Dust-Obscured Galaxy Protocluster and Cluster Survey (DOGPACS): Identifying Large-Scale Structures 9−10 Billion Light-Years Away
Date
2022Metadata
[+] Show full item recordAbstract
Primordial density fluctuations have grown over time due to gravitational
instability to form galaxies and, eventually, large-scale structures, such as clusters
of galaxies. Galaxy clusters are the most massive collapsed structures in the universe.
During cluster formation, the largest aggregation of gas, galaxies, and dark
matter passes through an intermediate phase called the protocluster. Over the past
decades, many studies have identified distant clusters and protoclusters due to
advanced observational strategies. However, the protocluster-to-cluster transformation
is still unclear, mainly due to the lack of large samples of early-stage clusters and
late-stage protoclusters. Our research has identified a large selection of nearly 300
galaxy cluster candidates at redshift 1.3 < z < 1.8 (9-10 billion light-years away)
during the formation epoch of the galaxy clusters. This study leverages the fact that high-z galaxy clusters and protoclusters exhibit enhanced star-formation and AGN
activity in their cores. The candidates are identified using a sample of highly star-
forming and/or AGN Ultra-Luminous Infrared Galaxies called the Dust-Obscured
Galaxies (DOGs) as signposts in the Spitzer Deep Wide-Field Survey (SDWFS) in
Bootes. A two-point correlation function analysis demonstrates that the sample has
a mass scale of the galaxy clusters. Using a more multi-wavelength SDWFS catalog,
this study has also uncovered a proto-supercluster structure at z = 1.75 (10 billion light-years away). This proto-supercluster is a bound structure hosting dozens
of protoclusters and clusters of galaxies, including the most massive galaxy cluster
(IDCS J1426.5+3508) found to date at z > 1.5. Follow-up studies of this supercluster
will provide a comprehensive picture of protocluster-to-cluster transition and the
evolution of its constituent gas and galaxies. Finally, we develop and implement a
novel machine learning technique to determine the photometric redshift (photo-z)
of the galaxies using a decision tree-based architecture as part of the mission work
for the upcoming Euclid Space Mission (2023). The photo-z results are comparable
with many other template fitting- and machine learning-based methods.
Table of Contents
Galaxy cluster candidates signposted by dust-obscured galaxies as 1.3≤z≤1.8 -- A proto-supercluster candidate hosting a massive galaxy cluster at z-1.75 -- Estimating photometric redshift probability distribution function using gradient boosted regression tree (GBIT)
Degree
Ph.D. (Doctor of Philosophy)