Fast gravitational waveform models for quasi-circular coalescences of neutron star--black hole binaries

This paper introduces the first frequency-domain gravitational waveform models for quasi-circular neutron star–black hole binaries that incorporate higher-order modes and tidal effects, demonstrating improved accuracy over predecessors through numerical relativity comparisons and consistent parameter estimation on real observational data.

The Gist

Imagine the universe is a giant, silent concert hall. For years, we've been listening to the "heavy metal" of this concert: the deep, booming crashes of two black holes smashing together. But recently, we started hearing a different kind of music: the collision of a black hole (the heavy, invisible giant) and a neutron star (a super-dense, tiny city made of matter).

This paper is about building better "microphones" and "sheet music" to listen to these specific collisions more clearly.

Here is the breakdown of what the scientists did, using simple analogies:

1. The Problem: The Old Microphones Were Too Simple

For a long time, the models scientists used to predict these collisions were like listening to a song with only the bass drum playing. They could hear the main beat (the dominant "quadrupole" mode), but they missed the hi-hats, the guitar riffs, and the complex harmonies (called "higher-order modes").

Furthermore, when a black hole eats a neutron star, the star can get ripped apart by gravity before it disappears. This is like a cookie crumbling in milk. The old models mostly treated the neutron star like a solid rock that just got swallowed whole. They didn't account for the "crumbs" (tidal effects) or the fact that the black hole might be spinning in a way that makes the whole system wobble (precession).

Because of these missing details, when scientists tried to figure out exactly how heavy the stars were or how fast they were spinning, they sometimes got the wrong answer.

2. The Solution: New, High-Fidelity Models

The authors of this paper built three new, super-accurate models (which they named IMRPhenomXHM NSBH, SEOBNRv5HM ROM NRTidalv3 NSBH, and IMRPhenomXPHM NSBH).

Think of these models as upgrading from a basic AM radio to a high-definition surround-sound system.

• They hear the whole orchestra: Instead of just the bass drum, these models capture the "higher-order modes"—the complex harmonics that happen when the masses are very different or when the stars are spinning.
• They taste the crumbs: They include "tidal effects." If the neutron star gets ripped apart, the model knows how that changes the sound of the crash.
• They handle the wobble: One of the models can even handle cases where the black hole is spinning sideways, causing the whole system to wobble like a top (precession).

3. How They Built It: The "Hybrid" Recipe

To make these models accurate, the scientists didn't just guess. They used a "hybrid" recipe:

• The Early Part (The Warm-up): They used math based on Einstein's theories to describe the slow approach of the stars.
• The Crash (The Climax): For the actual moment of impact, they used data from super-computer simulations (called Numerical Relativity). These simulations are like running a video game physics engine to see exactly what happens when a black hole eats a neutron star.
• The Calibration: They tuned their new models to match these super-computer simulations perfectly, ensuring that the "sound" of their models matched the "reality" of the simulations.

4. The Test Drive: Do They Work?

The scientists tested their new models in two ways:

1. Against Simulations: They compared their models to the super-computer data. The new models matched the simulations much better than the old ones, especially when the stars were very different sizes or when the neutron star got ripped apart.
2. Against Real Events: They used the new models to re-analyze real signals that the LIGO and Virgo detectors have already caught (like GW200105 and GW230529).

The Results:

• Consistency: When they looked at real events, the new models gave results that were very similar to what we already knew, which is good news—it means the old data wasn't "wrong," just less precise.
• Improvement: In some cases, the new models gave slightly different (and likely more accurate) answers about the mass and spin of the stars. For example, they were better at figuring out the exact mass ratio when the stars were of similar size.
• Speed: Even though these models are more complex, they are still fast enough to be used in real-time. They are like a Ferrari that is also a family minivan; they have high performance but are still practical for daily use.

5. Why It Matters

The paper concludes that as our detectors get more sensitive (like upgrading from a standard microphone to a studio-grade one), we will hear these cosmic collisions more clearly. To make sense of that clearer sound, we need these new, detailed models.

Without them, we might miss the subtle clues about how these stars formed, how they die, and what happens to the matter when a black hole devours a neutron star. The paper doesn't claim these models will cure diseases or predict the weather; their only job is to help us understand the physics of these violent cosmic crashes more accurately.

In short: The authors built better "earbuds" for listening to black hole and neutron star collisions, allowing us to hear the full symphony of the crash rather than just the bass drum.

Technical Summary

Technical Summary: Fast Gravitational Waveform Models for Quasi-Circular NSBH Coalescences

Problem Statement
The detection of neutron star–black hole (NSBH) binaries by the LIGO–Virgo–KAGRA (LVK) collaboration has highlighted a critical gap in gravitational-wave (GW) modeling. While existing inspiral-merger-ringdown (IMR) models for NSBH coalescences, such as IMRPhenomNSBH and SEOBNRv4 ROM NRTidalv2 NSBH, capture the dominant quadrupolar mode and aligned spins, they neglect higher-order harmonic modes (HMs) and spin precession. Furthermore, these models often lack accurate descriptions of tidal disruption effects in the amplitude and phase, which are crucial for distinguishing NSBH mergers from binary black hole (BBH) events and for inferring the equation of state of neutron stars. Time-domain (TD) models that include these features (e.g., TEOBResumS-GIOTTO, TEOBResumS-Dalí) exist but suffer from high computational costs, making them less suitable for low-latency applications and efficient parameter estimation (PE) compared to frequency-domain (FD) models.

Methodology
The authors present three new FD waveform models designed to address these limitations:

1. IMRPhenomXHM NSBH: A phenomenological model for quasi-circular, aligned-spin NSBH binaries.
2. SEOBNRv5HM ROM NRTidalv3 NSBH: An effective-one-body (EOB) model for the same configuration, utilizing a reduced-order model (ROM) for efficiency.
3. IMRPhenomXPHM NSBH: An extension of IMRPhenomXHM NSBH to include spin precession, constructed by "twisting up" the aligned-spin waveform in a co-precessing frame.

Construction and Calibration:

• Amplitude: The models incorporate tidal effects and disruption physics into the GW strain amplitudes.
  • IMRPhenomXHM NSBH uses a piecewise approach: PN tidal corrections for the inspiral, a suppressing function calibrated to Numerical Relativity (NR) simulations for the merger-ringdown, and a smooth transition. The calibration employs a "collocation points" approach, fitting the amplitude ratio of NR simulations against the underlying BBH model.
  • SEOBNRv5HM ROM NRTidalv3 NSBH extends the SEOBNRv4 ROM NRTidalv2 NSBH amplitude corrections, applying multiplicative factors to the underlying BBH amplitude (SEOBNRv5HM ROM) based on the effective tidal deformability.
  • Both models utilize a fitting strategy that augments PN information with direct tuning to NR simulations, including subdominant modes up to $\ell=4$.
• Phase: The phase models augment the underlying BBH phases (IMRPhenomXHM and SEOBNRv5HM ROM) with tidal contributions from the NRTidalv3 model. This includes a closed-form expression for the dominant mode and a scaling relation for HMs. To prevent unphysical divergences in the post-merger regime, a Taylor extrapolation algorithm is implemented.
• Precession: For IMRPhenomXPHM NSBH, precession is handled via the standard twisting-up procedure, mapping the aligned-spin signal to the inertial frame using time-dependent Euler angles derived from PN spin-precession equations.
• Data: The amplitude models are calibrated using a dataset of NR simulations from SpEC, SACRA, and BAM codes. This includes 162 SACRA simulations (dominant mode) and a subset of 25 BAM and SXS simulations including HMs. Short NR waveforms were hybridized with TEOBResumS-Dalí for calibration and NRHybSur3dq8Tidal for validation.

Key Results

• Accuracy: Time-domain comparisons and mismatch calculations against NR simulations (hybridized with NRHybSur3dq8Tidal) demonstrate that the new models significantly outperform their predecessors (IMRPhenomNSBH, SEOBNRv4 NSBH) and other existing models (DALI, GIOTTO). The inclusion of HMs reduces mismatches and decreases dependency on extrinsic parameters (inclination, polarization).
  • For aligned-spin systems, XHM NSBH and SEOBHMv5 NSBH show consistent agreement with NR, with mismatches generally below $10^{-2}$ for a wide range of parameters.
  • The precessing model (XPHM NSBH) reproduces GW emission from precessing systems with good accuracy (median mismatch $\sim 0.009$) in the single-spin case.
• Computational Efficiency: Despite including HMs and tidal effects, the new FD models remain computationally efficient.
  • IMRPhenomXHM NSBH is comparable in speed to the dominant-mode-only IMRPhenomNSBH.
  • SEOBNRv5HM ROM NRTidalv3 NSBH is roughly 3.7 times slower than its BBH baseline (SEOBNRv5HM) but remains efficient enough for PE.
• Parameter Estimation (PE):
  • Injection Studies: For simulated signals at SNR $\sim 26$, the new models recover injected parameters (mass ratio, tidal deformability) more accurately than BBH models or dominant-mode NSBH models. The inclusion of HMs is strongly preferred over PhenomNSBH for asymmetric mass systems.
  • Real Events: The models were applied to GW200105, GW200115, GW200115, GW230518, and GW230529. Results are consistent with previous LVK analyses and literature. While tidal constraints remain weak due to moderate SNRs, the inclusion of HMs and tidal effects leads to noticeable shifts in posteriors for mass ratio and luminosity distance in some events (e.g., GW230518). Bayes factors generally favor the new models over PhenomNSBH for events with asymmetric masses.

Significance and Claims
The paper claims to present the first frequency-domain models for quasi-circular, aligned-spin NSBH binaries that simultaneously include higher-order modes, spin precession, and calibrated tidal effects. The authors emphasize that these models offer an unprecedented combination of accuracy and computational efficiency. By providing fully self-consistent analyses that account for all relevant physical effects (HMs, precession, tides), the models aim to eliminate systematic biases in parameter estimation that arise from neglecting these effects. The work suggests that as detector sensitivities improve and higher-SNR signals are observed, the inclusion of these sub-dominant physical effects will become increasingly critical for reliable astrophysical inference, particularly for understanding the population of NSBH systems and the nature of tidal disruption. The authors modestly note that conclusions regarding specific populations (e.g., comparable-mass NSBHs) remain tentative given the current number of observations, but the new tools provide the necessary framework for future, more precise studies.