Reversible-jump MCMC reveals binary black hole subpopulations with distinct redshift evolution

Abstract

Analyses of the growing catalog of binary black hole (BBH) mergers observed by the LIGO-Virgo-KAGRA detectors are beginning to resolve features in their population-level mass, spin, and redshift distributions, revealing imprints of the astrophysical processes driving their formation and evolution. We present a novel method to search for subpopulations in the data using reversible-jump Markov chain Monte Carlo, providing interpretable results while making minimal prior assumptions. We find evidence for three subpopulations: a narrow subpopulation in primary mass at $\sim 10~M_\odot$ with preferentially aligned spins and unequal masses, consistent with isolated binary evolution; a subpopulation broadly distributed around $\sim 30~M_\odot$ with isotropically-distributed spins and a strong preference for equal mass ratios, consistent with dynamical formation in clusters; and a high-spin subpopulation spanning the continuum in mass, which we interpret as the confluence of multiple subdominant formation channels. When we allow for the independent redshift evolution of each subpopulation, we find that the subpopulation encompassing the $10~M_\odot$ peak evolves more quickly than the $30~M_\odot$ subpopulation, with implications for the delay-time distribution and metallicity-dependent BBH formation efficiency. Our work lays the foundation for a novel data-driven framework to infer the formation mechanisms of BBHs.