Hunting the first Cosmic Giants: formation and detectability of Direct Collapse Black Holes around high-redshift quasars

TL;DR

This work investigates the formation and detectability of direct collapse black holes (DCBHs) in the overdense environments that host bright quasars at high redshift. By coupling high-resolution dark matter merger trees with the Cosmic Archaeology Tool semi-analytic model, the authors show that Local Lyman–Werner radiation fields can trigger DCBH formation as early as $z\sim22$, yielding tens of seeds per quasar overdensity, though metal pollution later suppresses new seeding. A significant fraction of these seeds survive as satellites to $z\sim7$, with some remaining in the outskirts and others migrating inward; the study then assesses their Spectral Energy Distributions and evaluates JWST detectability, highlighting that AGN variability and host-galaxy contamination complicate identification but that targeted JWST imaging and spectroscopy could reveal these elusive remnants. The results provide a theoretical framework to test heavy-seed formation around the most massive high-z halos and to constrain the early SMBH seeding channel through forthcoming JWST surveys. Overall, the paper links seed formation physics to observable signatures in quasar environments, offering practical observational strategies and clarifying the role of local radiation fields and metal enrichment in shaping the early BH population.

Abstract

The rapid emergence of supermassive black holes (SMBHs) in the early Universe poses a challenge to current models of black hole growth. One promising formation pathway is the direct collapse black hole (DCBH) scenario, in which gas in pristine, low-metallicity halos forms supermassive (or quasi-) stars leading to massive black holes seeds under specific environmental conditions. In this work, we investigate the potential host environments of DCBHs by coupling a semi-analytic model tracing BH formation and galaxy co-evolution with high-resolution N-body dark matter merger trees. This allows us to trace the population of DCBHs formed during the hierarchical assembly of a $\sim 10^{12} ~\rm M_\odot$ dark matter halo hosting a bright $10^9 ~\rm M_\odot$ quasar at redshift $z \approx 7$. We find that, when accounting for local fluctuations in the UV radiation field within this early cosmic structure, massive BH seeds can form via direct collapse as early as $z \approx 22$. Even under more stringent conditions for heavy seed formation, tens of DCBHs are predicted to emerge within the simulated overdensity down to $z \sim 14$, at which point metal enrichment of the intergalactic medium inhibits further episodes of direct collapse. A significant fraction of the massive black hole population formed at $z > 14$ is expected to survive in satellite galaxies that do not merge with the central halo down to $z \approx 7$. We show that the existence of such a population of ungrown heavy BH seeds can be probed through deep JWST observations targeting regions surrounding bright high-redshift quasars, and we discuss tailored observational strategies to detect and identify these elusive systems.

Figures (8)

• Figure 1: Evolution of the properties of the main simulated galaxy. From top to bottom we show the BH mass, stellar mass, SFR, and BH accretion rate. cat predictions (orange lines) are compared with the results of the hydrodynamical simulation by lupi2019 (gray lines). For reference, the stellar mass panel also shows the DM halo mass growth and the total stellar mass assembled in the simulated overdensity with dotted and dashed lines, respectively.
• Figure 2: Projections in the XY and XZ plane of the spatial distribution of all the quasar-host progenitor halos at $z=15$ (for a total number of $\rm N_{\rm halos} = 30649$), colour coded according to the LW flux illuminating each galaxy. Distances are reported in physical kpc. Regions where the LW intensity is above the threshold for DCBH seed formation, $J_{\rm crit} = 300 ~ J_{ \rm 21}$ are marked with red colours. Green points show the position of halos hosting a heavy seed descendant.
• Figure 3: Redshift distribution of newly formed heavy BH seed for different model variants. Panel (a) shows the results for a homogeneous LW field (blue), while the results including the resolved local flux seen by each galaxy are shown in panels (c) and (e) for a critical threshold of, respectively, $\rm J_{LW, crit} = 300 ~J_{21}$ (orange) and $1000 ~J_{21}$ (red). Panel (b) shows the comparison between the corresponding cumulative seed distributions. The distribution of incident LW fluxes at DCBH formation vs redshift is shown in panels (d) and (f) for, respectively, local LW flux models with $\rm J_{LW,crit} = 300$ (orange points) and $1000$ (red points), where the blue-dashed line marks the background LW flux evolution within the simulated overdensity.
• Figure 4: Number of heavy seeds formed in the simulation as a function of redshift (upper panel) and related cumulative distribution (lower panel). Orange and yellow histograms show, respectively, the fraction of DCBH host halos having - at the time of seed formation - the closest star forming galaxy at a distance larger than 2 and 7 physical kpc.
• Figure 5: Projection of the simulated overdensity at $z \approx 7.5$, with distance from the centre of the structure marked in physical kpc and arcseconds. Star forming galaxies are marked with yellow points, while halos hosting heavy seed descendants or SMBHs ($\log(\rm M_{\rm BH}/M_\odot) > 6.5$) are highlighted as red circles and cyan pentagons, respectively. For illustrative purposes, the size of each point is scaled according to the stellar mass of the corresponding galaxy. The white smooth background points represent the underlying distribution of dark matter halos, most of which do not meet the conditions for triggering star formation. The green square highlights the system ID48871, for which we analysed the potential AGN detectability in the following sections. For reference, we overplot as a white box the $3" \times 3"$ field of view of IFU observations centred on the central quasar, and we also report the scale corresponding to $0.1 \times$ the NIRCam FoV, to highlight how the entire simulated overdensity would be encompassed within the $2' \times 2'$ coverage of a single NIRCam module.