Biased parameter inference of eccentric, spin-precessing binary black holes

This study demonstrates that analyzing gravitational wave signals from eccentric, spin-precessing binary black holes using quasi-circular waveform models leads to significant biases in inferred source parameters, thereby highlighting the critical need for comprehensive waveform models that simultaneously account for both eccentricity and spin precession.

The Gist

Imagine the universe is a giant, silent concert hall, and black holes are the musicians. When two black holes dance toward each other and crash, they create ripples in space-time called gravitational waves. Scientists use giant detectors (like LIGO and Virgo) to "listen" to these ripples and figure out who the musicians are: how heavy they are, how fast they spin, and how they are moving.

Usually, scientists assume these black holes dance in a perfect circle, like a figure skater spinning on a smooth spot of ice. This is the "standard model" they use to decode the music. However, this paper argues that sometimes, the black holes aren't dancing in a circle at all; they might be moving in a wobbly, oval path (eccentricity), or their spin axes might be wobbling like a top about to fall over (spin-precession).

Here is what the researchers found, explained simply:

1. The "Wrong Map" Problem

The scientists ran a massive experiment. They created fake gravitational wave signals in a computer. Some of these signals had black holes moving in perfect circles, but others had black holes moving in wobbly, oval paths (eccentric) or spinning in wobbly ways (precessing).

Then, they tried to "decode" these signals using the standard tools that assume everything is a perfect circle.

• The Analogy: Imagine you are trying to identify a car by listening to its engine. You have a manual that only describes a car driving in a straight line. If the car is actually driving in a tight, wobbly circle, your manual will get confused. It might tell you the car is a different model, or that the engine is spinning wildly, just because it's trying to force a circle-shaped explanation onto a wobbly reality.

2. The Big Mistakes (Biases)

When the scientists used the "perfect circle" tools to analyze the "wobbly" signals, the results were wrong in specific ways:

• Fake Spinning: If the black holes were just moving in an oval path but not wobbling on their axes, the standard tools often lied and said, "Hey, these black holes must be wobbling!" They mistook the oval shape of the orbit for a wobble in the spin.
• Wrong Weights: The tools also got the weight (mass) of the black holes wrong. The heavier the oval shape (eccentricity), the more wrong the weight calculation became.

3. The "Smoking Gun"

The researchers tested different "decoder" tools. They found that when a signal had a strong oval shape, the tool that assumed a "wobbly spin" (the standard tool) was a terrible fit.

• The Analogy: It's like trying to fit a square peg into a round hole. The math (called the "Bayes factor") showed a huge preference for a tool that actually accounts for the oval shape. The data was screaming, "I am an oval!" but the standard tool was insisting, "No, you are a circle, just a really weird one."

4. The Double Trouble

The most complex part of the study looked at black holes that were both moving in an oval path and wobbling on their axes.

• When they used the standard "circle" tool, it got the spin completely wrong, inventing a wobble that wasn't there or exaggerating one that was.
• However, when they used a tool designed for oval paths (even if it didn't account for the wobble), it could still correctly identify the oval shape.
• The Lesson: If you ignore the oval shape, you will get the spin wrong. If you ignore the spin, you might still get the shape right. Ignoring the shape is the bigger problem.

The Bottom Line

The paper concludes that as our detectors get more sensitive (hearing quieter and more complex sounds), we can no longer pretend all black hole dances are perfect circles. If we keep using the "perfect circle" map for "wobbly" black holes, we will keep making mistakes about what these cosmic objects actually are.

To get the right answer, we need new, more flexible tools that can handle both the oval paths and the wobbling spins simultaneously. Until we have those ready-to-use tools, our measurements of these cosmic crashes will remain biased and inaccurate.

Technical Summary

Technical Summary: Biased Parameter Inference of Eccentric, Spin-Precessing Binary Black Holes

Problem Statement
While the majority of gravitational wave (GW) events detected by LIGO and Virgo are consistent with binary black hole (BBH) mergers on quasi-circular orbits, there is growing evidence that some systems may possess non-zero orbital eccentricity or spin-precession, or both. These features often arise from dynamical formation channels (e.g., in globular clusters or active galactic nuclei). A critical challenge exists in parameter estimation (PE): current catalogs predominantly rely on waveform models that assume quasi-circularity. When signals containing eccentricity are analyzed with quasi-circular models, or when signals containing spin-precession are analyzed with aligned-spin models, degeneracies between these physical effects can lead to biased inferences of source parameters. Specifically, there is a risk that eccentricity might be misinterpreted as spin-precession (and vice versa), leading to incorrect astrophysical conclusions regarding the formation history of the binaries.

Methodology
The authors conducted a comprehensive study using Bayesian inference to quantify biases in source parameters when signals are recovered with waveform models that neglect either eccentricity or spin-precession. The study utilized three distinct injection strategies in zero-noise environments at a luminosity distance of 410 Mpc:

1. Non-spinning Eccentric Injections: Hybrid waveforms were constructed by combining eccentric, non-spinning Numerical Relativity (NR) simulations from the SXS catalog with post-Newtonian inspiral waveforms.
2. Aligned-Spin Eccentric Injections: Signals were generated using the TEOBResumS-DALI waveform model and NR simulations, varying eccentricity from 0 to 0.35 while keeping spin magnitudes fixed (aligned or anti-aligned).
3. Spin-Precessing Eccentric Injections: New NR simulations were performed using the SpEC code to generate signals with both eccentricity and spin-precession.

These signals were recovered using two primary classes of waveform models:

• Quasi-circular, spin-precessing models: Specifically IMRPhenomXP.
• Eccentric, aligned-spin models: Specifically TaylorF2Ecc (inspiral-only) and IMRESIGMA (inspiral-merger-ringdown).

The analysis focused on the posterior distributions of the chirp mass ($M_c$), the effective spin-precession parameter ($\chi_p$), and the eccentricity ($e$). Bayes factors were calculated to compare the evidence for eccentric, aligned-spin models versus quasi-circular, precessing-spin models.

Key Results

• Biases in Non-Spinning Systems: When non-spinning, eccentric signals were recovered using a quasi-circular, precessing-spin model (IMRPhenomXP), significant biases were observed. The inferred chirp mass ($M_c$) showed deviations that generally increased with the injected eccentricity. More critically, the posterior for the spin-precession parameter ($\chi_p$), which should be zero, peaked at non-zero values for high eccentricities. This indicates that the quasi-circular precessing model falsely interprets orbital eccentricity as spin-precession.
• Biases in Aligned-Spin Systems: For aligned-spin eccentric signals, the same trend was observed. Recovery with a quasi-circular precessing model resulted in an overestimation of $\chi_p$ as eccentricity increased. However, Bayes factor calculations ($B_{E/C}$) demonstrated that for sufficiently high eccentricity, an eccentric, aligned-spin model is strongly preferred over the quasi-circular precessing model.
• Biases in Spin-Precessing Systems: When signals possessing both eccentricity and spin-precession were analyzed:
  • Recovery with a quasi-circular precessing model led to a consistent overestimation of $\chi_p$, with the bias becoming significant as eccentricity increased.
  • Conversely, recovery with an eccentric, aligned-spin model (IMRESIGMA) successfully recovered the injected eccentricity values within 90% credible intervals, even when the injected $\chi_p$ was high.
  • Bayes factors consistently favored the eccentric, aligned-spin model over the quasi-circular precessing model as eccentricity increased.

Significance and Claims
The paper concludes that there is a complex interplay between eccentricity and spin-precession effects in GW signals. Neglecting one effect while modeling the other leads to significant biases in parameter estimation. Specifically:

1. Misinterpretation Risk: High eccentricity can mimic spin-precession in quasi-circular models, potentially leading to false claims of precession in eccentric systems.
2. Model Preference: For signals with detectable eccentricity, eccentric waveform models are statistically preferred over quasi-circular precessing models, even when the latter includes precession physics.
3. Need for Comprehensive Models: The authors emphasize the urgent need for readily usable, complete (inspiral-merger-ringdown) waveform models that simultaneously incorporate both eccentricity and spin-precession. Such models are essential for unbiased parameter estimation of GW signals that exhibit signs of either or both effects, particularly as detector sensitivity improves and signal-to-noise ratios increase.

The study does not propose new detection pipelines or future experimental setups but highlights the necessity of upgrading current inference frameworks to include these combined physical effects to avoid systematic errors in the interpretation of the BBH population.