Evolutionary Tracks and Spectral Properties of Quasi-stars and Their Correlation with Little Red Dots

TL;DR

The study tackles the origin of rapid supermassive black hole growth and the nature of JWST-detected Little Red Dots (LRDs) by modeling quasi-stars—radiation-supported envelopes around accreting black holes—within the MESA-QUEST framework. It generates synthetic spectral energy distributions (SEDs) using MESAcolors/ATLAS9, couples these to observed high-redshift host galaxies, and compares the continua to two LRDs, finding that a fiducial late-stage quasi-star with total mass ~10^6 M⊙ can reproduce key features such as large Balmer breaks and flat NIR slopes; matching more extreme LRDs may require BH masses of a few×10^6 M⊙. The work provides scaling relations to estimate stellar mass from bolometric luminosity and argues that the short quasi-star lifetimes could explain the observed LRD number density, supporting the direct-collapse route as a plausible SMBH formation channel. However, the current SED modeling is limited by LTE atmospheres and the absence of detailed emission-line physics, pointing to future non-LTE radiative-transfer work to refine the connections between quasi-stars and LRDs. If confirmed, LRDs could represent a fleeting yet pivotal phase in the birth of the first supermassive black holes in the early universe.

Abstract

JWST has revealed a population of red, compact, high-redshift (${z\sim3-10}$) objects referred to as ``Little Red Dots'' (LRDs). These objects exhibit unusual spectral features reminiscent of stellar spectra with blackbody-like SEDs, large hydrogen Balmer breaks, Balmer line absorption, and classical stellar absorption features such as calcium H&K and the calcium triplet. Following the recent suggestion that these may be actively accreting direct-collapse black holes in the process of assembly, i.e. quasi-stars, we present evolutionary models of quasi-stars using our recently released, publicly available MESA-QUEST modeling framework. We compute a grid of models spanning a range of black hole masses and predict the luminosities, temperatures, surface gravities, and lifetimes of these objects. We find that these models lie along a Hayashi track once they hit their ``late-stage'' which constitutes the majority of their lives ($\sim 20$~Myr). We present scaling relations for estimating the mass of a quasi-star as a function of the bolometric luminosity, as well as the bolometric luminosity as a function of the effective temperature for the Hayashi track. The short lifetimes in tandem with the observed number density of LRDs imply the possibility that every supermassive black hole was once a quasi-star. We compare synthetic spectra of our quasi-star models to observations of LRDs, and show that these models are broadly capable of reproducing the continuum spectra of observed LRDs. These results indicate that quasi-stars are promising candidates for the origin of supermassive black holes via direct collapse in the early universe.

Figures (4)

• Figure 1: Evolution of a theoretical quasi-star model in Hertzsprung-Russel (left) and Kiel (right) diagrams with initial parameters $M_\mathrm{BH}=10$ M$_\odot$, M$_\star$$=10^6$ M$_\odot$, and $Z=0$. After an initial contraction phase lasting $\sim0.1$ Myr, the model spends the remainder of its lifetime ($\sim$20 Myr) at $\sim 3.72\cdot 10^{10}~\textrm{L}_\odot$ and $\sim6000$ K before becoming a $\sim10^6$ M$_\odot$ SMBH. Black hole mass is shown with a color gradient. Lines of constant radius are shown in gray, labeled above the top axis. Note that while these lines appear vertical, this is simply due to the small luminosity range. The gray region in the Kiel diagram represents the log($g$)--$T_\textrm{eff}$ space covered by the ATLAS9 atmosphere models. The star represents the beginning of the quasi-star's late-stage, approximately where the luminosity peaks on the HR diagram and where it spends the majority of its life.
• Figure 2: Evolution of three quasi-stars with masses $10^4$ M$_\odot$ (blue), $10^5$ M$_\odot$ (green), and $10^6$ M$_\odot$ (red). A Hertzsprung-Russel diagram is shown on the left, and the black hole mass growth with age on the right. The start and end points in the HR diagram are marked with circles and squares, respectively. The beginning of the late-stage period tracing a Hayashi track is shown with star symbols, and the fit is depicted with a black dashed line. Lines of constant radius are shown in gray and labeled above the top axis.
• Figure 3: A set of quasi-star models at $M_\textrm{BH} \approx 0.1 M_\star$, which is the point at which the quasi-star has reached what we consider the "late-stage" where it remains predominantly static within the HR diagram. The gray shaded region indicates the space covered by the ATLAS9 grids. Additionally, these grids span $-5 \leq \textrm{[Fe/H]} \leq 1.5$.
• Figure 4: Predicted SEDs of a $10^6$ M$_\odot$ total mass, late-stage quasi-star both on its own (black dashed line) and embedded in a host galaxy (solid blue line) compared with JWST observations of LRDs (red) that includes the emission from the host stellar component with masses $\sim10^{9.5}$ M$_\odot$. The top row shows comparisons with UNCOVER-45924 labbe2024unambiguousagnbalmerbreak and the bottom row with MoM-BH*-1 Naidu2025. We show models with (left) and without (right) dust. Semi-synthetic models of quasi-stars embedded in a host galaxy are shown in blue, using hosts properties inferred for UNCOVER-24996 and MoM-UDS-948311 respectively. The faded colors surrounding the lines represent $1\sigma$ uncertainties in JWST measurements. Gray dashed lines indicate key wavelengths: the Balmer break (H$\infty$), H$\gamma$, H$\beta$, [OIII], and both CaII lines. Discrepancies in emission lines likely arise from the use of LTE atmospheres at nonzero metallicity, particularly in the H$\beta$ emission and CaII absorption lines - these are discussed in Sec. \ref{['sec:sed']}. Shaded regions indicate the far-UV and NIR approach, highlighting a few of the well fit continuum regions discussed in Sec. \ref{['sec:results']}. The stellar parameters of the quasi-star model at this point are $T_\textrm{eff}=6241$ K, $\textrm{log}(g)=-0.003$, and $Z=0$.