Science

LIGO’s black holes cluster at 10 Suns, a hint their stars never exploded

Nadia Okonkwo
Add us on Google

Most of the black holes that gravitational-wave detectors catch colliding seem to come in one size. In about three of every four merging pairs, the heavier black hole weighs close to ten times the mass of the Sun, while black holes only a few solar masses heavier are almost missing from the record.

That lopsided tally is the clue. A new analysis of the catalog built by the LIGO, Virgo and KAGRA observatories argues that the crowd near ten solar masses fits a population of stars that never exploded. Instead of blowing apart as supernovae, their cores fell straight in on themselves and became black holes without a flash. Theory has predicted such failed supernovae for decades, but no one has confirmed a single one by watching a star disappear.

The study was led by Katerina Chatziioannou, a physicist at the California Institute of Technology, with first author Isaac Legred, now a postdoctoral researcher at the University of Illinois Urbana-Champaign, and Jacob Golomb. “A vast majority of the merging black holes in the universe are about 10 times the mass of our Sun,” Chatziioannou said in a Caltech statement. “Our data are showing that this group of black holes may have a common origin.”

The standard account of a black hole’s birth runs through a supernova: a massive star burns through its fuel, its core collapses, and a shock wave throws the outer layers into space. If most black holes in colliding pairs skipped that explosion, a large share of massive stars die out of sight, and gravitational waves are the main way to count them.

How a list of collisions became a scale for dead stars

Gravitational waves are ripples in space-time sent out when two black holes spiral together and merge. LIGO’s two detectors in Louisiana and Washington, Virgo in Italy and KAGRA in Japan record them, and the shape of each signal carries the masses of the two objects and how fast they spin. The network has picked up more than 300 black hole collisions since the first detection.

The team took 153 well-measured pairs from GWTC-4.0, the collaboration’s catalog covering events up to the first part of its fourth observing run, and fitted them with a population model built for one question. The model has two parts: a narrow peak meant to stand for failed-supernova black holes, and a second component for everything heavier. Unlike most earlier models, it allows an empty stretch between the two and lets each group have its own spin and mass-ratio pattern. The analysis also corrects for the fact that lighter, fainter pairs are harder to detect.

That design choice matters. Flexible “nonparametric” methods, which assume no shape for the distribution, smooth it over a few solar masses and can blur a sharp hole out of existence. The authors wanted a model able to register a hundredfold drop within a narrow range, which is what supernova simulations predict if one band of stellar masses fails to explode.

While the paper was under review, the collaboration released its larger GWTC-5.0 catalog. The team reran the main analysis on the 257 binary black holes in it that pass their detection threshold, and the peak, the near-empty band and the spin difference all held.

A hundredfold drop within a few solar masses

The peak sits at 9.95 solar masses. The model assigns 76 percent of all pairs to it, with a plausible range of 59 to 83 percent. Between 10 and about 14 solar masses, the merger rate falls by roughly a factor of 100, and for heavier members between 12.0 and 16.1 solar masses it becomes statistically indistinguishable from zero. With the GWTC-5.0 data that empty band narrows slightly to 12.3 to 15.7 solar masses. The heavier population starts again around 16 solar masses, though that point is loosely pinned down.

“We see 10 solar-mass black holes merging all the time,” Legred said. “But why are, say, 14 solar-mass black holes so rare?”

The spins point the same way. Black holes in the ten-solar-mass group fall within a narrower range of spin rates than the rest of the sample, which suggests they formed through one shared route.

One older explanation for the peak is that stars of a certain size grow unusually dense cores that resist exploding. The paper argues that effect accounts for a rate boost of two to three times at most, while the observed excess near 10 solar masses is at least 50 times. Computer simulations of collapsing stars offer a better match: in several recent models, stars born with roughly 11 to 14 times the Sun’s mass at solar metallicity are the ones that fail to launch an explosion, while slightly lighter and heavier stars blow up and leave neutron stars or much smaller black holes.

For scale, the heaviest known neutron stars weigh a little over two Suns, and the merger known as GW190521 left a black hole of about 142 solar masses. The crowded corner sits near the light end of that range.

What the study doesn’t settle

No failed supernova has been confirmed directly, and the Caltech statement says plainly that the study found no direct evidence for one. The best candidates are two stars that faded from view, one in the galaxy NGC 6946 and one in Andromeda, catalogued as M31-2014-DS1. Both remain disputed, because a merger of two stars wrapped in dust could produce a similar fade. “We can look up at the sky and try to study historical records of massive stars that failed to explode, but it’s challenging because they just vanish without exploding, leaving a dark patch of sky,” Legred said.

The gap itself is not yet secure. The authors put the probability of a real separation between the peak and the heavier black holes at 88 percent with GWTC-4.0 and 85 percent with GWTC-5.0. When they compared their model with the collaboration’s standard population model, a Bayes factor of about 0.21 gave a mild preference to the standard model, even though the best fit of the new model was about 547 times more likely. The new model fits the data better, but it pays for that with extra parameters, and the statistics penalize it.

The result also depends on how the model is built. Earlier flexible analyses found no compelling gap, and a 2024 study that searched the same mass range found the rate drops by 20 to 30 times but never reaches zero. When Legred’s team forced both black holes in each pair to come from the same population, the sharp minimum just above 12 solar masses disappeared. Labelling the heavier member of each pair as a distinct class has its own pitfalls, because sorting any two masses by size skews the distribution of each.

The sample has blind spots too: pairs whose lighter member could be a neutron star were left out, and small black holes from successful supernovae are not modelled.

Common questions about failed supernovae and black holes

What is a failed supernova?

It is the death of a massive star in which the core collapses but the outward shock that normally blows the star apart stalls. Most of the star then falls inward and forms a black hole, with little or no visible explosion. From outside, the star would simply fade and vanish.

How do scientists weigh a black hole they cannot see?

As two black holes spiral together, the gravitational waves they emit rise in frequency and strength at a rate set by their masses and spins. Matching the recorded signal against models gives the mass of each one, usually to within a few solar masses.

Why are 10-solar-mass black holes so common?

Nobody knows for sure. This study shows the data are consistent with a special group of stars, born at around 11 to 14 solar masses, that collapse without exploding and leave black holes near 10 solar masses. Denser stellar cores alone seem too weak an effect to explain a crowd that size.

Has anyone seen a star disappear without exploding?

Two candidates exist: a red supergiant in the galaxy NGC 6946 and a supergiant in Andromeda called M31-2014-DS1, both of which faded from view. Neither is confirmed, because a dusty merger of two stars could look much the same.

The next test is already recorded

The paper was published on November 10, 2026, in Physical Review D, and Caltech announced it on November 17. The fourth observing run ended on November 18, 2025, and the detections from its final stretch, January to November 2025, have not yet been released as a catalog. If the empty band between 12 and 16 solar masses survives the next catalog, the case for a separate population of failed-supernova black holes gets much stronger. If it fills in, the crowd at 10 solar masses will need another explanation.

Reference: Legred et al., “Low-mass failed supernovae and the 10 M⊙ peak in the merging black hole mass distribution,” Physical Review D, 2026. DOI: 10.1103/mnzq-w1b5

Tags: , , , , ,

Add us on Google

Discussion

There are 0 comments.