The interplay of astrophysics and nuclear physics in determining the properties of neutron stars

TL;DR

This work tackles how neutron-star observables encode both dense-matter microphysics and astrophysical population effects by jointly inferring a universal EoS and separate mass distributions for Galactic EM NSs and GW-detected BNSs. It introduces a Gaussian-process-based EoS prior and a reweighting approach to coherently combine radio, X-ray (NICER), and gravitational-wave data, while allowing distinct maximum masses for the two NS populations. The analysis yields $R_{1.4}=12.2^{+0.8}_{-0.9}$ km and $M_{ m TOV}=2.28^{+0.41}_{-0.21}\,M_\odot$, with a 90% lower bound on the maximum $c_s^2$ around $0.59$, and population maxima $M_{ m pop, EM}=2.05^{+0.11}_{-0.06}\,M_\odot$ and $M_{ m pop, GW}=1.85^{+0.39}_{-0.16}\,M_\odot$. Crucially, the study finds no evidence that astrophysical processes push the maximum NS mass beyond the nuclear-physics TOV limit, though the EM and GW populations have distinct mass distributions. This demonstrates the potential of joint mass–EoS inference to disentangle microphysical and macrophysical factors and to tighten constraints on NS interior physics.

Abstract

Neutron star properties depend on both nuclear physics and astrophysical processes, and thus observations of neutron stars offer constraints on both large-scale astrophysics and the behavior of cold, dense matter. In this study, we use astronomical data to jointly infer the universal equation of state of dense matter along with two distinct astrophysical populations: Galactic neutron stars observed electromagnetically and merging neutron stars in binaries observed with gravitational waves. We place constraints on neutron star properties and quantify the extent to which they are attributable to macrophysics or microphysics. We confirm previous results indicating that the Galactic and merging neutron stars have distinct mass distributions. The inferred maximum mass of both Galactic neutron stars, $M_{\rm pop, EM}=2.05^{+0.11}_{-0.06}\,M_{\odot}$ (median and 90\% symmetric credible interval), and merging neutron star binaries, $M_{\rm pop, GW}=1.85^{+0.39}_{-0.16}\,M_{\odot}$, are consistent with the maximum mass of nonrotating neutron stars set by nuclear physics, $M_{\rm TOV} =2.28^{+0.41}_{-0.21}\,M_\odot$. The radius of a $1.4\,M_{\odot}$ neutron star is $12.2^{+0.8}_{-0.9}\,$km, consistent with, though $\sim 20\%$ tighter than, previous results using an identical equation of state model. Even though observed Galactic and merging neutron stars originate from populations with distinct properties, there is currently no evidence that astrophysical processes cannot produce neutron stars up to the maximum value imposed by nuclear physics.

Figures (13)

• Figure 1: Posterior on the mass distribution of the GW BNS (orange) and the Galactic NS (blue) population. We plot the median and 90% highest-probability credible regions. The EM population is constrained to much better precision than the GW one due to the low number of GW BNS detections. With the caveat that they correspond to the astrophysical BNS and observed Galactic NS distributions respectively, we find that the two distribution are inconsistent, in agreement with Ref. KAGRA:2021duu. Faint lines are random draws from the GW mass distribution, illustrating the bimodal uncertainties in the mass distribution.
• Figure 2: Marginalized posterior for the power-law slope $\alpha$ and maximum mass $M_{\rm pop,GW}$ of the GW population. The slope $\alpha$ is poorly constrained and thus its posterior rails against the upper prior bound, in turn affecting the $M_{\rm pop,GW}$ posterior.
• Figure 3: One- and two-dimensional posteriors for select EoS macroscopic and microscopic parameters: the TOV mass, $M_{\rm TOV}$, the radius and tidal deformability of a canonical $1.4\,M_{\odot}$ NS, $R_{1.4}$ and $\Lambda_{1.4}$ respectively, the radius of a $1.8\,M_{\odot}$ NS, $R_{1.8}$, and the log-base-10 pressure (divided by the speed of light squared) at twice and 6 times nuclear saturation, $p_{2.0}$ and $p_{6.0}$ respectively, when measured in $\rm{g}/\rm{cm}^3$. Two-dimensional contours denote the boundaries of the 90% credible regions. We show the prior (black), the posterior from the main analysis that marginalizes over the mass distribution (blue), and the analogous posterior that arises from additionally including the mass-radius measurement of J0437-4715 in the analysis of Ref. Legred:2021.
• Figure 4: Mass-radius inference, we show the $90 \%$ symmetric credible region for the radius at each mass. We plot the prior (black), posterior from the main analysis that marginalizes over the mass distribution (blue), and posterior from Ref. Legred:2021 that fixes the mass distribution to flat and does not include J0437-4715. The upper limit on the radius decreases by $\sim 0.5\,\rm{km}$ for all masses.
• Figure 5: Mass-central density inference, we show the $90 \%$ symmetric credible region for the NS mass at each value of the central density $\rho_c$. We plot the prior (black), posterior from the main analysis that marginalizes over the mass distribution (blue), and posterior from Ref. Legred:2021 that fixes the mass distribution to flat and does not include J0437-4715. Vertical lines denote multiples of the nuclear saturation density. Maroon and red contours mark $1$ and $2$-$\sigma$ credible regions, respectively, for the joint posterior on $\rho_{c}$-$M_{\rm TOV}$.