Things That Go Bump in the Night: Mapping Black Hole Collisions by Mass

69 observed mergers, 500 simulated catalogs, and one question: why do black holes prefer certain masses when they collide?

Farah, Amanda M.; Edelman, Bruce; Zevin, Michael; Fishbach, Maya; Ezquiaga, Jose Maria; Farr, Ben; Holz, Daniel|2022|27,068|View on Zenodo →
69observed binary black hole events
500mock catalogs for statistical validation
27Kdownloads by gravitational wave researchers+8K in last year
~35 M☉mass peak suggesting pair-instability physics

The universe has a preferred recipe for building black holes

On September 14, 2015, the LIGO detectors in Louisiana and Washington state simultaneously shuddered. The cause was a gravitational wave — a ripple in the fabric of spacetime — generated by two black holes spiraling into each other 1.3 billion light-years away. That first detection opened a new window on the universe. By 2022, the LIGO-Virgo-KAGRA network had recorded 69 binary black hole mergers, each one a pair of invisible objects revealed only by the spacetime distortion of their final collision.

Amanda Farah and her collaborators at the University of Chicago asked a deceptively simple question: what do the masses of these merging black holes tell us about how they formed? If black holes were produced randomly, their masses would follow a smooth, featureless distribution. Instead, the team found bumps and gaps — statistically significant features in the mass spectrum that suggest nature has preferred channels for building black holes. Using a flexible Power Law + Spline model, they fitted the mass distribution with enough freedom to detect deviations from simple power-law behavior without imposing artificial structure.

The dataset released here — 500 mock catalogs, injection sets, and hyperposterior samples — is the infrastructure of that analysis. It has been downloaded over 27,000 times by gravitational wave researchers who use it to test their own population models against the same observational evidence. The bumps in the mass spectrum near 10 and 35 solar masses may encode the physics of stellar evolution: how massive stars explode, which ones leave behind black holes, and which ones are destroyed entirely. Each feature in the spectrum is a fossil record of processes that occurred millions of years before the merger itself.

Primary black hole mass distribution

The mass spectrum of the heavier black hole in each merger, showing peaks and gaps that deviate from a simple power law

Mass ratio distribution of merging pairs

How symmetric are black hole mergers? A mass ratio of 1.0 means equal-mass partners

The bumps in the mass spectrum are fossil records — echoes of stellar deaths that occurred millions of years before the black holes collided.
01
A statistically significant excess of black holes near 35 solar masses may be linked to pair-instability supernovae — explosions so powerful they prevent heavier stellar cores from collapsing into black holes
02
Most mergers involve roughly equal-mass partners, with 35% of events having mass ratios above 0.9 — suggesting that formation channels preferentially pair similar black holes
03
The Power Law + Spline model detects features that simple parametric models miss, revealing structure in the mass spectrum that encodes stellar evolution physics

Stellar Physics

The features in the black hole mass spectrum constrain how massive stars end their lives. The peak near 35 solar masses and the apparent gap above 50 solar masses are predicted by pair-instability supernova theory — and this dataset provides the most direct observational test of those predictions.

🔬

Methodology

The 500 mock catalogs enable rigorous testing of population inference methods. As LIGO's catalog grows from 69 to hundreds of events in the coming observing runs, the statistical framework validated against this dataset will be essential for extracting increasingly precise population measurements.

🌌

Cosmological Implications

Understanding the black hole mass spectrum is a prerequisite for using gravitational wave events as cosmological probes. The mass distribution affects how gravitational wave signals are detected and interpreted, with direct consequences for measuring the expansion rate of the universe using 'standard sirens.'

Share this story

View on Zenodo →