GW250114 reveals black hole horizon signatures

TL;DR

The study targets direct observational access to black hole horizons by identifying horizon-driven direct waves in GW250114. After removing dominant quasinormal modes with a rational filter, the authors model the residual merger signal as a damped oscillator and corroborate this with a matched-filter analytic template anchored to the remnant properties. They find a direct-wave component with a frequency near $2\Omega_H$ and damping tracking $\kappa$, consistent with horizon-frame dragging and redshift, providing robust observational support for near-horizon physics. This establishes a new channel to probe horizon dynamics in dynamical, strong-gravity regimes and complements traditional black hole spectroscopy in testing general relativity and horizon physics.

Abstract

The horizon of a black hole, the "surface of no return," is characterized by its rotation frequency $Ω_H$ and surface gravity $κ$. A striking signature is that any infalling object appears to orbit at $Ω_H$ due to frame dragging, while its emitted signals decay exponentially at a rate set by $κ$ as a consequence of gravitational redshift. Recent theoretical work predicts that the merger phase of gravitational waves from binary black hole coalescences carries direct imprints of the remnant horizon's properties, via a "direct wave" component that (i) oscillates near $2Ω_H$, reflecting the horizon's frame dragging and the quadrupole nature of the gravitational radiation, and (ii) decays at an increasing rate characterized by $κ$, with additional screening from the black hole's potential barrier. In this paper, we report observational evidence for the direct wave in GW250114 with a matched-filter signal-to-noise ratio of $14.0^{+0.2}_{-0.1}$ ($13.5^{+0.1}_{-0.2}$) in the LIGO Hanford (Livingston) detector. The measured properties are in full agreement with theoretical predictions. These findings establish a new observational channel to directly measure frame-dragging effects in black hole ergospheres and explore (near-)horizon physics in dynamical, strong-gravity regimes.

Figures (9)

• Figure 1: Wave emission near the merger stage of a binary black hole coalescence, modeled as a point particle (small filled circle) spiraling into the remnant Kerr black hole, following the widely used Effective One-Body formalism. The red-shaded region indicates the potential barrier near the light ring, enclosing the ergosphere (green area), where the trajectory experiences strong frame dragging. Wave emission driven by the particle's motion persists from the early inspiral (blue arrow) through the final plunge, where waves (gray arrows) are gradually silenced by the remnant horizon. These waves are further screened by the potential barrier while propagating toward distant observers (black arrows). Simultaneously, quasinormal modes are excited during the barrier crossing (orange arrows). The merger portion of a gravitational-wave signal is a superposition of source-driven direct waves (black arrows) and free quasinormal-mode oscillations (orange arrows).
• Figure 2: Whitened strain data from the GW250114 event in the LIGO Hanford detector. Top: Schematic illustration of the three stages of a binary black hole coalescence: inspiral, merger, and ringdown. (a) Observed strain data (grey) and the corresponding NRSur7dq4 waveform reconstruction (red), bandpass-filtered between 20--2000 Hz. (b) Residual strain after removing the dominant quasinormal modes $(\ell=m=2,n=0,1,2)$. (c) Residual strain after removing both the quasinormal modes and the direct wave component. The merger time is aligned to $t=0$.
• Figure 3: Comparison between waveform components. Shown are the real part of the quadrupolar harmonic $(\ell= m =2)$ of the NRSur7dq4 waveform (grey), the waveform after removing the $(\ell= m=2, n=0,1,2)$ quasinormal modes (black), and an analytic model characterizing the direct wave signal (red). The analytic model closely matches the quasinormal-mode-removed waveform as early as $t=-7M_\mathrm{f}^\mathrm{det}$.
• Figure 4: Frequency and damping rate inferred from the data after removing dominant quasinormal modes, revealing the direct-wave signal. Results are shown as a function of analysis starting time (color scale) $t_\mathrm{start} \in [-9,-3]M_\mathrm{f}^\mathrm{det}$, covering and extending slightly beyond the theoretically supported regime for direct waves (see Fig. \ref{['fig:theoretical_comparison_7m']}). Plus symbols indicate fits to the reconstructed NRSur7dq4 waveform, while colored solid contours represent 90% credible regions derived from the real event data. Colored dashed contours mark the expected frequencies and damping rates of various quasinormal modes (90% credible), none of which intersect with the inferred direct-wave parameters. The red shaded region indicates the predicted horizon mode from Eq. \ref{['eq:omega_H']}. While the direct-wave frequencies are not expected to precisely match the horizon mode, they are anticipated to lie nearby.
• Figure 5: Inferred frequency and damping rate of the direct-wave signal at $t_\mathrm{start} = -7M_\mathrm{f}^\mathrm{det}$ under different analysis settings for $L_\mathrm{seg}$, $f_\mathrm{samp}$, and sky localization (with "max-L" denoting maximum-likelihood values). The main results presented in the paper correspond to the settings in blue. The analysis is robust against variations in these settings.