🔭 astrophysics

Supermassive Black Hole Growth in Massive Galaxies at Cosmic Dawn

Using cosmological hydrodynamic simulations of massive galaxies at Cosmic Dawn, this study demonstrates that while supermassive black hole growth is frequently stochastically suppressed by stellar feedback-driven turbulence, it can achieve self-regulation through AGN feedback once the black hole reaches a critical mass, though no evidence of galaxy-scale quenching of star formation was found.

Original authors: James Sunseri, Zachary L. Andalman, Romain Teyssier

Published 2026-06-12

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: James Sunseri, Zachary L. Andalman, Romain Teyssier

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: A Cosmic Mystery

Imagine the very early universe, about 13 billion years ago (a time astronomers call "Cosmic Dawn"). The James Webb Space Telescope (JWST) has recently spotted some massive, bright galaxies that shouldn't exist yet according to our old rulebooks. Even stranger, many of these galaxies seem to have supermassive black holes (SMBHs) at their centers that are already huge.

The big question is: How did these black holes grow so big, so fast?

This paper is like a set of "what-if" experiments run on a supercomputer. The authors built a digital universe to watch how these black holes eat gas, how they interact with the stars around them, and whether they can eventually slow themselves down.

The Setup: A Digital Sandbox

The researchers used a sophisticated simulation code called RAMSES. Think of this as a high-end video game engine, but instead of rendering graphics for a console, it renders the physics of gas, dark matter, and stars.

The Stage: They focused on a specific, crowded region of space that would eventually collapse into a massive galaxy halo (about 100 billion times the mass of our Sun).
The Actors:
- The Galaxy: A swirling, turbulent disk of gas and stars.
- The Black Hole: A "sink" particle placed in the center that eats whatever gas gets too close.
- The Stars: They explode as supernovae and blast energy out, churning up the gas.

The Main Characters: Feast and Starvation

The paper's most interesting discovery is how the black hole eats. The gas in these early galaxies isn't a smooth, calm soup; it's a chaotic, turbulent storm with two very different types of weather:

The "Feast" Mode (Cold, Dense Patches): Imagine the black hole diving into a thick, cold cloud of gas. It's like a vacuum cleaner sucking up a pile of spaghetti. The black hole eats furiously, growing at its maximum possible speed.
The "Starvation" Mode (Hot, Diffuse Patches): Suddenly, the black hole drifts into a hot, thin pocket of gas. It's like trying to eat soup with a fork that has huge holes in it. There's barely anything to grab. The black hole stops growing.

The Analogy: Imagine a person trying to eat at a buffet that is constantly being rearranged by a tornado. Sometimes the tornado pushes a giant steak right in front of them (Feast), and sometimes it blows the food away, leaving only air (Starvation).

The paper found that in these early galaxies, the black hole spends about 50% of its time starving and 50% feasting. This happens because the stars in the galaxy are so active (exploding as supernovae) that they keep churning the gas around, creating these random pockets of food and air.

The "Self-Regulation" Problem

In the modern universe, black holes have a "brake pedal." When they get too big, they shoot out powerful winds and heat that push gas away, stopping them from eating more. This is called self-regulation.

The researchers asked: Do these early black holes have a brake pedal?

The Result: It depends on how big the black hole is before the chaos starts.
- If the black hole starts small: It gets stuck in the "feast and starvation" cycle. It never gets big enough to turn on its "brake pedal" (AGN feedback) effectively. The stars keep churning the gas, and the black hole just keeps eating whenever it can.
- If the black hole starts big (or eats super-fast): It can eventually get massive enough that its own heat pushes back against the gas. It successfully turns on the brake pedal, stops eating, and regulates its own growth.

The Catch: The paper found that for a black hole to successfully turn on its brake pedal in this chaotic environment, it needs to be very massive very quickly. If it doesn't, it stays in the "wild eating" phase.

The Surprise: The Black Hole Doesn't Stop the Stars

Usually, we think that if a black hole gets too active, it heats up the whole galaxy and stops stars from forming (like turning off the lights in a factory).

The paper's surprising finding: In these early, rapidly growing galaxies, the black hole cannot stop the stars from forming.

The Analogy: Imagine a factory (the galaxy) that is being supplied by a giant, endless conveyor belt of raw materials (cold gas from the cosmic web).

The black hole is a worker who tries to heat up the factory floor to stop the machines (star formation).
But the conveyor belt is moving so fast, and bringing in so much fresh, cold material, that the heat from the black hole is instantly washed away.
No matter how hard the black hole tries to "quench" (stop) the factory, the fresh supply keeps the machines running. The galaxy keeps making stars regardless of what the black hole does.

The Conclusion: How to Match the Observations

The authors compared their digital experiments to the real galaxies JWST is seeing.

The Mismatch: Most of their simulations produced black holes that were too small compared to what JWST sees.
The Solution: To match the real universe, the simulations needed two specific ingredients:
1. Super-Eddington Accretion: The black hole must be able to eat gas at a rate faster than the theoretical "speed limit" (Eddington limit).
2. Weak Feedback: The black hole's "brake pedal" (feedback) must be very weak. If the brake is too strong, the black hole stops growing too early. If it's weak, the black hole can grow huge before it finally manages to regulate itself.

Summary in One Sentence

In the chaotic, star-filled nurseries of the early universe, supermassive black holes grow in a wild "feast-or-starve" cycle driven by turbulence, and they are often too small or too weak to stop their host galaxies from making stars, meaning they must eat incredibly fast to reach the massive sizes we see today.

Technical Summary: Supermassive Black Hole Growth in Massive Galaxies at Cosmic Dawn

Problem Statement
Recent observations by the James Webb Space Telescope (JWST) have revealed a population of massive, bright galaxies at Cosmic Dawn ( $z \gtrsim 9$ ) with high stellar masses ( $M_\star \gtrsim 10^8 M_\odot$ ) and star formation rates (SFRs). Several of these galaxies exhibit spectral features indicative of active galactic nuclei (AGN), suggesting the presence of overmassive supermassive black holes (SMBHs) relative to local scaling relations. Standard galaxy formation models struggle to explain the rapid growth of these SMBHs and the high star formation efficiencies observed. While theoretical proposals invoke heavy seeds or super-Eddington accretion, the interplay between SMBH growth, stellar feedback, and AGN feedback in the turbulent, multiphase interstellar medium (ISM) of early massive galaxies remains poorly understood.

Methodology
The authors conducted a suite of 11 cosmological hydrodynamic "zoom-in" simulations using the adaptive mesh refinement (AMR) code ramses. The simulations focus on a specific overdensity that collapses into a $\sim 10^{11} M_\odot$ halo by $z \sim 9$ , hosting a massive galaxy (Massive Dawn Galaxy or MDG) with $M_\star \gtrsim 10^8 M_\odot$ .

Key methodological components include:

Star Formation Model: Unlike standard simulations using the Kennicut-Schmidt law, this work employs a physically motivated, turbulence-based subgrid multi-freefall model. This allows for variable local star formation efficiencies (SFE) driven by the turbulent kinetic energy of the ISM, capable of reaching $\gtrsim 10\%$ .
SMBH Implementation: SMBHs are modeled as sink particles seeded at $z \sim 15$ when local gas and stellar masses exceed $10^5 M_\odot$ within a $10^8 M_\odot$ halo. Seed masses of $10^4 M_\odot$ and $10^5 M_\odot$ were tested.
Accretion Physics: Accretion follows a modified Bondi-Hoyle-Lyttleton formula with a boost factor to account for unresolved density fluctuations. The accretion rate is capped by an Eddington limit, which is varied across simulations via a factor $\lambda_{\text{edd}}$ (values: 1, 3, and $\infty$ for no limit).
Feedback Mechanisms:
- Stellar Feedback: Includes thermal photoionization (early-type) and Type II supernovae (late-type). A "low stellar feedback" variant was tested by removing photoionization to isolate its impact on ISM structure.
- AGN Feedback: Modeled as isotropic thermal energy injection into the sink sphere with a coupling efficiency $\epsilon_c$ (tested at 0% and 15%).
Simulation Suite: The 11 runs systematically varied AGN feedback strength, seed mass, maximal accretion limits, and stellar feedback efficiency to isolate their effects on SMBH-galaxy co-evolution.

Key Results

ISM-Driven Stochastic Accretion (Feast and Starvation):
In the fiducial MDG environment, stellar feedback generates a turbulent, multiphase ISM consisting of cold-dense ( $n_H \sim 10^3 \text{ cm}^{-3}$ ) and hot-diffuse ( $n_H \sim 10 \text{ cm}^{-3}$ ) patches. The SMBH sink sphere stochastically samples these phases.
- Feast Mode: When the sink enters a cold-dense patch, the sound speed drops, allowing accretion to reach the (super-)Eddington limit.
- Starvation Mode: When the sink enters a hot-diffuse patch, the sound speed rises, and the Bondi accretion rate drops significantly.
- In the fiducial model, the SMBH spends approximately 50% of the time in starvation mode ( $\alpha \approx 0.5$ ) after the onset of star formation ( $t \sim 300$ Myr), effectively suppressing the average growth rate by half compared to a smooth ISM.
Conditions for Self-Regulation:
AGN feedback can only self-regulate the SMBH growth if the black hole becomes massive enough for its heating to dominate cooling within the sink sphere ( $M_{\text{sink}} > M_{\text{sink,crit}}^{\text{cool}}$ ).
- This critical mass depends on gas density and metallicity. In the turbulent MDG environment, the SMBH must exceed this threshold for $\gtrsim 50\%$ of the time spent in hot-diffuse pockets to achieve runaway heating and unbind gas from the halo.
- Simulations with Eddington-limited accretion ( $\lambda_{\text{edd}}=1$ ) and standard seed masses often fail to reach self-regulation because the critical mass threshold rises with metallicity faster than the SMBH can grow, or the SMBH spends insufficient time above the threshold.
- Self-regulation is achieved in simulations with higher seed masses ( $10^5 M_\odot$ ) or super-Eddington accretion ( $\lambda_{\text{edd}} \ge 3$ ), provided the thermal coupling efficiency is sufficiently low ( $\epsilon_c \lesssim 0.15\%$ ) to match observed SMBH masses.
Lack of Galaxy-Scale Quenching:
Contrary to expectations from low-redshift models, AGN feedback in these simulations does not quench the global star formation rate (SFR) or halt the growth of the host galaxy's stellar mass.
- The halo accretes cold gas from cosmic filaments at a rate ( $\dot{M}_{\text{halo}}$ ) that significantly exceeds the maximal growth rate of the SMBH ( $\dot{M}_{\text{SMBH}}$ ).
- Because the halo growth timescale ( $\tau_{\text{halo}}$ ) is comparable to or shorter than the SMBH growth timescale, fresh cold gas continuously replenishes the ISM, overwhelming any heating from AGN feedback. The galaxy continues to grow exponentially regardless of SMBH activity.
Comparison to Observations:
The simulation suite that best matches the observed population of overmassive SMBHs at Cosmic Dawn (e.g., UHZ-1, GN-z11) is the super_edd_no_AGN run ( $M_{\text{seed}} \sim 10^4 M_\odot$ , $\lambda_{\text{edd}} = 3$ , no AGN feedback).
- To match observed masses with self-regulated SMBHs, the thermal coupling efficiency must be extremely low ( $\epsilon_c \lesssim 0.15\%$ ).
- The results suggest that super-Eddington accretion is likely necessary for SMBHs to grow massive enough to enter the self-regulated regime without requiring seed masses as high as $10^6 M_\odot$ .

Significance and Conclusions
The paper argues that the growth of SMBHs in massive galaxies at Cosmic Dawn is primarily regulated by the stochastic nature of the turbulent, multiphase ISM driven by stellar feedback, rather than by AGN feedback alone. The authors conclude that:

SMBHs in these environments are frequently "starved" by hot-diffuse gas pockets, limiting their growth unless they can sustain super-Eddington accretion.
AGN feedback is ineffective at quenching star formation on galactic scales during this epoch because the halo's rapid gas accretion continuously replenishes the fuel supply.
Achieving the massive SMBHs observed by JWST likely requires a combination of super-Eddington accretion and inefficient AGN feedback coupling, or very massive seeds, to overcome the rising critical cooling mass threshold before the galaxy evolves into a more stable, low-redshift-like state.

The authors note that their results are resolution-dependent and that future work incorporating non-equilibrium thermochemistry, radiative transfer, and higher spatial resolution ( $\sim 1$ pc) is required to fully validate these findings and refine the physical scales of feedback deposition.