GW200105: A detailed study of eccentricity in the neutron star-black hole binary

TL;DR

This paper reanalyzes GW200105, the first confidently identified NSBH merger, using state-of-the-art effective-one-body waveform models that incorporate orbital eccentricity and spin precession across the full IMR and include higher-order modes. Bayesian inference via the RIFT framework shows strong evidence for nonzero eccentricity $e_{20}$ (≈0.12–0.14) with a multimodal eccentricity posterior, and reveals mass-ratio and spin parameters that shift relative to precession-only analyses. Comparisons across eccentric-only and eccentric-plus-precession models indicate only a modest gain from including precession when eccentricity is present, highlighting waveform-model systematics as a likely source of discrepancies with some prior studies. Numerical-relativity simulations corroborate that eccentricity affects merger timing but does not substantially alter final-state properties, with implications for NSBH formation channels, particularly dynamical formation scenarios.

Abstract

GW200105_162426 is the first neutron star-black hole merger to be confidently confirmed through either gravitational-wave or electromagnetic observations. Although initially analyzed after detection, the event has recently gained renewed attention following a study [Morras et al. arXiv:2503.15393] that employed a post-Newtonian inspiral-only waveform model and reported strong evidence for orbital eccentricity. In this work, we perform a detailed analysis of GW200105 using state-of-the-art effective-one-body waveform models. Importantly, we present the first study of this event utilizing a physically complete model that incorporates both orbital eccentricity and spin precession across the full inspiral, merger, and ringdown stages, along with higher-order gravitational wave modes. Our results support the presence of eccentricity in the signal, with zero eccentricity excluded from the 99% credible interval, but yielding a mass ratio closer to the original LIGO-Virgo-KAGRA analysis, differing from the findings of [Morras et al. arXiv:2503.15393]. Additionally, similar to a previous eccentric-only analysis [de Lluc Planas et al. Astrophys. J. 995, 47 (2025).], we observe a multimodal structure in the eccentricity posterior distribution. We conduct targeted investigations to understand the origin of this multimodality and complement our analysis with numerical relativity simulations to examine how the inclusion of eccentricity impacts the merger dynamics.

Figures (5)

• Figure 1: Comparative corner plot for GW200105. This figure shows one- and two-dimensional marginal posterior distributions for $m_{1,2}$, $\chi_\text{eff}$, $\chi_{p}$, $D_L$ and $e_{20}$. Diagonal panels show the one-dimensional marginal posterior distributions, while contours in the off diagonal panels show the 90% credible intervals for the joint two-dimensional marginal posterior distributions. An additional panel illustrates the marginal posterior distributions for $q$. Different colors and linestyles represent results from various waveform models, grouped according to three distinct hypotheses for clearer comparison. Two publicly available analysis results are also included for comparison.
• Figure 2: Comparison of ${\cal M}_c$ and $e_{20}$ distributions from eccentric-only hypothesis. This figure shows the marginal one- and two-dimensional posterior distributions for ${\cal M}_c$ and $e_{20}$ obtained from TEOBResumS-E and SEOBNRv5EHM analyses, with Planas:TEHM included for comparison. The two-dimensional contours represent the $90\%$ credible interval. The scatter plot uses a color gradient to highlight points that lie within $15$ of the maximum $\ln{\cal L}_{\rm marg}{}$, based on the TEOBResumS-E results. For clarity, a reduced eccentricity range is shown.
• Figure 3: Impact of eccentricity prior. This figure compares the $e_{20}$ posterior distributions obtained under the eccentric-precessing hypothesis with TEOBResumS, using both the uniform and log-uniform priors. The log-uniform prior is evaluated with two different lower bounds: $10^{-2}$ and $10^{-4}$. Even in the more extreme case, where the lower bound is $10^{-4}$, the posterior retains substantial support at nonzero $e_{20}$ values.
• Figure 4: Systematics study. This figure shows the $e_{20}$ posterior distribution obtained using parameter estimation settings matched to those employed in morras2025orbitaleccentricityneutronstar. For comparison, results from the main analyses are also shown. Despite similar settings, the multimodal structure persists.
• Figure 5: Gravitational wave comparison. This figure shows the $(2,2)$ mode of the Weyl scalar $\Psi_4^{(2,2)}$ extracted at a radius of $r = 130M$ for both the circular and eccentric cases of NRq0.13, plotted against retarded time. The dashed vertical lines at $t_{\text{qc}} = 3231M$ and $t_{\text{ecc}} = 2796M$ mark the merger times, defined by the peaks in the waveform amplitude.