betaFeedback

WorksPhysics and Astronomy

Extending multi-messenger constraints on neutron star matter through the inclusion of direct Urca cooling

Abstract

The properties and behavior of strongly interacting matter at extreme densities can not only be determined through terrestrial experiments or through theoretical considerations, but also astrophysical observations of neutron stars (NSs) and binary neutron star systems become increasingly important to obtain complementary information. In light of these findings, we perform a Bayesian study inferring the NS equation of state (EoS) by incorporating constraints from the direct Urca (dUrca) process, thereby including information on matter composition in addition to traditionally used macroscopic observables. We implement a self-consistent treatment of ββ-equilibrium and charge neutrality in the jester framework, including the calculation of the proton fraction and the onset of nucleonic dUrca in the presence of electrons and muons. We combine constraints from chiral effective field theory, astrophysical measurements of NS masses, radii, and tidal deformabilities, and observations of rapidly and slowly cooling neutron stars. The dUrca constraint seems to be in favor of stiffer EoSs, but –at the current stage– has only a minor impact on macroscopic NS properties once we include other nuclear and astrophysical constraints. In contrast, the dUrca information provides a constraint on the composition of canonical-mass NSs favoring a higher proton fraction inside the star, illustrating the complementary information carried by cooling observations. Our results demonstrate the potential of incorporating composition-sensitive observables into multimessenger inference and provide a step toward a more comprehensive treatment of dUrca constraints with microscopically motivated EoS models.

Introduction

Despite noticeable progress over the last years, cf. e.g. [1, 2, 3, 4] for reviews and references, understanding the properties of strongly interacting matter at the highest densities in our Universe is still an unanswered question of modern physics. Although nuclear physics computations can provide important information up to about two times nuclear saturation density (2 nsat2\,n_{\mathrm{sat}}), i.e., twice the densities reached typically inside atomic nuclei, through methods such as chiral effective field theory [5, 6], or at extremely high densities (≳40nsat\gtrsim40n_{\mathrm{sat}}), where quantum chromodynamics can be solved perturbatively [7, 8], the intermediate regime of a few times nuclear saturation density remains theoretically uncertain and is difficult to probe directly in terrestrial experiments [2, 6, 9]. Neutron stars (NSs), among the densest objects known in our Universe, offer a unique opportunity to probe this density region and provide important complementary information. Luckily, observations of NSs or NS mergers can happen through numerous observational channels, which –when combined– provide a wealth of information that can be used to constrain the properties of strongly interacting matter. Such observations include radio observations of massive pulsars [10, 11, 12] revealing the maximum mass of NSs, X-ray pulse-profile modelling of rotating NSs through facilities such as NICER providing mass-radius measurements [13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23], as well as multimessenger observations of merging binary NSs, for which GW170817/GRB170817A/AT2017gfo [24, 25, 26, 27] is an exemplary case. We refer further to [4] for a review of many of the existing observational constraints.

The direct Urca (dUrca) process, consisting of neutron beta decay and its inverse electron-capture reaction, can occur only when the proton fraction exceeds a threshold value. For npe matter, i.e., matter consisting of neutrons, protons, and electrons, this threshold is ≃11.1%\simeq11.1\%, increasing to ∼14.8%\sim14.8\% when muons are present [28]. Since the proton fraction in β\beta-equilibrated NS matter is governed by the density dependence of the nuclear symmetry energy, the dUrca threshold directly connects NS cooling to the composition and nuclear EoS [2, 29]. Importantly, rapid cooling is not unique to nucleonic dUrca: fast neutrino emission may also arise from exotic degrees of freedom such as hyperons or deconfined quarks, while nucleonic dUrca can be substantially suppressed by nucleon superfluidity [2, 30, 31, 32]. However, superfluidity itself could also accelerate cooling, which makes the identification of a rapidly cooling source with nucleonic dUrca is not unique [33, 34].

Several studies have investigated the relation between dUrca, the nuclear symmetry energy, and NS properties, including Bayesian analyses based on nucleonic meta-models and nuclear energy-density functionals [29, 35, 36, 37]. Very recently, Ref. [38] combined nuclear-structure and isospin-transport data with χ\chiEFT and astrophysical constraints within an asymptotically causal metamodel [39], finding that nucleonic dUrca onset at or below 1.4 M⊙1.4\,M_\odot is strongly disfavored. Recent multimessenger EoS analyses have also used nuclear and astronomical constraints to infer the proton fraction and assess Urca cooling [9]. Here, we take a complementary step by incorporating the dUrca information explicitly into the likelihood of a multimessenger Bayesian EoS inference.

In contrast to macroscopic observables such as masses, radii, and tidal deformabilities, evaluating dUrca requires a model that provides the composition of β\beta-equilibrated matter. We therefore extend the JESTER framework [40] to consistently compute the proton fraction and dUrca threshold within different nucleonic models, while retaining an agnostic high-density sound-speed extension. This allows us to quantify how dUrca information affects both the inferred macroscopic properties of NSs and the proton fraction of canonical-mass stars.

EOS construction

We extend the available EoS classes in the JESTER framework and we use a metamodel [41] with a sound-speed extension [4, 42, 43] and also include the extended-Skyrme energy-density functionals [44, 45]. In the nucleonic model, we further incorporate muons in the lepton sector and solve β\beta-equilibrium self-consistently,1 and compute direct Urca thresholds for both electron and muon channels. These enable a more complete and microphysically consistent assessment of the impact of dUrca.

In this work, we keep the NS crust fixed to the DH(SLy4) EoS [46] and use a metamodel and extended-Skyrme EoS for the homogeneous nuclear matter. We consider two assumptions about the nucleonic EoS breakdown density. First, we use a variable breakdown density, where we sample nbreakn_{\mathrm{break}} randomly between 1 nsat1\,n_{\mathrm{sat}} to 4 nsat4\,n_{\mathrm{sat}}. This captures a wide range of EoSs on the speed-of-sound branch and is a widely adopted assumption [4]. Second, we determine nbreakn_{\mathrm{break}} as the density when the speed of sound within the nucleonic model reaches either zero or one, i.e., when the model breaks down. This provides an upper limit on nbreakn_{\mathrm{break}}, enabling a wider parameter space of the nucleonic EoS.

Generally, observations suggest that at sufficiently high densities and above a given proton fraction, Yp≥YDUY_p \ge Y_{\mathrm{DU}}, dUrca can set in. YDUY_{\mathrm{DU}} is determined by enforcing β\beta-equilibrium, charge-neutrality for zero-temperature hadronic matter consisting of protons (pp), neutrons (nn), electrons (ee), and muons (μ\mu).

Enforcing β\beta-equilibrium and momentum conservation, the Fermi surfaces of the particles fulfill pF,n≤pF,p+pF,ep_{F,n} \le p_{F,p} + p_{F,e}, which allows for the direct Urca n→p+e−+νˉen \to p + e^- + \bar{\nu}_e and its inverse reaction to occur. Using the relation pF,i=ℏ(3π2ni)1/3p_{F,i} = \hbar(3\pi^2 n_i)^{1/3}, with particle fractions defined by Yi=ni/nY_i = n_i/n, this condition equates to: (1−Yp)1/3≤Yp1/3+Ye1/3(1-Y_p)^{1/3} \le Y_p^{1/3} + Y_e^{1/3}, which can be simplified into:

Yp≥(1+(1+(Ye/Yp)1/3)3)−1=YDU,(1)Y_p \ge\left(1 + \left(1 + \left(Y_e/Y_p\right)^{1/3}\right)^3\right)^{-1} = Y_{\mathrm{DU}}, \tag*{(1)}

with Yp=Ye+YμY_p = Y_e + Y_\mu, where YDUY_{\mathrm{DU}} defines the onset of the dUrca process [28, 58]. If this condition is fulfilled in the center of the star, thermal energy will be quickly converted into neutrino radiation that escapes the NS, hence, the temperature will decrease rapidly. Therefore, for a certain EoS, there might be a density nDUn_{\mathrm{DU}} at which the proton fraction exceeds YDUY_{\mathrm{DU}} and the direct Urca process sets in. Figure 1 shows YDUY_{\mathrm{DU}} together with the prior proton fraction for various setups as further discussed throughout this paper.

Figure 1. Proton fraction YpY_p versus baryon density n/nsatn/n_{\mathrm{sat}} of extended-Skyrme EoS. The gray shaded band denotes the 95% prior region for the direct Urca threshold YDUY_{\mathrm{DU}}. The shaded regions show the 95% credible interval of posteriors using dUrca only (blue), Astro + χ\chiEFT (green), and Astro + χ\chiEFT + dUrca (red). Details about the individual Astro and χ\chiEFT constraints are outlined in Sec. III.

We can have the following scenarios. In the simplest case, dUrca sets in within the nucleonic part of the EOS described by the nucleonic model; see label (a) in Figure 2. In this case, we can directly compute the NS mass at which dUrca sets in and compare to observational constraints. Another possibility is that the nucleonic description terminates at low densities, either because the randomly chosen breakdown density nbreakn_{\mathrm{break}} is reached before the dUrca density, or because the sampled metamodel parametrization becomes unphysical (e.g., reaches its causality limit). In these cases, the nucleonic branch does not extend to densities high enough to determine whether dUrca took place; see label (b) in Figure 2. Hence, when the density exceeds nbreakn_{\mathrm{break}}, we use a sound-speed extension (CSE). In this work, we use a six-segment extension up to 25nsat25n_{\mathrm{sat}}. This provides greater flexibility, covers a wider range of stiffness, and reaches massive NSs in an agnostic fashion. However, no microscopic description can be deduced within the CSE region.

Schematic of the dUrca likelihood evaluation on mass–radius curves

Figure 2. Schematic of the dUrca likelihood evaluation on mass–radius curves for representative EOS construction. Left panel: Evaluated mass–radius branches under distinct onset regimes. Curve (a) shows nucleonic dUrca onset at nDU<nbreakn_{\mathrm{DU}} < n_{\mathrm{break}} (black star), with NSs satisfying nc≥nDUn_{\mathrm{c}} \ge n_{\mathrm{DU}} highlighted in red, followed by the CSE extension past nbreakn_{\mathrm{break}} (black circle, dark red). Curve (b) shows models without nucleonic dUrca (nDU>nbreakn_{\mathrm{DU}} > n_{\mathrm{break}} or absent); the yellow CSE segment indicates where the dUrca likelihood is marginalized from M(nbreak)M(n_{\mathrm{break}}) (star-in-circle) to M(n∗)M(n^*) across n∗∈[4,10] nsatn^* \in[4,10]\,n_{\mathrm{sat}} (white stars). Background shading showing the joint dUrca likelihood, with lighter color marking higher probability. Middle panel: Mass probability distributions for rapidly cooling sources [47, 48, 49, 50, 51, 52, 53, 54, 55] and the slow-cooling source HESS J1731−347 [56, 57]. Right panel: Cumulative distribution functions showing joint exclusion limits based on slow-cooling and rapid-cooling regimes, with transition zone defining the likelihood of dUrca mass M(nDU)M(n_{\mathrm{DU}}) point.

Bayesian inference setup and observational constraints

In addition to our EOS construction, the right panels of Figure 2 indicate the measured masses of objects whose temperature and age place them below the predictions of the minimal cooling paradigm [59], i.e., objects inferred to be rapidly cooling, for which we therefore assume dUrca is active. These objects are SAX J1808.4−3658 [47, 48] for which the cooling profile and the mass are independently measured. On the other hand, we have Cassiopeia A [49, 50, 51], PSR B2334+61 [52], and the Vela pulsar (PSR B0833−45) [53, 54, 55], for which their masses have been derived from their cooling profile or spectra. At lower masses, the temperature of a compact object is high enough to indicate the absence of rapid cooling, so we assume the dUrca process is not occurring in this region. We use observations of XMMU J173203.3−344518 [56] central compact object associated with HESS J1731-347 whose mass is independently measured. We emphasize that the mass and radius of HESS J1731-347 measured in [57] is strongly under debate, as using more recent observational data or distance measurements noticeably shifts the mass and radius towards larger values. However, we incorporate this as a relaxed constraint here as a proof of principle on how observations of non-dUrca for low mass NSs can be incorporated in the framework.

Using the TOV equations [60, 61], i.e., working under the assumption that the considered stars are stationary, spherically symmetric and can be described by an ideal fluid in general relativity, a given central density ncn_{\mathrm{c}} can be mapped to a given NS mass. This enables us to use mass estimates of both rapidly (labeled as data A) and slowly cooling objects (labeled as data B) to determine ndUrcan_{\mathrm{dUrca}}. Following this approach, the transition region (Figure 2) is computed following

LDU,trans=∏A(1−CDF⁡(M(A)))⏟upper limit×∏BCDF⁡(M(B))⏟lower limit,(2)\mathcal{L}_{\mathrm{DU,trans}} = \underbrace{\prod_A \left(1-\operatorname{CDF}(M_{(A)})\right)}_{\text{upper limit}} \times \underbrace{\prod_B \operatorname{CDF}(M_{(B)})}_{\text{lower limit}}, \tag*{(2)}

where M(A,B)M_{(A,B)} describe the masses of the rapidly or slowly cooling objects, respectively. The onset of dUrca (nc≥nDUn_{\mathrm{c}} \ge n_{\mathrm{DU}}) corresponds to a threshold mass MDUM_{\mathrm{DU}} within the ‘Transition zone’ LDU,trans\mathcal{L}_{\mathrm{DU,trans}}.

Given that the transition can happen within the nucleonic part of the EOS or within the agnostic extension, the overall likelihood of the dUrca constraint has to be split:

P(data∣MDU)={LDU−Nucl,if nDU<nbreak,LDU−CSE,otherwise.P(\mathrm{data}\mid M_{\mathrm{DU}}) = \begin{cases} \mathcal{L}_{\mathrm{DU-Nucl}}, & \text{if } n_{\mathrm{DU}} < n_{\mathrm{break}},\\ \mathcal{L}_{\mathrm{DU-CSE}}, & \text{otherwise.} \end{cases}

The first case is simply the likelihood of the transition zone,

LDU−Nucl=LDU,trans(nDU,EOSi).(3)\mathcal{L}_{\mathrm{DU-Nucl}} = \mathcal{L}_{\mathrm{DU,trans}}(n_{\mathrm{DU}},\mathrm{EOS}_i). \tag*{(3)}

Since it is not possible to compute the proton fraction within the CSE branch, we use the marginalized probability

LDU−CSE=∫LDU,trans(nDU,EOSi) π(nDU)π(n∗) dnDU dn∗(4)\mathcal{L}_{\mathrm{DU-CSE}} = \int\mathcal{L}_{\mathrm{DU,trans}}(n_{\mathrm{DU}},\mathrm{EOS}_i)\,\pi(n_{\mathrm{DU}})\pi(n^*)\,dn_{\mathrm{DU}}\,dn^* \tag*{(4)}

where nDUn_{\mathrm{DU}} represents the density at which the dUrca sets in. For our analysis, we use a uniform prior on nDUn_{\mathrm{DU}}, π(nDU)=U(nbreak,n∗)\pi(n_{\mathrm{DU}})=\mathcal{U}(n_{\mathrm{break}},n^*) with n∗∈π(n∗)=U(4nsat,10nsat)n^* \in\pi(n^*)=\mathcal{U}(4n_{\mathrm{sat}},10n_{\mathrm{sat}}). When nTOV<nDUn_{\mathrm{TOV}}<n_{\mathrm{DU}} and/or nbreak>4nsatn_{\mathrm{break}}>4n_{\mathrm{sat}}, the probability is set to zero.

Beyond the dUrca constraint, we combine our analysis with χ\chiEFT [4] alongside multiple astrophysical observations. The astrophysical constraints incorporate tidal deformability measurements from gravitational-wave event GW170817 under the GWTC−-1 low-spin posterior [24, 26], joint mass-radius estimates of PSR J0030+0451 [13, 14], PSR J0740+6620 [16, 19], PSR J0437−-4715 [18, 22], and PSR J0614−-3329 [20], heavy radio pulsar mass measurements of PSR J1614−-2230 [11] and PSR J0348+0432 [10], and the maximum mass upper limit from [12]. We construct the joint astrophysical likelihood as Lastro=LΛLMRLMTOV\mathcal{L}_{\mathrm{astro}}=\mathcal{L}_{\Lambda}\mathcal{L}_{MR}\mathcal{L}_{\mathrm{MTOV}}, where LΛ\mathcal{L}_{\Lambda} and LMR\mathcal{L}_{MR} are evaluated from posterior combinations and mixture distributions taken from the referenced sources, while LMTOV\mathcal{L}_{\mathrm{MTOV}} follows the formulation in [43]. The total likelihood is defined as the product across individual contributions, Ltotal=∏iLi\mathcal{L}_{\mathrm{total}}=\prod_i\mathcal{L}_i with Li∈{LDU,Lastro,LχEFT}\mathcal{L}_i\in\{\mathcal{L}_{\mathrm{DU}},\mathcal{L}_{\mathrm{astro}},\mathcal{L}_{\chi\mathrm{EFT}}\}, and each combination of constraints is evaluated both with and without the direct Urca condition.

Employing the Bayesian setup described above, we use the JESTER framework [40] to obtain our results. JESTER utilizes a JAX-based backend to enable native GPU acceleration [62, 63]. Posterior distributions are sampled using a sequential Monte Carlo algorithm configured with 4000 particles [64, 65, 66].2

Results and discussions

We show the mass-radius diagrams for the metamodel within the causal (random) nbreakn_{\mathrm{break}} construction in top (bottom) panels of Figure 3. From left to right, we display the prior, the inclusion of constraints from χ\chiEFT, Astro, and combined χ\chiEFT + Astro. Light red curves show NS where the dUrca threshold was achieved by the nucleonic EoS, with dark red showing its CSE extension. Yellow curves show cases in which the dUrca sets in within the CSE branch. We write in each plot the results for the maximum mass and the fiducial radius of a 1.4M⊙1.4M_{\odot} star, R1.4R_{1.4}, with (without) the dUrca constraint in black (blue). The left plots show that the dUrca constraint leads to a larger NS radius, therefore a stiffer EoS. Including χ\chiEFT disfavors stiff EoSs, while astrophysical constraints are the most stringent and favor EoSs leading to smaller radii. The combined posterior on the right panels shows that random (causal) nbreakn_{\mathrm{break}} constraints the R1.4R_{1.4} and MTOVM_{\mathrm{TOV}} to ≈11.89\approx11.89 km and ≈2.18 M⊙\approx2.18\,M_{\odot} (≈11.91\approx11.91 km and ≈2.27 M⊙\approx2.27\,M_{\odot}). Comparing the two treatments of the nucleonic metamodel branch, we find that extending the metamodel up to its causal limit (upper panels) favors larger NS masses and radii compared to randomly sampling nbreakn_{\mathrm{break}} (bottom panels). While previous studies have pointed out that extending the nucleonic metamodel to high densities can bias EoS inference [39, 67], our results quantify the corresponding impact on global NS properties and demonstrate the sensitivity of the inferred MTOVM_{\mathrm{TOV}} and R1.4R_{1.4} to the assumed breakdown density.

Mass-radius posterior predictive for causal and random $n_{\mathrm{break}}$ with metamod­el EoS

Figure 3. Mass-radius posterior predictive for causal (top) and random (bottom) nbreakn_{\mathrm{break}} with metamod­el EoS. Blue lines showing posterior samples with no cooling constraints. Samples with dUrca constraints are shown in red if the dUrca sets in inside the nucleonic branch and yellow if it sets in the CSE branch. Dark red refers to the continuation of the nucleonic branch (red). Upper subplot shows R1.4R_{1.4} histograms and right subplot shows MTOVM_{\mathrm{TOV}}. Black lines in the histograms shows overall cooling constraints. Light gray lines showing prior distribution as comparison while dark gray lines at the lower masses showing values below dUrca density for dUrca branch or nbreakn_{\mathrm{break}} for CSE branch.

Figure 4 shows the mass-radius diagram for the nucleonic matter modeled by the extended-Skyrme functional. The general trend is equivalent to the metamodel case. In the upper panels, we note that Skyrme can probe more massive NS under the nucleonic assumption than the metamodel. This enables us to probe higher densities and to reduce noticeably the cases in which nDUn_{\mathrm{DU}} happens within the CSE branch. We also note that the Skyrme combined posterior with dUrca leads to R1.4≈12.2R_{1.4}\approx12.2 km and MTOV≈2.16 M⊙M_{\mathrm{TOV}}\approx2.16\,M_{\odot}, more or less regardless of the nbreakn_{\mathrm{break}} construction. Compared to the metamodel, the Skyrme framework leads to a slightly larger canonical NS radius and a slightly lower TOV mass, although both are consistent within the given uncertainties. Generally, the extended-Skyrme mass-radius results agree well with the metamodel case.

Figure 4. Same as Figure 3 but for extended-Skyrme.

We show in Figure 5 the results for the tidal deformability of a 1.4 M⊙1.4\,M_{\odot} NS in the left panels, the nuclear empirical parameter LsymL_{\mathrm{sym}} in the center, and the proton fraction in the center of the canonical 1.4 M⊙1.4\,M_{\odot} on the right panels. Again, the bottom panels show the case where we pick nbreakn_{\mathrm{break}} randomly, and the top panels show the case where the nucleonic model is used up to its causal limit. The priors are shown as dashed lines, where the dUrca (blue) prior tends toward higher Λ1.4\Lambda_{1.4}, LsymL_{\mathrm{sym}}, and YpY_p than χ\chiEFT (green) and Astro (red). Solid lines show the posterior distributions for the different constraints. χ\chiEFT (green) and astro (red) are known to prefer softer EoS, and the obtained posteriors peak at lower values of Λ1.4\Lambda_{1.4} and LsymL_{\mathrm{sym}}. The dUrca constraint provides a rather broad distribution for tidal deformability and LsymL_{\mathrm{sym}}, showing negligible impact on NS macroscopic properties, as the χ\chiEFT and astrophysical constraints are substantially more informative for the inference. In contrast, the proton-fraction distribution on the right shows that the combined χ\chiEFT++ Astro ++ dUrca posterior is concentrated around Yp≈0.15Y_p\approx0.15, indicating that dUrca provides the dominant constraint on the composition of a 1.4 M⊙1.4\,M_{\odot} NS. This contrasts with Ref. [38], where laboratory symmetry-energy data yield Yp≈0.07Y_p \approx0.07–0.090.09 at 1.4 M⊙1.4\,M_{\odot} and disfavor dUrca onset below this mass. The difference is driven by the cooling data: there, dUrca is evaluated a posteriori within a single nucleonic metamodel, whereas here it enters the likelihood and may also

Figure 5. Distributions of (left) tidal deformability for a canonical 1.4 M⊙1.4\,M_\odot, (center) the slope of the symmetry energy LsymL_{\mathrm{sym}}, and (right) the proton fraction at the center of the canonical star with metamodel (top) and Skyrme (bottom) construction. Top (bottom) sub-panels on each EoS show the results for causal-limit (random) nbreakn_{\mathrm{break}}. Dashed (solid) lines show the prior (posterior) distributions. We show the different constraints (χ\chiEFT, Astro, dUrca) using different colors.

set in beyond nbreakn_{\mathrm{break}}. We note that through the inclusion of Cassiopeia A, PSR B2334+61, and PSR B0833-45, based on the assumption that dUrca is active within these objects, it is to be expected that our analysis provides a higher proton fraction within the stars’ center compared to [38]. Without the dUrca likelihood, our χ\chiEFT + Astro posterior is compatible with theirs. If both results hold, rapid cooling near 1.4 M⊙1.4\,M_\odot then requires larger source masses or a non-nucleonic fast-cooling channel.

As previously, the use of the metamodel up to its causal limit leads to a moderate shift in the tidal deformability and the slope of the symmetry energy. The combined distribution (solid red) favors Λ1.4=376−168+209(357−153+191)\Lambda_{1.4} = 376^{+209}_{-168}(357^{+191}_{-153}) for random (causal) nbreakn_{\mathrm{break}}, and Lsym=39.8−27.3+44.9 MeV (52.7−37.5+45.6 MeV)L_{\mathrm{sym}} = 39.8^{+44.9}_{-27.3}\,\mathrm{MeV}\ (52.7^{+45.6}_{-37.5}\,\mathrm{MeV}) for random (causal) nbreakn_{\mathrm{break}}, both for 95% CI. The distributions for the same quantities using the extended-Skyrme model for the nucleonic EoS are in good agreement with the metamodel case. The Skyrme posteriors show larger fractions of the nucleonic branch compared to the CSE branch than the metamodel. Constraints on the three quantities are tighter from the Skyrme EoS and also less sensitive to the choice of nbreakn_{\mathrm{break}} compared to the metamodel, with Lsym=43.7−34.6+32.4 MeV(43.6−30.2+29.9 MeV)L_{\mathrm{sym}} = 43.7^{+32.4}_{-34.6}\,\mathrm{MeV}(43.6^{+29.9}_{-30.2}\,\mathrm{MeV}) for random (causal) nbreakn_{\mathrm{break}}, both for 95% CI.

Conclusions

In this work, we have performed a multimessenger Bayesian analysis combining nuclear physics information from χ\chiEFT, macroscopic NS observables from gravitational waves and NICER observations, and constraints on rapid cooling through the dUrca process. To incorporate the latter consistently, we extended the JESTER framework to compute the composition of β\beta-equilibrated matter self-consistently and determine the corresponding dUrca threshold. This provides a more complete assessment of the information available from multimessenger observations than analyses based solely on macroscopic EoS observables, since dUrca constraints require explicit information about the microscopic composition, in particular the proton fraction YpY_p. Within the assumptions adopted here for the nucleonic models and the dUrca likelihood, we find that the dUrca constraint has only a minor impact on macroscopic NS properties such as the tidal deformability or radius, while additional information on the composition can be provided. This study represents a step towards incorporating composition information systematically into multimessenger EoS inference.

Notably, the qualitative results remain consistent across the four EoS constructions: metamodel and extended-Skyrme, each with random or causal breakdown of the nucleonic hypothesis. This consistency matters for the composition results, since YpY_p at a given mass is only defined where the nucleonic description still holds. In this respect, the extended-Skyrme functional retains the nucleonic hypothesis up to higher densities and NS masses than the metamodel, and is correspondingly less sensitive to the nbreakn_{\mathrm{break}} construction, so that a larger fraction of the posterior provides a well-defined composition for canonical-mass stars. We stress that this concerns the availability of a nucleonic composition rather than its value, which the cooling constraint sets as discussed above.

The constraint of Eq. (2) enters entirely through the comparison between the measured mass of a cooling object and the threshold mass MDUM_{\mathrm{DU}}, yet for most rapidly cooling sources the mass is itself derived from the thermal modeling we wish to interpret. Independent and complementary determinations of mass, surface temperature and age for the same objects would remove this circularity [36, 54], with low-mass NS carrying particular weight, because they are not expected to reach YDUY_{\mathrm{DU}} for any EoS compatible with current nuclear physics constraints [29, 31].

Acknowledgements

M. D. D. acknowledges support from a doctoral research grant from the German Academic Exchange Service (DAAD). T.W. is supported by the research program of the Netherlands Organization for Scientific Research (NWO) under grant number OCENW.XL21.XL21.038. T. D. acknowledges funding from the European Union (ERC, SMArt, 101076369). Views and opinions expressed are those of the authors only and do not necessarily reflect those of the European Union or the European Research Council. Neither the European Union nor the granting authority can be held responsible for them. The simulations were performed on DFG-funded research cluster Jarvis at the University of Potsdam (INST 336/173-1; project number: 502227537).

References

References

  1. [1]J. M. Lattimer, The nuclear equation of state and neutron star masses, Ann. Rev. Nucl. Part. Sci. 62, 485 (2012), arXiv:1305.3510 [nucl-th].DOI
  2. [2]G. F. Burgio, H. J. Schulze, I. Vidana, and J. B. Wei, Neutron stars and the nuclear equation of state, Prog. Part. Nucl. Phys. 120, 103879 (2021), arXiv:2105.03747 [nucl-th].DOI
  3. [3]K. Chatziioannou, H. T. Cromartie, S. Gandolfi, I. Tews, D. Radice, A. W. Steiner, and A. L. Watts, Neutron stars and the dense matter equation of state, Rev. Mod. Phys. 97, 045007 (2025), arXiv:2407.11153 [nucl-th].arxiv.org/abs/2407.11153
  4. [4]H. Koehn et al., From existing and new nuclear and astrophysical constraints to stringent limits on the equation of state of neutron-rich dense matter, Phys. Rev. X 15, 021014 (2025), arXiv:2402.04172 [astro-ph.HE].arxiv.org/abs/2402.04172
  5. [5]R. Machleidt and D. R. Entem, Chiral effective field theory and nuclear forces, Phys. Rept. 503, 1 (2011), arXiv:1105.2919 [nucl-th].
  6. [6]C. Drischler, J. W. Holt, and C. Wellenhofer, Chiral Effective Field Theory and the High-Density Nuclear Equation of State, Ann. Rev. Nucl. Part. Sci. 71, 403 (2021), arXiv:2101.01709 [nucl-th].DOI
  7. [7]A. Kurkela, P. Romatschke, and A. Vuorinen, Cold Quark Matter, Phys. Rev. D 81, 105021 (2010), arXiv:0912.1856 [hep-ph].DOI
  8. [8]O. Komoltsev and A. Kurkela, How Perturbative QCD Constrains the Equation of State at Neutron-Star Densities, Phys. Rev. Lett. 128, 202701 (2022), arXiv:2111.05350 [nucl-th].DOI
  9. [9]C. Y. Tsang, M. B. Tsang, W. G. Lynch, R. Kumar, and C. J. Horowitz, Determination of the equation of state from nuclear experiments and neutron star observations, Nat. Astron. 8, 328 (2024), arXiv:2310.11588 [nucl-ex].DOI
  10. [10]J. Antoniadis et al., A Massive Pulsar in a Compact Relativistic Binary, Science 340, 6131 (2013), arXiv:1304.6875 [astro-ph.HE].DOI
  11. [11]Z. Arzoumanian et al. (NANOGrav), The NANOGrav 11-year Data Set: High-precision timing of 45 Millisecond Pulsars, Astrophys. J. Suppl. 235, 37 (2018), arXiv:1801.01837 [astro-ph.HE].DOI
  12. [12]L. Rezzolla, E. R. Most, and L. R. Weih, Using gravitational-wave observations and quasi-universal relations to constrain the maximum mass of neutron stars, Astrophys. J. Lett. 852, L25 (2018), arXiv:1711.00314 [astro-ph.HE].DOI
  13. [13]T. E. Riley et al., A NICER View of PSR J0030+0451: Millisecond Pulsar Parameter Estimation, Astrophys. J. Lett. 887, L21 (2019), arXiv:1912.05702 [astro-ph.HE].DOI
  14. [14]M. C. Miller et al., PSR J0030+0451 Mass and Radius from NICER Data and Implications for the Properties of Neutron Star Matter, Astrophys. J. Lett. 887, L24 (2019), arXiv:1912.05705 [astro-ph.HE].DOI
  15. [15]T. E. Riley et al., A NICER View of the Massive Pulsar PSR J0740+6620 Informed by Radio Timing and XMM-Newton Spectroscopy, Astrophys. J. Lett. 918, L27 (2021), arXiv:2105.06980 [astro-ph.HE].DOI
  16. [16]M. C. Miller et al., The Radius of PSR J0740+6620 from NICER and XMM-Newton Data, Astrophys. J. Lett. 918, L28 (2021), arXiv:2105.06979 [astro-ph.HE].DOI
  17. [17]S. Vinciguerra, T. H. J. Salmi, A. L. Watts, et al., An Updated Mass–Radius Analysis of the 2017–2018 NICER Data Set of PSR J0030+0451, The Astrophysical Journal 961, 62 (2024).DOI
  18. [18]D. Choudhury, T. H. J. Salmi, S. Bogdanov, A. L. Watts, et al., A NICER View of the Nearest and Brightest Millisecond Pulsar: PSR J0437−4715, The Astrophysical Journal Letters 971, L20 (2024).
  19. [19]T. Salmi, D. Choudhury, Y. Kini, T. E. Riley, S. Vinciguerra, A. L. Watts, M. T. Wolff, Z. Arzoumanian, S. Bogdanov, D. Chakrabarty, K. Gendreau, S. Guillot, W. C. G. Ho, D. Huppenkothen, R. M. Ludlam, S. M. Morsink, and P. S. Ray, The Radius of the High-mass Pulsar PSR J0740+6620 with 3.6 yr of NICER Data, The Astrophysical Journal 974, 294 (2024).DOI
  20. [20]L. Mauviard, S. Guillot, T. Salmi, D. Choudhury, B. Dorsman, D. González-Caniulef, M. Hoogkamer, D. Huppenkothen, C. Kazantsev, Y. Kini, J.-F. Olive, P. Stammler, A. L. Watts, M. Mendes, N. Rutherford, A. Schwenk, I. Svensson, S. Bogdanov, M. Kerr, P. S. Ray, L. Guillemot, I. Cognard, and G. Thureau, A NICER View of the 1.4 M⊙ Edge-on Pulsar PSR J0614-3329, The Astrophysical Journal 995, 60 (2025).DOI
  21. [21]Y. Kini et al., A NICER View of PSR J0030+0451: Updated Constraints from 6 yr of NICER Observations, Astrophys. J. 1005, 201 (2026), arXiv:2602.23743 [astro-ph.HE].
  22. [22]M. C. Miller et al., The Radius of PSR J0437−4715 from NICER Data, Astrophys. J. Lett. 1000, L48 (2026), arXiv:2512.08790 [astro-ph.HE].DOI
  23. [23]M. C. Miller et al., The Radius of the Neutron Star PSR J0614-3329 from NICER Data, arXiv:2609.00965 (2026).arxiv.org/abs/2609.00965
  24. [24]B. P. Abbott et al. (LIGO Scientific, Virgo), GW170817: Observation of Gravitational Waves from a Binary Neutron Star Inspiral, Phys. Rev. Lett. 119, 161101 (2017), arXiv:1710.05832 [gr-qc].
  25. [25]P. S. Cowperthwaite et al., The Electromagnetic Counterpart of the Binary Neutron Star Merger LIGO/Virgo GW170817. II. UV, Optical, and Near-infrared Light Curves and Comparison to Kilonova Models, Astrophys. J. Lett. 848, L17 (2017), arXiv:1710.05840 [astro-ph.HE].
  26. [26]B. P. Abbott et al. (LIGO Scientific, Virgo), Properties of the binary neutron star merger GW170817, Phys. Rev. X 9, 011001 (2019), arXiv:1805.11579 [gr-qc].DOI
  27. [27]B. P. Abbott et al. (LIGO Scientific, Virgo), GW170817: Measurements of neutron star radii and equation of state, Phys. Rev. Lett. 121, 161101 (2018), arXiv:1805.11581 [gr-qc].
  28. [28]T. Klähn, D. Blaschke, S. Typel, E. N. E. van Dalen, A. Faessler, C. Fuchs, T. Gaitanos, H. Grigorian, A. Ho, E. E. Kolomeitsev, M. C. Miller, G. Röpke, J. Trümper, D. N. Voskresensky, F. Weber, and H. H. Wolter, Constraints on the high-density nuclear equation of state from the phenomenology of compact stars and heavy-ion collisions, Phys. Rev. C 74, 035802 (2006).DOI
  29. [29]J. Margueron, R. Hoffmann Casali, and F. Gulminelli, Equation of state for dense nucleonic matter from meta-modeling. II. Predictions for neutron star properties, Phys. Rev. C 97, 025806 (2018), arXiv:1708.06895 [nucl-th].DOI
  30. [30]D. G. Yakovlev, A. D. Kaminker, O. Y. Gnedin, and P. Haensel, Neutrino emission from neutron stars, Phys. Rept. 354, 1 (2001), arXiv:astro-ph/0012122.DOI
  31. [31]D. G. Yakovlev and C. J. Pethick, Neutron star cooling, Ann. Rev. Astron. Astrophys. 42, 169 (2004), arXiv:astro-ph/0402143.DOI
  32. [32]A. Y. Potekhin, J. A. Pons, and D. Page, Neutron stars—cooling and transport, Space Sci. Rev. 191, 239 (2015), arXiv:1507.06186 [astro-ph.HE].DOI
  33. [33]D. Page, M. Prakash, J. M. Lattimer, and A. W. Steiner, Rapid Cooling of the Neutron Star in Cassiopeia A Triggered by Neutron Superfluidity in Dense Matter, Phys. Rev. Lett. 106, 081101 (2011), arXiv:1011.6142 [astro-ph.HE].DOI
  34. [34]P. S. Shternin, D. G. Yakovlev, C. O. Heinke, W. C. G. Ho, and D. J. Patnaude, Cooling neutron star in the Cassiopeia A supernova remnant: evidence for superfluidity in the core, Mon. Not. Roy. Astron. Soc. 412, L108 (2011), arXiv:1012.0045 [astro-ph.HE].DOI
  35. [35]T. Malik, B. K. Agrawal, and C. Providencia, Inferring the nuclear symmetry energy at suprasaturation density from neutrino cooling, Phys. Rev. C 106, L042801 (2022), arXiv:2206.15404 [nucl-th].DOI
  36. [36]M. V. Beznogov and D. G. Yakovlev, Statistical theory of thermal evolution of neutron stars – II. Limitations on direct Urca threshold, Mon. Not. Roy. Astron. Soc. 452, 540 (2015), arXiv:1507.04206 [astro-ph.SR].
  37. [37]M. V. Beznogov and A. R. Raduta, Bayesian Survey of the Dense Matter Equation of State Built upon Skyrme Effective Interactions, Astrophys. J. 966, 216 (2024).
  38. [38]G. Montefusco, P. Klausner, M. Antonelli, C. Ciampi, D. Gruyer, and F. Gulminelli, Ruling out nucleonic direct Urca cooling in low-mass neutron stars using nuclear data, arXiv:2609.24940 (2026).arxiv.org/abs/2609.24940
  39. [39]G. Montefusco, M. Antonelli, and F. Gulminelli, An Asymptotically Causal Metamodel for Neutron Star Equations of State, arXiv:2604.00196 (2026).arxiv.org/abs/2604.00196
  40. [40]T. Wouters, P. T. H. Pang, H. Koehn, H. Rose, R. Somasundaram, I. Tews, T. Dietrich, and C. Van Den Broeck, Leveraging differentiable programming in the inverse problem of neutron stars, Phys. Rev. D 112, 043037 (2025), arXiv:2504.15893 [astro-ph.HE].arxiv.org/abs/2504.15893
  41. [41]J. Margueron, R. Hoffmann Casali, and F. Gulminelli, Equation of state for dense nucleonic matter from meta-modeling. I. Foundational aspects, Phys. Rev. C 97, 025805 (2018), arXiv:1708.06894 [nucl-th].DOI
  42. [42]I. Tews, J. Margueron, and S. Reddy, Critical examination of constraints on the equation of state of dense matter obtained from GW170817, Phys. Rev. C 98, 045804 (2018), arXiv:1804.02783 [nucl-th].DOI
  43. [43]T. Dietrich, M. W. Coughlin, P. T. H. Pang, M. Bulla, J. Heinzel, L. Issa, I. Tews, and S. Antier, Multimessenger constraints on the neutron-star equation of state and the Hubble constant, Science 370, 1450 (2020), arXiv:2002.11355 [astro-ph.HE].DOI
  44. [44]N. Chamel, S. Goriely, and J. M. Pearson, Further explorations of Skyrme-Hartree-Fock-Bogoliubov mass formulas. XI. Stabilizing neutron stars against a ferromagnetic collapse, Phys. Rev. C 80, 065804 (2009), arXiv:0911.3346 [nucl-th].DOI
  45. [45]G. Grams, W. Ryssens, G. Scamps, S. Goriely, and N. Chamel, Skyrme-Hartree-Fock-Bogoliubov mass models on a 3D mesh: III. From atomic nuclei to neutron stars, Eur. Phys. J. A 59, 270 (2023), arXiv:2307.14276 [nucl-th].arxiv.org/abs/2307.14276
  46. [46]F. Douchin and P. Haensel, A unified equation of state of dense matter and neutron star structure, Astron. Astrophys. 380, 151 (2001), arXiv:astro-ph/0111092.
  47. [47]C. O. Heinke, P. G. Jonker, R. Wijnands, and R. E. Taam, Constraints on Thermal X-ray Radiation from SAX J1808.4-3658 and Implications for Neutron Star Neutrino Emission, Astrophys. J. 660, 1424 (2007), arXiv:astro-ph/0612232.DOI
  48. [48]B. Dorsman, T. Salmi, A. L. Watts, M. Ng, A. Bobrikova, V. Loktev, J. Poutanen, and J. Wilms, Pulse profile modelling of the accretion-powered millisecond pulsar SAX J1808.4−3658 using NICER data from its 2019 and 2022 outbursts, Mon. Not. Roy. Astron. Soc. 545, star1983 (2025), arXiv:2501.07152 [astro-ph.HE].
  49. [49]W. C. G. Ho and C. O. Heinke, A Neutron Star with a Carbon Atmosphere in the Cassiopeia A Supernova Remnant, Nature 462, 71 (2009), arXiv:0911.0672 [astro-ph.HE].DOI
  50. [50]W. C. G. Ho, Y. Zhao, C. O. Heinke, D. L. Kaplan, P. S. Shternin, and M. J. P. Wijngaarden, X-ray bounds on cooling, composition, and magnetic field of the Cassiopeia A neutron star and young central compact objects, Mon. Not. Roy. Astron. Soc. 506, 5015 (2021), arXiv:2107.08060 [astro-ph.HE].DOI
  51. [51]P. S. Shternin, D. D. Ofengeim, C. O. Heinke, and W. C. G. Ho, Constraints on neutron star superfluidity from the cooling neutron star in Cassiopeia A using all Chandra ACIS-S observations, Mon. Not. Roy. Astron. Soc. 518, 2775 (2022), arXiv:2211.02526 [astro-ph.HE].DOI
  52. [52]K. E. McGowan, S. Zane, M. Cropper, W. T. Vestrand, F. A. Cordova, and C. Ho, Evidence for surface cooling emission in the xmm-newton spectrum of the x-ray pulsar psr b2334+61, Astrophys. J. 639, 377 (2006), arXiv:astro-ph/0508439.DOI
  53. [53]D. Viganò, N. Rea, J. A. Pons, R. Perna, D. N. Aguilera, and J. A. Miralles, Unifying the observational diversity of isolated neutron stars via magneto-thermal evolution models, Mon. Not. Roy. Astron. Soc. 434, 123 (2013), arXiv:1306.2156 [astro-ph.SR].DOI
  54. [54]S. Beloin, S. Han, A. W. Steiner, and K. Odbadrakh, Simultaneous Fitting of Neutron Star Structure and Cooling Data, Phys. Rev. C 100, 055801 (2019), arXiv:1812.00494 [nucl-th].DOI
  55. [55]P. S. Negi, A starquake model for vela pulsar, Mon. Not. Roy. Astron. Soc. 366, 73 (2006), arXiv:astro-ph/0510562.DOI
  56. [56]D. Klochkov, V. Suleimanov, G. Pühlhofer, D. G. Yakovlev, A. Santangelo, and K. Werner, The neutron star in HESSJ1731-347: Central compact objects as laboratories to study the equation of state of superdense matter, Astron. Astrophys. 573, A53 (2015), arXiv:1410.1055 [astro-ph HE].
  57. [57]V. Doroshenko, V. Suleimanov, G. Pühlhofer, and A. Santangelo, A strangely light neutron star within a supernova remnant, Nature Astron. 6, 1444 (2022).DOI
  58. [58]J. M. Lattimer, M. Prakash, C. J. Pethick, and P. Haensel, Direct URCA process in neutron stars, Phys. Rev. Lett. 66, 2701 (1991).pubmed.ncbi.nlm.nih.gov/10043594
  59. [59]D. Page, J. M. Lattimer, M. Prakash, and A. W. Steiner, Minimal cooling of neutron stars: A New paradigm, Astrophys. J. Suppl. 155, 623 (2004), arXiv:astro-ph/0403657.DOI
  60. [60]R. C. Tolman, Static solutions of Einstein’s field equations for spheres of fluid, Phys. Rev. 55, 364 (1939).DOI
  61. [61]J. R. Oppenheimer and G. M. Volkoff, On massive neutron cores, Phys. Rev. 55, 374 (1939).DOI
  62. [62]R. Frostig, M. J. Johnson, and C. Leary, Compiling machine learning programs via high-level tracing, Systems for Machine Learning 4 (2018).
  63. [63]A. Cabezas, A. Corefins, J. Lao, and R. Louf, Blackjax: Composable Bayesian inference in JAX (2024), arXiv:2402.10797 [cs.MS].arxiv.org/abs/2402.10797
  64. [64]P. Del Moral, A. Doucet, and A. Jasra, Sequential monte carlo samplers, Journal of the Royal Statistical Society Series B: Statistical Methodology 68, 411 (2006).
  65. [65]N. Chopin and O. Papaspiliopoulos, An Introduction to Sequential Monte Carlo, Springer Series in Statistics (Springer Cham, 2020).DOI
  66. [66]C. Dai, J. Heng, P. E. Jacob, and N. Whiteley, An invitation to sequential monte carlo samplers, Journal of the American Statistical Association 117, 1587 (2022).arxiv.org/abs/2007.11936
  67. [67]L. Suleiman, A. F. Fantina, F. Gulminelli, and J. Read, Nuclear parameter inference with semi-agnostic priors, arXiv:2512.05315 10.1051/0004-6361/202558416 (2025).DOI

Paper details

vlnc:W_s6wyv5d4
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 5

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents