Searching for gravitational wave echoes from black hole binary events in the third observing run of LIGO, Virgo, and KAGRA collaborations

Gravitational wave echo signals have been proposed as evidence for the modification of the spacetime structure near the classical event horizon. These signals are expected to occur after the mergers of compact binaries as a sequence of weak pulselike signals. Some studies have shown evidence of the echo signals from several binary black hole merger events. On the other hand, the other studies have shown the low significance of such signals from various events in the first, second, and third observing runs (O1, O2, and O3). Our previous study also shows the low significance of echo signals from events in O1 and O2, though, we observe that more than half of the events have p-value smaller than 0.1 when the simply modeled waveform is used for the analysis. Since there are only nine events appropriate for this analysis in O1 and O2, it is necessary to analyze more events to evaluate the significance statistically. In this study, we search for echo signals from binary black hole events observed during O3 operated by LIGO, Virgo, and KAGRA collaborations. We perform the template-based search by using two different models for echo signal templates: simply modeled one and physically motivated one. Our results show that the distributions of p-values for all events analyzed in this study are consistent with the noise distribution. This means that no significant echo signals are found for both models from O3 events.

Figure 1

Spectra of the best-fit echo templates of the Simple (blue) and BHP (orange) models for GW190412 are shown. Detector’s noise for this event for Hanford (red) and Livingston (cyan) are also shown. Best fit values are $(\mathrm{\Delta }{t}_{\mathrm{echo}},\gamma )=(0.1129s,0.9)$ for the Simple model and $(a,M,\varphi )=(0.68,45.2M_{\odot },2.356)$ for the BHP model.

Figure 2

Left: distributions of a cumulative probability of p-values for the O3 events for the Simple model (blue) and the BHP model (orange). P-values are calculated from the first background estimation method (BG1). Black dashed line corresponds to the uniform distribution (null hypothesis). Shaded regions correspond to the 90% error regions. Right: The same as the left panel, but for the improved background estimation method case (BG2).

Figure 3

Distributions of a cumulative probability of p-value for injected signals for the Simple model (blue) and the BHP model (orange). P-values are calculated from the improved background estimation method (BG2).