Black hole mergers of AGN origin in LIGO–Virgo’s O1–O3a observing periods

The origin of the black hole mergers detected by LIGO and Virgo remains an open question. While the unusual mass and spin of a few events constrain their possible astrophysical formation mechanisms, it is difficult to classify the bulk of the observed mergers. Here we consider the distribution of masses and spins in LIGO/Virgo's first and second observing catalogs, and find that for a significant fraction (25%) of these detected events, an AGN-disk origin model is preferred over a parametric mass-spin model fit to the full GWTC-2 merger sample (Bayes factor B>10). We use this to estimate the black hole merger rate in AGNs to be about 2.8±1.8\, Gpc−3yr−1, comparable to theoretical expectations. We find that AGNs can explain the rate and mass distribution of the observed events with primary black hole mass in the pair-instability mass gap (M≳50\, M⊙).