Every black hole merger that gravitational-wave detectors have ever caught tells a story that ends the same way: two dense objects spiral together and become one. But a new study argues that for a meaningful slice of these events, the story does not start at the beginning. Some of the black holes now colliding have already collided before. They are the survivors of earlier smash-ups, second-generation objects carrying the mass and spin of their own violent histories into the next merger.
The work, published in Physical Review Letters and led by a team at MIT, mines the latest LIGO-Virgo-KAGRA gravitational-wave catalog, GWTC-4.0, for statistical fingerprints of this kind of repeated, or hierarchical, merging. The authors — Cailin Plunkett, Thomas Callister, Michael Zevin, and Salvatore Vitale — conclude that roughly 14% of the merging black holes in the dataset show evidence of having been through at least one prior collision event — that is, of being second-generation objects. In other words, a substantial minority of the population has a past life.
What "second-generation" actually means
Stars end their lives in a fairly predictable way. A massive star collapses and leaves behind a black hole in a characteristic mass range — the MIT team's framing puts these first-generation, stellar-born black holes at roughly 10 to 30 solar masses. That is the raw material the universe manufactures directly.
But black holes do not always stay put. In dense stellar environments — the crowded cores of star clusters, where objects are packed tightly enough to interact repeatedly — a black hole that forms from one merger can sink toward the center, find another partner, and merge again. Do that once and you get a second-generation object. The study places these products of previous smash-ups at 20-plus solar masses, with some pushing past 40 to 45 solar masses. The mass tells the tale: to build a black hole that heavy, it helps to have started from something that was already the sum of two smaller ones.
This is the mechanism the researchers model explicitly. Their analysis tracks the spin dynamics of repeated black-hole mergers inside dense stellar clusters and asks whether that machinery can populate the parts of the mass distribution that ordinary stellar collapse struggles to fill.
The 46-solar-mass line
The most striking result is a transition. When the team fits the population's spin behavior, they find a threshold near 46 solar masses. Below it, the picture is a mix — mostly first-generation black holes, with hierarchical objects woven in. Above it, the population becomes almost entirely hierarchical. Cross that line, and you are essentially looking at black holes with histories.
That number is not arbitrary. It sits at what astrophysicists call the pair-instability mass gap: a range where the physics of dying stars is expected to forbid black-hole formation outright. Very massive stars are thought to be torn apart by runaway electron-positron pair production before they can collapse into a black hole, leaving a gap in the mass spectrum where stellar processes simply should not deliver objects. If black holes exist in and above that gap — and detectors keep finding them — something other than a single star's death has to be putting them there. Repeated mergers are the leading candidate, and the sharp transition near 46 solar masses is exactly the signature you would expect if hierarchical assembly is the process filling in the forbidden zone.
There is a second feature in the data as well: a peak in the hierarchical merger rate near 16 solar masses. That lower-mass bump suggests the repeated-merger channel is not confined to the exotic heavyweight end of the population. It leaves a mark down in the range where first-generation black holes also live, which is part of why teasing the two populations apart requires careful statistical modeling rather than a simple mass cutoff.
Why spin is the tell
Mass alone does not prove a black hole has a past. A heavy object could, in principle, come from an unusually massive star. What lets the researchers separate generations is spin. When two black holes merge, the orbital angular momentum of the pair gets baked into the spin of the remnant, tending to leave second-generation black holes with a distinctive spin signature that first-generation, collapse-born objects do not share. By modeling how spin is distributed across the population — and how it correlates with mass — the team can infer which events are more likely to involve objects that have merged before. That spin-mass relationship is what anchors both the 46-solar-mass transition and the estimate that about 14% of merging black holes have been through repeated collisions.
Why It Matters
Gravitational-wave astronomy began as a way to confirm that black holes merge at all. It is now mature enough to do population genetics on them. Distinguishing black holes born directly from stars from those assembled through repeated collisions turns the growing catalog of detections into a probe of where these mergers happen: the dense stellar clusters that make hierarchical assembly possible in the first place. A population that is "almost entirely hierarchical" above 46 solar masses gives astronomers a concrete, testable handle on the pair-instability mass gap — one of the sharpest predictions stellar physics makes about what black holes should and should not exist. And the claim that roughly one in seven merging black holes has a hidden history reframes the whole dataset: many of the events LIGO, Virgo, and KAGRA record are not first chapters but sequels. As GWTC catalogs keep growing, that statistical portrait will only sharpen, and the boundary between what stars build and what gravity assembles afterward will come into clearer focus.
