Do post-merger gravitational-wave echoes reveal structure at the horizon scale?
In plain words
If a black hole had a reflecting surface just outside where its horizon should be, part of the ringing would bounce back and forth and produce delayed repeated pulses called echoes. Finding or excluding them tests whether horizons are what general relativity says.
Precise statement
For a nonspinning remnant of mass $M$ with a partially reflecting surface at areal radius $r_s+\delta$, echoes repeat with delay Delta t ~ (4 G M/c^3) |ln(delta/r_s)|; a surface one Planck length $l_P=\sqrt{G\hbar/c^3}$, approximately 1.6 x 10^-33 cm, of proper distance outside the horizon has $\delta\sim l_P^2/r_s$, so Delta t ~ (8 G M/c^3) ln(r_s/l_P). Determine whether data contain such echoes and bound the surface reflectivity $R$ as a function of the proper distance of the surface from the horizon, down to $l_P$. An answer is a detection or upper limits on R.
What would settle it
Template-based and model-independent echo searches in loud events with upper limits on R for a surface at Planck-scale proper distance.
Status in the literature
Tentative echo claims (2016-2017) were not confirmed by later searches, including collaboration analyses through GWTC-3.