betaFeedback

WorksPhysics and Astronomy

NS 1987A: A Potential Probe of Crustal Superfluidity with Future X-ray Observations

Abstract

Several indirect observations imply that the central compact object in the supernova SN 1987A is likely to be the youngest neutron star identified to date, NS 1987A. We investigate how a future detection of its thermal X-ray emission could constrain neutron-star interiors, using cooling models and a generic high-resolution, large-area X-ray observatory with Lynx-like capabilities as a sensitivity benchmark. For the adopted sensitivity and observing window, NS 1987A is detectable at t∼60 yrt\sim60~{\rm yr} (roughly 2047) unless rapid neutrino cooling operates, such as the direct Urca process. If the neutron superfluidity in the inner crust is weak, the temperature profile inside NS 1987A is likely not to be isothermal at the detection. In that case, the future direct observation of NS 1987A is a potential probe of the properties inside the inner crust. We compare alternative X-ray observatories and emphasize that both effective area and angular resolution, needed to reduce contamination from the surrounding remnant, determine whether this diagnostic can be realized.

Introduction

The central compact object (CCO) in supernova (SN) 1987A, which was confirmed through the detection of neutrinos by Kamiokande ([21, 6]), is a unique astrophysical laboratory because of its youth, enabling us to study the formation of supernova remnant (SNR 1987A) and a neutron star (NS 1987A). Thanks to such attractions, various wavelength observations have been operated for more than 30 years.

Multi-wavelength observations across the electromagnetic spectrum, have detected emission probably associated with NS 1987A. ALMA observations ([56]) have revealed a blob of reprocessed emission possibly stemming from NS 1987A. NuSTAR observations ([19] (Greco et al. 2022)) indicated that the hard (≳8\gtrsim8 keV) X-ray spectra of SN1987A require a non-thermal emission component, likely originating from a very young pulsar wind nebula (PWN). Black-body radiation is emitted by NS 1987A if a thick layer of light elements encloses the NS (e.g. [42]). More recently, JWST ([14]) have found an excess of emission associated with either this thermal emission from NS 1987A or with the nonthermal emission from the PWN. All these findings imply that the CCO is likely to be the youngest neutron star. Naturally, it is possible that the emission observed by JWST and ALMA is due to a combination of nonthermal and thermal emission, but for our goals, we will focus on the latter.

In Dohi et al. [11] (hereafter Paper I), we showed the diagnostic potential of thermal X-ray observations of NS 1987A using the Lynx concept, with a characteristic angular resolution of 0.5 arcsec and a mirror effective area of ∼2 m2\sim2\ \mathrm{m}^{2} at a photon energy of 1 keV, as a benchmark. Lynx currently has no funded implementation or scheduled launch, and may never be realized; it should therefore not be regarded as an expected observing facility. NASA is reassessing the concept for the next decadal survey rather than preparing it for launch (Lynx2030 Science Analysis Group et al. [31]). Hereafter, “Lynx-like” denotes a generic observatory with the benchmark imaging and sensitivity capabilities adopted in Paper I, without implying a particular mission or launch date. Paper I compared cooling models with limiting intrinsic luminosities at t=50t = 50 yr (2037), including absorption by supernova ejecta (Orlando et al. [39]). Extending this comparison to later ages allows us to study how the epoch of a future thermal detection would constrain the stellar interior, independently of whether Lynx itself is launched.

Through the comparison between the observed NS luminosity or surface temperature and the theoretical one based on so-called cooling theory, one can extract information on neutron-star interior, such as the EOS and nucleon superfluidity (for reviews, see Yakovlev & Pethick [61]; Page et al. [43]; Burgio et al. [7]). While most observed isolated NSs are older than the relaxation time scale as trelax≃10−100t_{\mathrm{relax}} \simeq10 - 100 yrs (Lattimer et al. [28]; Gnedin et al. [17]; Sales et al. [49]), they are likely to be isothermal (Yakovlev et al. [60]). On the other hand, NS 1987A is comparable with trelaxt_{\mathrm{relax}}. Due to its youth, the crust may be hotter than the core, implying that neutrino emission in the inner crust is significant due to its strong temperature dependence. In fact, Paper I showed the importance of crust superfluidity for the detectability by a Lynx-like observatory. As the detection age is delayed, the neutrino emission in the core becomes more significant, such as the nucleon direct Urca (DU) process, which is a rapid cooling process strongly dependent on the EOS (Lattimer et al. [27]). Thus, the future detection of NS 1987A must provide various information on NS properties depending on the detection age.

In this work, we investigate the thermal evolution of NS 1987A and the information on its interior that could be obtained from a future thermal X-ray detection. We use the Lynx-like sensitivity benchmark to calculate illustrative detection ages, including models with the DU process, which Paper I did not consider. The physical connection between crustal superfluidity, thermal relaxation, and the observable luminosity is independent of the telescope; the numerical detection ages depend on its sensitivity and availability. We discuss alternative observatories and their limitations in § 4.

Cooling models and observations

To describe thermal luminosity, there are basically four kinds of parameters: EOS, mass MNSM_{\mathrm{NS}}, nucleon superfluidity, and envelope mass MenvM_{\mathrm{env}}. In this work, we employ two EOSs of APR (Akmal et al. [2]) and Y+K (Muto et al. [35] (2022); Muto [34]), which have different radii, 11.6 km and 12.7 km at 1.40 M⊙1.40\,M_{\odot} stars, respectively. The corresponding threshold masses of the DU process are MDU/M⊙=1.96M_{\mathrm{DU}}/M_{\odot} = 1.96 and 1.31, respectively. We consider only light NSs with MNS≃1.2M_{\mathrm{NS}} \simeq1.2–1.6 M⊙1.6\,M_{\odot} according to model comparison with observed supernova lightcurves of SN1987A (see Page et al. [42] for details). Hence, the cooling models with APR EOS always obey the so-called minimal cooling scenario (Page et al. [44]). In the Y+K EOS, degrees of hyperons and K−K^{-} condensation are included, and the Λ\Lambda-induced DU process occurs with MNS=1.5 M⊙M_{\mathrm{NS}} = 1.5\,M_{\odot}, but this effect is minor due to the presence of nucleon DU process. Note that K−K^{-} condensation does not appear in the mass range we consider, regardless of the strength of interaction between nucleons and kaons (Bhat et al. [5]). Although we select two EOS and limited mass ranges, their impacts on cooling curves are notably small in slow cooling scenario, as in a comprehensive study (Lim et al. [30]).

In young neutron stars such as NS 1987A, the neutron 1S0{}^{1}S_{0} superfluidity in the inner crust is one of the important ingredients for its surface temperature. In this work, we consider six kinds of superfluidity: CCDK (Chen et al. [9]), SFB (Schwenk et al. [50]), WAP (Wambach et al. [58]), GC (Gezerlis & Carlson [16]), GIPSF (Gandolfi et al. [15]), and T (Takatsuka [54]) roughly in order of strength (the same as the right panel of Figure 6 in Page et al. [42]), in addition to no superfluidity (we refer to “All Unpaired”). Regarding the superfluidity in the core, we fix the “0.5a” model for 3P2{}^{3}P_{2} neutrons (Page et al. [46]), which reproduces the observed cooling rate of Cassiopeia A, while we fix the CCDK model for 1S0{}^{1}S_{0} protons (Chen et al. [9]). The corresponding maximum critical temperatures are Tc,n3=5×108T_{\mathrm{c,n3}} = 5 \times10^{8} K and Tc,p1=6×109T_{\mathrm{c,p1}} = 6 \times10^{9} K, respectively. With the ages up to 200 yr, we confirm that the thermal luminosity remains largely unchanged for Tc,n3≲7×108T_{\mathrm{c,n3}} \lesssim7 \times10^{8} K, since such a low critical temperature delays the onset of the pair breaking and formation process until after 200 yr (e.g., Page et al. [46] (2009)). On the impact of proton superconductivity on young NSs, see Appendix § 6.

Figure 6. Cooling curves with and without CCDK proton superconductivity, assuming the CCDK neutron 1S0{}^{1}\mathrm{S}_{0} superconductivity model, Y+K EOS and 1.6M⊙1.6M_{\odot} stars. Observational properties plotted are the same as Figure 1, but assuming the adopted mass and EOS.

In terms of envelope mass, we consider η≡g14Menv/MNS=10−8\eta\equiv g_{14} M_{\mathrm{env}}/M_{\mathrm{NS}} = 10^{-8} and 10−710^{-7}, which connects the surface temperature and outermost crust temperature at ρ=1010\rho= 10^{10} g cm−3^{-3}, in accordance with the accreted envelope model of Potekhin et al. [47]. For η≲10−9\eta\lesssim10^{-9}, thermal luminosity is too low to account for the ALMA observations, even considering the uncertainties on extra heating mechanisms such as 44Ti{}^{44}\mathrm{Ti} radioactive heating [10, 42].

All sensitivity curves below retain the instrument response and remnant-background treatment of Paper I; they are a fixed Lynx-like benchmark, not sensitivity forecasts for other missions. The subscript “Lynx” in Eq. (1) identifies this benchmark. With the use of NSCool code [41], we systematically calculate thermal luminosities and compare with ALMA observations and sensitivity limits of a Lynx-like observatory, the latter of which corresponds to the critical luminosity between non-detection and detection of NS 1987A. We take their concrete value from our previous work at t=50t = 50 yr, L(50)L(50) (see Table 1–6 in Paper I), assuming the conservative parameter sets of exposure time of 0.2 Ms and 3σ3\sigma significance. The NS kick velocity is chosen to be vkick=500 km s−1v_{\mathrm{kick}} = 500\ \mathrm{km\ s^{-1}} inferred from recent observations [26, 55]. For the Y+K models, the correction of gravitational redshift is applied to the original values with the APR. We note that, in all cases, Lynx-like sensitivity limits with vkick≲500 km s−1v_{\mathrm{kick}} \lesssim500\ \mathrm{km\ s^{-1}} are higher than luminosities from cooling models at t=50t = 50 yr.

NS mass203720472047 (w/ approx.)
1.20 M⊙1.20\ M_{\odot}1.380.5570.799
1.40 M⊙1.40\ M_{\odot}1.420.5730.822
1.60 M⊙1.60\ M_{\odot}1.450.5800.839

Table 1. Lynx-like benchmark sensitivity limits in units of 1035 erg s−110^{35}\ \mathrm{erg\ s^{-1}}, at 2037(t=50t = 50 yr) and 2047(t=60t = 60 yr) calculated without and with the approximation of Eq. (1) (see texts for details).

To consider the detection beyond 50 years, we extrapolate the Lynx-like sensitivity limit as follows: Paper I verified the following time evolution of Lynx-like benchmark sensitivity limits (see their footnote 23):

LLynx(t)=LLynx(50)t−3 for t≲200 yr,(1)L_{\mathrm{Lynx}}(t) = L_{\mathrm{Lynx}}(50)t^{-3}\ \text{for } t \lesssim200\ \mathrm{yr}, \tag*{(1)}

which we assume for simplicity. This nearly mimics that the SNR 1987A freely expands with the original density distribution for t≲200t \lesssim200 yr. To confirm the validity of such an approximation more, we calculate the Lynx-like benchmark sensitivity limits at t=60t = 60 yr in the self-consistent manner same as in Paper I. Namely, they are obtained by synthesizing the 2047 Lynx-like spectra of the faint inner region of SN 1987A, taking into account for the absorption pattern of cold ejecta based on the latest state-of-art magnetohydrodynamic (MHD) simulation models broadly comparable with the observed morphology [39].1 Table 1 presents a comparison of sensitivity limits with and without the approximation of Eq. (1) in the APR EOS. We found that Lynx-like sensitivity limits are slightly overestimated with our approximation, leading to the longer putative detection age of NS 1987A. Thus, our above assumption

is a conservative one at least up to t≲60t \lesssim60 yr, since the detection obtained in our analysis can be regarded as the upper limit value. This compensates for the high uncertainty on the structure of the ambient medium, currently unknown, in which the shock will expand in the next decades.

The detection age also depends on how thermal luminosity of NS 1987A evolves, that is, L(t)∝t−kL(t) \propto t^{-k} with power-law index kk. A simple cooling theory shows k∼1/3−1/2k \sim1/3 - 1/2 [44], which is much smaller than the expansion rate of SNR, as in Eq. (1). Thus, in the absence of rapid cooling, the declining absorption can bring thermal emission above the adopted sensitivity limit. To retain a concrete observing window, we use the same illustrative interval, 2045–2075 (58≤t≤8858 \leq t \leq88 yr), corresponding to 30 years of operation. The plotted detection ages and detection/non-detection classifications are conditional on this sensitivity and window; a later start could miss a short-lived detectable phase. Labels referring to Lynx in the existing figures denote this benchmark.

Results

Cooling curve with young NSs

Figure 1 presents the EOS and mass dependence of the comparison. In APR EOS, all NSs cool slowly according to the minimal cooling scenario, leading to the detection of NS 1987A by a Lynx-like observatory around t≈58t \approx58 yrs with η=10−7\eta= 10^{-7} while t≈63t \approx63 yrs with η=10−8\eta= 10^{-8}. In Y+K EOS, NSs cool faster than those with the APR EOS, except with 1.20 & 1.30M⊙M_\odot stars, due to the presence of the DU processes. We found that most of the rapid cooling models do not intersect with the sensitivity curves by a Lynx-like observatory, implying the non-detection of NS 1987A by a Lynx-like observatory. Thus, the DU process generally disfavors detection within the adopted benchmark window.

Figure 1. Mass dependence of cooling curves and Lynx-like benchmark sensitivity limits with Y+K and APR EOS. Sensitivity limits without the approximation are also plotted as triangle symbols, taking into account the uncertainties of EOS and mass.

The exceptional models are those with 1.36 M⊙M_\odot stars, where cooling curves are temporarily located in higher-luminosity regions than the sensitivity curves by a Lynx-like observatory. In this case, NS 1987A is detectable for some years, like a shooting star2. This is because the mass is close to the threshold mass MDUM_{\rm DU}, implying that the volume where the DU process is active is not large enough to significantly affect surface luminosity (e.g., Yakovlev & Pethick [61]). Although how close they are strongly depends on EOS ([30, 49]), such a ”weak” DU process may realize the detection of NS 1987A.

Figure 2 presents the impact of crust superfluidity on the detection of NS 1987A, assuming 1.40 M⊙M_\odot stars. As already reported by Page et al. [42], the crust superfluidity reduces thermal luminosity of young NSs because of the suppression of the neutron specific heat (Page et al. 2009). This becomes more significant with the Y+K compared to the APR, which allows the DU process to occur; one can see that stronger crust superfluidity leads to earlier rapid cooling. Except for the case without crust superfluidity, one can see no intersection between the cooling curve and the sensitivity curve by a Lynx-like observatory, showing the non-detection of NS 1987A with the DU process.

Figure 2. Same as Figure 1, but neutron 1S0^1S_0 superfluid-model dependence of cooling curves assuming 1.40 M⊙M_\odot stars.

While the APR EOS realizes the detection in all parameter regions we considered, the Y+K EOS does not always do so due to the presence of the DU process. Figure 3 displays how the detectability of the NS 1987A changes with the Y+K EOS. If the NS mass is lower than the critical mass of the DU process, all models show the detection, as with the APR, but for MNS≳MDUM_{\rm NS} \gtrsim M_{\rm DU}, the detectability depends on the crustal superfluidity. In general, a stronger crust superfluid model tends to make the DU cooling more efficient due to the reduction of neutron specific heat, finally shortening the thermal relaxation timescale (Gnedin et al. [17]). One may think that tobst_{\rm obs} is smaller with MNS/M⊙M_{\rm NS}/M_\odot=1.35 and 1.36 compared with that with MNS/M⊙M_{\rm NS}/M_\odot=1.31–1.34 with weak superfluid models (CCDK, SFB, and WAP). This seems to be attributed to the difference in crust thickness, which is connected to the thermal relaxation timescale. In fact, the crust thickness is longer for 1.35 M⊙M_\odot stars by 1.1% than that for 1.34 M⊙M_\odot stars.

Figure 3. NS mass, η\eta, and crust superfluid model dependences of the detection or non-detection of NS 1987A with the Y+K EOS: ● denotes the detection while × denotes non-detection within 88 years after the birth. △ denotes the shooting-star scenario. The contour indicates the age when NS 1987A is detected by a Lynx-like observatory. The dashed line corresponds to the critical mass of the DU process with the Y+K EOS, MDU=1.31 M⊙M_{\rm DU}=1.31\,M_\odot.

If MNS−MDUM_{\rm NS}-M_{\rm DU} becomes larger, due to the large-volume core of DU cooling, it becomes more impossible to detect NS 1987A due to rapid cooling before the putative detectable age (see also Figure 1). The higher envelope mass also makes detection more possible, but except for mass close to MDUM_{\rm DU}, it has little impact on the criterion of detection and non-detection unless we consider a small envelope mass enough not to account for the ALMA observations.

Detection age of NS 1987A v.s. Relaxation time

As we see the above, by comparing cooling curves with sensitivity limits by a Lynx-like observatory that consider the X-ray absorption due to supernova ejecta, one can know the detection age of NS 1987A, tobst_{\rm obs}. Also, we calculate the effective thermal relaxation timescale trelaxt_{\rm relax}, defined as the age when the cooling rate ∣dln⁡Lγ∞dln⁡t∣\left|\frac{d\ln L_\gamma^\infty}{d\ln t}\right| is maximum for 1<t<2001<t<200 yr, which mimics when isolated NSs become isothermal (Lattimer et al. [28]). Most isolated NSs observed are thought to be isothermal, i.e., tobs>trelaxt_{\rm obs}>t_{\rm relax}, such as the case of Cassiopeia A (e.g., Yakovlev et al. [60]), but except for our target young neutron star, NS 1987A: If tobs<trelaxt_{\rm obs}<t_{\rm relax}, the inner crust should have hotter than the core, implying that the neutrino cooling in the inner crust plays the most important role on the surface luminosity. In other words, NS 1987A is a potentially unique target for probing the properties of inner-crust matter using future luminosity observations.

Figure 4 presents EOS and mass dependence of tobst_{\rm obs} and trelaxt_{\rm relax} with the CCDK of crust superfluidity, which implies tobs∼60t_{\rm obs}\sim60 yr except for Y+K EOS with MNS≳1.40 M⊙M_{\rm NS}\gtrsim1.40\,M_\odot, where the DU process operates. The behavior of trelaxt_{\rm relax} is complicated by superfluidity [17], but, except for 1.20M⊙M_\odot stars, higher-mass models tend to yield smaller trelaxt_{\rm relax} (this feature can also be seen with APR). This derives from slightly early cooling around 100 yr even with the slow cooling scenario (e.g., Yakovlev & Pethick [61]; Lim et al. [30]; Dohi et al. [12]).

Figure 4. Mass dependence of observation time and relaxation time with APR and Y+K EOSs, assuming neutron 1S0{}^{1}S_{0} superfluid model as the CCDK. The observational ages with Y+K EOS for 1.40, 1.50 and 1.60M⊙M_\odot are tobs≈202t_{\rm obs}\approx202 yr regardless of η\eta values, respectively.

One can also see two cases of tobs<trelaxt_{\rm obs}<t_{\rm relax} and tobs>trelaxt_{\rm obs}>t_{\rm relax}, which imply that NS 1987A is not isothermal (where the crust is hotter than the core) and isothermal at the detection age, respectively. For instance, if the DU process occurs (With Y+K EOS and MNS≳1.40 M⊙M_{\rm NS}\gtrsim1.40\,M_\odot), it corresponds to the latter case. On the other hand, the APR models with MNS=1.30M_{\rm NS}=1.30–1.50 M⊙1.50\,M_\odot show the former case. Therefore, future luminosity observations of NS 1987A by a Lynx-like observatory can probe the physical properties in either the inner crust or the core, respectively, depending on model parameters.

The most important ingredient on trelaxt_{\rm relax} under the minimal cooling scenario is the crust superfluidity, as shown in Figure 5: The stronger the crust superfluidity is, the shorter trelaxt_{\rm relax} is. Meanwhile, tobst_{\rm obs} does not drastically increase by crust superfluid models. Therefore, if the crust superfluidity is weak, such as the CCDK model, one can show trelax>tobst_{\rm relax}>t_{\rm obs}, implying that NS 1987A is a potential site to probe crustal properties.

Figure 5. Neutron 1S0{}^{1}S_{0} Superfluid dependence of observation time and relaxation time with APR EOS at 1.4M⊙M_\odot.

Discussion and Conclusion

We investigated how the thermal evolution of the putative NS 1987A connects future luminosity measurements to its interior physics. For the adopted Lynx-like sensitivity and observing window, rapid neutrino cooling, such as the DU process, generally disfavors detection because it reduces the surface luminosity before the thermal emission becomes accessible. However, for a mass close to the threshold for rapid cooling, the star can remain detectable for a short interval, as in the shooting-star scenario. These are conditional observational diagnostics, rather than predictions for the launch of a specific telescope.

In the standard cooling scenario, NS 1987A can be detected at t∼60t \sim60 yr if a facility reaching the adopted Lynx-like sensitivity is operating at that epoch. This is because (i) the cooling rate is smaller than the decrease of Lynx-like sensitivity limit, assuming that SNR 1987A is freely expanding, and (ii) ALMA observations by Cigan et al. [10] imply high-mass envelope with η≳10−8\eta\gtrsim10^{-8} at t≃28t \simeq28 yr, resulting in high thermal luminosity enough to enable the future detection of NS 1987A.

Table 2 compares representative X-ray facilities with the adopted benchmark. Angular resolution is essential as well as effective area: the bright equatorial ring is only about 1.6 arcsec across (Orlando et al. [39]), so subarcsecond imaging is needed to reduce its contamination in an extraction region around the compact object. This is not a universal sharp resolution threshold, since the point-spread-function wings, source position, and remnant brightness also matter. Chandra provides such imaging but has a much smaller effective area. AXIS would have increased the count rate, but its broader point-spread function would have mixed more remnant emission into the central region; moreover, it became ineligible for the Probe selection in March 2026 (Astronomy and Astrophysics Advisory Committee [3]), and NASA subsequently selected the far-infrared mission PRIMA (NASA Jet Propulsion Laboratory [37]). NewAthena offers a large effective area, but its ∼9\sim9 arcsec imaging cannot spatially separate the central source from the ring. Nevertheless, among the available satellites in the 2040s, NewAthena would be the most promising candidate to detect thermal emission from NS 1987A. Its X-IFU spectroscopy could constrain the shocked plasma and help search for an additional central component in the integrated spectrum, although isolating thermal NS emission would remain model dependent. HUBS, designed for diffuse soft X-ray spectroscopy, likewise cannot spatially isolate the central source. Thus, these alternatives offer complementary constraints but do not provide a demonstrated replacement for the benchmark thermal detection.

FacilityAngular resolutionArea at 1 keV (cm2^2)Launch / status
Chandra/ACIS-S<0.5′′< 0.5^{\prime\prime} (mirror FWHM)∼555\sim555 (nominal, BI)1999; operating
Lynx-like benchmark0.5′′0.5^{\prime\prime} (HPD)∼20,000\sim20{,}000 (mirror)No scheduled launch; unfunded concept
AXIS concept1.5′′1.5^{\prime\prime} (on-axis HPD)∼3,900\sim3{,}900 (incl. detector)Ineligible in 2026
NewAthena<9′′< 9^{\prime\prime} (on-axis HEW)>11,000> 11{,}000 (mirror req.)2039 target; adoption pending
HUBS concept<60′′< 60^{\prime\prime} (HPD)∼500\sim500 (FoV-averaged)Proposed; launch not confirmed

Table 2. Representative X-ray capabilities and mission status (October 2026).

In particular, evidence for a nonthermal PWN component, as in Chandra–NuSTAR studies (Greco et al. [19] (2022)), is distinct from the surface thermal luminosity needed for our cooling diagnostic.

Our cooling curves therefore remain useful even if Lynx is never launched: a measured thermal luminosity and age, compared with the relaxation timescale, can distinguish an interior that has thermally relaxed from a hotter crust and test crustal superfluidity within the model uncertainties discussed here. Whether a particular facility can supply that luminosity must be assessed with its full energy-dependent response, exposure, ejecta absorption, and remnant contamination. A nondetection is informative only when its thermal-luminosity upper limit reaches the relevant cooling models. We do not transfer the numerical detection ages to NewAthena, AXIS, or HUBS by scaling their areas; such forecasts would require new spectral simulations. The benchmark therefore provides a reference for assessing future thermal measurements against the cooling models.

We also discussed the implications when NS 1987A is detected. It is found that NS 1987A is likely to be isothermal at the detection, unless the neutron superfluidity in the crust is weak. This implies that the observed luminosity and age of NS 1987A can extract the information on the NS core, but not the inner crust, except for the CCDK model of 1S0{}^{1}S_{0} neutron superfluidity.

The 1S0{}^{1}S_{0} neutron superfluidity has been discussed in terms of condensed-matter physics (for reviews, see Sedrakian & Clark [51]; Ohashi et al. [38]). Under the standard mean-field (BCS) theory, TcT_{\mathrm{c}} value is around 20 GK (Tajima et al. [53]), which is also supported by the lattice Monte-Carlo calculation Abe & Seki [1]. This implies that NS 1987A could not be a probe for the properties inside the inner crust. However, if the beyond mean-field effects, such as the pairing fluctuation and many-body effects, are considered, the efficiency of cooling inside the inner crust will be changed, such as the enhancement of specific heat (Inotani et al. [25]), and reduction of TcT_{\mathrm{c}} due to the so-called Gork’ov and Melik-Barkhudarov (GMB) effects (Gor’kov & Melik-Barkhudarov [18]). To be slightly optimistic, investigating more sophisticated neutron superfluid models in the inner crust beyond BCS theory may make the detection of NS 1987A lead to a unique probe of crustal matter.

The present work serves as a general approach to specify the physics of NS 1987A from future detection of thermal emission. Although the physical link between cooling, relaxation, and luminosity does not depend on Note. HPD and HEW denote the diameter enclosing half the energy; FWHM is a different measure. Areas are published reference values, not a uniform set of detector-folded, epoch-specific responses. The Chandra value is the nominal ACIS-S back-illuminated (BI) value, not its 2026 area; contamination reduces its soft response with time. Lynx and NewAthena values refer to mirror areas, before detector/filter losses. The AXIS value is the updated 2025 on-axis baseline, superseding the 4,200 cm2^2 value on the mission website. NewAthena entries are optical design requirements (req.); ideal-optics ray tracing predicts ∼1.24\sim1.24 m2^2 at 1 keV, about 62% of the Lynx mirror benchmark, rather than a detector-inclusive area or a sensitivity ratio. Future dates are targets, not guaranteed launches. The current English HUBS roadmap lists pathfinder and engineering development without a launch date. Sources: Chandra X-ray Center (Chandra X-ray Center [8]); NASA’s Lynx concept reassessment (Lynx2030 Science Analysis Group et al. [31]); the AXIS camera team (Miller et al. [33]), the AAAC status report (Astronomy and Astrophysics Advisory Committee [3]), and NASA’s Probe selection (NASA Jet Propulsion Laboratory [37]); NewAthena optics studies (’S Jegers et al. 2024; INAF Brera Astronomical Observatory [24]) and ESA’s current launch target (European Space Agency [13]); HUBS mission description (HUBS Collaboration [23]).

the observing facility, the detection ages are conditional on our benchmark; it should be mentioned about the following two issues: One is the limitation of the number of cooling models, in particular with the fast cooling scenario. Then, it would be worth doing more systematic calculations to judge the detectability, such as Figure 3 with different EOSs (threshold mass for the onset of the DU process) and stellar mass. Another issue is to assume Eq. (1) for t>50t > 50 yr, whose approximation becomes worse as NS 1987A gets older due to an increase in the swept mass from the circumstellar medium. Our imposed approximation is somewhat conservative, as shown in Table 1. Fortunately, this means that the detection would be achieved more easily. Access to sensitive X-ray observations during the thermal-relaxation epoch would therefore be particularly valuable, irrespective of the mission that provides them.

To determine the detection age of NS 1987A more accurately, systematic calculations of sensitivity limits with realistic ejecta profiles are necessary in the future. However, they take a huge computational cost due to time-to-time spectrum synthesis with 3D MHD models (Orlando et al. [40]). Moreover, the morphology described by 3D MHD models would still be affected by the large uncertainty on the structure and composition of the outer ambient medium.

Let us briefly mention the putative impact of the internal magnetic field which we did not consider. If the dipole magnetic field at the pole is as high as ∼3×1015\sim3 \times10^{15} G, the Joule heating increases the thermal luminosity of NS 1987A to Lγ∞∼5×1035L_{\gamma}^{\infty} \sim5 \times10^{35} erg s−1^{-1} (Viganò et al. [57]), making the detection of NS 1987A earlier. Because the free precession period that slightly differs from the rotational one can be measured in magnetars enclosed in SNR (Makishima et al. [32]), one can judge whether NS 1987A is a magnetar or not before future deep thermal X-ray observations3. If NS 1987A turns out to be a magnetar, our scenario to detect thermal emission from NS 1987A would become more testable.

Acknowledgements

We thank T. Muto for providing his EOS data. We also thank D. Patnaude and P. Slane for helpful discussion on Lynx, and R. Hirai on SNR ejecta. This work is supported by JSPS KAKENHI (JP23K25874, JP25H00675, JP25K17403, and JP26K17149). S.N. is also supported by the JST ASPIRE Program ”RIKEN-Berkeley mathematical quantum science initiative”.

Software: Gnuplot (Williams & Kelley [59]).

On the impact of proton superconductivity on NS 87A

Around the age of NS 1987A, the proton superconductivity should have already been realized in the core. As time passes, thermal relaxation proceeds, and neutrino cooling in the core eventually becomes dominant. Here, we discuss the role of proton superconductivity on cooling curves of young NSs.

Figure 6 shows the cooling curves with and without CCDK proton superconductivity, fixing the Y+K EOS with 1.6 M⊙1.6\ M_{\odot}. In the fast cooling case where the DU process occurs, the proton superconductivity makes the cooling timescale shorter by ∼50\sim50 yr. In the minimal cooling case (i.e., when the DU process is artificially switched off), the behavior of cooling curves is similar, but the detection age by Lynx-like observatory is changed by ∼10\sim10 yr. Considering that the CCDK is a relatively strong model ([29] (Ho et al. 2015)), this 10-yr difference may be a rough upper estimate of the uncertainty in the detection age associated with proton superconductivity, at least under a minimal cooling scenario.

The reason why proton superconductivity affects cooling curves of young NS is mainly due to the change of electron thermal conductivity in the core, which is given by κe=π2kB2Tne3me∗νe\kappa_e = \frac{\pi^2 k_B^2 T n_e}{3m_e^*\nu_e}, where kBk_B is the Boltzmann constant, nen_e is the electron number density, and me∗m_e^* is the effective electron mass at the Fermi surface. νe\nu_e denotes the effective collision frequency, and neglecting the contribution from muons, it is expressed as νe≃νee+νep\nu_e \simeq\nu_{ee}+\nu_{ep}, where νee\nu_{ee} and νep\nu_{ep} denote the contributions from electron-electron and electron-proton scattering, respectively. In the presence of proton superconductivity, νep\nu_{ep} is suppressed due to fewer protons available to scatter. νee\nu_{ee} is also suppressed by so-called Meissner screening ([52, 4]). Both effects therefore increase κe\kappa_e, which shortens the cooling timescale as expected in Figure 6. We note that how the cooling timescale is changed by proton superconductivity depends nontrivially on NS mass, since it is proportional to the crust thickness squared and inversely proportional to the relativistic factor cubed ([28, 17]). Generally, both the crust thickness and the relativistic factor decrease as the NS is heavier. Moreover, Sales et al. [49] shows that, in the NS mass regions just above MDUM_{\mathrm{DU}}, the thermal relation becomes much slower under non-superfluid cooling models.

References

  1. [1]Abe, T., & Seki, R. 2009, PhRvC, 79, 054002, doi: 10.1103/PhysRevC.79.054002DOI
  2. [2]Akmal, A., Pandharipande, V. R., & Ravenhall, D. G. 1998, PhRvC, 58, 1804, doi: 10.1103/PhysRevC.58.1804DOI
  3. [3]Astronomy and Astrophysics Advisory Committee. 2026, 2026 Annual Report of the Astronomy and Astrophysics Advisory Committee, https://nsf-gov-resources.nsf.gov/files/2026- AAAC-Annual-Report.pdf
  4. [4]Bertoni, B., Reddy, S., & Krapaj, E. 2015, PhRvC, 91, 025806, doi: 10.1103/PhysRevC.91.025806DOI
  5. [5]Bhat, B., Dohi, A., Muto, T., & Noda, T. 2026, arXiv e-prints, arXiv:2605.09723, doi: 10.48550/arXiv.2605.09723DOI
  6. [6]Bionta, R. M., Blewitt, G., Bratton, C. B., et al. 1987, PhRvL, 58, 1494, doi: 10.1103/PhysRevLett.58.1494DOI
  7. [7]Burgio, G. F., Schulze, H. J., Vidaña, I., & Wei, J. B. 2021, PPNP, 120, 103879, doi: 10.1016/j.ppnp.2021.103879DOI
  8. [8]Chandra X-ray Center. 2026, Chandra Pocket Guide, https://cxc.harvard.edu/cdo/pocket_guide.pdf
  9. [9]Chen, J. M. C., Clark, J. W., Davé, R. D., & Khodel, V. V. 1993, NuPhA, 555, 59, doi: 10.1016/0375-9474(93)90314-NDOI
  10. [10]Cigan, P., Matsuura, M., Gomez, H. L., et al. 2019, ApJ, 886, 51, doi: 10.3847/1538-4357/ab4b46DOI
  11. [11]Dohi, A., Greco, E., Nagataki, S., et al. 2023, ApJ, 949, 97, doi: 10.3847/1538-4357/accef3
  12. [12]Dohi, A., Nakazato, K., Hashimoto, M.-a., Yasuhide, M., & Noda, T. 2019, PTEP, 2019, 113E01, doi: 10.1093/ptep/ptz116DOI
  13. [13]European Space Agency. 2026, NewAthena Factsheet, https://www.esa.int/Science_Exploration/Space_Science/NewAthena_factsheet
  14. [14]Fransson, C., Barlow, M. J., Kavanagh, P. J., et al. 2024, Science, 383, 898, doi: 10.1126/science.adj5796DOI
  15. [15]Gandolfi, S., Illarionov, A. Y., Fantoni, S., Pederiva, F., & Schmidt, K. E. 2008, PhRvL, 101, 132501, doi: 10.1103/PhysRevLett.101.132501DOI
  16. [16]Gezerlis, A., & Carlson, J. 2008, PhRvC, 77, 032801, doi: 10.1103/PhysRevC.77.032801DOI
  17. [17]Gnedin, O. Y., Yakovlev, D. G., & Potekhin, A. Y. 2001, MNRAS, 324, 725, doi: 10.1046/j.1365-8711.2001.04359.xDOI
  18. [18]Gor’kov, L., & Melik-Barkhudarov, T. 1961, Sov. Phys. JETP, 13, 1018
  19. [19]Greco, E., Miceli, M., Orlando, S., et al. 2021, ApJL, 908, L45, doi: 10.3847/2041-8213/abdf5aDOI
  20. [21]Hirata, K., Kajita, T., Koshiba, M., et al. 1987, PhRvL, 58, 1490, doi: 10.1103/PhysRevLett.58.1490DOI
  21. [23]HUBS Collaboration. 2026, Hot Universe Baryon Surveyor: Mission Introduction, https://hubs-mission.cn/en/introduction.html
  22. [24]INAF Brera Astronomical Observatory. 2026, BEaTriX: The Calibration Challenge for NewAthena, https://brera.inaf.it/pagine/beatrix-facility/
  23. [25]Inotani, D., Yasui, S., & Nitta, M. 2020, PhRvC, 102, 065802, doi: 10.1103/PhysRevC.102.065802DOI
  24. [26]Larsson, J., Fransson, C., Kavanagh, P. J., et al. 2025, ApJ, 991, 130, doi: 10.3847/1538-4357/adf741DOI
  25. [27]Lattimer, J. M., Pethick, C. J., Prakash, M., & Haensel, P. 1991, PhRvL, 66, 2701, doi: 10.1103/PhysRevLett.66.2701DOI
  26. [28]Lattimer, J. M., van Riper, K. A., Prakash, M., & Prakash, M. 1994, ApJ, 425, 802, doi: 10.1086/174025DOI
  27. [29]Lim, Y., & Holt, J. W. 2021, PhRvC, 103, 025807, doi: 10.1103/PhysRevC.103.025807DOI
  28. [30]Lim, Y., Hyun, C. H., & Lee, C.-H. 2017, Int. J. Mod. Phys. E, 26, 1750015, doi: 10.1142/S021830131750015XDOI
  29. [31]Lynx2030 Science Analysis Group, Bandler, S. R., Brenneman, L. W., et al. 2026, arXiv e-prints, arXiv:2609.00033, doi: 10.48550/arXiv.2609.00033DOI
  30. [32]Makishima, K., Uchida, N., & Enoto, T. 2026, PASJ, doi: 10.1093/pasj/psag011DOI
  31. [33]Miller, E. D., Grant, C. E., Goeke, R., et al. 2025, arXiv e-prints, doi: 10.48550/arXiv.2508.14157
  32. [34]Muto, T. 2025, PhRvC, 111, 045802, doi: 10.1103/PhysRevC.111.045802DOI
  33. [35]Muto, T., Maruyama, T., & Tatsumi, T. 2021, Phys. Lett. B, 820, 136587, doi: 10.1016/j.physletb.2021.136587DOI
  34. [37]NASA Jet Propulsion Laboratory. 2026, NASA Selects Far-Infrared Telescope as First in New Mission Class, https://www.jpl.nasa.gov/news/nasa-selects-far-infrared-telescope-as-first-in-new-mission-class/
  35. [38]Ohashi, Y., Tajima, H., & van Wyk, P. 2020, PPNP, 111, 103739, doi: 10.1016/j.ppnp.2019.103739DOI
  36. [39]Orlando, S., Ono, M., Nagataki, S., et al. 2020, A&A, 636, A22, doi: 10.1051/0004-6361/201936718DOI
  37. [40]Orlando, S., Miceli, M., Ono, M., et al. 2025, A&A, 699, A305, doi: 10.1051/0004-6361/202554862DOI
  38. [41]Page, D. 2016, NSCool: Neutron star cooling code. http://ascl.net/1609.009
  39. [42]Page, D., Beznogov, M. V., Garibay, I., et al. 2020, ApJ, 898, 125, doi: 10.3847/1538-4357/ab93c2DOI
  40. [43]Page, D., Geppert, U., & Weber, F. 2006, NuPhA, 777, 497, doi: 10.1016/j.nuclphysa.2005.09.019DOI
  41. [44]Page, D., Lattimer, J. M., Prakash, M., & Steiner, A. W. 2004, ApJS, 155, 623, doi: 10.1086/424844DOI
  42. [46]Page, D., Prakash, M., Lattimer, J. M., & Steiner, A. W. 2011, PhRvL, 106, 081101, doi: 10.1103/PhysRevLett.106.081101DOI
  43. [47]Potekhin, A. Y., Chabrier, G., & Yakovlev, D. G. 1997, A&A, 323, 415. https://arxiv.org/abs/astro-ph/9706148
  44. [49]Sales, T., Lourenço, O., Dutra, M., & Negreiros, R. 2020, A&A, 642, A42, doi: 10.1051/0004-6361/202038193DOI
  45. [50]Schwenk, A., Friman, B., & Brown, G. E. 2003, NuPhA, 713, 191, doi: 10.1016/S0375-9474(02)01290-3DOI
  46. [51]Sedrakian, A., & Clark, J. W. 2019, Eur. Phys. J. A, 55, 167, doi: 10.1140/epja/i2019-12863-6DOI
  47. [52]Shternin, P. S., & Yakovlev, D. G. 2007, PhRvD, 75, 103004, doi: 10.1103/PhysRevD.75.103004DOI
  48. [53]Tajima, H., Hatsuda, T., van Wyk, P., & Ohashi, Y. 2019, Scientific Reports, 9, 18477, doi: 10.1038/s41598-019-54010-7DOI
  49. [54]Takatsuka, T. 1984, PTP, 71, 1432, doi: 10.1143/PTP.71.1432DOI
  50. [55]Tegkelidis, C., Larsson, J., & Alp, D. 2025, A&A, 704, A272, doi: 10.1051/0004-6361/202555479DOI
  51. [56]Utrobin, V. P., Wongwathanarat, A., Janka, H. T., et al. 2019, A&A, 624, A116, doi: 10.1051/0004-6361/201834976DOI
  52. [57]Viganò, D., Rea, N., Pons, J. A., et al. 2013, MNRAS, 434, 123, doi: 10.1093/mnras/stt1008DOI
  53. [58]Wambach, J., Ainsworth, T. L., & Pines, D. 1993, NuPhA, 555, 128, doi: 10.1016/0375-9474(93)90317-QDOI
  54. [59]Williams, T., & Kelley, C. 2022, Gunplot 5: an interactive plotting program54, v5.4
  55. [60]Yakovlev, D. G., Ho, W. C. G., Shternin, P. S., Heinke, C. O., & Potekhin, A. Y. 2011, MNRAS, 411, 1977, doi: 10.1111/j.1365-2966.2010.17827.xDOI
  56. [61]Yakovlev, D. G., & Pethick, C. J. 2004, ARA&A, 42, 169, doi: 10.1146/annurev.astro.42.053102.134013DOI

Paper details

vlnc:W_8gaj5434
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

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