Have We Found the Black Hole Desert?

Astronomers disagree on whether they’ve found evidence that stars don’t make certain sizes of black hole.

Scientists expect a gap in the range of masses a black hole can have. Due to the vagaries of nuclear fusion, some goliath stars — we’re talking 100 to 260 times the Sun’s mass — should blow themselves to smithereens, leaving no remnant behind. This stellar annihilation will leave its mark on the population of black holes that are born when stars die: No black holes should exist with masses between roughly 50 and 130 Suns. (The exact bounds are uncertain.)

Yet over the last few years, a global consortium of gravitational-wave detectors called LIGO-Virgo-KAGRA (LVK) has spotted several events that seem to be mergers involving black holes with masses in this forbidden range.

The hefty black holes might be second-generation objects, formed by an earlier melding of two black holes. These objects then meet up with another partner and merge. The second merger is the event that sends out the spacetime ripples we see.

One indicator that this multi-generation scenario holds true would be how the two merging black holes’ spins are oriented with respect to the plane of their orbit around each other. Fabio Antonini (Cardiff University, UK) and others have previously calculated that black holes that meet up well after they’re born would have a wide range of spins and/or spin orientations. Black hole binaries that formed together as stars, on the other hand, tend to have lower individual spins and to dance around each other upright.

Antonini and his collaborators saw evidence of this pattern in LVK’s previous catalog of gravitational-wave signals: Mergers in which the bigger black hole (called the primary) is at least 45 solar masses or so tend to have a variety of spins and orientations, suggesting those binaries came together after the black holes themselves had already formed.

Now, Hui Tong (Monash University, Australia) and colleagues say they’ve taken things a step further and found evidence of the mass gap itself by looking at the smaller black hole of each pair.

Measuring the mass of a black hole involved in a collision isn’t easy. Each mass estimate isn’t an exact number; it’s a range of possible values, inferred from the properties of the spacetime ripples the merger created. Even after taking that into account, tallying up how many black holes exist of each mass is tricky, because gravitational-wave detectors don’t see all merging black holes equally well — they’re sensitive to certain events more than others.

“Because of all this, simply counting the masses can be misleading,” Tong says. Instead, researchers use statistical models to combine the data and detectors’ limitations with various assumptions, in order to step back and see what the larger black hole population might be like.

Tong and his collaborators looked at the 153 clearest events from the fourth, latest LVK catalog. Unlike in previous work, they focused on the smaller black hole in each pair, called the secondary. Their analysis, reported April 1st in Nature, indicates a sharp gap begins in the secondaries’ masses around 45 Suns.

This cutoff is the same mass at which a wider range of spin and spin orientations begins, they note. Thus, it’s likely that the too-big primaries are second-generation objects from a previous merger, and their smaller partners came straight from stars, with masses constrained by the way big stars blow up.

However, another team disagrees. Anarya Ray and Vicky Kalogera (both Northwestern University) analyzed the same data, using models they say are more flexible than those that Tong’s team employed. (Tong’s team disputes that comparison.) While both groups agree that there’s a drop-off in the number of secondaries with masses above roughly 40 Suns, Ray and Kalogera say it’s a far gentler decline, not a cliff. There’s a hint of a gap starting around 60 solar masses, but we haven’t detected enough events to know for sure, they conclude.

Furthermore, the duo’s analysis suggests that many of these high-mass binaries pair objects with more equal sizes than expected if a second-generation and first-generation black hole had joined up. The study appears in the February 10th Astrophysical Journal Letters. (The publication date is before the Nature paper because this debate began when the pro-gap study was still in draft form on the arXiv preprint server.)

Both teams are sticking to their guns.

As for Antonini, he says the Tong paper is interesting, but he’s unconvinced that the evidence for the gap in secondary masses is strong. He’d want to see more models detect the gap in the data.

“It is likely that we have to wait some time for a definitive answer,” Ray says — at least until the LVK collaboration finishes analyzing the latest observing run, which ended last November, and releases its findings. But if there are not enough high-mass mergers in that crop to shed substantial light on the question, we’ll have to wait longer.

LVK will likely announce another 200 or so detections in the next year. The collaboration hopes to do a six-month observing run starting this September or October. Then it will go offline for major upgrades, coming back online hopefully in 2028 for a longer run.

References:

Hui Tong et al. “Evidence of the Pair-Instability Gap from Black-Hole Masses.” Nature. April 1, 2026. Draft preprint here.

Anarya Ray and Vicky Kalogera. “Reexamining Evidence of a Pair-Instability Mass Gap in the Binary Black Hole Population.” Astrophysical Journal Letters. February 10, 2026.

Fabio Antonini et al. “Star Cluster Population of High Mass Black Hole Mergers in Gravitational Wave Data.” Physical Review Letters. January 7, 2025.https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.134.011401