Postmerger Spectra and Dense-Matter Signatures

If a binary-neutron-star merger avoids immediate collapse, its remnant can emit a short, high-frequency gravitational-wave signal whose peaks reflect bulk rotation, oscillations, and nonlinear couplings. Peak frequency, damping, and collapse time can carry information beyond inspiral tidal deformability, but their interpretation is calibrated on finite simulation sets and is strongly affected by thermal physics and detector noise. The data and evidence boundary for this page is 2026-08-18.

Helpful background. Stress–Energy and Conservation supplies remnant energy budgets; the Einstein equations supply strong-field dynamics; Relativistic Two-Body Dynamics supplies inspiral parameters; Relativistic Matter and Compact Objects supplies dense-matter models; Gravitational Radiation and Waveforms supplies strain and extraction conventions; and Computational Relativity supplies waveform validation.

Spectral features and physical interpretation

Choose a merger time $t_m$, extract a specified strain or Weyl-scalar mode, and window the postmerger interval before Fourier transformation. Long-lived material remnants often show a dominant quadrupolar peak, commonly called $f_2$, with secondary peaks associated in particular model classes with spiral patterns or nonlinear combinations of radial and quadrupolar motion. A damped template illustrates the measured quantities,

$$h(t)=A e^{-(t-t_m)/\tau} \cos\!\left[2\pi f_2(t-t_m)+\phi\right], \qquad t\geq t_m .$$

Real postmerger signals are nonstationary: contraction and angular-momentum loss can chirp the frequency, multiple components overlap, and collapse truncates the waveform. Baiotti and Rezzolla review spectral structures, remnant classes, and EOS correlations in Baiotti and Rezzolla 2017, §V.B.2, Figs. 7–11.

Relations between $f_2$ and a stellar radius, compactness, or maximum density are empirical and conditional on total mass, mass ratio, spin, EOS sample, thermal prescription, and remnant class. A strong first-order phase transition could shift a peak or collapse time relative to a hadronic calibration, but thermal-EOS, neutrino, magnetic, resolution, and waveform-model changes can produce competing shifts. No single peak is a model-independent phase-transition diagnosis.

Versioned strain and simulation data

The event-level example is the Gravitational Wave Open Science Center (GWOSC) GW170817 strain data release and technical-data version, which provides calibrated H1, L1, and V1 products and parameter-estimation links. The event page specifically warns that Virgo data are noisier above $2\,{\rm kHz}$, illustrating why detector and frequency-dependent response cannot be replaced by a broadband sensitivity slogan. The targeted analysis by Abbott and collaborators reported no significant postmerger candidate and set model-dependent constraints in Abbott et al. 2017, §§3–4, Tables 1–3, and Fig. 1. That nondetection is tied to the event, calibration, windows, pipelines, and waveform families used; it is not evidence that the remnant emitted no signal.

Simulation-calibrated inference needs a named training release. The CoRe second release provides multipolar waveforms, simulation metadata, and multiple resolutions through its CoRe simulation repositories. The accompanying article documents methods, coverage, and waveform accuracy in Gonzalez et al. 2023, §§2–4 and Appendices A–B. Its separate Gonzalez et al. 2022, Zenodo record 7253784, version v1 contains the data and notebooks for merger and postmerger fits, as identified in the article’s Appendix C; that small archive is not the full waveform database. A reproducible analysis must pin the simulation identifiers, repository revisions, extraction radii, and resolution files actually used. A training set is not a population prior: its EOS, masses, spins, thermal models, resolutions, and survival times must be compared with the intended inference domain.

Search likelihood and selections

For calibrated detector data $\tilde d_k(f)$, a Gaussian-noise likelihood uses

$$\ln{\cal L}(\vartheta,\lambda) =-\frac12\sum_k \left\langle d_k-h_k(\vartheta,\lambda) \mathrel{\big|}d_k-h_k(\vartheta,\lambda)\right\rangle_k+\text{constant},$$

where the inner product uses a stated high-frequency noise spectrum. Parameters $\vartheta$ may include amplitude, $f_2$, damping, chirp, start time, sky position, and polarization; nuisance parameters $\lambda$ include calibration and noise uncertainty. An unmodeled excess-power search replaces the waveform family with a time–frequency statistic but still needs a background distribution and detection threshold.

Selections include the inspiral-event definition, detector/network inclusion, data-quality vetoes, the postmerger start and stop times, frequency band, spectral window, search resolution, and trials over time–frequency volume. Nondetections require efficiency from injections or a likelihood-based upper limit. A catalog population analysis must include the probability that an inspiral is detected and that usable high-frequency data exist; selecting only nearby loud events changes the mass and orientation population.

Priors on remnant survival, $f_2$, damping, amplitude, EOS family, phase transition, masses, spin, distance, and orientation can dominate a weak signal. A simulation relation must include training uncertainty and ideally a leave-one-EOS-family-out test. Report likelihood normalization, interval convention, sampler and convergence diagnostics, and posterior-predictive checks rather than only a maximum-score spectrum.

Systematics and reproduction boundary

Instrument systematics include high-frequency strain calibration, timing, spectral lines, nonstationary noise, data-quality segments, and noise-spectrum estimation. Analysis choices add window leakage, merger-time uncertainty, short-duration frequency resolution, look-elsewhere effects, and waveform interpolation. Simulation systematics include finite grid and extraction radius, residual eccentricity, atmosphere, thermal extension, neutrino and magnetic treatment, mass-ratio and spin coverage, and numerical collapse time.

Dense-matter claims should be stress-tested by changing the spectral window and noise model, marginalizing calibration, repeating with alternate waveform families, and withholding whole EOS families from calibration. A peak relation that only interpolates individual simulations but fails this out-of-family test does not establish inference beyond its training set.

No strain or simulation was reanalyzed for this page. Reproduction would pin the GWOSC event version or other dataset DOI, file checksums, calibration release, quality segments, detector list, PSD settings, time–frequency window, software repository and environment, random seeds, waveform or CoRe release, training selection, likelihood, priors, injection set, and posterior samples. It would regenerate the search statistic or upper limit under alternate noise and windows and perform leave-one-EOS-family-out validation. A new data release or high-frequency calibration after 2026-08-18 triggers review.

Check finite-window resolution

A postmerger spectrum uses an unpadded rectangular interval of duration $T=20\,{\rm ms}$. What is its natural Fourier-bin spacing, and can two narrow peaks separated by $30\,{\rm Hz}$ be resolved by bin location alone?

Solution

The natural spacing is

$$\Delta f=\frac{1}{T} =\frac{1}{0.020\,{\rm s}} =50\,{\rm Hz}.$$

Two peaks separated by $30\,{\rm Hz}$ are closer than the natural bin spacing and are not resolved by bin location alone. Zero-padding can interpolate the plotted spectrum but cannot create information. A justified parametric model may infer sub-bin frequencies at sufficient signal-to-noise, but then the likelihood, window, damping model, calibration, and prior determine the uncertainty.

References

• Abbott, B. P., et al. “Search for Post-merger Gravitational Waves from the Remnant of the Binary Neutron Star Merger GW170817.” Astrophysical Journal Letters 851 (2017): L16. DOI.
• Baiotti, Luca, and Luciano Rezzolla. “Binary Neutron Star Mergers: A Review of Einstein’s Richest Laboratory.” Reports on Progress in Physics 80 (2017): 096901. arXiv.
• Gonzalez, Alejandra, et al. “Second Release of the CoRe Database of Binary Neutron Star Merger Waveforms.” Classical and Quantum Gravity 40 (2023): 085011. DOI.
• Gonzalez, Alejandra, et al. “Second Release of the CoRe Database of Binary Neutron Star Merger Waveforms.” Companion fit data and notebooks, version v1 (2022). Zenodo archive.
• Gravitational Wave Open Science Center. “GW170817 Event Data Release.” LIGO Laboratory, 2017, accessed 2026-08-18. Data release.