Stellar Morphology of Optically Dark or Faint Galaxies at $z>3$ with JWST

TL;DR

This work uses JWST/NIRCam imaging to characterize the stellar morphology of optically dark or faint galaxies (OFGs) at $3<z<4$. By constructing a large photometric catalog and performing Bayesian Sérsic fits on both OFGs and a parent star-forming population, the authors quantify differences in mass, dust attenuation, sizes, and shapes, and place OFGs within the mass-size and Sigma$_{R_e}$–mass planes. They find OFGs are ~8–9× more massive and ~4× more dust-obscured than typical SFGs, while having similar effective radii and Sérsic indices, though they tend to be rounder. The results indicate that dust attenuation is primarily driven by stellar mass, with OFGs likely representing a diverse, high-mass extension of the SFG population rather than a distinct, rapidly evolving transitional pathway toward quiescence. This underscores the importance of JWST in revealing the structure and dust properties of massive high-$z$ galaxies and refining our understanding of early galaxy evolution.

Abstract

JWST offers an unprecedented view of optically dark or faint galaxies (OFGs), previously missed by HST. They are likely massive, heavily dust-obscured star-forming galaxies (SFGs) that substantially contribute to the cosmic SFR density at $z>$3. To identify drivers of their high dust attenuation and their role in early universe galaxy evolution, we analyse the stellar morphology of 65 OFGs (from 1892 SFGs at 3$<z<$4) using NIRCam/F444W imaging from the PRIMER and CEERS fields. We study correlations between dust attenuation ($A_v$) and galaxy properties, like stellar mass, size, and orientation, and compare scaling relations between OFGs and typical SFGs. We find that OFGs are ~8-9 times more massive and ~4 times more dust attenuated than the parent sample. Structurally, OFGs resemble parent SFGs in median $R_e$ and median $Σ_{R_e}$ but may be slightly rounder on average. While $A_v$ strongly correlates with stellar mass, it does not show significant dependence on stellar mass-normalised effective radius and stellar mass surface density, Sérsic index, axis ratio, or SFR surface density. The mass-size and mass-surface density relations place OFGs as a higher-mass extension of SFGs, with no concrete proof of evolutionary differences between them. This suggests that OFGs are heavily dust-obscured primarily due to their high stellar masses, which facilitates dust production and retention, with older stellar populations likely contributing as well. Although some OFGs exhibit high $Σ_\mathrm{R_e}$ and occupy regions of the mass-size plane similar to quiescent galaxies, the overall sample is not representative of this. Their current structures resemble typical SFGs, with no concrete signs of rapid compaction. Diversity in their physical properties shows that OFGs span a range of evolutionary states with few showing reduced star formation, while most remain actively star-forming.

Figures (10)

• Figure 1: The sample selection from the CEERS, PRIMER-COSMOS, and PRIMER-UDS fields. Sources are identified based on their red colors in F150W-F444W versus their brightness observed in the F444W band. The red lines show the colour and magnitude cuts in F150W and F444W used to select OFGs. The grey empty circles indicate the entire parent sample of SFGs at $z > 3$. The red empty circles indicate all OFGs at $z > 3$ which are very well detected in both F150W and F444W bands while the red empty upward triangles indicate sources with $2\sigma$ lower limits in their color. The final parent sample for $3< z < 4$ is shown with black filled circles and the final OFG in red filled circles and filled upward triangles. Note that all OFGs are part of the parent sample. The shaded area denotes the magnitude limit of F444W $=26$ mag used for the final selection of galaxies based on results from simulation tests (see appendix \ref{['pysersic_limit']}). This cut allows us to recover reliable morphological measurements.
• Figure 2: UVJ diagram Williams2009 for the final sample of OFGs (diamonds) and parent SFGs (dots). Sources that were found within the quiescent locus (in red) of the UVJ selection criteria, based on their rest frame colors, were removed from the final sample. The sources are colour coded with respect to their dust attenuation magnitudes ($A_{v}$ mag) obtained through methods mentioned in Section \ref{['photometric catalogue']}.
• Figure 3: Top panel: Histogram showing stellar mass (left), dust attenuation (middle), and redshift (right) of the parent SFG (grey) and OFG (red) samples. Bottom panel: Histogram showing effective radius (left), Sérsic index (middle), and axis ratio (right) of the parent SFG (grey) and OFG (red) samples. For both panels, the 50th (median), 16th, and 84th percentiles of the distributions are shown in the insets above each histogram. The y-axis on both panels represent the fraction of the number of sources attributed to each histogram bin with respect to the total number of sources for the given population.
• Figure 4: Left panel: Rest-frame optical mass-size relation for parent SFGs (grey points and black lines) and OFGs (red) for $3 < z < 4\, (F444W)$. The down-pointing arrows represent sources with $R_e$ smaller than the PSF HWHM or the resolution limit in F444W. These arrows indicate that the resolution limit is treated as an upper limit on their size measurements. The teal lines represent the mass-size relation at $3 < z < 4\, (F444W)$ obtained from JWST Rest-frame UV/Optical sizes allen2024galaxysizemassbuildup, while the orange lines denote those from Varadaraj2024 at $z = 3\, (F356W)$ for reference. The purple lines indicate the mass-size relation of quiescent galaxies (QGs) at $z< 3.3$ito2024sizestellarmass. Dashed lines show the 1$\sigma$ scatter of the lines from the literature. The shaded light red and grey regions show the $1\sigma$ scatter of the best-fit relations of OFGs and parent SFGs, respectively. The light green line highlights the resolution limit of the NIRCam PSF in the F444W band. Right panel: Here are sample RGB images from JWST/NIRCam (F115W, F277W, F444W) that effectively illustrate the distinctions between parent SFGs and OFGs. The mass-size relation derived for the parent sample, OFGs and QGs is also underlaid for comparison. Notably, the differences in morphology and color are striking across these categories. OFGs exhibit significantly redder colors compared to the parent sample. The parent sample displays a diverse range of morphologies, spanning from compact to more extended structures. In contrast, OFGs are characterised by predominantly round or ellipsoidal shapes.
• Figure 5: Redshift-binned effective stellar surface mass density $\Sigma_\mathrm{R_{e}}$ vs. stellar mass. Grey points represent SFGs, while red points represent OFGs. The upward pointing arrows represent the same sources that lie below the resolution limit in F444W as in Figure \ref{['fig:combined-mass-size']}. The purple solid lines indicate the $\Sigma_\mathrm{R_{e}}$-mass relation of QGs at $2 < z < 3$, with the dashed purple lines showing the intrinsic scatter for this population, as derived from Barro2017. The green solid and dashed lines represent the redshift-extrapolated best-fit lines and intrinsic scatter for SFGs from the same study. Black lines denote the $\Sigma_\mathrm{R_{e}}$-mass relation for the parent SFG sample at the indicated redshifts (see legend). The red line represents the $\Sigma_\mathrm{R_{e}}$-mass relation for OFGs ($3 < z < 4$), with the shaded light red and grey regions showing the $1\sigma$ scatter of the best-fit relations for OFGs and parent sample, respectively.