass distribution function, star formation rate) of the stellar system where a given star was formed. Thus, studying a variety of elements is fundamental in order to properly identify the progenitor systems of different stars, to constrain models of galactic formation, and to understand stellar nucleosynthesis itself. In particular, measurements of heavier species and robust ages for stellar tracers are severely lacking. The RR Lyrae stars (RRLs) are old (≥ 10 Gyr) radial pulsators with well defined period-luminosity relations in near-infrared (NIR) bands. These relations allow for precise distance determinations and make the RRLs excellent probes of the chemical structure of the Halo. The RRLs are classified according to their pulsation mode, with the RRab pulsating in the fundamental mode, the RRc in the first overtone, and the RRd, very few in number, in a mixed mode. High resolution spectroscopic studies of these stars are scarce. Such studies are fundamental for the investigation of α elements, Fe-peak elements, and neutron(n)-capture elements. They are also necessary for the calibration of photometric indexes and low resolution (LR) spectroscopic estimates of metallicity that can be applied to large samples. In this work, we used data collected with nine high-resolution (HR) spectrographs installed at eight different large telescopes and by two LR surveys (SEGUE-SDSS, LAMOST). We applied the methodology of HR spectroscopy to a sample of 162 RRLs (138 RRab, 23 RRc, 1 RRd) covering a wide range in metallicity ([Fe/H] from -3.2 to 0.2). Our spectroscopic investigaton is based on a list of atomic lines with updated transition parameters. We meticulously removed lines that showed evidence of blending or of dependence on effective temperature. Using this line list, we obtained measurements of the light odd-Z elements Na and Al; the α elements Mg, Si, S, Ca, and Ti; the Fe-peak elements Sc, V, Cr, Mn, Fe, Co, Ni, Cu, and Zn; and the n-capture elements Sr, Y, Zr, Ba, La, Ce, Pr, Nd, and Eu. In particular, our investigation includes the bona fide tracer of the s-process Ba, its r-process counterpart Eu, and the poorly studied rare earth Pr. Using stars in common with other HR literature investigations, we brought abundance measurements of another 85 RRLs (65 RRab, 20 RRc) into our chemical abundance scale. We used these HR, high quality measurements to probe the chemical evolution history of the Halo, and also to compare it to the other Galactic components and nearby dwarf galaxies. In this regard, the RRLs present the unique advantage of being old tracers with precise distance measurements, while other commonly employed stellar populations have very poorly constrained ages and distances. We found that a sample of metal-rich, α-poor RRLs is present at low Galactic heights. In the [α/Fe] versus [Fe/H] plane, this sample displays a smooth continuity with the rest of the RRLs in the Halo, pointing to a shared origin. Furthermore, this sample traces the metal-rich tail of the Halo that is poorly covered by other stellar tracers and displays a behavior with metallicity that is at odds with the other Galactic components (Bulge and Disk), but similar to the Sagittarius dwarf spheroidal. We performed the same comparison for the other chemical species and discussed the effects of NLTE corrections, different line lists, and systematics due to evolutionary effects. By applying the same HR analysis to a sample of six non-variable giants and seven field dwarfs, we showed that differences in abundance, especially for α elements, between field stars of varied ages and the RRLs are intrinsic. The ∆S method provides metallicity estimates for RRLs with LR spectra. The number of available LR spectra for RRLs is about two orders of magnitude larger than HR spectra. Therefore, a finely tuned LR metallicity estimator is an extremely valuable counterpart to the more accurate but scarcer HR measurements. The ∆S method results are widely used in the calibration of photometric methods such as Fourier parameter decomposition. Until recently, this method was only applied to RRab stars outside the phase interval of the rising branch of the light curve and relied on the metallicity scale of Zinn & West (1984), which is not linear with modern HR measurements. Using the HR results for 143 RRLs (111 RRab, 32 RRc), we developed a brand new calibration of the ∆S method in the same metallicity scale as our HR results. It can be applied to the whole pulsation cycle of both RRab and RRc pulsators, with preliminary evidence that it is valid for the RRd as well. We applied this new calibration to a sample of 7768 RRLs (5196 RRab, 2572 RRc) for which no HR spectra are available and found excellent agreement with the results derived in HR. The complete sample with 247 RRLs studied in HR is the largest and most homogeneous data set of old stellar tracers studied in the literature. It provides constrains on all major chemical families for the oldest stellar component in the Galaxy covering 3 dex in metallicity. Our sample covers a significant fraction of the Halo starting at a Galactocentric distance of approximately 4 kpc, with HR measurements reaching 26 kpc, and LR metallicity estimates extending as far as 150 kpc. These results are crucial not only for Galactic formation modeling but also in the investigation of nucleosynthetic processes due to its homogeneity in both abundance scale and age, its spatial distribution, and its wide metallicity coverage. Furthermore, the new calibration of the ∆S method is applicable, for the first time, to RRLs of all pulsation modes observed at any phase, without the requirement of detailed photometric studies or carefully timed observations. The work performed in this thesis has also been partially or fully employed by our group in various investigations, including probing the fine structure of the Bailey diagram and its relation to the Oosterhoff dichotomy, the development of new barycentric velocity estimators and velocity curve templates, and a new calibration of the Fourier parameter decomposition method using our new metallicity scale.