⚛️ general relativity

Gravitational-wave background from supermassive black holes without merger trees: Tidally regulated mergers and direct inference from the NANOGrav 15 yr data

This paper presents a fast, resolution-free semi-analytic framework that bypasses traditional merger trees to directly infer the gravitational-wave background from supermassive black holes using NANOGrav 15-year data, revealing a population where tidal stripping stalls 38% of mergers and requiring a black-hole mass normalization significantly higher than local calibrations to explain the observed signal.

Original authors: Shin'ichiro Ando

Published 2026-10-06

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Shin'ichiro Ando

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

Deep in the silence between the stars, a faint, rhythmic hum is rising. It is not a sound we can hear with our ears, but a ripple in the fabric of space-time itself, a gravitational wave background generated by the most massive objects in the universe. For decades, astronomers have suspected that supermassive black holes, the invisible giants sitting at the centers of galaxies, should be merging as their host galaxies collide. When two such black holes spiral together, they should shake the cosmos, creating a low-frequency roar that a global network of pulsar timing arrays has recently begun to detect. This signal offers a unique window into how galaxies grow and how black holes evolve, but to read the story written in these ripples, scientists must first understand the complex journey a black hole takes from the moment its host galaxy falls into a larger one until the two black holes finally collide.

A new study by Shin'ichiro Ando offers a fresh, streamlined way to calculate this cosmic journey, bypassing the messy, computer-heavy methods that have dominated the field. Instead of simulating millions of individual galaxy histories, the researcher built a mathematical framework that tracks the average behavior of black holes as they drift through the centers of merging galaxies. The core discovery is that the path to a collision is not guaranteed. As a smaller galaxy falls into a larger one, its black hole is supposed to sink toward the center, pulled by gravity. However, the smaller galaxy is also being stripped of its outer layers by the gravitational tug of the larger host. The study finds that for a significant number of black holes, this stripping happens so fast that the black hole loses the very mass needed to pull it inward. These black holes get stuck, wandering the outskirts of their new home for billions of years, never reaching the center to merge.

This realization changes the expected strength of the gravitational wave background. If many black holes are stuck and never merge, there are fewer collisions happening today than previously thought. The new calculation shows that this "stalling" effect reduces the predicted amplitude of the gravitational wave signal by about eleven percent compared to older models that assumed all black holes eventually merge. While this difference is small enough that it fits within the current margin of error in the data, it has a profound implication for what we see in the universe. The model predicts that roughly thirty-eight percent of the black holes that have been accreted by larger galaxies are still in transit or permanently stalled by the present day (with about 21% being truly stalled, owing more than a Hubble time of sinking). This population of wandering black holes, which the model calculates without needing to fit any extra numbers to the data, matches the number of such objects seen in detailed computer simulations of galaxy formation.

When the researcher compared this new, physically grounded model against the latest data from the North American Nanohertz Observatory for Gravitational Waves, the results pointed to a specific conclusion about the black holes themselves. The data suggests that the black holes involved in these mergers are significantly more massive than what is predicted by the standard relationship between a black hole's mass and the size of its host galaxy's central bulge. Specifically, the best fit requires the black holes to be about twice as massive as the local calibration suggests. This finding is robust; it holds true whether the researcher uses the new stalling model or the older, simpler models, and it remains consistent even when accounting for uncertainties in how long it takes for black holes to finish their final approach.

The study also explored whether the black holes are moving in perfect circles or on stretched, oval paths, and whether they are surrounded by dense clusters of stars that might speed up their final collision. The data allows for the possibility of eccentric orbits or dense environments, but it does not strictly require them. The signal is compatible with a wide range of scenarios, meaning the universe could be full of these complex interactions, or it could be simpler. What is certain is that the new framework provides a direct, fast, and accurate way to test these ideas against the data without relying on complex computer shortcuts. By treating the delay between galaxy mergers and black hole collisions as a continuous physical process rather than a fixed guess, the research offers a clearer picture of the cosmic population. It confirms that the gravitational wave background is indeed loud, likely driven by black holes that are more massive than previously thought, and it predicts a hidden population of wandering giants that future telescopes may one day spot directly.

Technical Summary: Gravitational-Wave Background from Supermassive Black Holes without Merger Trees

Problem Statement
Pulsar Timing Arrays (PTAs), particularly the NANOGrav 15-year dataset, have reported evidence for a stochastic gravitational-wave background (GWB) at nanohertz frequencies, likely arising from inspiraling supermassive black hole (SMBH) binaries. Extracting astrophysical parameters from this signal requires population models connecting halo mergers to binary coalescence. Existing approaches face a trade-off: Monte Carlo merger trees (based on Extended Press–Schechter theory or N-body catalogs) are flexible but computationally expensive, requiring precomputed model banks or emulators for inference; conversely, fully analytic models are fast but rely on raw merger rates with known $\mathcal{O}(1)$ inaccuracies and often treat merger delays as fixed constants, ignoring the complex interplay of tidal stripping and dynamical friction. Furthermore, standard models often fail to accurately predict the population of "wandering" SMBHs that have not yet reached their host galaxy centers.

Methodology
The author presents a semi-analytic framework that computes the GWB via direct integration, bypassing Monte Carlo merger trees entirely. The framework consists of three core components:

Tree-Free, N-body-Calibrated Merger Integral:
Instead of generating discrete merger trees, the model uses a continuous accretion kernel derived from the subhalo accretion-rate distribution of Y11 (recalibrated against N-body simulations). This kernel is integrated over the Sheth–Tormen mass function and host mass history. This approach is resolution-free and deterministic, counting every branch–branch crossing exactly once without realization noise. The kernel is calibrated to reproduce N-body merger rates (F10) within 3.4%, a significant improvement over the raw LC93 EPS kernel which deviates by 14%.
Tidally Regulated Merger Delays:
The model couples the satellite's tidal mass loss (calibrated via the sashimi semi-analytic code) continuously to Chandrasekhar dynamical friction. Unlike standard fixed-mass delay formulas, this approach integrates the orbit while the satellite's mass decreases.
- Mechanism: As a satellite sinks, tidal forces strip its dark matter envelope. The drag force, proportional to the surviving bound mass, weakens.
- Critical Threshold: A sharp, nearly redshift-independent critical mass ratio $\xi \simeq 0.04$ (satellite-to-host mass) emerges. Satellites below this threshold are stripped faster than they sink; their drag collapses onto the surviving stellar core, and orbital decay effectively stalls.
- Outcome: At $z=0$ , 21% of accreted satellites owe more than a Hubble time of sinking (true "stall"), and 38% are unmerged in total (including those still in transit). This contrasts with fixed-mass models which predict fewer stalled satellites and different selection criteria based on epoch rather than mass ratio.
Direct Inference without Emulators:
The framework is computationally efficient (milliseconds per likelihood evaluation after caching), allowing the full five-parameter posterior to be sampled directly against the NANOGrav 15-year Hellings–Downs correlated free spectrum using Markov Chain Monte Carlo (MCMC). No emulator or precomputed model bank is used.
- Parameters: The model samples the black-hole mass normalization offset ( $\Delta_{KH}$ ), log-normal scatter ( $\sigma_{\text{dex}}$ ), residual hardening delay ( $\tau_{\text{hard}}$ ), transition eccentricity ( $e_t$ ), and nuclear density factor ( $\rho_\times$ ).
- Data Handling: The analysis excludes the 15.8 nHz frequency bin, as subsequent re-analysis attributes its excess power to pulsar noise (chromatic transients in PSR J1713+0747).

Key Results

Black Hole Mass Normalization: The data prefer a black-hole mass normalization $\Delta_{KH} = +0.31^{+0.20}_{-0.23}$ dex above the locally calibrated Kormendy–Ho relation (rising to $+0.42^{+0.23}_{-0.24}$ dex if the 15.8 nHz bin is retained). This indicates the nanohertz background is "loud" relative to local SMBH calibrations. The result is robust to the choice of delay model, merger kernel, and local mass calibration.
Spectral Shape: The measured spectral index $\gamma \approx 3.5$ is consistent with the standard gravitational-wave-driven value of $13/3$ within the 95% region. The data are compatible with, but do not strictly require, eccentric binaries or dense stellar environments, though the posterior shows a preference for higher eccentricity ( $e_t \approx 0.67$ ) and/or dense nuclei.
Wandering Black Holes: The tidally regulated delay predicts that 38% of accreted satellites are unmerged by $z=0$ . Specifically, the model predicts a wandering black-hole mass fraction of $\sim 10\%$ in galaxy-scale halos ( $M_{200c} < 10^{13} M_\odot$ ), matching cosmological hydrodynamic simulations (e.g., Romulus) without tuning. In contrast, fixed-mass delay models predict only $\sim 5.6\%$ .
Amplitude Reduction: The inclusion of tidal regulation reduces the GWB amplitude by 11.2% compared to fixed-mass formulas. This reduction is absorbed by the mass normalization parameter in the fit, making the spectra of the two delay models nearly indistinguishable to current PTA data.

Significance and Claims
The paper claims to establish a "middle path" in GWB modeling that combines the speed of analytic integration with the physical fidelity of N-body-calibrated dynamics. Its primary significance lies in:

Direct Sampling: Demonstrating that a physically detailed, N-body-calibrated model can be sampled directly against PTA data without emulators, eliminating interpolation errors and model-bank biases.
Physical Origin for Stalling: Providing a first-principles derivation of the stalled SMBH population and the critical mass ratio threshold ( $\xi \simeq 0.04$ ), rather than imposing these as external parameters.
Independent Electromagnetic Prediction: Offering a parameter-free prediction for the wandering black-hole census (mass budget and host distribution) that is independent of the GWB fit. This serves as a falsifiable electromagnetic test of the delay physics, distinct from the GWB amplitude constraints.
Resolution of Tension: The framework confirms the tension between the loud nanohertz background and local SMBH mass relations, showing that this excess is robust even when using a more accurate, N-body-calibrated merger kernel and physically motivated delay physics.

The author notes that while the current data cannot spectrally distinguish between the tidal and fixed-mass delay models (as the mass normalization absorbs the amplitude difference), the tidal model's prediction for the wandering population aligns better with hydrodynamic simulations, suggesting the tidal coupling is the more physically accurate description. The framework is designed to be extensible to the LISA band, where occupation physics becomes the dominant observable.

Technical Summary: Gravitational-Wave Background from Supermassive Black Holes without Merger Trees

More like this