Primordial Black Hole Clusters Get Torn Apart in the Galaxy, Complicating Dark Matter Searches
Date: 2026-07-14
Featured image: [Artist’s impression of a cluster of primordial black holes in the Milky Way halo; credit: NASA/JPL-Caltech]
If dark matter is made of tiny black holes formed in the first instants after the Big Bang, they probably did not stay in dense clusters for long. A new study from the Lebedev Physical Institute in Moscow shows that clusters of primordial black holes are steadily torn apart by gravitational encounters as they orbit the Milky Way, scattering individual black holes across the galactic halo.
The finding, submitted to arXiv on July 10, has direct consequences for how astronomers search for primordial black holes using gravitational microlensing. The appeal of clustering had been that it might help primordial black holes evade existing microlensing limits by producing complex light curves that survey telescopes cannot easily identify. The new results show that this evasion is only partial.
“The original appeal of clustering as a way to evade microlensing limits is therefore only partially realized,” the authors write. “A substantial fraction of the inner-halo cluster population is disrupted over the galaxy’s lifetime regardless of initial cluster mass.”
How Clusters Break Apart
The study, led by M.V. Tkachev and S.V. Pilipenko, combines cosmological N-body simulations with 72 dedicated binary collision simulations to track the fate of primordial black hole clusters from redshift 9 to the present day. They simulate clusters of two masses: 1 million solar masses, containing roughly 33,000 individual black holes of 30 solar masses each, and 10 million solar masses, containing roughly 330,000 black holes.
The disruption mechanism is cumulative. Each time a cluster passes near another cluster, a small fraction of its member black holes are stripped away. Individual encounters are weak, typically ejecting less than 1 percent of a cluster’s mass per event. But over 13.8 billion years, the hits add up.
At the Sun’s distance from the galactic center, a 1-million-solar-mass cluster retains about 50 percent of its original mass by the present day. A 10-million-solar-mass cluster retains only 4 percent, because larger clusters present a bigger gravitational target and collide more frequently.
Half of the total mass loss occurs before redshift 2, during the early, cold phases of galaxy assembly, a channel that standard zero-redshift analytic estimates entirely miss.
What This Means for Dark Matter Searches
The surviving smooth fraction of primordial black holes, measured toward the Large and Small Magellanic Clouds, is about 49 percent for the smaller cluster size and more than 90 percent for the larger. Those free-flying black holes are subject to standard point-lens microlensing limits from surveys like OGLE, EROS, MACHO, and the Subaru Hyper Suprime-Cam.
The paper emphasizes that reanalyses of microlensing survey data must adopt a radially varying smooth fraction rather than treating clusters as static lenses. A population that is 90 percent smooth in the inner halo (where most surveys look) is not meaningfully evading detection.
The results are consistent with a companion paper by Toshchenko and colleagues from April 2026, which found that up to 93 percent of primordial black hole dark matter in clusters could evade microlensing detection due to complex light curves, but that a significant population of isolated black holes remains, meaning constraints are not completely lifted.
Gravitational Wave Connections
The stripping of primordial black holes from clusters into isolated objects also affects binary formation rates and merger dynamics. The same group’s prior work showed that primordial black hole binary merger rates could be enhanced by a factor of 6 to 8 due to clustering effects, which would imply a lower abundance of primordial black holes needed to explain LIGO and Virgo’s observed merger rate.
Earlier work from the group also showed that even a tiny primordial black hole fraction of one part in 10,000 leads to strong concentration in the centers of dwarf galaxies, tightening astrophysical constraints by two orders of magnitude.
The paper does not settle whether primordial black holes constitute dark matter, but it sharpens the question. Future microlensing surveys will need to account for a mixed population of clustered and smooth primordial black holes, with the smooth fraction varying by location in the galaxy. The answer to whether dark matter is made of black holes will depend on getting that radial distribution right.