Population Properties of Binary Black Holes with Eccentricity

TL;DR

The paper tackles whether the BBH population contains a measurable eccentric fraction by performing the first population inference that jointly fits mass, spin, redshift, and eccentricity using SEOBNRv5EHM waveforms within a hierarchical Bayesian framework. It extends the GWTC-4 population model with a two-component eccentricity mixture and assesses four priors through Monte Carlo integration, finding no strong eccentricity evidence and placing a 90% CL bound on the eccentric fraction at $\zeta_\epsilon \leq 0.0239$, with results highly model-dependent for more extreme assumptions. Across models, the analysis demonstrates robustness of the quasi-circular population conclusions while highlighting how prior choices influence inferences about rare high eccentricity tails. The work establishes a framework for incorporating eccentricity into population studies and provides benchmarks for future, larger BBH catalogs and more flexible eccentricity models.

Abstract

The improved sensitivity of Gravitational-Wave detectors and the development of eccentric waveform models enable us to explore the growing catalog of gravitational-wave events with measurable eccentricity. This opens new opportunities to gain insight into the formation channels and evolutionary pathways of compact binary systems using eccentricity. However, most recent population analyses have been limited to quasi-circular binaries, primarily due to constraints in waveform modeling and sensitivity estimates. We are now entering an era where both of these limitations are being addressed, allowing for a more comprehensive investigation of eccentric binary populations. In this work, we perform the first population inference analysis that simultaneously fits the mass, spin, redshift, and eccentricity distribution. Specifically, we use source-parameter estimation provided by the Rapid Iterative FiTting (RIFT) framework using the SEOBNRv5EHM waveform model, and a default O4a population model extended to include eccentricity. We find population properties broadly consistent with conclusions obtained in previous analyses assuming quasi-circular binaries. Consistent with our conclusions about each event, we bound the branching ratio for eccentric events to be below $0.0239$ at $90\%$ confidence with our fiducial eccentricity mixture models. Using four different parametric population models for eccentricity, we point out that the rate of eccentric events is weakly constrained by observations and highly model-dependent.

Figures (11)

• Figure 1: PE Comparison: The blue color shows the parameter estimation with quasi-circular assumption, published under GWTC-4 catalog on zenodo vs orange color shows parameter estimation performed with RIFT using eccentric waveform SEOBNRv5EHM. We have showed only top twenty events from the \ref{['tab:event_summary']} where they are sorted based on the Bayes factor.
• Figure 2: PE Corner Plot: This plot shows the posterior distributions of the chirp mass $\mathcal{M}_c$, effective spin $\chi_{\mathrm{eff}}$, eccentricity $\epsilon$, and redshift $z$ for a representative eccentric BBH merger. The contours represent the $50\%$ credible intervals for each event, providing insights into the correlations between these parameters.
• Figure 3: Eccentricity Support: This CDF shows the low eccentricity support in 139 BBHs whose parameter estimation performed with RIFT using SEOBNRv5EHM.
• Figure 4: Mixture Eccentricity Model: This shows the comparison studies for four models: Nonoverlapping Mixture, Overlapping Mixture, High Eccentricity Truncated, and Low Eccentricity Truncated.
• Figure 5: Mixture Eccentricity Model: This shows the comparison in log scale for four models: Nonoverlapping Mixture, Overlapping Mixture, High Eccentricity Truncated, and Low Eccentricity Truncated.