Growth of Light Seed Black Holes in the Early Universe

TL;DR

The paper demonstrates that light seed black holes (LSBHs) formed from Population III remnants can grow rapidly in the early universe to masses of order $\sim 10^4\,M_{\odot}$ within a few Myr when the Bondi-Hoyle-Lyttleton radius is resolved in fully cosmological zoom-in simulations. Using high-resolution AREPO simulations that include PopIII/PopII star formation, SNe and BH feedback, a redshift-dependent Lyman-Werner background, and vorticity-corrected accretion, the study finds bursts of super-Eddington accretion can drive rapid growth, with some BHs reaching up to about $1.4\times 10^4\,M_{\odot}$ by $z\sim21.5$, though feedback often halts subsequent accretion. Growth is predominantly via gas accretion rather than mergers, and direct-collapse seeds have a systematic advantage in the early growth phase; metal-enriched environments still permit multiple growth episodes. These results bridge the gap between PopIII remnants and SMBHs, implying LSBHs can be viable progenitors of the high-redshift SMBH population and yielding testable predictions for JWST and LISA observations.

Abstract

Observations from the James Webb Space Telescope (JWST) have uncovered supermassive black holes (SMBHs) with masses exceeding $10^6 \mathrm{M}_{\odot}$ at redshifts $z > 8$, posing significant challenges to existing models of early black hole formation and growth. Here we show, in a fully cosmological setting, that light seed black holes (LSBHs), remnants of Population III stars, can grow rapidly to $\sim10^4 \mathrm{M}_{\odot}$ in the early Universe. This growth is enabled by our novel black hole seeding prescription and the unprecedented resolution of our zoom-in cosmological simulations, which resolve the dense environments necessary for efficient accretion. Our results provide robust evidence that LSBHs can attain the masses required to serve as the dominant progenitors of the SMBH population observed at later cosmic epochs. These findings have far-reaching implications for the interpretation of JWST observations and future gravitational wave detections with LISA.

Figures (12)

• Figure 1: Mass growth history of BHs: We show all BHs that accreted more than 0.1 $\rm{M_{\odot}}~$ (dashed lines) and all BHs that doubled their initial mass (solid lines). The simulations L13, L14, L15 and L15_BHFB are coloured red, blue, violet and yellow respectively. Firstly, we see that the number of growing BHs increases with resolution. Secondly, we see that the BHs accrete very rapidly and their rapid-accretion episodes are often short-lived, lastly at most a few million years. In this time, many BHs are able to realise super-Eddington accretion levels with many BHs growing to more than $10^4$$\rm{M_{\odot}}~$ in only a few Myr.
• Figure 2: Gas density projection for PopIII star formation, BH formation, supernova feedback and thermal feedback for the most massive BH in the L15_BHFB simulation: a, b, c: Within cosmological filaments, small mini-halos collapse into galaxies and begin metal-free PopIII star formation. The galaxy forms a disk like structure with clear spiral arms. d: The first PopIII transitions to a BH through direct collapse and accretes mass, reaching 335 $\rm{M_{\odot}}~$. The injected thermal energy causes the galaxy to distort in the centre. e: The galaxy reassembles and collapses back onto the BH, triggering a period of intense star formation. f: SNe feedback from another PopIII star injects significant energy into the galaxy, once again distorting it. Eventually, other PopIII stars undergo SNe combined with the thermal BH feedback completely expel the gas from the galactic centre.
• Figure 3: Black Hole growth in terms of the Eddington fraction: We show the Eddington factors of all BHs which are growing sufficiently to double their mass. The simulations L13, L14, L15 and L15_BHFB are coloured red, blue, violet and yellow respectively. We see that regardless of resolution, BHs are able to achieve accretion rates in excess of the canonical Eddington limit. For several time periods in our highest resolution simulations maximum accretion rates can exceed Eddington factors of $10^3$ albeit only for very short durations. Nonetheless, this clearly demonstrates that super-Eddington accretion is possible and indeed likely a necessary condition to drive rapid growth at high redshift.
• Figure 4: Gas properties within 10 pc of growing BHs. We show the density-weighted gas properties within a 10 pc sphere surrounding all BHs that grew more than 0.1 $\rm{M_{\odot}}~$ (dashed) and for BHs that double their initial seed mass (solid lines). The simulations L13, L14, L15 and L15_BHFB are coloured red, blue, violet and yellow respectively. In (a), we show the total mass of the gas surrounding the BHs as the BHs evolve with redshift. We notice that for almost all BHs, there is a drastic decrease in mass as rapid accretion begins. This is because SNe feedback expel gas from the mini-halos, raising the gas temperature beyond $10^5$ K (b) and driving relative radial velocities to 1000 km/s (c), higher than the escape velocities of the mini-halos. All of these factors cause the accretion rate onto the BHs to sharply decrease (d), completely stopping accretion onto the BHs.
• Figure 5: Extended Data Figure 1 | Population density of PopIII stars. We show the IMF of PopIII stars from all our simulations. The simulations L13, L14, L15 and L15_BHFB are coloured red, blue, violet and yellow respectively. This is the top-heavy IMF with a characteristic mass of 20 $\rm{M_{\odot}}$. The minimum mass of PopIII stars in 1 $\rm{M_{\odot}}~$ and the maximum mass is 300 $\rm{M_{\odot}}$.