• Probing Conditions for Lyman-Continuum Escape in the Reionization Epoch
![[ML]](./research_files/LyC.png)
JWST observations during the epoch of reionization (EoR) have unveiled a high abundance of galaxies, an elevated ionizing photon production rate, and ubiquitous active galactic nuclei. These findings indicate an overabundance of ionizing photons beyond the required budget for reionization. However, the unknown escape fraction of Lyman-continuum (LyC) photons from galaxies complicates assessing the ability of these sources to supply ionizing photons to the intergalactic medium. The best practice is to use lower-redshift star-forming galaxies to comprehend their conditions conducive to LyC escape and calibrate a set of proxy measurements that might reliably predict the emergent ionizing continuum from a galaxy. I analyzed HST imaging observations of gravitationally-lensed, intrinsically faint galaxies at z = 1.3 - 3.0 (Jung et al. 2024b, ApJ, 971, 175). The study provides essential constraints on escaping LyC fluxes from UV faint-end galaxies (down to MUV ~ −14), revealing no significant LyC flux. These findings challenge the expectation of higher LyC escape fractions from UV-faint sources. Additionally, I work on the HST spectroscopic data of local compact star-forming galaxies with extreme [OIII]/[OII] ratios to investigate favorable physical conditions for LyC escape (Jung et al. 2024c, arXiv:2409.11460). These targets exhibit intense CIV emission with high CIV/CIII] ratios, indicative of hard ionizing radiation and substantial LyC leakage. Collectively, these studies underscore the importance of local analog galaxies in understanding the physical conditions conducive to LyC escape and identify potential ionization sources (e.g., X-ray binaries, AGNs, very massive stars) of these extreme galaxies that resemble reionization-era galaxies.
• Lyman-alpha Emitters in the Epoch of Reionization with JWST Observations
![[ML]](./research_files/CEERS_LAE.png)
In the JWST/ERS CEERS collaboration (PI: Finkelstein), I spearheaded the analysis of the rest-frame UV to optical spectra for three LAEs at z > 7 (Jung et al. 2024a, ApJ, 967, 73). Our research employed two distinct methodologies to assess the size of ionized bubbles surrounding these galaxies. Initially, we analyzed the Lyman-alpha transmission to determine the minimum size of ionized bubbles necessary for the escape of Lyman-alpha photons. Subsequently, we calculated the ionizing photon output from the LAEs and derived their Strömgren radii based on emission-line properties. By employing both approaches, we were able to cross-verify and validate our findings regarding the ionized bubble sizes. Our analysis indicates that the source with the highest ionization condition might be capable of ionizing its own surrounding bubble, whereas the other two appear incapable of ionizing such extensive regions without additional ionizing photon sources. The detection of Lyman-alpha emission from these galaxies points to varied scenarios concerning the escape of Lyman-alpha during the epoch of reionization (EoR). This detailed case study of LAEs during the EoR underscores the need for further statistical analyses to more fully understand the mechanisms behind Lyman-alpha escape and to identify the primary sources of ionizing photons, whether they be bright or faint galaxies, that contribute to the formation of ionized bubbles in the neutral intergalactic medium.
• Predicting Lyman-alpha from Reionization-Era Galaxies by Machine Learning
![[ML]](./research_files/ML.png)
Accurately predicting emergent Lyman-alpha emission from galaxies during the reionization era is crucial for measuring Lyman-alpha transmission in the intergalactic medium (IGM). This task requires a deep understanding of Lyman-alpha systematics. In this context, a supervised machine learning (ML) approach emerges as a potent tool for investigating these systematics in the high-redshift universe and for making precise predictions of Lyman-alpha emissions from reionization-era galaxies. Such advancements are pivotal in refining the use of Lyman-alpha as a reionization probe. Leveraging an extensive dataset of Lyman-alpha observations from the post-reionization era, I am currently developing ML models that infer Lyman-alpha emissions based on the physical properties of galaxies (refer to the accompanying flowchart). A proof-of-concept analysis using a Random Forest Regressor has already demonstrated the potential of this ML approach. It successfully reproduces observed Lyman-alpha emissions in the MUSE-Wide survey, indicating the feasibility of building an effective ML inference model. The fine-tuning of these ML systems is expected to yield accurate predictions of intrinsic Lyman-alpha emissions from reionization-era galaxies. This will enable us to place more precise estimates of Lyman-alpha transmission in the IGM, thereby providing a critical constraint for understanding the reionization process.
• Earlier Reionization in Regions around UV-luminous Galaxies
![[T_IGM]](./research_files/Boosted_T_IGM.png)
Using the combined spectroscopic dataset of HST grism and Keck DEIMOS+MOSFIRE observations of 275 reionization-era galaxies, we constrained Lyman-alpha equivalent-width (EW) distribution at z > 6 as a function of UV luminosity. We provided the empirical measurements of the IGM transmission, defined as the relative ratio of observed-to-expected (or intrinsic) EWs of Lyman-alpha. We found that the IGM transmission appears to be higher for the UV-luminous galaxies, suggesting that reionization proceeds faster in regions around such galaxies (Jung et al. 2022a, ApJ, 933, 87). This observational finding is consistent with theoretical predictions on Lyman-alpha transmission in the IGM from reionization simulations (Park, IJ et al. 2021, ApJ, 922, 263). The theoretical study shows that UV-luminous galaxies tend to reside in “ionized bubbles” as they ionize large volumes and are located in overdense regions, making it easier for Lyman-alpha photons to escape.
• Texas Spectroscopic Search for Lyman-alpha Emission at the End of Reionization
![[Lya_Survey]](./research_files/TexasSurvey.png)
My doctoral thesis at the University of Texas at Austin, entitled “Constraining the End of Reionization with Lyman-alpha Spectroscopy,” was a pioneering study that utilized Lyman-alpha emission to map the ionization state of the IGM during reionization. I performed statistical analyses on their Lyman-alpha emission properties by analyzing spectroscopic observations of distant galaxies. This work led to a series of publications in the Astrophysical Journal, developing a method to estimate the neutral hydrogen fraction in the IGM, discovering new Lyman-alpha emitters (LAEs) at the highest redshifts, and advancing our understanding of the timeline and topology of reionization (Jung et al. 2018, ApJ, 864, 103; Jung et al. 2019, ApJ, 877, 146; Jung et al.. 2020, ApJ, 904, 144).
• A Spatially-Resolved Stellar Population Study Using the HST Imaging: Reduced star-formation rates in the centers of massive galaxies at z = 4
![[Reduced_SFR]](./research_files/Reduced_SFR.png)
In the PhD program at the University of Texas at Austin, I have dealt with the Hubble Space Telescope imaging data for studying spatially-resolved star formation histories of distant galaxies, which was published, Evidence for Reduced Specific Star Formation Rates in the Centers of Massive Galaxies at z = 4 (Jung et al. 2017, ApJ, 834, 81). In this work, I developed my own galaxy spectral energy distribution fitting IDL procedure based on a Markov Chain Monte Carlo (MCMC) algorithm, and performed a spatially-resolved stellar population study of z = 4 galaxies. A key discovery of the study is that the galaxies with the highest central mass densities have reduced star formation in their centers, possibly observing the earliest phases of bulge formation.
• Building Dark Matter Halo Merger Trees (ySAMtm)
![[Dark Matter Halo Merger Trees]](./research_files/MergerTree.png)
(Left) schematic diagram showing the processes of a dark matter halo merger (Jung et al. 2014, ApJ, 794, 74).
During my master's program, I studied the theoretical understanding of galaxy evolution using N-body simulations and a semi-analytic model with Prof. Sukyoung K. Yi at Yonsei University in Seoul, South Korea. Particularly, I developed a dark matter (DM) halo merger tree builder, called ySAMtm, which constructs DM halo merger trees from cosmological N-body simulations and provides a DM halo backbone for the semi-analytic model of Prof. Yi's group (ySAM; Lee & Yi, 2013). Jung et al. (2014) also study large-scale environmental effects on galaxy assembly by analyzing the semi-analytic model galaxies.