Ultra High-Redshift or Closer-by, Dust-Obscured Galaxies? Deciphering the Nature of Faint, Previously Missed F200W-Dropouts in CEERS

TL;DR

This study targets faint F200W-dropouts in the CEERS field to disentangle ultra-high-redshift galaxies from heavily dust-obscured, low-mass systems. By combining two JWST photometric catalogs with CosMix-based stacking and multi-code SED fitting (Bagpipes, CIGALE, EAZY), the authors identify five potential $z>\!8$ candidates (extending to $z>15$ in best-fit solutions) and three nearby HELM-like dusty dwarfs, plus a strong-line emitter CURION at $z\sim5$. They compute a UV luminosity function at $z\sim17$ under high-$z$ assumptions and show results broadly consistent with other studies, while critically examining degeneracies and the need for longer-wavelength data. The work emphasizes the importance of MIR/(sub-)mm photometry and spectroscopy to confirm the nature of these dropouts and to constrain dust production, early galaxy formation, and dark matter implications in the high-redshift regime. The methodology, including CosMix stacking and a multi-SED-fitting approach, provides a robust framework for future high-redshift object searches and contamination assessment in JWST surveys.

Abstract

The James Webb Space Telescope (JWST) is revolutionizing our understanding of the Universe by unveiling faint, near-infrared dropouts previously beyond our reach, ranging from exceptionally dusty sources to galaxies up to redshift $z \sim 14$. In this paper, we identify F200W-dropout objects in the Cosmic Evolution Early Release Science (CEERS) survey which are absent from existing catalogs. Our selection method can effectively identify obscured low-mass ($\log \text{M}_* \leq 9$) objects at $z \leq 6$, massive dust-rich sources up to $z \sim 12$, and ultra-high-redshift ($z > 15$) candidates. Primarily relying on NIRCam photometry from the latest CEERS data release and supplementing with Mid-Infrared/(sub-)mm data when available, our analysis pipeline combines multiple SED-fitting codes, star formation histories, and CosMix - a novel tool for astronomical stacking. Our work highlights three $2<z<3$ dusty dwarf galaxies which have larger masses compared to the typical dusty dwarfs previously identified in CEERS. Additionally, we reveal five faint sources with significant probability of lying above $z>15$, with best-fit masses compatible with $Λ$CDM and a standard baryons-to-star conversion efficiency. Their bi-modal redshift probability distributions suggest they could also be $z<1.5$ dwarf galaxies with extreme dust extinction. We also identify a strong line emitter galaxy at $z \sim 5$ mimicking the near-infrared emission of a $z \sim 13$ galaxy. Our sample holds promising candidates for future follow-ups. Confirming ultra high-redshift galaxies or lower-z dusty dwarfs will offer valuable insights into early galaxy formation, evolution with their central black holes and the nature of dark matter, and/or cosmic dust production mechanisms in low-mass galaxies, and will help us to understand degeneracies and contamination in high-z object searches.

Figures (20)

• Figure 1: RGB mosaic of CEERS DR v.1.0 NIRCam data showing the position of our sample of dropouts, which is described in Section \ref{['4|sec:sampleselection']}. Sources belonging to the Ultra High-Redshift (UHR) galaxy candidates sample are represented by colored star markers (see Section \ref{['uhrproperties']}), while the strong line emitter CURION is depicted with a diamond marker (see Section \ref{['sec:curion']}). All other F200W-dropouts are instead indicated by gray circles. The available CEERS DR 0.6 MIRI pointings overlapping with the EGS NIRCam-covered area are shown as red contours.
• Figure 2: Updated CEERS catalog photometry errors vs CEERS ASTRODEEP-JWST photometry errors for the sources in common between the two catalogs (displayed as gray dots) for a sample band (F356W). The bisector of the plot is represented as a black dashed line surrounded by a 1$\sigma$ confidence interval (shaded black area) representing the typical dispersion of the data points around the bisector. A power-law fit is displayed to showcase how the updated CEERS catalog's errors tend to be higher with respect to CEERS ASTRODEEP-JWST ones.
• Figure 3: [F277W - F444W] color-magnitude diagram for our F200W-dropouts. The F200W-dropout sample's objects are shown as colored stars (UHR candidates), pea green diamonds (CURION, a strong line emitter), cyan crosses (stacked UHR galaxies) or gray circles (all other objects in the sample). For CURION, U-100588 and U-106373 we report 1$\sigma$ upper limits. Black dots in the background represent the colors of CEERS galaxies, with gray-shaded contours enclosing 50% (inner), 80% (middle), and 95% (outer) of the total CEERS source density. The F444W 5$\sigma$ depth is marked as a black dashed line, while the brown shaded area represents the color-magnitude space occupied by LRDs in 2024ApJ...968....4P (albeit such selection requires also F150W-F200W<0.5). Gray error bars represent the average errors for objects with colors that are not upper limits.
• Figure 4: [F150W - F444W] color-magnitude diagram for our F200W-dropouts. The F200W-dropout representation scheme is the same adopted in Figure \ref{['fig:colormagplot1']}, as well as the F444W 5$\sigma$ depth threshold and x-axis average errors. In addition, the HST-dark galaxy selection technique from 2024MNRAS.530..966G is represented by a brown continuous line.
• Figure 5: [F277W - F356W] versus [F200W - F277W] color-color diagram for our sources. The F200W-dropout representation scheme is the same adopted in Figure \ref{['fig:colormagplot1']} and Figure \ref{['fig:colormagplot2']}, as well as the x-axis average errors. The black rectangle highlights the UHR LBG selection for $15 < z < 20$ galaxies by Castellano2025, whereas the blue shaded area corresponds to the $15 < z < 20$ LBG selection adopted in 2024arXiv241113640K.