The Emergence of Little Red Dots from Binary Massive Black Holes

TL;DR

The v-shaped spectral energy distributions of little red dots (LRDs) at $z\sim6-8$ are explained by a binary massive black hole system surrounded by a circumbinary disk and two mini-disks. The turnover near the Balmer limit arises from the Wien tail of the inner edge of the circumbinary disk, while the UV is powered by the hotter mini-disks, allowing modest dust attenuation ($A_V\lesssim1$) and avoiding the need for heavy reddening. Analytic and numerical SEDs reproduce observed LRD colors and predict spectral evolution as binaries inspiral, with testable multi-messenger signatures in gravitational waves detectable by LISA/TianQin and in the stochastic GW background, potentially reconciling the Sołtan argument with LRD demographics. The model also accounts for the high LRD abundance at $z\sim6-8$, fits individual extreme cases (e.g., A2744-45924), and provides a framework linking early AGN triggering to later GW-dominated coalescences, thereby enriching the co-evolution narrative of BH growth and cosmic structure.

Abstract

Little red dots (LRDs) are a newly identified class of broad-line active galactic nuclei (AGN) with a distinctive v-shape spectrum characterized by red optical and blue UV continuum emission. Their high abundance at redshifts of $z\sim6-8$ and decline at lower redshifts suggest a transient origin. We propose that the spectral shape of LRDs originates from compact binary black hole systems, where each black hole is surrounded by a mini-disk and embedded in a larger circum-binary disk. With a binary separation of $\lesssim 10^3$ Schwarzschild radii, the Wien tail of a $T\simeq 5000~{\rm K}$ blackbody spectrum at the inner edge of the circum-binary disk produces the red optical emission, while the mini-disks power the UV continuum. Binary torques carve out a gap between the circum-binary disk and mini-disks, setting the turnover wavelength of the v-shaped spectrum around the Balmer limit. This scenario naturally reproduces LRD spectra requiring only modest dust attenuation ($A_V\lesssim 1$ mag), resolving overestimated luminosities for LRDs in previous studies and alleviating a tension with the so-called Soltan argument. This model predicts a distinct spectral evolution as the binary orbit decays through binary-disk interactions and gravitational waves (GWs), linking early-stage "proto-LRD" binaries to the broader AGN population and late-stage "LRD-descendants" to coalescing binaries detectable in GW experiments.

Figures (7)

• Figure 1: A schematic illustration of a binary BH system accreting from the circum-binary disk and mini-disk around each BH. The expected SED, composed of the colder circum-binary disk and the hotter mini-disks, naturally produces a characteristic v-shape, consistently observed in LRDs.
• Figure 2: Turnover wavelength ($\lambda_{\rm t}$) characterizing the v-shaped LRD spectrum for each model. The red shaded region and thick curves present the binary separations where $\lambda_{\rm t}=3000-5000~{\rm \AA}$.
• Figure 3: Radial structures of the surface temperature of accretion disks around a binary BH with a total mass of $M=10^7~M_\odot$, mass ratio $q=0.1$, and orbital separation $a=300~R_{\rm S}$, for three components: the primary mini-disk (purple), secondary mini-disk (green), and circum-binary disk (cyan). Each component is shown for different mass accretion rates of $\dot{m}_0=1$ (solid), $3$ (dashed), and $10$ (dotted). For the mini-disks, the horizontal axis denotes the radial distance from each BH, normalized by the Schwarzschild radius of the total BH mass. For $1\leq \dot{m}_0 \leq 10$, the temperature profiles of both the circum-binary disk and the primary mini-disk follow $T_{\rm eff} \propto r^{-5/8}$, consistent with a thin disk model including mass loss characterized by $p = 0.5$. The secondary mini-disk follows $T_{\rm eff} \propto r^{-1/2}$ in the slim disk regime where advection cooling dominates radiative cooling. Its inner-edge shifts depending on the accretion rate (see text).
• Figure 4: Spectra of a binary BH system with $M=10^7~M_\odot$, $q=0.1$, $\dot{m}_0=1.0$, and $p=0.5$. Left: The case with $a=10^3~R_{\rm S}$. The attenuation-free SED for the primary mini-disk (purple), secondary mini-disk (green), circum-binary disk (cyan), and the total system (black dashed). The black solid curve shows the SED with modest attenuation ($A_V=0.3$ mag), yielding UV and optical slopes of $\beta_{\rm UV}= -1.2$ and $\beta_{\rm opt}= 0$ (gray lines), consistent with the color selection criteria for LRDs. Right: SEDs for various binary separations in the range over $a=10^2-10^4~R_{\rm S}$ (solid curves). For reference, the dashed curves present the circum-binary disk spectra for the smallest and largest separations. As the orbital separation shrinks, the surface temperature of the circum-binary disk increases, shifting the peak wavelength toward shorter values.
• Figure 5: A spectral model for the brightest LRD, A2744-45924, at $z_{\rm spec}=4.46$ with NIRCam photometry and PRISM data Labbe_2024b and MIRI photometry and ALMA flux upper limits Setton_2025b. The binary BH-disk SED model assumes parameters $M=2\times 10^8~M_\odot$, $q=0.1$, $\dot{m}_0=3.0$, $p=0.5$, $a=300~R_{\rm S}$, and $A_V=0.75$ mag, and reproduces the observed SED with only moderate attenuation, as the Wien-tail spectrum of the circum-binary disk is intrinsically red in the rest-frame optical bands. Blackbody spectra with dust temperatures of $T_{\rm dust} = 400~{\rm K}$ (orange) and $150~{\rm K}$ (brown) are overlaid, with normalizations keeping photon energy conservation. Given that the total IR luminosity is limited to $\sim 2\times 10^{11}~L_\odot$, the IR spectrum of warm dust emission is compatible with the MIRI and ALMA observations.