The Balmer Break and Optical Continuum of Little Red Dots from Super-Eddington Accretion

TL;DR

This work demonstrates that the Balmer break and red optical continua of Little Red Dots can arise naturally from super-Eddington accretion onto relatively low-mass black holes. By combining analytic atmosphere models with gray and multigroup radiation transport, the authors compare two simple geometries: a truncated thin disk and a geometrically thick, roughly spherical inflow/outflow. The disk can produce a Balmer break only with fine-tuned inner temperatures around Teff ≈ 5000 K, while the sphere scenario yields a robust 4000–6000 K photosphere and a strong Balmer break across a broad range of densities, matching observed LRD spectra without invoking dust or external absorption. These results imply a physically plausible LRD population powered by high accretion rates onto low-mass black holes and point to future three-dimensional radiative-transfer studies to refine predictions and test against observations across the UV to infrared.

Abstract

The physical origin of Little Red Dots (LRDs)--compact extragalactic sources with red rest-optical continua and broad Balmer lines--remains elusive. The redness of LRDs is likely intrinsic, suggesting optically thick gas emitting at a characteristic effective temperature of $\sim5000{\rm~K}$. Meanwhile, many LRD spectra exhibit a Balmer break, often attributed to absorption by a dense gas shell surrounding an AGN. Using semi-analytical atmosphere models and radiation transport calculations, we show that a super-Eddington accretion system can give rise to a Balmer break and a red optical color simultaneously, without invoking external gas absorption for the break or dust reddening. The break originates from a discontinuity in opacity across the Balmer limit, similar to that of early-type stars, but the lower photosphere density of super-Eddington systems, $ρ<10^{-9}{\rm~g~cm^{-3}}$, implies a significant opacity contrast even at a cool photosphere temperature of $\sim5000{\rm~K}$. Furthermore, while accretion in the form of a standard thin disk requires fine tuning to match the optical color of LRDs, an alternative scenario of a geometrically thick, roughly spherical accretion flow implies an effective temperature $4000{\rm~K}\lesssim T_{\rm eff}\lesssim6000{\rm~K}$ that is very insensitive to the accretion rate (analogous to the Hayashi line in stellar models). The continuum spectra from the latter scenario align with the Balmer break and optical color of currently known LRDs. We discuss predictions of our model and the prospects for more realistic spectra based on super-Eddington accretion simulations.

Figures (12)

• Figure 1: Cartoon showing the two scenarios considered in this work. Straight or curved arrows indicate direction of gas flow. Dashed lines represent the photosphere. Wiggle arrows indicate photons. Left: A standard thin disk in super-Eddington accretion, which we investigate in Section \ref{['sec:disk']}. It reproduces the red optical color and Balmer break of LRDs if the disk has an inner truncation radius where the effective temperature is $\sim5000{\rm~K}$, which is fine-tuned. Right: A spherical gas profile undergoing turbulent accretion, which we investigate in Section \ref{['sec:sphere']}. This model gives rise to a red optical continuum and the Balmer break over a large range of gas densities (a proxy for a wide range of accretion rates). Components with question marks, i.e., truncation in the disk scenario and outflow and clumps in the sphere scenario, are discussed in Section \ref{['sec:conclusions']}.
• Figure 2: The ratio of two continuum effective opacities (accounting for absorption and scattering), $\kappa_{{\rm eff}}(4000{\rm~\AA})/\kappa_{{\rm eff}}(3600{\rm~\AA})$, for different temperature and density. A ratio significantly below one tends to produce a Balmer break. Star symbols mark the parameters corresponding to the star Vega and the Sun. The optimal temperature to produce a Balmer break increases with density.
• Figure 3: Workflow in this paper. Methods are outlined in yellow blocks, and results are outlined in blue blocks. "Sec.", "app.", and "cond." stand for "section", "appendix", and "condition".
• Figure 4: Model continuum spectra in the disk scenario. Left: the $M_{\rm BH}=10^7~M_\odot, \lambda_{\rm Edd}=10^2$ case. Each curve shows the flux integrated over disk annuli from $R_{\rm in}$ to $R_{\rm out}=10^{4.3}~R_{\rm g}$. The legend marks the inner radius, $R_{\rm in}$, and the effective temperature at the inner radius, $T_{\rm eff,in}$, derived from Equation (\ref{['eq:disk_Teff']}). A blackbody curve at $T=5000{\rm~K}$ and the spectrum of an LRD, RUBIES-UDS-31747, are shown for comparison. A collection of observed LRD spectra are shown in gray in the background. Right: Same as left, but for the $M_{\rm BH}=10^7~M_\odot, \lambda_{\rm Edd}=10^0$ case, with $R_{\rm out}=10^{3.6}~R_{\rm g}$.
• Figure 5: Effective temperature of the sphere model with $L=L_{\rm Edd}(10^6~M_\odot)$. The horizontal axis is the reference density, $\rho_{\rm ref}$, a proxy for the Eddington ratio (Equation (\ref{['eq:sphere_Eddington_ratio']})) and the photosphere density ($\rho_{\rm ref}\sim\rho_{\rm ph}$ if $T_{\rm eff}\sim5000{\rm~K}$). Curves show analytical estimates using the effective opacity at $5000\rm~\AA$ (solid) or the Rosseland mean opacity (dotted). These estimates depend on the dimensionless photosphere scale height, $H_{\rm ph}/R_{\rm ph}$, as a free parameter. Plus signs mark measurements of the gas temperature at $\tau_{\rm eff}(5000{\rm~\AA})=1/2$ from the gray simulations, where the photosphere scale height is determined self-consistently by radiation energy transport. The analytical effective temperature is very insensitive to the reference density for a given luminosity and $H_{\rm ph}/R_{\rm ph}$, although numerical measurements suggest a somewhat steeper $T_{\rm eff}-\rho_{\rm ref}$ relation.