Euclid: The first statistical census of dusty and massive objects in the ERO/Perseus field

TL;DR

This study uses Euclid ERO imaging in combination with Spitzer IRAC data to identify HST-to-IRAC extremely red objects (HIEROs) in the Perseus field and construct a galaxy stellar mass function (GSMF) at $3.5<z<5.5$. By cross-matching multi-band photometry, applying rigorous contaminant removal, and performing SED fitting with Bagpipes that account for nebular emission, the authors derive physical properties for 42 robust HIEROs and estimate their contribution to the high-mass end of the GSMF. They find very dusty, massive galaxies at early times, with a mean redshift near $z\sim4$ and mean $\log_{10}(M_*/M_\\odot)\approx10.9$, including a subset of overmassive systems that drive the high-mass GSMF. The results imply that dust-obscured star formation could be more abundant at early epochs than previously thought and highlight the need for follow-up spectroscopy and deeper imaging (e.g., Euclid DR1) to robustly constrain the high-mass end and the role of AGN and line emission biases.

Abstract

Our comprehension of the history of star formation at $z>3$ relies on rest-frame UV observations, yet this selection misses the most dusty and massive sources, yielding an incomplete census at early times. Infrared facilities such as Spitzer and the James Webb Space Telescope have revealed a hidden population at $z=3$-$6$ with extreme red colours, named HIEROs (HST-to-IRAC extremely red objects), identified by the criterion $H_{\mathrm{E}}-\mathrm{ch2}>2.25$. Recently, Euclid Early Release Observations (ERO) have made it possible to further study such objects by comparing Euclid data with ancillary Spitzer/IRAC imaging. We investigate a $232$ arcmin$^2$ area in the Perseus field using VIS and NISP photometry, complemented by the four Spitzer channels and ground-based MegaCam bands ($u$, $g$, $r$, ${\rm H}α$, $i$, $z$). Applying the colour cut yields $121$ HIEROs; after removing globular clusters, brown dwarfs, and unreliable cases through visual inspection of multiband cutouts, we obtain a final sample of $42$ robust HIEROs. Photometric redshifts and physical properties are estimated with the SED-fitting code Bagpipes. From the resulting $z_{\mathrm{phot}}$ and $M_*$ values, we compute the galaxy stellar mass function at $3.5<z<5.5$. Even after excluding possible AGN hosts or systems where the stellar mass may be overestimated, the high-mass end remains comparable to previous determinations, suggesting the true abundance could be higher. These results highlight the importance of further study of this obscured population to assess its role in the cosmic star-formation rate density and its consistency with galaxy-formation models, demonstrating Euclid's capability to advance our understanding of dust-hidden star formation across early epochs.

Figures (17)

• Figure 1: RGB image of the Perseus cluster made by combining the images taken through the three NISP filters. The coloured solid lines show the extent of the two Spitzer/IRAC data sets, with the cyan and magenta lines showing the extent of the first and second pointings, respectively. The images have been aligned with the WCS coordinate system, with north up and east to the left. In the low right corner, a scale bar representing $5$ arcmin is shown.
• Figure 2: Colour-magnitude ($\HE-\mathrm{ch2}$ versus ch2) distribution of the matched and Spitzer sample. The yellow star symbols show the objects that satisfy the HIERO colour criterion (shown as a dashed red line). The triangle symbols show lower limits for the -undetected objects, calculated using the $1\,\sigma$ magnitude limit. The other points show the parent sample.
• Figure 3: Example multi-wavelength cutout of a discarded object, due to an artefact in the VIS image. The cutout sizes are $15" \times 15"$.
• Figure 4: Bagpipes fits to the SEDs for three representative objects, IDs $42$, $28$, and $11$. The posterior probability distributions for the redshifts, PDF($z$), are shown as insets. Photometric detections are shown by the coloured circles and 3$\,\sigma$ upper limits are plotted as triangles. The coloured lines show the three different fits described in the text. The PDF for ID $42$ peaks at low redshift, while ID $28$ is bimodal and ID $11$ has a single high-redshift peak, making it representative of the objects that are likely to be HIEROs.
• Figure 5: Properties of the HIEROs. Panel a) shows the redshift versus stellar mass. The red circles show the most massive candidates, with stellar mass greater than $10^{11.7} M_\odot$. These objects are from here on defined as 'overmassive' and shown with the same symbol, i.e., red circles, in all other panels. The solid lines report the minimum observable stellar mass producing an IRAC ch2 magnitude of $22.7$. All three curves represent star-forming galaxies, the green with an age between the redshift epoch and the burst peak of $200$ Myr, the teal one of $100$ Myr, and the dark blue one of $50$ Myr. In panel b) the redshift versus dust attenuation distribution is reported. Panel c) shows the dust attenuation versus stellar mass. The cyan solid line shows the relation from mclure2018dust, while the magenta line delimits the area identifying the so-called HELM galaxies bisigello2025spectroscopic. Different symbols report values from previous studies, as indicated in the legend.