Towards mitigation of apparent tension between nuclear physics and astrophysical observations by improved modeling of neutron star matter

Observations of neutron stars (NSs) by the LIGO-Virgo and NICER collaborations have provided reasonably precise measurements of their various macroscopic properties. In this paper, we employ a Bayesian framework to combine them and place improved joint constraints on the properties of the NS equation of state (EOS). We use a hybrid EOS formulation that employs a parabolic expansion-based nuclear empirical parametrization around the nuclear saturation density augmented by a generic 3-segment piecewise polytrope model at higher densities. Within the 90% credible level this parametrization predicts $R_{1.4}={12.57}_{-0.92}^{+0.73}\mathrm{km}$ and ${\mathrm{\Lambda }}_{1.4}={550}_{-225}^{+223}$ for the radius and dimensionless tidal deformability, respectively, of a $1.4M_{\odot }$ NS. Finally, we show how the construction of the full NS EOS based solely on the nuclear empirical parameters at saturation density leads to certain tension with the astrophysical data, and how the hybrid approach provides a resolution to it.

Figure 1

Bayesian evidence of the astrophysical data compared between a set of nested models with varying transition densities where the first polytrope is attached. We find 1.25 ${\rho }_0$ is the most favored transition density.

Figure 2

In the top panel, the marginalized posterior distribution of the pressure in the NS interior as a function of energy is shown, using a nuclear physics informed prior: (i) on the left, PSR $\mathrm{J}0740+6620$ alone, (ii) in the middle, two GW observations combined, and (iii) on the right, with NICER data added. Some standard EOS curves like APR [50], SLy [51], WFF1 [52], MPA1 [53], H4 [54], and TW99 [55] are also overlaid. In the bottom panel corresponding mass-radius posterior distributions are shown. In the left panel (both top and bottom) 90% CR of priors is shown in gray.

Figure 3

Similar to Fig. 2 but with piecewise-polytrope parametrization.

Figure 4

Posterior distributions of nuclear parameters after adding the PSR $\mathrm{J}0740+6620$ observation, shown using the prior ranges from Table 1. In the marginalized one-dimensional plots, the median and $1\sigma$ CR are shown.

Figure 5

Posterior distributions of nuclear parameters after adding PSR $\mathrm{J}0740+6620$ and two GW observations, shown using the prior ranges from Table 1. In the marginalized one-dimensional plots, the median and $1\sigma$ CR are shown.

Figure 6

Posterior distributions of nuclear parameters after adding PSR $\mathrm{J}0740+6620$, two GW, and NICER observations, shown using the prior ranges from Table 1. In the marginalized one-dimensional plots, the median and $1\sigma$ CR are shown.

Figure 7

Same as Fig. 6 but with uniform priors on $K_0$, $e_{\mathrm{sym}}$, and $L$.

Figure 8

Similar to Fig. 2 but with the (inadequate) empirical EOS parametrization.