Frequency-domain gravitational waveform models for inspiraling binary neutron stars

TL;DR

This paper presents a frequency-domain gravitational waveform model for inspiraling binary neutron stars, calibrated against high-precision numerical and analytical waveforms, achieving high accuracy in phase, mismatch, and parameter estimation.

Contribution

The authors develop and validate a new frequency-domain waveform model for binary neutron stars, with improved accuracy and systematic error control over previous models.

Findings

Phase difference less than 0.1 rad for key parameters.

Mismatch below 1.1×10⁻⁵ for SNR ≤ 50.

Systematic error in tidal deformability measurement under 20.

Abstract

We develop a model for frequency-domain gravitational waveforms from inspiraling binary neutron stars. Our waveform model is calibrated by comparison with hybrid waveforms constructed from our latest high-precision numerical-relativity waveforms and the SEOBNRv2T waveforms in the frequency range of $10$--$1000\,{\rm Hz}$. We show that the phase difference between our waveform model and the hybrid waveforms is always smaller than $0.1\, {\rm rad}$ for the binary tidal deformability, ${\tilde \Lambda}$, in the range $300\lesssim{\tilde \Lambda}\lesssim1900$ and for the mass ratio between 0.73 and 1. We show that, for $10$--$1000\,{\rm Hz}$, the distinguishability for the signal-to-noise ratio $\lesssim50$ and the mismatch between our waveform model and the hybrid waveforms are always smaller than 0.25 and $1.1\times10^{-5}$, respectively. The systematic error of our waveform model in the measurement of ${\tilde \Lambda}$ is always smaller than $20$ with respect to the hybrid waveforms for $300\lesssim{\tilde \Lambda}\lesssim1900$. The statistical error in the measurement of binary parameters is computed employing our waveform model, and we obtain results consistent with the previous studies. We show that the systematic error of our waveform model is always smaller than $20\%$ (typically smaller than $10\%$) of the statistical error for events with the signal-to-noise ratio of $50$.