A quasi-star is born: formation and evolution of accreting quasi-stars as a metallicity-independent pathway to Little Red Dots

J. Roman-Garza, D. Schaerer, C. Charbonnel, T. Fragos, E. Cenci, R. Marques-Chaves, P. Oesch, M. Xiao

First listed 2026-03-23 | Last updated 2026-03-23

Abstract

To investigate the rest-frame optical emission of "Little Red Dots", we model the formation of and evolution of quasi-stars, i.e. stellar envelopes supported by the accretion luminosity onto a central black hole, originating from rapidly accreting proto-stars reaching the supermassive star regime ($>10^4$ M$_{\odot}$) and undergoing general relativistic instability. We compute stellar evolution models with net mass gain rates $=0.01$, 0.1, and 1 M$_{\odot}$/yr and metallicities $Z=0$-0.01. For the mass gain rates $\ge 0.1$ M$_{\odot}$/yr, stars remain nearly fully convective with $T_\mathrm{eff}\sim4000$-9000~K. The general relativistic instability leading to central BH formation occurs at $M_\star\sim3.5\times10^4$ M$_{\odot}$ ($6.6\times10^4$ M$_{\odot}$) for $\dot{M}_{\rm acc}=0.1$ M$_{\odot}$/yr (1 M$_{\odot}$/yr), at luminosities $L \sim 10^9$ L$_{\odot}$. The lifetime of quasi-stars is estimated to be $10^7$-$10^8$~yr, $\sim$100-1000 times longer than their progenitors. In an environment allowing for rapid accretion the formation, evolution, and properties of quasi-stars are found be essentially independent of metallicity. Comparing the luminosities of our models with those of Little Red Dots at $z<4.5$ ($L_\mathrm{bol}\sim10^{9.5}$-$10^{11.5}$ L$_{\odot}$) yields quasi-star masses $10^{4.5}$-$10^{6.5}$ M$_{\odot}$. The observed minimum luminosity of $\sim10^{9.5}$~\Lsun\ implies accretion rates $\gtrsim0.1$ M$_{\odot}$/yr for Little Red Dots progenitors. Our models offer a metallicity-independent framework supporting quasi-stars as the source of Little Red Dot optical emission, and provide insights into their lifetimes, composition, and progenitor environment.

Short digest

Using MESA, the authors follow rapidly accreting proto-stars (Ṁ=0.01–1 M_sun/yr; Z=0–0.01) into supermassive stars and then quasi-stars supported by a central black hole to explain the rest-frame optical output of Little Red Dots. For Ṁ≥0.1 M_sun/yr the envelopes stay near the Hayashi regime (Teff≈4–9 kK) until general-relativistic instability at M*≈3.5×10^4–6.6×10^4 M_sun (L≈10^9 L_sun), after which quasi-stars live 10^7–10^8 yr with evolution essentially independent of metallicity; surface CNO abundances should vary with brief high-N phases. Matching model luminosities to LRDs at z<4.5 (L_bol≈10^9.5–10^11.5 L_sun) implies quasi-star masses ≈10^{4.5}–10^{6.5} M_sun and progenitor accretion rates ≳0.1 M_sun/yr. This provides a metallicity-independent pathway to LRD optical emission and long duty cycles for early black-hole growth.

Key figures to inspect

• Fig. 1 Kippenhahn diagram: read off when the general-relativistic instability triggers BH formation, how the convective envelope and nitrogen mass fraction evolve, and how BH and envelope masses grow for accreting vs non‑accreting quasi-stars—including the crash/end points.
• Fig. 2 HR diagram: compare color-coded accretion-rate tracks against the de Graaff et al. (2025) LRD locus; verify the Teff≈4–9 kK Hayashi-like regime and the extrapolated luminosity maxima relative to observed LRD luminosities.
• Accreting vs non-accreting QS comparison (dashed/orange vs solid tracks across Figs. 1–2): assess how continued external accretion shifts final QS/BH masses and where models terminate.
• Appendix D summary table (with Appendix B extrapolations): extract lifetimes (10^7–10^8 yr) and final BH seed masses as functions of Ṁ and the BH accretion-efficiency parameter to gauge duty cycle and seed growth.