Insights into Star Formation and Stellar Death Using Star-Forming Galaxies and Supernova Remnants
Loading...
Citations
Altmetric:
Altmetric::
Download
Associated Records
Authors
File Licence
Editors
Date
2025
Source Title
Journal Title
Journal ISSN
Volume Title
Publisher
Melbourne Authors
Affiliation
School of Physics
Document Type
PhD thesis
Subject Code
Citation
Access Status
Open Access
DOI
Access URL
Open Access URL
Open Access at PMC
Linked Resource URL
NHMRC Grant Code
ARC Grant Code
Description
© 2025 Sonja Rose Panjkov
Abstract
The processes of star formation and stellar death are central to many areas of astrophysics, from galaxy formation and evolution, to the production of the elements. This thesis delves into these phenomena using star-forming galaxies and supernova remnants (SNRs).
Previous studies have indicated that cosmic star formation rate density peaks at a redshift of ~2. Such studies hinge on the identification of large, uncontaminated galaxy samples, which can be a fraught exercise, as typical spectral techniques may be costly or arduous to implement, and certain wavelength bands suffer significantly from dust extinction. This thesis explores an alternative approach, which leverages machine-learning techniques to classify galaxies and quasars using wide-field photometric catalogues and the 1.4 GHz radio emission as a dust-agnostic, star-formation tracer. An XGBoost model was trained on ~58,000 sources using 105 features from RACS-mid1 and WISE, achieving a weighted F1 score of ~93%. The star-formation rates of the 28,501 low-redshift sources were calculated using a modified 1.4-GHz calibration based on the infrared-radio correlation of the dataset, yielding a local star formation rate density of (1.4 +- 0.2) x 10-2 solar masses yr-1 Mpc-3, consistent with previous results.
Massive stars with M >= 8 M solar masses end their lives as core-collapse supernovae (SNe), while lower mass stars may explode as Type Ia SNe. To improve our understanding of stellar death, this thesis contains a study that relates SNR properties to the progenitor stars by investigating the progenitor properties of the LMC core-collapse SNe. Of the 78 LMC SNRs, 45 likely arose from the core collapse of a massive star based on independent constraints and the local star-formation histories (SFHs). For these sources, the most likely progenitor mass was determined by relating bursts in the SFH to the progenitor lifetime and mass. ~90% are associated with mid- or high-mass progenitors, consistent with theoretical predictions and in contrast to the dearth of high-mass progenitors identified in pre-explosion images. Through a comparison to other galaxies with varying metallacities, the cumulative core-collapse progenitor mass distribution was shown to become steeper with decreasing metallicity, indicating that environment likely plays an important role in the production of massive progenitors.
Notes
Terms of Use
Keywords
Supernova remnants, Star formation, Machine learning, Large Magellanic Cloud