Dark Recipe for the First Giants: From Population III Stars to Early Supermassive Black Holes via Dark Matter Capture

TL;DR

The paper tackles the origin of supermassive black holes at high redshift by proposing that non-annihilating dark matter with non-gravitational Standard Model interactions accumulates inside Population III stars, triggering premature collapse into BH seeds of mass $M_{\rm seed} \sim \mathcal{O}(10-100)\,M_\odot$. Seeds then grow through Bondi accretion under various regimes (Eddington to super-Eddington) to explain SMBHs observed at $z\gtrsim 5$, with the seed-formation timescale largely set by thermalization, $\tau_{\therm}$, and a detailed account of DM capture via SD DM–proton scattering. The authors construct a SMBH mass function by embedding seed formation in a gNFW halo population, calibrating with UniverseMachine and halo growth fits, achieving good agreement with current high-$z$ SMBH and LRD observations after applying a normalization factor $f_{\rm norm}$. They further predict gravitational-wave signals from SMBH mergers—both individual events detectable by LISA and a stochastic background within PTA sensitivities—providing a multi-messenger test for this DM-induced seeding scenario. Collectively, the framework yields testable implications for upcoming direct detection (e.g., XLZD), JWST- and GW-driven surveys, and future GW observatories, linking early-Universe DM physics to observable SMBH demographics and gravitational waves.

Abstract

The presence of supermassive black holes (SMBHs) at high redshifts ($z>5$), as revealed by James Webb Space Telescope (JWST), challenges standard black hole (BH) formation scenarios. We propose a mechanism in which non-annihilating dark matter (DM) with non-gravitational interactions with the Standard Model (SM) particles accumulates inside Population III (Pop III) stars, inducing their premature collapse into BH seeds having the same mass as the parent star. Owing to their early formation, these seeds can accrete for longer periods and grow into the SMBHs observed at early cosmic times. Focusing on spin-dependent (SD) DM-proton interactions, we identify regions of parameter space that account for the observed high-redshift SMBH population, their mass function, and the SMBH-stellar mass relation. Portions of this parameter space are testable by forthcoming direct detection experiments. The scenario may lead to distinctive gravitational wave (GW) signatures from SMBH mergers, accessible to Laser Interferometer Space Antenna (LISA) and pulsar timing array (PTA) observations.

Figures (7)

• Figure 1: Contours of seed BH formation time of $10^5 \, {\rm yrs}$ for the lowest (blue) and the highest (red) mass Pop III stars for DM–proton SD interaction ($\sigma^{\rm SD}_{\chi p}$) are shown along with competitive limits in the parameter space. These two masses represent the lowest and highest masses across all redshifts from Hirano:2015wxa. The gray, pink, and orange shaded regions represent the exclusion limits from combined PICO-60 PICO:2019vsc and LZ LZ:2024zvo, bounds from neutron star transmutation Garani:2018kkd, and the neutrino fog OHare:2021utq, respectively. The purple point denotes the chosen DM benchmark point. The gray dashed curve represents the projected $1 \,$ktonne-year (kt-y) SD sensitivity of XLZD XLZD:2024nsu, obtained by scaling the current SD LZ limits.
• Figure 2: Left: SMBH masses as a function of observed redshifts for our chosen DM benchmark point, shown for Eddington (blue: $\eta_{\rm acc}=1$), sub-Eddington (green: $\eta_{\rm acc}= 0.5$ and gray: $\eta_{\rm acc}= 0.75$), and super-Eddington (red: $\eta_{\rm acc}=1.4$) accretion rates, along with observed SMBHs Inayoshi:2019funKovacs:2024zfhMaiolino:2023zduTaylor2025CAPERSLRDz9AGYang:2021imtGoulding:2023gqaLarson2023ACDKokorev2023UNCOVERANMaiolino:2023bpiFurtak2023AHBNaidu:2025rpoLin2025BridgingQAAkins2024StrongRE. Right: SMBH mass and host galaxy stellar mass relation at $z_{\rm obs} = 6$ is shown for Eddington (blue, $\eta_{\rm acc}=1$), sub-Eddington (green: $\eta_{\rm acc}= 0.5$ and gray: $\eta_{\rm acc}= 0.75$), and super-Eddington (red: $\eta_{\rm acc}=1.4$) accretion rates, together with the observed SMBH and stellar mass data at similar redshifts Stone2023UndermassiveHGYue2023EIGERVCMaiolino:2023bpiHarikane2023AJFDing2022DetectionOSJuodbalis2024ADO.
• Figure 3: SMBH mass function calculated at $z_{\rm obs} = 6.0$ (dashed) and $z_{\rm obs} = 6.5$ (solid) for Eddington (blue: $\eta_{\rm acc}=1$), sub-Eddington (green: $\eta_{\rm acc}= 0.5$ and gray: $\eta_{\rm acc}= 0.75$), and super-Eddington (red: $\eta_{\rm acc}=1.4$) accretion rates, with the normalization factor ($f_{\rm norm}$) equal to $10^{-5.9}$, $10^{-5.6}$, $10^{-6.3}$, and $10^{-6.8}$, respectively. Measurements of the SMBH mass function at similar redshifts Wu:2022njoMatthee:2023utnKokorev2024censusTaylor2025broad are also displayed.
• Figure 4: Characteristic strains from typical SMBH mergers are shown with solid (lowest Pop III mass) and dashed (highest Pop III mass) lines for two values of $z_{\rm in}$ and $z_{\rm obs}$ motivated from the redshift-dependent IMF in Hirano:2015wxa. We have evolved seed BHs with Eddington accretion to form SMBHs with masses of $1.5 \times 10^6 \, M_\odot$ (blue, dashed), $9.7 \times 10^7 \, M_\odot$ (blue, solid), $1.3 \times 10^6 \, M_\odot$ (red, dashed), and $3.3 \times 10^6 \, M_\odot$ (red, solid). The stochastic GW background from Eddington-accreting SMBHs is represented by the olive shaded band. The GW detector sensitivities are illustrated with different black lines. The mass ratios for the individual mergers and stochastic background are taken to be $1$ and $1/3$, respectively.
• Figure S1: The density (left) and temperature (right) profiles as a function of radial distance, obtained using MESA, are shown for Pop III stars with masses ranging from $10 \, M_\odot$ to $800 \, M_\odot$.