Submillimeter galaxy overdensities around physically associated quasar pairs

TL;DR

This study uses JCMT/SCUBA-2 $450$ and $850\,\mu$m imaging of nine physically associated quasar pairs at $z=2.45$–$3.83$ to probe their megaparsec-scale environments. By constructing differential $850\,\mu$m counts with forward Monte Carlo modeling (Schechter function with $S_0=2.5$ mJy) and validating against blank-field benchmarks, the authors quantify robust SMG overdensities: $\delta_{ m scaled} \approx 2.7 \pm 0.1$ and $\delta_{ m cumul} \approx 3.4 \pm 0.3$, in all fields. They derive cumulative star formation rate densities around $1700 \pm 100\ M_\odot\,\mathrm{yr}^{-1}\mathrm{Mpc}^{-3}$ (integrated) and a central-aperture value of order a few $10^2\ M_\odot\,\mathrm{yr}^{-1}\mathrm{Mpc}^{-3}$, with molecular gas masses broadly matching protocluster simulations such as Magneticum. The SMG overdensities around quasar pairs appear higher than those around single quasars and show little alignment with the quasar-pair axis or Ly$\alpha$ nebulae, suggesting SMGs trace filaments on scales distinct from Ly$\alpha$ emission. The results substantiate the reliability of quasar pairs as tracers of protocluster environments, though spectroscopic follow-up is needed to confirm SMG membership and map the full filamentary structure.

Abstract

A commonly employed method to detect protoclusters in the young universe is the search for overdensities of massive star forming galaxies, such as submillimeter galaxies (SMGs), around high-mass halos, including those hosting quasars. In this work, we study the Megaparsec environment surrounding nine physically associated quasar pairs between $z=2.45$ and $z=3.82$ with JCMT/SCUBA-2 observations at 450 $μ$m and 850 $μ$m covering a field of view of roughly 13.7 arcmin in diameter (or 32 Mpc$^2$ at the median redshift) for each system. We identify a total of 170 SMG candidates and 26 non-SMG and interloper candidates. A comparison of the underlying 850 $μ$m source models recovered with Monte Carlo simulations to the blank field model reveals galaxy overdensities in all fields, with a weighted average overdensity factor of $δ_{\rm cumul} = 3.4 \pm 0.3$. From this excess emission at 850 $μ$m, we calculate a star formation rate density of $1700 \pm 100$ M$_{\odot}$ yr$^{-1}$ Mpc$^{-3}$, consistent with predictions from protocluster simulations and observations. Compared to fields around single quasars, those surrounding quasar pairs have higher excess counts and more centrally peaked star formation, further highlighting the co-evolution of SMGs and quasars. We do not find preferential alignment of the SMGs with the quasar pair direction or their associated Ly$α$ nebulae, indicating that cosmic web filaments on different scales might be traced by the different directions. Overall, this work substantiates the reliability of quasar pairs to detect overdensities of massive galaxies and likely sites of protocluster formation. Future spectroscopic follow-up observations are needed to confirm membership of the SMG candidates with the physically associated quasar pairs and definitively identify the targeted fields as protoclusters.

Figures (10)

• Figure 1: Magnitude and projected distance of the quasar pairs presented in this work (squares) and found in catalogs (circles). The points are color-coded by redshift.
• Figure 2: SNR maps at 850 $\mu$m. The position of the brighter (fainter) quasar is indicated by a darkgreen (lightgreen) star. Identified catalog SMGs are marked by black circles. The grey dashed line indicates the edge of the effective area (2.5 $\times$ the central rms). Sources inconsistent with the expected SMG SED at the quasar redshift are marked with grey rectangles.
• Figure 3: Completeness percentage of the source extraction as a function of injected source flux in the 450 $\mu$m band (left) and the 850 $\mu$m band (right).
• Figure 4: Differential number counts at 850 $\mu$m in the data (black points) and the simulated maps (green points). The underlying source model of the quasar pair fields (blue dashed) is shown in comparison to the field model (black dashed) with the 1$\sigma$ spread of the differential number counts from 500 realizations of the source model (yellow shaded area) and the fiducial model (grey shaded area).
• Figure 5: Cumulative SFR of the quasar pair fields compared to the ten protocluster zoom-in regions simulated in FOREVER22 at $z = 3$Yajima2022, the 32 most massive halos from Magneticum at $z=2.79$Remus2023Dolag2025, and known protoclusters from the literature (Wang2025 and references therein). The dotted line shows the cumulative SFR of the mean blank field in FOREVER22 Yajima2022.