Poking Holes: Looking for Gaps in LIGO/Virgo's Black Hole Population

TL;DR

This paper investigates the existence of an upper mass gap in black hole populations using gravitational wave data, extending models to include a potential gap and analyzing its evidence with Bayesian methods.

Contribution

It introduces extended mass distribution models allowing for a mass gap and applies them to GWTC-2 data to assess the presence of the upper mass gap.

Findings

Mild evidence for an upper mass gap with the Truncated model.

No significant preference for a gap with the Powerlaw+Peak model.

Results depend on the interpretation of GW190521's mass measurements.

Abstract

Stellar evolution models predict the existence of a gap in the black hole mass spectrum from $\sim55 M_\odot - 120 M_\odot$ due to pair-instability supernovae (PISNe). We investigate the possible existence of such an "upper" mass gap in the second gravitational wave transient catalog (GWTC-2) by hierarchically modeling the astrophysical distribution of black hole masses. We extend the Truncated and Powerlaw+Peak mass distribution families to allow for an explicit gap in the mass distribution, and apply the extended models to GWTC-2. We find that with the Truncated model there is mild evidence favoring an upper mass gap with log Bayes Factor $\ln \mathcal{B} = 2.79$, inferring the lower and upper bounds at $56.12_{-4.38}^{+7.54} M_\odot$, and $103.74_{-6.32}^{+17.01} M_\odot$ respectively. When using the Powerlaw+Peak model, we find no preference for the gap. When imposing tighter priors on the gap bounds centered on the expected PISN gap bounds, the log Bayes factors in favor of a gap mildly increase. These results are however contingent on the parameter inference for the most massive binary, GW190521, for which follow up analyses showed the source may be an intermediate mass ratio merger that has component masses straddling the gap. Using the GW190521 posterior samples from the analysis in Nitz & Capano (2021), we find an increase in Bayes factors in favor of the gap. However, the overall conclusions are unchanged: There is no preference for a gap when using the Powerlaw+Peak model. This work paves the way for constraining the physics of pair-instability and pulsational pair-instability supernovae and high-mass black hole formation.