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

The SEEDZ simulations reveal that massive black holes at cosmic dawn grow rapidly to $\sim10^6$ M$_\odot$ via super-Eddington accretion before self-regulating through feedback that evacuates host halos, implying that observed high-redshift galaxies likely formed through a two-step process of quenching followed by baryonic reconstitution, and that black hole masses at $z > 12.5$ are fundamentally capped by feedback efficiency and halo binding energy.

The Gist

The Big Picture: A Cosmic Race Against Time

Imagine the early Universe as a giant construction site. Astronomers have recently found some massive "black hole monsters" (black holes with millions of times the mass of our Sun) that appeared very early in the Universe's history. This is a puzzle: How did these monsters get so big, so fast?

The SEEDZ team ran a series of super-computer simulations to answer this. They wanted to see how these black holes grow and what stops them from getting even bigger. Think of their simulation as a high-definition movie of the early Universe, tracking the life story of these black holes from their "birth" until they are about 500 million years old (a time astronomers call redshift 12.5).

The Story of the Black Hole Growth

In the movie, the black holes start as "seeds" (some light, some heavy). To grow, they eat gas from their surroundings, much like a vacuum cleaner sucking up dust.

1. The Feast: For a short, intense burst of time (about 5 to 30 million years), these black holes go on a binge-eating spree. They eat gas so fast that they exceed their "speed limit" (the Eddington limit), growing from a few thousand suns to over a million suns.
2. The Stop Sign: Suddenly, the feeding stops. The black hole doesn't run out of food, and it isn't stopped by exploding stars nearby. Instead, the black hole stops itself.

The "Self-Sabotage" Analogy

Here is the most important discovery of the paper, explained with a metaphor:

Imagine a black hole is a very hungry child sitting at a table full of food.

• The Problem: As the child eats, they get so excited and energetic that they start kicking the table.
• The Result: The kicking is so violent that it throws all the food off the table and into the next room.
• The Outcome: The child is now full (or at least, has eaten as much as they can), but the table is empty. The child can't eat anymore because the food is gone.

In the simulation, the "kicking" is feedback. As the black hole eats, it releases massive amounts of energy (heat). This energy is so strong that it blows the gas (the food) out of the galaxy entirely, pushing it into the empty space between galaxies. Once the gas is gone, the black hole has nothing left to eat, and its growth stops.

What This Means for the "Monsters" We See

The paper concludes that in these simulations, a black hole simply cannot grow much larger than 1 million suns at this early stage of the Universe. The reason isn't a lack of food; it's that the act of eating destroys the food supply.

The authors offer two possibilities to explain why we might see even bigger black holes with the James Webb Space Telescope (JWST):

1. The Model is Too Strong: Maybe our "kicking" (feedback) is too violent in the simulation, and in reality, black holes are more polite and don't blow all the food away.
2. The "Two-Step" Dance: Maybe the black hole did blow all the food away and stopped growing. But later, the galaxy merged with another galaxy that had fresh food, or new gas flowed in from the cosmic web. The black hole then "woke up" and grew again. It's like the child clearing the table, waiting for a delivery truck to bring a new feast, and then eating again.

The Bottom Line

The paper suggests that the size of the biggest black holes in the early Universe is limited by the size of their "home" (the galaxy halo). If the black hole gets too big, its own energy blows the house apart.

Unless the black holes are very good at hiding their energy or the Universe works differently than our models suggest, we shouldn't expect to find black holes much heavier than 1 million suns at this specific early time. If we do find them, it means our understanding of how these cosmic monsters behave needs a major update.

Technical Summary

Technical Summary: Black Hole Feedback, Galaxy Quenching, and Outflows at Cosmic Dawn (SEEDZ Simulations)

Problem Statement
The discovery of massive black holes (MBHs) with masses exceeding $10^6 M_\odot$ within the first billion years of the Universe ($z \gtrsim 6$) by the James Webb Space Telescope (JWST) challenges standard models of compact object formation. A central unresolved question is how these black holes grew from initial seed masses to such large scales within the allotted timeframe. Furthermore, observations suggest a discrepancy between high-redshift MBH-to-stellar mass ratios ($M_{BH}/M_{star} \gtrsim 10^{-2}$) and local Universe trends ($M_{BH}/M_{star} \simeq 10^{-4}$), implying that stellar components may build up after the formation of the MBH seed. This study investigates the physical mechanisms governing the growth and eventual cessation of accretion for MBHs in the early Universe, specifically focusing on the interplay between super-Eddington accretion and feedback.

Methodology
The analysis utilizes the SEEDZ simulation suite, a set of fully 3D hydrodynamic cosmological simulations performed with the moving-mesh code Arepo2. The study focuses on three "zoom-in" regions (Rarepeak, Normal1, and Normal2) within a parent box, evolving from high redshift down to $z = 12.5$.

Key numerical and physical components include:

• Seeding and Formation: The simulations model both light seeds (from Population III stars) and heavy seeds ($5 \times 10^3 - 10^5 M_\odot$) formed via runaway stellar collisions or supermassive star collapse. Heavy seeds form in metal-enriched gas without metallicity restrictions.
• Accretion Physics: Black hole accretion is modeled using a Bondi-Hoyle-Lyttleton formalism with a kernel radius to account for resolution limits. The model incorporates vorticity adjustments to suppress the accretion of high-angular-momentum gas.
• Feedback Mechanisms: Thermal feedback is injected isotropically into gas cells within the accretion radius. The radiative efficiency ($\epsilon$) is dynamic: it is calculated based on black hole spin for sub-Eddington accretion and transitions to a "slim disc" model for super-Eddington accretion, where photon trapping reduces effective efficiency and feedback strength.
• Dynamics: The model includes dynamical friction corrections for black holes moving through dark matter and gas, and a merger criterion based on gravitational binding and relative motion.
• Resolution: The simulations resolve gas cells down to $\sim 6$ pc (physical) at $z=15$ and dark matter particle masses of $\sim 6.9 \times 10^4 M_\odot h^{-1}$, allowing for the resolution of halos down to a few times $10^6 M_\odot$.

Key Results

1. Growth Trajectories: The most massive MBHs in the simulations grow from initial heavy seed masses to $\sim 10^6 M_\odot$ by $z = 12.5$. This growth occurs in short, intense bursts of super-Eddington accretion lasting 5–30 Myr.
2. Quenching Mechanism: The primary factor halting MBH growth is feedback from the MBH itself, not the exhaustion of gas reservoirs, nearby supernovae, or the exhaustion of available gas.
   • During super-Eddington bursts, the energy injected into the surrounding gas exceeds the binding energy of the host halo.
   • This feedback completely evacuates the gas from the host halo (masses $\sim 10^7 - 10^8 M_\odot$), ejecting it into the intergalactic medium (IGM).
   • The simulations show that even with a relatively weak feedback model (thermal coupling factor $f_c = 0.05$ and reduced efficiency during super-Eddington phases), the energy injection is sufficient to unbind the halo.
3. Galaxy Disruption: The rapid accretion and subsequent feedback effectively quench the host galaxy. In halos containing the most massive MBHs, star formation is suppressed after the onset of the MBH's rapid growth phase. Conversely, halos with dominant stellar populations but sub-dominant black holes do not exhibit such thermal disruption.
4. Outflows and Recovery: The feedback drives powerful outflows with velocities reaching $10^3$ km s$^{-1}$ and temperatures up to $10^7$ K. While some halos show signs of gas recovery (inflow rates returning to pre-outflow levels) on timescales of 20–100 Myr, the MBHs in the current simulation timeframe do not experience a second major accretion phase.
5. Mass Limits: The maximum mass attainable by MBHs at $z \approx 12.5$ is effectively set by the combination of the feedback model and the gravitational binding potential of the host halo. Unless feedback is extremely ineffective (e.g., contained within a small region or growth is merger-dominated), MBH masses should not significantly exceed $10^6 M_\odot$ at these redshifts.

Significance and Claims
The paper posits that the observed population of massive black holes at high redshifts presents a tension with current theoretical models if those models assume efficient, continuous growth. The authors propose two potential resolutions to reconcile their findings with JWST observations:

1. Weaker Feedback: MBH feedback in the early Universe may be significantly less efficient than assumed in these simulations (perhaps due to anisotropic jets or different coupling efficiencies).
2. Two-Step Formation: High-redshift galaxies observed by JWST may have formed via a two-step process. An MBH initially quenches its host galaxy by evacuating the gas reservoir. Subsequently, the galaxy reconstitutes its baryonic reservoir through mergers with gas-rich galaxies or accretion from the cosmic web, allowing the stellar component to grow and eventually match the local $M_{BH}-M_{star}$ relation.

The study concludes that unless feedback is highly suppressed or growth is dominated by mergers (which the current resolution cannot fully resolve), the maximum mass of black holes at $z > 12.5$ is capped near $10^6 M_\odot$. Future work will utilize higher-resolution simulations to better resolve merger dynamics and the specific conditions of "Little Red Dot" systems to test these hypotheses.