Revisiting the Cosmic String Origin of GW190521

For the first time we analyze gravitational-wave strain data using waveforms constructed from strong gravity simulations of cosmic string loops collapsing to Schwarzschild black holes; a previously unconsidered source. Since the expected signal is dominated by a black-hole ringdown, it can mimic the observed gravitational waves from high-mass binary black hole mergers. To illustrate this, we consider GW190521, a short duration gravitational-wave event observed in the third LIGO-Virgo-KAGRA observing run. We show that describing this event as a collapsing cosmic string loop is favored over previous cosmic string analyses by an approximate log Bayes factor of 22. The binary black hole hypothesis is still preferred, mostly because the cosmic string remnant is nonspinning. It remains an open question whether a spinning remnant could form from loops with angular momentum, but if possible, it would likely bring into contention the binary black hole preference. Finally, we suggest that searches for ringdown-only waveforms would be a viable approach for identifying collapsing cosmic string events and estimating their event rate. This Letter opens up an important new direction for the cosmic-string and gravitational-wave communities.

Figure 1

Gravitational waveform $h_{\ell m}^{+}$ of a collapsing cosmic string loop with radius $R_0=2.85\mathrm{AU}$ and dimensionless string tension $G\mu /c^2={10}^{-7}$. This corresponds to an initial mass of $M_0\approx 181M_{\odot }$ that collapses to form a remnant black hole of mass $M\approx 175M_{\odot }$ in the source frame, radiating $\approx 6M_{\odot }$ in gravitational waves. The vertical dashed line depicts the transition between the infall and black hole formation stages, where there is a sudden change in the frequency content of the signal.

Figure 2

The time-domain gravitational wave strain data for the LIGO-Livingston detector [90, 92] around the time of GW190521 [22]. The top, middle and bottom panels plot the best-fit detector-frame waveforms assuming a cosmic string cusp, loop collapse and BBH merger respectively. The cosmic string cusp waveform is the best-fitting template from the LVK’s cosmic string matched filter search pipeline [85, 86] and the BBH waveform is the reconstructed maximum likelihood waveform from the LVK analysis [22, 87]. The strain data and waveforms are whitened by the noise power spectral density provided as part of the LVK data release [87]. We shift the data and waveforms by the coalescence time in the LIGO-Livingston detector $t_{\mathrm{c}}$ as reported in the LVK BBH analysis.

Figure 3

Network signal-to-noise ratios (assuming a network of the LIGO-Hanford, LIGO Livingston and Virgo detectors) for the best-fit BBH, loop and cusp waveforms, $\rho =\{14,12,10\}$, respectively. As with Fig. 2, we shift the time series by the coalescence time in the LIGO-Livingston detector $t_{\mathrm{c}}$ as reported in the LVK BBH analysis.