🔭 astrophysics

Black Hole Feedback, Galaxy Quenching and Outflows at Cosmic Dawn: Analysis of the SEEDZ Simulations

The SEEDZ simulations demonstrate that massive black holes in the early Universe rapidly grow to approximately $10^6$ M $_\odot$ via super-Eddington accretion before self-regulating through feedback that evacuates host halos, thereby setting a mass limit that suggests either weaker early feedback than previously assumed or a two-step formation process for high-redshift galaxies observed by JWST.

Original authors: Lewis R. Prole, John A. Regan, Daxal Mehta, Rüdiger Pakmor, Sophie Koudmani, Martin A. Bourne, Simon C. O. Glover, John H. Wise, Ralf S. Klessen, Michael Tremmel, Debora Sijacki, Ricarda S. Beckmann

Published 2026-07-02

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Lewis R. Prole, John A. Regan, Daxal Mehta, Rüdiger Pakmor, Sophie Koudmani, Martin A. Bourne, Simon C. O. Glover, John H. Wise, Ralf S. Klessen, Michael Tremmel, Debora Sijacki, Ricarda S. Beckmann, Martin G. Haehnelt, John Brennan, Pelle van de Bor, Paul C. Clark

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 Race Against Time

Imagine the early Universe as a construction site. The builders are Supermassive Black Holes (MBHs), and they are trying to grow as big as possible before the "construction deadline" (a time in the Universe called redshift 12.5, which is very early in cosmic history).

Astronomers have recently spotted some of these black holes using the James Webb Space Telescope (JWST). They are surprisingly huge for their age. This paper asks a simple question: How did they get so big so fast, and what stopped them from getting even bigger?

The authors ran a series of complex computer simulations (called SEEDZ) to watch these black holes grow. Here is what they found.

1. The Growth Spurt: Eating Like a Vacuum

In the beginning, these black holes were "seeds"—some were light (like a small pebble), and some were heavy (like a boulder).

The simulations showed that the biggest black holes didn't grow slowly over millions of years. Instead, they went on a massive eating binge.

The Analogy: Imagine a vacuum cleaner that suddenly sucks up everything in the room in just a few seconds.
What happened: For a short burst (5 to 30 million years), these black holes ate gas at a rate far faster than physics usually allows (called "super-Eddington accretion"). They grew from their seed size to about 1 million times the mass of our Sun very quickly.

2. The Problem: The Black Hole's Own Burp

You might think the black hole would keep eating until it ran out of food. But the simulations showed something surprising: The black hole stopped itself.

The Analogy: Imagine a person eating a huge meal. As they eat, they get so hot and energetic that they accidentally blow the table apart, scattering all the food into the air. They can't eat anymore because the food is gone.
What happened: As the black hole ate, it released a massive amount of energy (heat) into the gas around it. This energy was so strong that it heated the gas so much that it exploded outward, leaving the black hole's home (its "halo") completely empty.
The Result: The black hole didn't stop because it ran out of food; it stopped because it cooked its own food supply and blew it away.

3. The Limit: The Size of the House

The paper found a strict limit on how big these black holes could get at this early stage.

The Analogy: Think of the galaxy as a house and the black hole as a tenant. If the tenant gets too rowdy and throws a party that is too loud, the energy blows the roof off the house. Once the roof is gone, the party stops.
The Finding: The maximum size a black hole could reach was determined by the strength of the house (the galaxy's gravity). If the black hole grew much bigger than 1 million solar masses, the energy it released would be strong enough to completely destroy the galaxy's gas reservoir.
Conclusion: In these simulations, black holes at this early time cannot get significantly bigger than 1 million suns unless something changes the rules.

4. The Mystery: What About the Giant Black Holes We See?

This creates a puzzle. The JWST has seen black holes that seem to be even bigger than this limit. The paper offers two possible explanations for this:

Our "Party" Model is too loud: Maybe real black holes are much quieter than our simulation suggests. Maybe they don't blow their food away as easily as we modeled.
The "Two-Step" Dance: Maybe the black holes we see today grew to their limit, blew away their galaxy's gas, and then waited. Later, they might have merged with other gas-rich galaxies or pulled in fresh gas from the cosmic web, allowing them to grow again after the initial explosion.

Summary

The paper concludes that in the early Universe, black holes are like greedy eaters who accidentally burn down their own kitchen.

They grow fast by eating gas.
The energy from eating blows the gas away.
This stops them from growing further, capping their size at roughly 1 million solar masses for that specific era.
If we see bigger ones, they either grew differently, or they had to wait for a "second chance" to rebuild their food supply later.

The authors emphasize that their model is a "best guess" based on current physics, and future, higher-resolution simulations will help figure out if real black holes are as destructive as the ones in their computer game.

Technical Summary: Black Hole Feedback, Galaxy Quenching, and Outflows at Cosmic Dawn

Problem Statement
Recent observations by the James Webb Space Telescope (JWST) have revealed a substantial population of massive black holes (MBHs) with masses exceeding $10^6 M_\odot$ existing within the first billion years of the Universe ( $z \gtrsim 6$ , extending to $z \sim 12.3$ ). These findings challenge current theoretical models regarding compact object formation and growth. Specifically, it remains unclear how black holes grew from initial seed masses to such large scales within the allotted timeframe. Furthermore, high-redshift observations indicate an over-massive black hole-to-stellar mass ratio ( $M_{BH}/M_{star} \gtrsim 10^{-2}$ ) compared to the local Universe relation ( $M_{BH}/M_{star} \simeq 10^{-4}$ ), suggesting a distinct evolutionary pathway where the stellar component builds up after the formation of the black hole seed. The SEEDZ simulation suite was designed to investigate these early galaxy and MBH evolution pathways, specifically focusing on the physical mechanisms that regulate or halt MBH growth at cosmic dawn.

Methodology
This study analyzes the SEEDZ simulation suite, a set of fully 3D hydrodynamic cosmological simulations performed using the moving-mesh code Arepo2. The simulations utilize a "zoom-in" approach, focusing on three distinct regions (Normal1, Normal2, and Rarepeak) within a parent box of $40 \, \text{cMpc} \, h^{-1}$ . The high-resolution regions have a dark matter particle mass of $\approx 6.9 \times 10^4 M_\odot h^{-1}$ and a minimum cell size of $\approx 6 \, \text{pc}$ (physical at $z=15$ ).

Key numerical and physical components include:

Seeding: The simulations model both "light seeds" (remnants of Population III stars) and "heavy seeds" (formed via runaway collisions in dense clusters, supermassive stars, or quasi-stars). Heavy seeds form with masses between $5 \times 10^3 M_\odot$ and $10^5 M_\odot$ based on a power-law distribution.
Accretion: Black hole accretion is modeled using the Bondi-Hoyle-Lyttleton formalism, adjusted for simulation resolution via a kernel radius and modified by gas vorticity to prevent over-accretion of high-angular momentum gas.
Feedback: Thermal energy is injected isotropically into gas cells within the accretion radius. The radiative efficiency ( $\epsilon$ ) is calculated dynamically: it follows standard sub-Eddington values ( $\sim 0.1$ ) but transitions to a "slim disc" model during super-Eddington accretion, where photon trapping reduces effective efficiency and weakens feedback.
Dynamics: The model includes dynamical friction corrections to account for the suppression of gravitational interactions due to softening lengths, and a merger criterion based on proximity and relative velocity.
Scope: The analysis focuses on the growth history of the most massive black holes in the suite up to redshift $z=12.5$ .

Key Results

Growth and Mass Limits: The most massive black holes in the simulations grow to masses of $\sim 10^6 M_\odot$ by $z=12.5$ . This growth occurs during short bursts of super-Eddington accretion lasting 5–30 Myr. The final mass is not determined by the initial seed mass but by environmental conditions.
Feedback-Driven Quenching: The primary mechanism halting MBH growth is feedback from the black hole itself, not supernovae or gas exhaustion. For the most actively accreting MBHs, the injected thermal energy exceeds the binding energy of the host halo. This results in the complete evacuation of gas from the host halo into the inter-galactic medium, effectively quenching the galaxy.
Outflow Characteristics: At the moment of accretion cut-off, gas outflows reach velocities of $\sim 10^3 \, \text{km s}^{-1}$ and temperatures of $10^7 \, \text{K}$ . These outflows expand to scales of $\sim 20 \, \text{kpc}$ before cooling. While some halos show signs of recovering gas inflow rates on timescales of 20–100 Myr, the black holes in these simulations do not experience a second major growth phase within the simulation timeframe.
Halo Mass Dependence: The maximum black hole mass is effectively set by the combination of the feedback model and the binding potential of the host halo. In the simulated halos (masses up to $\sim 10^{10} M_\odot$ ), the feedback energy required to unbind the gas limits black hole growth to approximately $10^6 M_\odot$ .
Star Formation History: Halos hosting the most massive black holes exhibit very little star formation prior to the formation of the heavy seed, allowing for efficient accretion. Once accretion begins, the resulting AGN feedback suppresses further star formation. In contrast, halos with dominant stellar populations show continuous star formation but lack the rapid, unimpeded accretion seen in the massive BH hosts.

Significance and Conclusions
The paper concludes that in the early $\Lambda$ CDM Universe, under the implemented feedback models, massive black holes are unlikely to exceed $\sim 10^6 M_\odot$ through accretion alone by $z=12.5$ . The feedback from rapid accretion is sufficiently energetic to unbind the gas reservoir of the host halo, thereby self-regulating the black hole's growth.

The authors propose two interpretations for the observed population of massive black holes at high redshifts that exceed this limit:

Weaker Feedback: MBH feedback in the early Universe may be significantly less efficient than modeled (e.g., if growth is merger-dominated rather than accretion-dominated, or if feedback is contained within a small region).
Two-Step Formation: Observed high-redshift galaxies may have formed via a two-step process where an MBH initially quenches its host galaxy by evacuating the gas reservoir, and later reconstitutes its baryonic reservoir through mergers with gas-rich galaxies or accretion from the cosmic web, allowing for subsequent growth.

The authors note that their current resolution limits the ability to fully resolve black hole mergers, which could be a dominant growth mechanism in later stages. Future higher-resolution simulations are planned to address this and to better constrain the feedback parameters, particularly in the context of "Little Red Dot" systems observed by JWST. The results suggest that unless feedback is extremely ineffective or contained, the maximum mass of black holes at redshifts before 12.5 should not significantly exceed $10^6 M_\odot$ .