🔭 astrophysics

Properties of black hole mergers in disks of active galactic nuclei

This paper utilizes one-dimensional N-body simulations and semi-analytical modeling to demonstrate that the observed properties of binary black hole mergers, including mass distributions and spin correlations, can be explained by black hole mergers within active galactic nucleus accretion disks through a combination of efficient gas accretion and hierarchical mergers.

Original authors: Hiromichi Tagawa, Zoltán Haiman, Bence Kocsis

Published 2026-04-30

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Hiromichi Tagawa, Zoltán Haiman, Bence Kocsis

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

Imagine the center of a galaxy as a bustling, chaotic dance floor. In the middle stands a massive, super-heavy celebrity (the Supermassive Black Hole), surrounded by a swirling, thick fog of gas (the accretion disk). Scattered throughout this fog are hundreds of smaller, invisible dancers (stellar-mass black holes).

This paper is a detailed simulation of how these invisible dancers interact, pair up, and eventually crash into each other to create the gravitational waves we detect on Earth. The authors, Tagawa, Haiman, and Kocsis, built a computer model to see if this "galactic dance floor" scenario explains the strange properties of the black hole mergers we've already observed.

Here is a breakdown of their findings using everyday analogies:

1. The Dance Floor Setup

In their "standard" model (the Fiducial Model), they set up the dance floor with specific rules:

The Fog: The gas disk isn't uniform. It has different densities and temperatures, much like a stormy ocean with calm patches and rough waves.
The Dancers: They start with a crowd of black holes. As the gas swirls, it drags the black holes inward (like a conveyor belt).
The Pairing: As the black holes get dragged through the gas, they bump into each other. Sometimes, the gas acts like a "glue," slowing them down just enough to get stuck together and form a binary pair (a dancing couple).

2. The "Family Tree" of Black Holes

One of the most fascinating discoveries is hierarchical merging.

First Generation: Two original black holes meet and merge.
Second Generation: The new, heavier black hole (the child) finds another partner and merges again.
Third Generation and Beyond: This process repeats. The paper suggests that the heaviest black holes we see (like the one in event GW231123) aren't just two original stars crashing; they are often "great-grandchildren" of previous collisions. They are the result of a family tree where black holes keep eating each other to get bigger.

3. Solving the "Spin" Mystery

Black holes spin. When two merge, their spins can be aligned (spinning in the same direction as the orbit) or misaligned (spinning in random directions). This creates a property called "effective spin" ( $\chi_{eff}$ ).

The paper explains two confusing patterns seen in real data:

The "Heavy and Aligned" Pattern: Observations show that the heaviest black holes often have spins that are well-aligned with their orbit. The authors explain this with gas accretion. Imagine a figure skater spinning; if you pour water (gas) onto them from a specific direction, it forces them to spin in that direction. In the AGN disk, the gas forces the black holes to align their spins before they merge.
The "Heavy and Low-Mass Ratio" Pattern: Sometimes, the heaviest mergers involve one very heavy black hole and one much lighter one. The paper suggests this happens because the heavy "child" black holes (from previous generations) are so massive that they easily pair up with lighter, "first-generation" black holes, creating an uneven couple.

4. The "Goldilocks" Model

The authors tested 20 different versions of their simulation, changing variables like:

How long the galaxy stays active (the "party" duration).
How thick the gas fog is.
How fast the black holes eat gas.

They found that Model M20 is the "Goldilocks" scenario—it's just right.

It assumes black holes eat gas, but not too much (otherwise, they get too heavy too fast).
It allows for hierarchical merging (the family tree).
This specific combination successfully reproduces the weird correlations seen in real data: heavy black holes with aligned spins, and heavy black holes paired with lighter ones.

5. Where Do They Crash?

The simulation shows that most crashes happen in two specific zones:

The Inner Ring (Close to the center): Here, the gas is thick, and black holes are dragged in quickly. They crash here because the gas forces them together.
The Outer Ring: Here, some black holes were already paired up before they even entered the gas disk. They drift in slowly and crash later.

6. The "Eccentric" Dancers

Most black hole pairs spiral in smoothly. However, some crash with a very "bumpy" or elliptical path (high eccentricity). The paper suggests that if the dance floor is flat (2D), black holes are more likely to crash into each other at weird angles, creating these bumpy orbits. This could be a signature that a merger happened in an AGN disk rather than in empty space.

The Bottom Line

The paper concludes that the "Active Galactic Nucleus" (AGN) disk is a very plausible factory for the black hole mergers we see. Specifically, a scenario where black holes eat gas (to align their spins) and merge in generations (to get very heavy) fits the data better than other theories.

However, they warn that if we find out the spins of the heaviest black holes are too perfectly aligned, our current model might need tweaking, suggesting the gas might not be as efficient at "spinning them up" as we thought.

In short: The universe's black holes aren't just lonely drifters; they are social dancers in a gas-filled ballroom, forming families, eating gas to change their spin, and crashing together in a way that perfectly matches the signals we hear from Earth.

1. Problem Statement

Ground-based gravitational wave (GW) observatories (LIGO, Virgo, KAGRA) have detected approximately 200 binary black hole (BH) mergers. The astrophysical origin of these events remains debated, with competing channels including isolated binary evolution, dynamical interactions in star clusters, and mergers within the accretion disks of Active Galactic Nuclei (AGN).

Key observational anomalies that challenge standard formation channels include:

Unusually massive BHs: Events like GW190521 and GW231123 fall within or above the pair-instability supernova mass gap, difficult to explain via isolated evolution.
Spin-Mass Correlations: Recent analyses suggest correlations between effective spin ( $\chi_{\text{eff}}$ ), chirp mass ( $M_{\text{chirp}}$ ), and mass ratio ( $q$ ). Specifically, a negative correlation between $q$ and average $\chi_{\text{eff}}$ , and a positive correlation between $\chi_{\text{eff}}$ and $M_{\text{chirp}}$ , have been reported.
Eccentricity: Some events may possess high eccentricity, a signature of dynamical environments.

The paper aims to investigate whether the AGN disk channel can reproduce these observed properties and to determine how model parameters (e.g., disk density, BH number, accretion efficiency) influence the merger population.

2. Methodology

The authors employ one-dimensional N-body simulations combined with a comprehensive semi-analytical model to track the evolution of BH populations in AGN disks.

Key Model Components:

Disk Structure: The AGN disk is modeled as a geometrically thin, optically thick Shakura-Sunyaev disk in the inner regions, transitioning to a gravitationally unstable region (Toomre $Q=1$ ) in the outer regions where stars form. Two disk prescriptions are tested: the fiducial Thompson et al. (2005, TQM) model and the Sirko & Goodman (2003, SG) model.
BH Dynamics:
- Capture: BHs are captured from the nuclear star cluster via gas dynamical friction and accretion torques.
- Migration: BHs migrate inward via Type I/II torques.
- Binary Formation: Binaries form via gas dynamical friction (gas capture), three-body interactions, and GW capture (GWC).
- Interactions: The model includes binary-single (BS) and binary-binary interactions, with updated prescriptions for encounter timescales and angular momentum randomization.
- Hardening/Merger: Binaries are hardened by gas torques and GW emission until merger.
Accretion & Spin: BHs grow via Eddington-limited (or super-Eddington) accretion. Spin evolution is tracked, including alignment via the Bardeen-Peterson effect and spin magnitude changes due to mergers.
Updates from Previous Work: The authors corrected computational bugs regarding radiative feedback (turning off gas dynamical friction under specific conditions) and encounter probabilities. They also updated the natal spin distribution ( $a=0.1$ ) and the initial BH mass distribution (up to $30-45 M_\odot$ ).

Simulation Scope:
The study runs 20 distinct models (M1–M20) varying parameters such as:

BH number density and initial velocity dispersion.
SMBH mass ( $4\times10^6$ to $4\times10^7 M_\odot$ ).
AGN inflow rates and disk models (TQM vs. SG).
Accretion prescriptions (standard, super-Eddington, ZEBRA flow, duty cycles).
Migration and interaction geometries (3D isotropic vs. 2D coplanar).

3. Key Contributions

Comprehensive Parameter Space: This is the most extensive study to date on AGN merger properties, systematically varying disk physics, accretion rates, and interaction geometries.
Correction of Previous Biases: The authors identified and fixed bugs in previous simulations (Tagawa et al. 2020c, 2021c) related to radiative feedback and encounter timescales, leading to more accurate merger rates and mass distributions.
Mechanism for Observed Correlations: The paper provides a physical explanation for the observed correlations between mass, spin, and mass ratio, linking them to the interplay between hierarchical mergers and efficient gas accretion.
Identification of a "Best-Fit" Model: By comparing simulation outputs with LVK observational data, the authors identify specific parameter regimes (specifically Model M20) that best reproduce the observed population.

4. Key Results

A. Mass and Mass Ratio Distributions

Fiducial Model (M1): Predicts mass and mass ratio distributions similar to LVK observations, but only if the AGN lifetime is short ( $\lesssim 3$ Myr). Longer durations lead to an overproduction of high-mass mergers due to hierarchical growth.
Hierarchical Mergers: Massive mergers (e.g., $M_{\text{chirp}} > 70 M_\odot$ ) are primarily driven by hierarchical mergers (mergers of merger remnants).
Accretion Impact: Models with highly efficient accretion (M17–M19) produce excessive high-mass mergers. However, introducing a low duty cycle (Model M20, $\sim 10\%$ ) or wind mass loss regulates growth, aligning the mass distribution with observations.

B. Spin Correlations ( $\chi_{\text{eff}}$ vs. $q$ and $M_{\text{chirp}}$ )

Negative $q$ - $\chi_{\text{eff}}$ Correlation: The observed trend where lower mass ratios correlate with higher effective spins is reproduced in models combining efficient gas accretion and hierarchical mergers.
- Mechanism: Accretion aligns spins with the orbital angular momentum (increasing $\chi_{\text{eff}}$ ). Hierarchical mergers between different generations (e.g., 1st gen + 2nd gen) naturally produce low mass ratios ( $q$ ).
Positive $\chi_{\text{eff}}$ - $M_{\text{chirp}}$ Correlation: Efficient accretion allows massive BHs to align their spins with the binary orbit, leading to higher $\chi_{\text{eff}}$ for high-mass systems.
Dispersion: A negative correlation between $q$ and the dispersion of $\chi_{\text{eff}}$ is found in most models, driven by hierarchical mergers producing high-spin, low- $q$ binaries.

C. Massive Mergers (GW190521 / GW231123 analogs)

Formation Channels: Events like GW231123 ( $M_1 \sim 137 M_\odot$ $M_{1} \sim 137 M_{⊙}$ ) can be formed via:
1. Hierarchical Mergers: $\ge 3$ generations of BHs (typical spin $\sim 0.7$ ).
2. Efficient Accretion: Low-generation BHs growing rapidly via gas accretion (typical spin $\sim 1.0$ ).
Location: Massive mergers predominantly occur in the inner disk ( $\sim 0.01$ –$0.05$ pc), where gas density is high and hierarchical interactions are frequent.

D. Eccentricity and Location

Eccentricity: The eccentricity distribution at 10 Hz is sensitive to the geometry of binary-single interactions. 2D coplanar interactions (Model M5) significantly increase the fraction of highly eccentric mergers ( $e_{10\text{Hz}} \gtrsim 0.4$ ), potentially explaining specific high-eccentricity candidates.
Radial Distribution: Mergers occur in two peaks:
1. Outer ( $\sim 3$ pc): Dominated by primordial binaries and low-generation mergers.
2. Inner ( $\sim 0.01$ –$0.05$ pc): Dominated by gas-capture binaries and hierarchical mergers.

5. Significance and Conclusions

The paper concludes that the AGN disk channel is a viable and likely significant contributor to the observed GW population, provided specific physical conditions are met.

Preferred Model (M20): The authors tentatively identify Model M20 as the most promising scenario. It incorporates:
- The TQM disk model.
- A transition to the ZEBRA (Zero-Bernoulli) accretion state.
- A low accretion duty cycle ( $\sim 10\%$ ).
- This model successfully reproduces the negative $q$ - $\chi_{\text{eff}}$ correlation, the positive $\chi_{\text{eff}}$ - $M_{\text{chirp}}$ correlation, and the observed mass distribution without overproducing high-mass events.
Constraints on Physics: The results suggest that if AGN disks contribute significantly to GW events, the accretion onto stellar-mass BHs must be regulated (e.g., by winds or duty cycles) to prevent runaway mass growth, and hierarchical mergers must be a dominant pathway for the most massive events.
Future Outlook: The study highlights the need for future GW observations to constrain the duty cycle of accretion and the efficiency of hierarchical mergers. It also suggests that detecting electromagnetic counterparts (flares) or measuring Doppler acceleration in GW signals could further validate the AGN channel.

In summary, this work provides a robust theoretical framework linking the complex dynamics of AGN disks to the statistical properties of the gravitational wave population, offering a compelling explanation for the most extreme merger events observed to date.