Gravitational waves expose hidden families of merging black holes

Not all black holes merge the same way. Two independent studies, both published on July 6, 2026 in Physical Review Letters, have arrived at the same striking conclusion: merging black holes do not belong to a single, uniform population. Instead, they separate into distinct subpopulations, each carrying the fingerprints of a different astrophysical formation channel. The convergence of these two analyses marks a turning point in gravitational-wave astronomy, transforming a growing catalog of detections into a tool for mapping the origins of the most extreme objects in the universe.

The question gravitational waves could not answer until now

Since the first detection of gravitational waves in September 2015 by LIGO, the catalog of observed binary black hole mergers has expanded rapidly. The fourth gravitational-wave transient catalog, known as GWTC-4, assembled by the LIGO-Virgo-KAGRA collaboration, contains over 150 confirmed detections. Each event encodes information about the masses of the two merging black holes and their spins, meaning both the rate and the direction at which each one rotates. Despite this wealth of data, a fundamental question had remained stubbornly unresolved: how do these black holes actually form, and do they all share a common origin?

Two research groups have now tackled that question using different methodologies and reached remarkably consistent answers.

Three distinct populations emerge from the data

Sharan Banagiri, Eric Thrane, and Paul D. Lasky, all from Monash University in Australia, adopted a phenomenological approach to analyze the GWTC-4 data. Rather than assuming a particular astrophysical model, they let the data speak for itself by searching for natural groupings in the distribution of masses, mass ratios, and spin magnitudes.

Their analysis reveals that the merging binary black hole population divides into at least three subpopulations, each occupying a distinct range of primary mass and exhibiting its own characteristic spin behavior.

The first subpopulation accounts for approximately 79 percent of all detected mergers. These systems cluster around a primary mass of roughly 10 solar masses, display small spin magnitudes, and show spins preferentially aligned with the orbital angular momentum. This combination of properties is a textbook signature of isolated binary evolution: two stars born together as a gravitational pair exchange mass during their lifetimes, collapse into black holes, and eventually spiral inward to merge without external gravitational interference. In these systems, the mass ratios are nearly flat, meaning the two merging black holes can have a wide range of relative sizes, and the absence of significant spin tilting indicates that no strong dynamical kicks disrupted the orbital alignment.

The second subpopulation represents about 14.5 percent of the sample. Its mass distribution peaks near 35 solar masses. Unlike the first group, these binaries exhibit chaotic spin behavior with misaligned orientations, and their mass ratios tend toward unity, meaning the two merging black holes have comparable masses. These characteristics are consistent with dynamical formation in dense stellar environments such as globular clusters, where gravitational encounters between multiple stars and black holes constantly shuffle orbital parameters. In such crowded settings, black holes do not form in situ as isolated pairs but are instead brought together through three-body interactions, exchange encounters, and gravitational focusing, all of which randomize spin orientations and preferentially pair objects of similar mass.

The third and rarest subpopulation comprises roughly 2.5 percent of detections. These are the most massive systems in the catalog, with primary masses exceeding approximately 40 solar masses. Their spins are high, with a nearly flat distribution between 0 and 1 that encompasses the characteristic value of 0.7 predicted for second-generation merger remnants. Their mass ratios are notably unequal, with the more massive component often about twice as heavy as its companion. This combination of high mass, high spin, and asymmetric pairing is the expected hallmark of hierarchical mergers: events in which at least one of the two colliding black holes is itself the product of a previous merger.

Hierarchical mergers and the pair-instability gap

The second study, led by Cailin Plunkett and Salvatore Vitale at the Massachusetts Institute of Technology, together with Thomas Callister at Williams College and Michael Zevin at Northwestern University and the Adler Planetarium, approached the problem from a more astrophysically motivated direction. Their analysis focused specifically on the spin signatures expected from hierarchical mergers, introducing a model that operates in the joint parameter space of two well-measured quantities: the effective inspiral spin and the effective precessing spin.

The effective inspiral spin captures the components of black hole spin parallel to the orbital angular momentum, while the effective precessing spin measures the perpendicular components. Previous studies had used the effective inspiral spin alone to estimate the fraction of hierarchical mergers in the catalog. Plunkett and collaborators are the first to incorporate precession information as well, arguing that it is essential for a reliable distinction between hierarchical and first-generation mergers.

Their results are decisive. Above roughly 45 solar masses, the binary black hole population transitions into a regime that is nearly entirely hierarchical. This threshold is precisely where stellar theory predicts the onset of the pair-instability mass gap, a forbidden range of black hole masses resulting from the physics of dying massive stars.

In the cores of very massive stars approaching the end of their nuclear burning, temperatures become high enough for energetic photons to convert into electron-positron pairs. This process softens the pressure support that holds the star against gravitational collapse, triggering explosive instabilities. For stars in a specific mass range, these pulsational pair-instability episodes strip away mass from the outer layers before the final collapse, preventing the formation of black holes heavier than approximately 45 solar masses through normal stellar evolution. For even more massive progenitors, the instability becomes so extreme that the entire star is disrupted in a pair-instability supernova, leaving no remnant at all. The result is a predicted gap in the mass distribution of black holes roughly spanning the range from about 50 to 130 solar masses.

Yet gravitational-wave observations have detected black holes within and above this forbidden range. Hierarchical mergers in dense stellar clusters provide a natural explanation. When two first-generation black holes merge in a sufficiently dense environment, such as the core of a globular cluster or a nuclear star cluster, the remnant can be gravitationally retained rather than ejected. That remnant, carrying a characteristic spin of approximately 0.7, can then pair with another black hole and merge again, producing a second-generation object whose mass exceeds the pair-instability limit imposed on single stellar collapse. The Plunkett et al. study also identifies a peak in the hierarchical merger rate at around 15 solar masses for the companion component, which they interpret as reflecting the global peak of the first-generation black hole mass function that feeds the hierarchical channel.

Convergence from independent methods

What makes these two results particularly compelling is their independence. The Monash group used a phenomenological mixture model with no prior assumption about formation channels, while the MIT group employed an astrophysically motivated spin model designed specifically to test the hierarchical merger hypothesis. Both groups analyzed the same GWTC-4 dataset, yet arrived at fully consistent conclusions: the high-mass end of the merging black hole population, above roughly 40 to 45 solar masses, is dominated by objects that were not born from collapsing stars but assembled through successive mergers in gravitationally dense environments.

This convergence across different modeling frameworks is significant because it reduces the risk that the detected subpopulations are artifacts of a particular set of assumptions. When two teams with distinct methodologies, different parameterizations, and different prior choices reach the same structural conclusion, the result gains substantial robustness.

Implications for stellar physics and future observations

The identification of multiple subpopulations has implications beyond black hole astrophysics. The location of the transition mass, near 45 solar masses, provides an observational constraint on nuclear burning in massive stars, specifically on the rate of the carbon-12 to oxygen-16 fusion reaction during helium burning. This nuclear reaction rate determines where exactly the pair-instability process activates and therefore sets the lower edge of the mass gap. Gravitational-wave population studies are now constraining nuclear physics in regimes that laboratory experiments cannot easily probe.

With the GWTC-5.0 catalog, released in late May 2026 and pushing the total number of gravitational-wave detections to 390, the statistical power available for these population analyses continues to grow. New detections are arriving several times per week during the ongoing fourth observing run. As detector sensitivity improves and the catalog expands further, astronomers will be able to resolve finer structure within each subpopulation, measure the redshift evolution of the hierarchical merger fraction, and potentially identify additional formation channels that are currently hidden beneath the noise.

The universe builds its black holes through multiple assembly lines. Gravitational waves have finally given us the tools to tell them apart.

© 2026 SKYCR.ORG | Homer Dávila Gutiérrez, FRAS. All rights reserved. Total or partial reproduction prohibited without express authorization. Original source: Physical Review Letters (2026). DOI: 10.1103/blyb-lqv6