Not all black hole collisions are created equal. A new analysis of the latest gravitational-wave detections has found that merging black holes appear to sort themselves into three distinct groups, each separated by sharp boundaries in mass and each exhibiting different physical behaviors. The result, drawn from the LIGO-Virgo-KAGRA (LVK) collaboration’s newest catalog of observed events, suggests that black hole mergers across the universe are not products of a single formation process but of at least three separate cosmic assembly lines.
The findings, posted to the arXiv preprint server in May 2025, arrive as gravitational-wave astronomy enters a new phase of statistical power. With dozens of fresh detections from upgraded instruments, researchers are no longer just cataloging individual collisions. They are mapping the demographics of black holes themselves.
Three groups, two sharp boundaries
The study builds on GWTC-4, the most recent gravitational-wave transient catalog compiled by the LVK collaboration. That catalog covers detections from O4a, the first segment of the collaboration’s fourth observing run, and reflects significant upgrades in detector sensitivity compared to earlier runs. The GWTC-4.0 catalog paper documents the instrument configuration, candidate selection criteria, and observing conditions for the run.
Working from that dataset, a separate research team conducted a population-level analysis and reported evidence for at least three binary black hole subpopulations. “We find evidence for at least three subpopulations of binary black holes, separated by sharp transitions” in the primary mass spectrum, the authors of the subpopulation analysis wrote. At 90% credibility, the boundaries fall at roughly 27.7 solar masses and 40.2 solar masses. That creates three bands:
- Below 27.7 solar masses: The most common group, potentially dominated by black holes born from the collapse of massive stars in isolated binary star systems.
- Between 27.7 and 40.2 solar masses: A middle tier that may reflect a different formation channel, such as dynamical encounters in dense stellar environments like globular clusters.
- Above 40.2 solar masses: The heaviest mergers, which could involve direct collapse of extremely massive stars or hierarchical mergers, where a black hole produced by a previous collision merges again.
Each subpopulation shows distinct inferred properties beyond mass alone, including differences in spin behavior. The LVK collaboration’s own population properties paper offers independent support for some of these patterns, reporting overdensities around specific primary masses in the overall mass distribution. That same analysis constrains the spins of merging black holes, finding non-extremal spins and measurable fractions of negative effective inspiral spin, a quantity that encodes how aligned or misaligned the black holes’ rotations are relative to their orbital motion. Negative values suggest at least one black hole spins opposite to the orbit, a signature that carries direct implications for how the binary formed.
Why the boundaries matter
If the black hole mass spectrum were smooth, it would suggest a single dominant formation pathway producing mergers across a continuous range of sizes. Sharp transitions at specific masses point to something more interesting: distinct physical processes that each operate within their own mass range and leave behind different signatures in spin and merger rate.
Astrophysicists have long debated whether most binary black holes form through isolated binary evolution (two stars orbiting each other, evolving, and each leaving behind a black hole that eventually spirals inward) or through dynamical assembly in crowded stellar environments. The spin data offers a way to distinguish between these channels. Isolated binary evolution tends to produce black holes whose spins are aligned with the orbital plane, while dynamical formation in dense clusters tends to randomize spin orientations. The detection of negative effective inspiral spin in a meaningful fraction of events suggests that at least some mergers form dynamically.
The three-subpopulation framework gives that debate a sharper structure. Rather than asking which single channel dominates, researchers can now ask whether different channels dominate in different mass ranges, and whether the boundaries between groups correspond to known thresholds in stellar physics, such as the pair-instability mass gap, a predicted range of masses where certain stellar explosions should prevent black hole formation entirely.
What could change
Several important caveats apply. The three-subpopulation model is a statistical inference, not a direct observation. The boundaries at 27.7 and 40.2 solar masses emerge from the data at 90% credibility, meaning there is roughly a one-in-ten chance the true boundaries lie elsewhere. In Bayesian statistics, these credible intervals depend on the prior assumptions built into the analysis. Independent teams using different statistical frameworks may find somewhat different boundaries or a different number of groups.
Selection effects also complicate the picture. Gravitational-wave detectors are more sensitive to heavier systems because more massive black holes produce stronger signals detectable at greater distances. That bias means high-mass mergers are overrepresented in the catalog relative to their true cosmic abundance. Population studies attempt to correct for this by modeling detection efficiency as a function of mass and distance, but any residual mismatch between the model and reality can imprint artificial structure on the inferred mass distribution.
Neither the subpopulation analysis nor the LVK population properties paper has completed peer review at a traditional journal. Both are hosted on arXiv, a preprint repository supported by Cornell University and partner institutions. That means the results have been made public for community scrutiny but have not yet passed formal refereeing, a process that could lead to refinements in statistical methods, alternative modeling choices, or updated interpretations. No official LVK press release has endorsed the three-subpopulation interpretation as a collaboration-wide finding.
What comes next for black hole demographics
The real test will come with more data. The LVK collaboration’s fourth observing run is ongoing as of May 2026, and the second segment, O4b, is expected to add substantially to the catalog. If the same mass boundaries continue to appear as the sample grows, and if independent analyses converge on similar values, confidence in the three-group model will strengthen. If the boundaries shift or blur, the field will revise the framework.
For now, the emerging structure in the black hole mass distribution marks a turning point for the field. Gravitational-wave astronomy spent its first decade proving it could detect individual collisions. With catalogs now large enough to support population-level statistics, the discipline is beginning to answer a bigger question: not just that black holes merge, but how the universe builds them in the first place.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
