Blind spots and biases: the dangers of ignoring eccentricity in gravitational-wave signals from binary black holes

Abstract

Most gravitational wave (GW) events observed by the LIGO and Virgo detectors are consistent with mergers of binary black holes (BBHs) on quasi-circular orbits. However, some events are also consistent with non-zero orbital eccentricity, which can indicate that the binary formed via dynamical interactions. Active GW search pipelines using quasi-circular waveform templates are inefficient for detecting eccentric mergers. Also, analysing eccentric GW signals with waveform models neglecting eccentricity can lead to biases in the recovered parameters. We explore the detectability and characterisation of eccentric signals when searches and analyses rely on quasi-circular waveform models. We find that for a reference eccentric population, the fraction of events having fitting factor (FF) $< 0.95$ can be up to $\approx 2.2\%$ compared to $\approx 0.4\%$ for the baseline population. This leads to the loss in signal recovery fraction for up to $6\%$ for parameter space with non-negligible eccentricity ($e_{10} > 0.01$) and high mass ratio ($q > 3$). We perform parameter estimation (PE) for non-spinning and aligned-spin eccentric GW injections from BBHs with a total mass $M=35 M_\odot$, based on numerical relativity simulations and an EOB based inspiral-merger-ringdown model (TEOBResumS). We recover these injections using both quasi-circular and eccentric waveform models. For cases with $e_{20} \sim 0.1$, quasi-circular models fail to estimate chirp mass within the 90% credible interval accurately. Further, for these low-mass injections, spin-induced precession does not mimic eccentricity. For injections of $e_{20}\sim 0.1$, PE conducted with an inspiral-only eccentric waveform model correctly characterises the injected signal to within 90% confidence, and recovers the injected eccentricities, suggesting that such models are sufficient for characterisation of low-mass eccentric BBH. (abridged)