Euclid Quick Data Release (Q1): Identification of massive galaxy candidates at the end of the Epoch of Reionisation

TL;DR

This study searches for the most massive galaxies at the end of the Epoch of Reionisation by identifying $M_* > 10^{10.25}$ M$_\odot$ systems at $z\in[5,7]$ over ~23 deg$^2$ in Euclid Q1 EDF-N and EDF-F, combining Euclid NISP+VIS with Spitzer/IRAC and ground-based data. Using LePHARE and BAGPIPES SED fitting, the authors derive photometric redshifts and stellar masses, validating redshifts against ~42k spectroscopic measurements and finding $f_{out} = 9.6\%$ with $\sigma_{NMAD}=0.045$; they identify 145 robust massive candidates, including 5 with $M_* > 10^{11}$, yielding a surface density of $\sim 6.3$ deg$^{-2}$ (a lower limit). The bulk of the sample sits on the star-formation main sequence in the SFR–$M_*$ plane, with a minority in the starburst or transition regions; the inferred properties include dust attenuation up to $E(B-V)\approx0.75$ and ages approaching the Universe's age at those redshifts. Extreme value statistics show no tension with standard halo mass functions or star-formation efficiencies, consistent with $\Lambda$CDM. The results imply that a non-negligible population of massive, dusty, and sometimes old galaxies existed by $z\sim5$–7, but deeper imaging and spectroscopy are needed to fully confirm their masses and star-formation histories.

Abstract

Probing the presence and properties of massive galaxies at high redshift is one of the most critical tests for galaxy formation models. In this work, we search for galaxies with stellar masses M* > 10^10.25 Msun at z in [5,7], i.e., towards the end of the Epoch of Reionisation, over a total of ~23 deg^2 in two of the Euclid Quick Data Release (Q1) fields: the Euclid Deep Field North and Fornax (EDF-N and EDF-F). In addition to the Euclid photometry, we incorporate Spitzer Infrared Camera (IRAC) and ground-based optical data to perform spectral energy distribution (SED) fitting, obtaining photometric redshifts and derived physical parameters. After applying rigorous selection criteria, we identify a conservative sample of 145 candidate massive galaxies with M* > 10^10.25 Msun at z in [5,7], including 5 objects with M* > 10^11 Msun. This makes for a surface density of about 6.3 deg^-2 at z in [5,7], which should be considered a lower limit because of the current depth of the Euclid data (H_E < 24, 5 sigma in Q1). We find that the inferred stellar masses are consistent with galaxy formation models with standard star-formation efficiencies. These massive galaxies have colour excess E(B-V) values up to 0.75, indicating significant dust attenuation in some of them. In addition, half of the massive galaxies have best-fit ages comparable to the age of the Universe at those redshifts, which suggests that their progenitors were formed very early in cosmic time. About 78% of the massive galaxies lie on the star-forming main sequence (MS) in the SFR-M* plane, ~12% are found in the starburst region, and 10% in the transition zone between the MS and starbursts. We find no significant evidence for outshining or AGN contamination that could account for the elevated specific star-formation rates (sSFR) observed in the ~12% of galaxies classified as starbursts.

Figures (6)

• Figure 1: Photometric redshifts that we derive in our work versus the spectroscopic redshifts from the literature available in EDF-N and EDF-F. For a more detailed description of the spectroscopic sample, we refer the reader to Q1-TP005. We report an outlier fraction of $9.6\%$ over $\sim 42\,000$ galaxies in $z\in(0,6)$ at all magnitudes. We show the identity as a red continuous line, together with the shaded area delimiting catastrophic outliers, defined as $|z_{\rm phot}-$$z_{\rm spec}|/$$(1+z_{\rm spec})>0.15$. These cases are highlighted in red colour.
• Figure 2: Distributions of our massive galaxy sample (red histograms, only for $M_* > 10^{10.25},\mathrm{M_\odot}$) compared to all Euclid Q1 galaxies (black-outlined histograms, not individually visually inspected) in the redshift range $z\in[5,7]$. From left to right and top to bottom, we show the distributions of photometric redshift, stellar mass, best-fit age, and colour excess.
• Figure 3: Size–stellar mass relation for VIS-detected galaxies in our sample (36 objects), showing the Sérsic effective radius from modeling as a function of stellar mass. Our high-mass galaxies are represented by red squares. Observational relations from recent JWST-based studies are included as shaded regions: blue for ward_evolution_2024 and purple for yang_cosmos-web_2025. The minimum effective radius () at the average redshift of our sample is shown as a gray line.
• Figure 4: Left panel: Stellar mass ($M_*$) versus photometric redshift. Our massive galaxy candidates are shown as red squares, overlaid on a gray hexagonal density plot representing the Q1 sample. The 5 candidates above $M_* >10^{11}\, \rm{M_\odot}$ are shown with a black outline. For comparison, we include massive galaxies reported in the literature: orange triangles from chworowsky_evidence_2024, green hexagons from xiao_accelerated_2024 (with black outlines highlighting the three most massive objects S1, S2, and S3) and a pink star from xiao_panoramic_2025. Right panel: Zoom-in of the left panel, focusing on our massive galaxy candidates shown as red squares. Confidence intervals from the Extreme Value Statistics (EVS) model lovell_extreme_2023, computed for a survey area comparable to ours, are shown as blue shaded regions, by assuming a lognormal distribution of the star-formation efficiency (SFE). The gray shaded area at the top represents the absolute upper limit under the assumption of 100% SFE.
• Figure 5: UVJ diagram for the massive galaxy sample (red squares), with the full Q1 population at $z\in[5,7]$ represented as a gray hexagonal kernel density overlay. The galaxy colours for this diagram have been computed from LePHARE-derived rest-frame magnitudes.