Follow-up analyses of the binary-neutron-star signals GW170817 and GW190425 by using post-Newtonian waveform models

We reanalyze the binary-neutron-star signals, GW170817 and GW190425, focusing on the inspiral regime to avoid uncertainties on waveform modeling in the postinspiral regime. We use post-Newtonian waveform models as templates, which are theoretically rigid and efficiently describe the inspiral regime. We study potential systematic difference in estimates of the binary tidal deformability $\tilde{\Lambda}$ by using different descriptions for the point-particle dynamics and tidal effects. We find that the estimates of $\tilde{\Lambda}$ show no significant systematic difference among three models for the point-particle parts: TF2, TF2g, and TF2+, when they employ the same tidal model. We compare different tidal descriptions given by different post-Newtonian orders in the tidal phase. Our results indicate that the estimates of $\tilde{\Lambda}$ slightly depend on the post-Newtonian order in the tidal phase and an increase in the tidal post-Newtonian order does not lead to a monotonic change in the estimate of $\tilde{\Lambda}$. We also compare the estimate of $\tilde{\Lambda}$ obtained by the post-Newtonian tidal model and numerical-relativity calibrated tidal models. We find that the post-Newtonian model gives slightly larger estimate of $\tilde{\Lambda}$ and wider posterior distribution than the numerical-relativity calibrated models. According to Bayesian model comparison, it is difficult to identify a preference among the post-Newtonian orders by relying on the GW170817 and GW190425 data. Our results indicate no preference among numerical-relativity calibrated tidal models over the post-Newtonian model. Additionally, we present constraints on equation-of-state models for neutron stars with the post-Newtonian model, which show that the GW170817 data disfavor less compact models, though they are slightly weaker constraints than the numerical-relativity calibrated tidal models.