Galaxy evolution through the lens of stellar population synthesis

Sextl, Eva Maria Theresia (2025): Galaxy evolution through the lens of stellar population synthesis. Dissertation, LMU München: Fakultät für Physik

Abstract

The study of galactic evolution remains one of the central questions in modern astronomy. Understanding the governing processes requires careful examination of both the dynamical and chemical evolution of galaxies, as these shape their structure, composition, and long-term behavior. Despite their significance, many previous studies have focused on global indicators or gradients, often assuming strict azimuthal symmetry. Such simplifications overlook that galaxies are complex systems influenced by localized feedback mechanisms and environmental effects that govern their development and diversity. This thesis investigates these issues through stellar population synthesis (SPS), a spectral analysis technique that decomposes the integrated light of a galaxy region into the contributions from stars of different ages and metallicities. Since individual stars are typically unresolvable beyond the Local Group, SPS is one of the most powerful tools for exploring the stellar content of distant systems. Through a fitting technique, this method determines key physical parameters such as stellar ages, metallicities, star formation histories, and the dust properties of the interstellar medium. Collectively, they provide fundamental insight into the processes that govern galaxy formation and evolution. Although SPS, including its most popular variant 'full-spectral fitting', has been developed and improved over several decades, a tremendous potential remains untapped. This work aims to refine and extend stellar population synthesis to make it a competitive alternative to established techniques such as HII region emission line studies and the spectral analysis of individual extremely bright stellar sources like supergiant stars and stellar clusters. Ultimately, we improve the methodological framework of SPS to derive robust and physically meaningful stellar population properties within the challenging environment of star-forming nuclear rings. As a first step, we analyzed stacked spectra constructed from $\sim$200,000 star-forming galaxies in the Sloan Digital Sky Survey (SDSS) with SPS. By using our own carefully tested spectral fitting algorithm, we measure metallicities, ages, and star formation histories of the young and old stellar populations together with interstellar reddening and extinction. We obtain the mass-metallicity relationship for young and old stellar populations, which aligns well with analytical look-back evolution models. The focus of this exploratory study extends to underlying assumptions about star formation and dust properties, which have been largely omitted in previous research. In an interdisciplinary extension to the previous traditional application of SPS, the first results of an ambitious and award-winning AI project are discussed. A generative machine learning framework is trained on SDSS data and capable of predicting optical galaxy spectra from photometric broadband images. We apply SPS to the predicted galaxy spectra to test the AI's capability. With further refinement, this approach could recover key galaxy properties, normally accessible only through spectroscopic observations, from photometric data alone. Next, we investigated the spatially resolved spectra of galaxies in the TYPHOON survey. In the grand design spiral NGC 1365, the stellar disk grows inside-out, but an unexpected central dip in metallicity signals the inflow of metal-poor gas and interruptions in star formation caused by active galactic nucleus (AGN) feedback. Similarly, M83, a massive barred spiral galaxy with an irregular nuclear region, shows a high and fairly uniform metallicity in its young stars with a localized central decrease linked to gas inflow or AGN effects, alongside a dust cavity correlated with molecular gas. For the first time, a spatial one-to-one comparison of metallicities derived from full-spectral fitting with those obtained from individual young stellar probes was carried out in the fourth and fifth publication. To do this, we moved beyond common metallicity definitions in the field to obtain a more physically meaningful and robust measure called 'physical metallicity'. This updated definition aligns with the conventions commonly used in chemical evolution studies, and the successful comparison confirms the robustness of our spectral fitting technique. In the fifth publication, the focus turns to one of the current key instruments for spatially resolved spectroscopy: the MUSE spectrograph at the $8$m VLT. This powerful tool delivers high-quality spectroscopic data across a two-dimensional field of view in a single exposure. Applying SPS, we challenge prior interpretations from the flagship MUSE-TIMER survey regarding the chemical evolution of prominent nuclear rings in the local universe. The results reveal that peculiar low-metallicity signatures in NGC 7552, NGC 613, NGC 1097, and NGC 3351 originate from older stellar populations, tracing past inflows of metal-poor gas, followed by normal enrichment to (super)solar levels. By separating young and old stellar contributions, the study avoids biases of conventional averaging methods and highlights the role of very young stars in accurate modeling. The findings, supported by reddening maps, confirm ongoing inflows and show that MUSE’s spatial and spectral resolution enables detailed investigations of nuclear star-forming regions despite lacking the blue part of the optical spectrum. Together, these five studies demonstrate that stellar population synthesis, when refined with physically motivated definitions and applied across statistical, spatially resolved, and machine learning domains, provides a physically grounded and predictive tool for tracing galactic evolution over cosmic time. With the wealth of high-quality data already available, a unique opportunity of investigation in both the local and distant universe lies ahead.

• Dokumententyp: Dissertationen (Dissertation, LMU München)
• Keywords: Galaxy Evolution, Stellar Populations, Metallicity, Nuclear Ring, SDSS
• Themengebiete: 500 Naturwissenschaften und Mathematik ; 500 Naturwissenschaften und Mathematik > 530 Physik
• Fakultäten: Fakultät für Physik
• Sprache der Hochschulschrift: Englisch
• Datum der mündlichen Prüfung: 15. Dezember 2025
• 1. Bericht­erstatter:in: Kudritzki, Rolf-Peter