Analysis of a subsolar-mass black hole trigger from the second observing run of Advanced LIGO

We perform an exhaustive follow-up analysis of a subsolar-mass black hole candidate from the second observing run of Advanced LIGO, reported by Phukon et al. in 2021. The origin of this trigger is unclear, because the reported signal-to-noise ratio (SNR) of $8.6$ and inverse false alarm rate of about $0.5$ yr are too low to claim a gravitational-wave origin, but large enough to be intriguing. When using more precise waveforms, extending the frequency range down to 20 Hz, removing a prominent blip glitch and marginalizing over all the model parameters, we find that the network signal-to-noise ratio posterior distribution lies mostly below the search value, with the 90\% confidence interval being $7.94^{+0.70}_{-1.05}$. If one assumes that the signal comes from a real gravitational-wave merger event, we find a light component $m_2 = 0.76^{+0.50}_{-0.14} M_\odot$, suggesting a compact object of mass below one solar mass at $83.8\%$ confidence level. Such a low mass for a compact object would suggest an unexpectedly light neutron star or a black hole of primordial or exotic origin. The primary mass would be $m_1 = 4.71^{+1.57}_{-2.18} M_\odot$, likely in the lower mass gap, for a mass ratio of $q=0.16^{+0.34}_{-0.06}$, at a distance of $D_{\rm L}=124^{+82}_{-48}$ Mpc. The improved sensitivity of the next observing runs would make it possible to observe similar signals with a higher SNR and to distinguish a sub-solar mass component.