From Source Properties to Strong-Field Tests: a multipronged analysis of GW250114 with an effective one-body model for generic orbits

Abstract

We present a detailed analysis of GW250114, the loudest gravitational-wave signal observed to date, using a waveform model capable of describing binary black holes in generic (eccentric and precessing) orbits. Our analysis builds on LIGO-Virgo-KAGRA (LVK)'s results, finding that the source is consistent at a probability of $\geq 96\%$ with the merger of two first-generation, nearly equal-mass, low-spin black holes, forming a remnant within the pair-instability mass gap. The signal's high signal-to-noise ratio ($\gtrsim 75$) enables the detection of the subdominant $(4,\pm4)$ multipoles, whose presence we confirm with higher evidence than previously reported by the LVK. Restricting the analysis even to post-peak data yields $\log_{10}B\gtrsim 1$ in favor of models including the $(4,\pm4)$ mode, demonstrating that this contribution remains detectable well into the post-merger phase. We further perform three independent tests of general relativity, complementary to those performed by the LVK: a modified residual analysis confirms that our semi-analytical model fully describes the signal without detectable discrepancies; a subdominant mode test finds that the amplitude of the $(4,\pm4)$ multipoles agrees with general-relativistic expectations; and a parameterised analysis of the plunge-merger-ringdown regime recovers the GR expectation within the 50\% credible region for the remnant mass and spin, and within the 90\% interval for the $(2,\pm2)$ peak amplitude. Collectively, these results reinforce GW250114 as a landmark event for a precision test of gravity.

Figures (9)

• Figure 1: Posterior distributions for orbital eccentricity $e$ and the waveform parameter $\zeta$ obtained with eccentric precessing (dash-dotted) and eccentric non-precessing (solid) models. Both analyses give consistent results, with median values $e \simeq 0.01$ and $\zeta \simeq 3.5$ for both cases.
• Figure 2: Posterior distributions of mass and spin parameters for GW250114. The 2D panels show the posteriors colored by the local posterior-to-prior density ratio $p(\boldsymbol{\xi}|d)/\pi(\boldsymbol{\xi})$, with warmer colors indicating greater support from the data. Black contours enclose 68% and 90% credible regions. The data favor an approximately equal-mass binary ($q \sim 1$) with low spin magnitudes ($\chi_{1,2} \sim 0.1$) and misaligned spins, as evidenced by the broad distributions of tilt angle cosines $\cos\theta_{1,2}$ measured relative to the orbital angular momentum.
• Figure 3: Posterior of azimuthal angle (measured in degrees at a reference frequency of 20 Hz) for various bbh models. The data favors a source with a bimodal azimuth separated by $\pi$, reflecting the binary's symmetry under rotation $\phi \rightarrow \phi + \pi$.
• Figure 4: Time-gated analysis of GW250114 to test the contribution of the $(4,\pm4)$ mode as a function of the start time $\Delta t^\mathrm{geo}_\mathrm{start}/M$, measured relative to the signal peak in units of remnant black-hole mass. The curves show the change in the log Bayes factor $\Delta \log_{10} \mathcal{B}$ (colored) and the improvement in matched-filter snr relative to a quadrupole-only model (black). The $(4,\pm4)$ contribution remains detectable even post signal's peak, thereby showing that GW250114 is the first gw event with a measurable $(4,\pm4)$ contribution persisting into the post-merger phase.
• Figure 5: Residual distributions for Hanford (left) and Livingston (right) detectors after subtracting the whitened TEOBResumS-Dalí posterior waveforms from whitened strain data. Individual curves show posterior samples, colored by their Jensen–Shannon divergence from the expected white noise distribution (black dashed). We find that the residuals are statistically indistinguishable from noise ($\text{JSD} = 0.088 \pm 0.003$, $p > 0.05$ for all samples), thereby indicating that the data is well described by gr-consistent bbh waveforms.