Dynamical Formation of Low-Mass Merging Black Hole Binaries like GW151226

TL;DR

This study uses numerical models of star clusters with varying ages and metallicities to show how different types of binary black hole mergers, like GW151226, can form dynamically in the local universe, emphasizing the role of cluster properties.

Contribution

It demonstrates for the first time that lower-mass BBHs like GW151226 can be formed dynamically in younger, higher-metallicity clusters, expanding understanding of BBH formation channels.

Findings

GW151226's mass aligns with models for clusters with Z/Zsun≥0.5.

Younger, higher-metallicity clusters can produce BBHs like GW151226.

Massive GW150914-like systems require even lower metallicity environments.

Abstract

Using numerical models for star clusters spanning a wide range in ages and metallicities (Z) we study the masses of binary black holes (BBHs) produced dynamically and merging in the local universe ($z\lesssim0.2$). After taking into account cosmological constraints on star-formation rate and metallicity evolution, which realistically relate merger delay times obtained from models with merger redshifts, we show here for the first time that while old, metal-poor globular clusters can naturally produce merging BBHs with heavier components, as observed in GW150914, lower-mass BBHs like GW151226 are easily formed dynamically in younger, higher-metallicity clusters.More specifically, we show that the mass of GW151226 is well within $1\sigma$ of the mass distribution obtained from our models for clusters with Z/Zsun$\gtrsim0.5$. Indeed dynamical formation of a system like GW151226 likely requires a cluster that is younger and has a higher metallicity than typical Galactic globular clusters. The LVT151012 system, if real, could have been created in any cluster with Z/Zsun$\lesssim0.25$. On the other hand, GW150914 is more massive (beyond $1\sigma$) than typical BBHs from even the lowest-metallicity (Z/Zsun=0.005) clusters we consider, but is within $2\sigma$ of the intrinsic mass distribution from our cluster models with Z/Zsun$\lesssim0.05$; of course detection biases also push the observed distributions towards higher masses.