Indication of a Pulsar Wind Nebula in the hard X-ray emission from SN 1987A
Emanuele Greco, Marco Miceli, Salvatore Orlando, Barbara Olmi, Fabrizio Bocchino, Shigehiro Nagataki, Masaomi Ono, Akira Dohi, Giovanni Peres
DDraft version January 25, 2021
Typeset using L A TEX twocolumn style in AASTeX63
Indication of a Pulsar Wind Nebula in the hard X-ray emission from SN 1987A
Emanuele Greco,
1, 2
Marco Miceli,
1, 2
Salvatore Orlando, Barbara Olmi, Fabrizio Bocchino, Shigehiro Nagataki,
3, 4
Masaomi Ono,
3, 4
Akira Dohi, and Giovanni peres
1, 2 Universit`a degli Studi di Palermo, Dipartimento di Fisica e Chimica, Piazza del Parlamento 1, 90134 Palermo, Italy INAF-Osservatorio Astronomico di Palermo, Piazza del Parlamento 1, 90134 Palermo, Italy Astrophysical Big Bang Laboratory (ABBL), RIKEN Cluster for Pioneering Research, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan RIKEN Interdisciplinary Theoretical and Mathematical Science Program (iTHEMS), 2-1 Hirosawa, Wako, Saitama 351-0198, Japan Department of Physics, Kyushu University, 744 Motooka, Nishi-Ku, Fukuoka Fukuoka 819-0395, Japan
ABSTRACTSince the day of its explosion, SN 1987A (SN87A) was closely monitored with the aim to studyits evolution and to detect its central compact relic. The detection of neutrinos from the supernovastrongly supports the formation of a neutron star (NS). However, the constant and fruitless search forthis object has led to different hypotheses on its nature. Up to date, the detection in the ALMA dataof a feature somehow compatible with the emission arising from a proto Pulsar Wind Nebula (PWN) isthe only hint of the existence of such elusive compact object. Here we tackle this 33-years old issue byanalyzing archived observations of SN87A performed by
Chandra and
NuSTAR in different years. Wefirmly detect nonthermal emission in the 10 −
20 kev energy band, due to synchrotron radiation. Thepossible physical mechanism powering such emission is twofold: diffusive shock acceleration (DSA)or emission arising from an absorbed PWN. By relating a state-of-the-art magneto-hydrodynamicsimulation of SN87A to the actual data, we reconstruct the absorption pattern of the PWN embeddedin the remnant and surrounded by cold ejecta. We found that, even though the DSA scenario cannotbe firmly excluded, the most likely scenario that well explains the data is the PWN emission.
Keywords:
X-rays: general - supernovae: individual (SN 1987A) INTRODUCTIONSN 1987A (SN87A) in the Large Magellanic Cloud(LMC) was a hydrogen-rich core-collapse supernova(SN) discovered on 1987 February 23 (West et al. 1987).It occurred approximately 51.4 kpc from Earth (Pana-gia 1999) and its dynamical evolution is strictly relatedto the very inhomogenous circumstellar medium (CSM),composed by a dense ring-like structure within a diffuseHII region (Sugerman et al. 2005). SN87A is the firstnaked-eye SN exploded since telescopes exist and its evo-lution has been deeply monitored in various wavelengths(McCray 1993; McCray & Fransson 2016). In particu-lar, the X-ray band is ideal to investigate the interactionof the shock front with the CSM and the emission of theexpected central compact leftover of the supernova ex-plosion.
Corresponding author: Emanuele [email protected]
Despite the unique consideration granted with deepand continuous observations, and the neutrinos detec-tion (Bionta et al. 1987) strongly indicating the forma-tion of a neutron star (Vissani 2015), the elusive com-pact object of SN87A is still undetected. The most likelyexplanation for this non-detection is ascribable to theabsorption due to ejecta, i.e. the dense and cold ma-terial ejected by the supernova (Fransson & Chevalier1987): because of the young age of SN87A, the ejectaare still very dense and the reverse shock generated inthe outer shells of the supernova remnant (SNR) hasnot heated the inner ejecta yet. Thus, photo-electricabsorption from this metal-rich material can hide theX-ray emission of a hypothetical compact object. Dur-ing the last years, many works (Orlando et al. 2015;Alp et al. 2018; Esposito et al. 2018; Page et al. 2020)investigated the upper limit on the luminosity of theputative compact leftover in various wavelengths, con-sidering the case of a neutron star (NS) emitting thermal(black-body) radiation and obscured by the cold ejectaand/or dust. However, the lack of strong constraints on a r X i v : . [ a s t r o - ph . H E ] J a n Greco et al. the absorption pattern in the internal area of SN87Aprevented either to further constrain the luminosity ofthe putative NS or to make predictions about its futuredetectability.On the other side, the X-ray emission from a young NSmay include a significant nonthermal component: thesynchrotron radiation arising from the pulsar wind neb-ula (PWN) associated with the rotating NS. Recently,
ALMA images showed a blob structure whose emission issomehow compatible with the radio emission of a PWN(Cigan et al. 2019). However, the authors themselveswarned that this blob could be associated with otherphysical processes, e.g. heating due to Ti decay. It isnatural, then, to look for the high-energy counterpart ofthe synchrotron radiation in the X-ray band.In this letter, we report on the analysis of observationsof SN87A performed between 2012 and 2014 by
Chan-dra and
NuSTAR . We also take advantage of the state-of-the-art MHD simulation from Orlando et al. (2020)(hereafter Or20) to reconstruct the absorption patternwithin SN87A, and link it to the observed spectra. Weprovide a single model describing the emission in theseyears from 0.5 to 20 keV, isolating the synchrotron ra-diation arising from the remnant. We then discuss thepossible origin of such emission and show how the pres-ence of a PWN appears to be the most likely scenario. X-RAY DATA ANALYSISWe used data collected in 2012, 2013 and 2014with
Chandra
ACIS-S and
NuSTAR
CZT (FPMA andFPMB). We reprocessed
Chandra and
NuSTAR datawith the standard pipelines available within CIAOv4.12.2 and NuSTARDAS, respectively. For details onthe observations and the data reduction see AppendixA.We extracted spectra from a circular region centeredat α = 5 h m s and δ = − ◦ (cid:48) (cid:48)(cid:48) with a radius of2 (cid:48)(cid:48) and 43 (cid:48)(cid:48) for Chandra and
NuSTAR data, respectively.These regions enclose all the remnant and were chosentaking into account the PSFs of the different telescopes(see Fig. 1).We simultaneously analyzed
Chandra and
NuSTAR spectra for each year considered by adopting a modelcomposed by a galactic absorption component (
TBabs model in XSPEC), two optically thin isothermal com-ponents in non-equilibrium of ionization ( vnei model)and a constant factor which takes into account cross-calibration between different detectors. The columndensity n H is fixed to 2 . × cm − (Park et al. 2006).We found cross-calibration factors <
2% within different
Chandra data sets of the same year, and <
8% between
Chandra and
NuSTAR , compatible with the character- istic corrections between
NuSTAR detectors and othertelescopes. Temperatures, emission measure, ionizationage and abundances of O, Ne, Mg, Si and S were leftfree to vary in the fitting process. All other abundanceswere kept fixed to those found by Zhekov et al. (2009).We report no significant variations in the chemical abun-dances in the time range considered (2012 to 2014).In agreement with previous works (Orlando et al.2015; Miceli et al. 2019, Or20), we found that the soft(0 . − kT and ionization parameters, τ ,are compatible with previous measures (Zhekov et al.2009). However, we found strong residuals in all NuS-TAR spectra at energies >
10 keV, clearly showing thatan additional component must be added to the model toproperly describe the hard X-ray emission (see Reynoldset al. 2015). The best-fit model and the residuals areshown in the left panel of Fig. 2. This additional com-ponent cannot be associated with thermal emission fromthe shocked plasma, since otherwise unrealistically hightemperatures would be necessary ( kT ∼
20 keV), oneorder of magnitude higher than the maximum electrontemperature predicted for SN 1987A (Orlando et al.2015) and never observed in SNRs.If the emission is unlikely to be thermal, it is plausi-ble to suppose that it is nonthermal. This nonthermalemission may arise from a possible compact object, mostlikely a NS embedded in its PWN (hereafter PWN87A),which emits synchrotron radiation. The main issue intackling this scenario is to isolate the radiation comingfrom this object.We can estimate the absorbing power of cold ejectasurrounding the PWN by taking into account the abun-dance and density pattern along the line of sight pro-vided by the MHD model by Or20 (see Appendix B) inthe considered years. To include this information in thespectral analysis, we added an absorbed power-law mod-eled with vphabs within XSPEC. Since its parameterswere derived from the model they are not free to varyin the fit. With this additional component, we obtaineda very good description of the observed spectra over thewhole 0 . −
20 keV energy band (right panel of Fig. 2).The best-fit value of photon index and normalizationof the absorbed power-law do not change significantlybetween the three years considered. We then fitted si-multaneously the 2012, 2013 and 2014 spectra in orderto decrease the uncertainties in the best-fit parametersof the power-law component. Temperatures, ionizationparameters and normalization of thermal emission werefree to change in time, while normalization and pho-ton index of nonthermal emission (and chemical abun- ndication of a pulsar wind nebula in SN 1987A Figure 1.
Chandra and
NuSTAR count images of SN87A. All the images are smoothed with a 1.5 σ -gaussian. Left panel . Chandra image in the 0.1-8 keV. The cyan and the red circular regions mark the 2 (cid:48)(cid:48) region used to extract the source and thecorresponding background spectra for the
Chandra data, respectively.
Central panel . Closeup of the left panel. The 0 . (cid:48)(cid:48) radiusblack ring marks the faint central region of SN87A. Right panel. NuSTAR image in the 3-30 keV band. The cyan and red regionsidentify the 43 (cid:48)(cid:48) region used to extract the source spectra and the corresponding background for the
NuSTAR data, respectively.A navigable 3D graphic showing the PWN position in the remnant interior and the comparison with the observations is availableat the link https://skfb.ly/6XZIU. − − − no r m a li z ed c oun t s s − k e V − Chandra+NuSTAR two vnei − all years1 102 5 − − ( da t a − m ode l ) / e rr o r Energy (keV) − − − no r m a li z ed c oun t s s − k e V − Chandra+NuSTAR PWN scenario − all years1 102 5 − − ( da t a − m ode l ) / e rr o r Energy (keV)
Figure 2.
Spectra extracted in the 0 . −
20 keV energy band from
Chandra and
NuSTAR in various years with the correspondingbest-fit model and residuals. A different color is associated with each of the twenty data set. On the left, the best-fit modelis composed by two thermal components. On the right, the best-fit model also takes into account the emission coming from aheavily absorbed PWN. The spectra have been rebinned for presentation purposes. dances) were left free to vary in the fitting procedure,but forced to be the same over the three years.The resulting PWN best-fit photon index is Γ =2 . +0 . − . and the X-ray luminosity in the 1 −
10 keV bandis L − = (2 . ± . × erg/s (Table 1). Our bestfit values are compatible with typical values found forPWNe (Sect. 3).We found an increase in flux in the 8 −
20 keV bandbetween 2012 and 2014 (as reported by Reynolds et al.2015). We point out that the emission from the PWNis consistent with being constant, and that the increasein the hard X-ray flux is related to the steadily grow-ing thermal emission arising from the interaction of theremnant with the HII region.Because of the high ejecta absorption, the PWN fluxis strongly suppressed below 8 keV. In particular, we found that the PWN emission is much dimmer ( ∼ Chandra in thefaint central region of SN87A (identified by a circularregion with radius R = 0 . (cid:48)(cid:48) , well inside the bright X-ray emitting ring, see central panel in Fig. 1).As an alternative scenario, we also considered the casein which the hard X-ray emission is associated with syn-chrotron radiation due to DSA occurring in the outerlayers of SN87A. We then removed the cold ejecta ab-sorption. The best-fit photon index is Γ DSA = 2 . ± . Greco et al. good description of the observed spectra ( χ = 1457with 1221 d.o.f. to be compared with χ = 1442with 1221 d.o.f. of the PWN scenario).Since from a statistical point of view an improvementof only 1% is not enough to favour the PWN case overthe DSA one, in Sect. 3 we present a comparison of thephysical implications in both cases. DISCUSSIONIn Sect. 2 we showed that a nonthermal component isneeded to properly fit the
NuSTAR data in the 10 − vphabs model, because of the presence of coldejecta surrounding the putative compact object. Theheavy absorption leads to a negligible contribution ofthe power-law at energies below 6 keV and thus to asteeper Γ. However, the photon index values in the twoscenarios are compatible with each other taking into ac-count the 90% confidence error bars. Moreover, the verysimilar values of χ does not allow us to exclude one ofthe two possible emission mechanisms under a merelystatistical point of view.In the DSA scenario, the flux is expected to varywith time in the same way both in the X-ray and radiobands. The synchrotron radio flux of SN87A increasedby ∼
15% between 2012 and 2014 (Cendes et al. 2018).Under the DSA hypothesis, this woud be at odd with ourfindings, since we observe a steady X-ray synchrotronemission, and a 15% increase of the nonthermal X-rayflux is discouraged at the 90% confidence level. However,we note that the DSA scenario could allow for spectralvariations in time that would make X-ray variations lesspredictable from radio variations.To further investigate the emission nature in the DSAscenario, we replaced the power-law component withthe XSPEC srcut model, which describes synchrotronemission from an exponentially cut off power-law distri-bution of electrons in the assumption of homogeneousmagnetic field (Reynolds 1998). For each of the threeyears considered, we constrained the normalization ofthe srcut component (i.e., its flux at 1 GHz, S GHz ),from the corresponding values observed at 9 GHz byCendes et al. (2018), taking into account the radio spec-tral index of SN87A ( α = 0 .
74, Zanardo et al. 2013). Wefixed S GHz and α in the srcut model and left the breakfrequency, ν b , free to vary. The radio-to-X-rays spec-tral index α can be assumed to be constant since the synchrotron cooling time is much longer than the ageof the system, thus no significant cooling is expected(i.e. τ sync ∼
125 yr ( E/
10 TeV) − ( B/ µ G) − ).We obtained hν b = 2 . +0 . − . keV, hν b = 2 . +0 . − . keVand hν b = 1 . +0 . − . keV, in 2012, 2013 and 2014 respec-tively. The break energy is compatible with 2 keV in thethree years considered, though the general trend seemsto point towards a decrease with time.The synchrotron emission of electrons peaks at energy hν = 1 . × E B , where B is the magnetic fieldin units of 100 µG and E is the electron energy inunits of 100 TeV. Considering our break energy hν b ∼ E spanning from ∼ . ∼ .
33, for B ranging from 6 to 1. This maxi-mum electron energy seems to be quite high, especiallyconsidering the relatively low shock speed. In fact, thesynchrotron radio emission originates in the HII region(Zanardo et al. 2013; Cendes et al. 2018; Orlando et al.2019), where the shock velocity is of only 2000 km/s(Cendes et al. 2018, Or20).In the DSA scenario, we can estimate the accel-eration time scale as (Parizot et al. 2006) τ acc =124 ηB − V − s E
100 43 yr, where η is the Bohm factor and V s is the shock velocity. In the hypothesis of maximumefficiency ( η = 1) and a standard magnetic field B =1, with V s = 2000 km/s, we would need τ acc ∼
390 yrto accelerate the X-ray emitting electrons up to the ob-served maximum energy, i.e., much more than the age ofSN87A. The observed maximum energy can be obtainedin 25 yr only by assuming that the downstream magneticfield is amplified by the SN87A slow shock up to ∼ µ G (in this case the maximum electron energy would beof ∼
14 TeV), and only assuming that the accelerationproceeds at the Bohm limit. The aforementioned issues(steady synchrotron flux and extremely large electronenergy in a relatively slow shock) concur in making theDSA scenario unconvincing.On the other hand, the PWN scenario has a strongphysical motivation, as we show below. In the absence ofa direct identification of the compact object eventuallypowering PWN87A, the only possible way to constrainits properties is to use the X-ray luminosity obtainedin our analysis to locate the putative PWN within thePWNe population.The properties of a generic PWN can be associatedto those of its progenitor SNR and surrounding ISMintroducing the characteristic time and luminosity scales(Truelove & McKee 1999): t ch = E − / M / ρ − / , (1) L ch = E sn /t ch , (2) ndication of a pulsar wind nebula in SN 1987A Table 1.
Best-fit parameters of the model adopted to describe
Chandra and
NuSTAR observations performed in 2012, 2013and 2014. Chemical abundances and power law parameters are kept constant along the various years. Uncertainties are at 90%confidence level. Component Parameter 2012 2013 2014TBabs n H (10 cm − ) 0.235 (fixed)kT (keV) 2.85 +0 . − . ± .
07 2.85 ± .
05O 0.32 +0 . − . Ne 0.57 ± . ± . +0 . − . S 0.65 ± . τ (10 s/cm ) 2.1 +0 . − . ± . +0 . − . EM (10 cm − ) 7.0 ± . ± . +0 . − . kT (keV) 0.60 ± .
04 0.65 +0 . − . +0 . − . vnei τ (10 s/cm ) 1.8 +0 . − . ± . ± . cm − ) 31 +2 − ± ± . +0 . − . L pwn0 . − (10 erg/s) 4 . +4 − . L pwn1 − (10 erg/s) 2 . ± . pwn10 − (10 erg/s) 0.32 +0 . − . Flux . − (10 erg/s/cm ) 93 ± +4 − +2 − Flux − (10 erg/s/cm ) 1.5 ± . +0 . − . +0 . − . χ (d.o.f.) 1442 (1221) where E sn is the supernova explosion energy, usually as-sumed to be 10 erg, M ej is the mass in the SNR ejectaand ρ ism the mass density of the ISM. These last param-eters have been considered to vary uniformly in: M ej ∈ [5 −
20] M (cid:12) (Smartt et al. 2009) and n ism ∈ [0 . − − (Berkhuijsen 1987; Magnier et al. 1997; Long et al.2010; Bandiera & Petruk 2010; Asvarov 2014), where ρ ism = m p n ism and m p is the proton mass. For thepulsar population the best choice is to consider young γ − ray emitting pulsars (Watters & Romani 2011; John-ston et al. 2020), better suited for describing pulsarspowering PWNe than the old radio-emitting ones (Kaspiet al. 2006).Considering the pulsar parameters from that popula-tion (namely the initial spin-down time τ and luminos-ity L ), with the choice of the canonical dipole brakingindex n = 3, the PWNe population can be then con-structed scaling time and luminosity with the charac-teristic ones defined in Eq. 1-2. In the ( τ /t ch , L /L ch )plane the PWNe population appears as an ellipsoidalsurface, where each point corresponds to various physi-cal sources with different combinations of τ , L , M ej and ρ ism . We use this PWNe population to discuss the possi- The PWNe population is liable of changes in the parametersplane, especially if considering a different braking index valuethan the standard dipole one (Parthasarathy et al. 2020.) ble location of PWN87A, by scaling the observed X-rayluminosity for the corresponding characteristic quanti-ties. In particular, the ejecta mass and kinetic energyare 18M (cid:12) and 2 × erg, respectively (Or20, and ref-erences therein). Given the complex structure of theremnant, the density of the material in which the ejectaexpand shows large inhomogeneities, varying from ∼ . − in the pre-shock blue supergiant wind, ∼
100 cm − in the HII region, up to 10 − cm − in the dense ring(Sugerman et al. 2005, Or20). We consider a value of ∼
100 cm − , representative of the equatorial zone of theHII region. Thus the two scaling are: t = 807 . (cid:18) E sn × erg (cid:19) − / (cid:18) M ej (cid:12) (cid:19) / (cid:18) ρ ism m p cm − (cid:19) − / , (3) L = 7 . × erg / s (cid:18) E sn × erg (cid:19) . (4)Taking into account the ejecta absorption, as calcu-lated from our MHD simulation, we derive from thespectral analysis the unabsorbed X-ray luminosity of thecentral source, finding L pwn X = 4 . +4 − . × erg/s in the0 . − . E through the conversion efficiency parame- Greco et al.
Table 2.
Summary of the most probable values for the pul-sar spin-down time for different values of the X-ray efficiency η X , and the associated spin-down luminosity (from the best-fit value). All values are given considering n ism = 100 cm − . η X L (erg/s) τ (yr)10 − . × − . × − . × − . × − . × ter η X = L X / ˙ E , that shows very large variations withinthe population, from ∼ − to ∼ − (Kargaltsev &Pavlov 2008).To maintain our analysis as general as possible, herewe have considered the entire range of measured η X .In Fig. 3 we show the population of PWNe as deter-mined associating the γ − ray emitting pulsars with thediscussed ranges of parameters for SNRs and ISM (inlight blue). The positioning of the putative PWN inSN87A for varius η X is shown as brown lines that in-tersect horizontally the distribution. The Crab nebula,the Vela nebula and the PWN in Kes 75 are shown forcomparison. As it can be easily seen, all the differentpossibilities lead to a location of PWN87A fully com-patible with the population. Assuming an ambient den-sity of 100 cm − might not be correct for SN 1987A,because of the complex density distribution. Therefore,we relaxed this assumption and, in the right panel ofFig. 3, we show that the putative positions for varying η X are perfectly consistent with the PWN population,even considering the two extreme density values for thesurrounding medium (0 . − cm − , Sugerman et al.2005, Or20).We verified that our estimate of the PWN luminosityin the X-rays is consistent with the radio luminosity de-rived by Cigan et al. (2019) under the (very reasonable)assumption that the low energy break frequency of thesynchrotron radiation is ν b (cid:38) Hz, while we expectthat the high absorption of the cold ejecta prevents thedetectability of the PWN emission in the optical band. An estimate (purely based on a statistical argument)of the most probable spin-down time for each η X can bedetermined by combining the PWNe probability distri-bution with the estimates of the pulsar spin-down lumi-nosity (Table 2, with n ism = 100 cm − ).In conclusion, a PWN seems to be the most likelysource of the synchrotron radiation in hard X-rays ofSN87A, even though the DSA scenario cannot be firmlyexcluded. A more conclusive way to discern between thetwo scenarios will be provided by future observations.An increase in the hard ( >
10 keV) X-ray flux, simi-lar to that observed in radio would be easily detectablewith
NuSTAR , thus supporting the DSA scenario. A de-crease would support the PWN scenario. On the otherhand, the rapid ejecta expansion and rarefaction will re-duce the soft X-ray absorption of an inner source. Inparticular, from our MHD simulation, we estimate thatthe PWN will become detectable by
Chandra and/orLynx (Appendix C) in the 2030s, definitely confirmingthe PWN scenario.With the present data and based on our findings, thePWN scenario seems the most likely (and appealing) toaccount for the nonthermal X-ray emission that we havedetected. Software:
CIAO (Fruscione et al. 2006), HEA-SOFT , XSPEC (Arnaud 1996), NuSTARDAS Facilities:
Chandra , NuSTAR (Harrison et al. 2013)ACKNOWLEDGEMENTSWe thank the anonymous referee for the helpful com-ments. EG, MM, SO, and FB acknowledge financialcontribution from the INAF mainstream program. MM,EG and SO acknowledges contribution from
NuSTAR (NARO18) in the framework of the ASI - INAF agreee-ment I/037/12/0. This work is supported by JSPSGrants-in-Aid for Scientific Research “KAKENHI A”Grant Numbers JP19H00693. SN, MO, AD wish to ac-knowledge the support from the Program of Interdisci-plinary Theoretical & Mathematical Sciences (iTHEMS)at RIKEN, and the support from Pioneering Programof RIKEN for Evolution of Matter in the Universe (r-EMU).REFERENCES By assuming a radio photon index Γ R = 1 . X = 2 . https://heasarc.gsfc.nasa.gov/docs/software/heasoft/ https://heasarc.gsfc.nasa.gov/docs/nustar/analysis/nustar swguide.pdf Alp, D., Larsson, J., Fransson, C., et al. 2018, ApJ, 864,174, doi: 10.3847/1538-4357/aad739 https://cxc.harvard.edu/index.html ndication of a pulsar wind nebula in SN 1987A Figure 3.
Left panel . Positioning of the putative PWN87A (brown lines) within the PWNe population (in light-blue color) inthe case of the representative density 100 cm − of the HII region. The PWNe population is represented with contours enclosingthe 99% of the entire population and the 20% (to better mark the center). Different possible positions of the PWN have beenobtained using the detected X-ray flux in the 0.5-8 keV band (with the best fit value shown as a solid-dashed line and theentire range of variation as a shaded area) and considering the widest possible variation for the X-ray efficiency η X (Kargaltsev& Pavlov 2008). Right panel . Variation of the predicted location of the PWN87A within the PWNe population, for differentvalues of η X , considering an extreme variation of the ambient density, from 0 . − (representative of the blue giant wind, inyellow) to 10 cm − (the maximum density in the dense ring in the HII region, in green).Arnaud, K. A. 1996, in ASP Conf. Ser. 101: AstronomicalData Analysis Software and Systems V, 17Asvarov, A. I. 2014, A&A, 561, A70,doi: 10.1051/0004-6361/201322774Bandiera, R., & Petruk, O. 2010, A&A, 509, A34,doi: 10.1051/0004-6361/200912244Berkhuijsen, E. M. 1987, A&A, 181, 398Bionta, R. M., Blewitt, G., Bratton, C. B., et al. 1987,PhRvL, 58, 1494, doi: 10.1103/PhysRevLett.58.1494Cendes, Y., Gaensler, B. M., Ng, C. Y., et al. 2018, ApJ,867, 65, doi: 10.3847/1538-4357/aae261Cigan, P., Matsuura, M., Gomez, H. L., et al. 2019, ApJ,886, 51, doi: 10.3847/1538-4357/ab4b46Esposito, P., Rea, N., Lazzati, D., et al. 2018, ApJ, 857, 58,doi: 10.3847/1538-4357/aab6b6Fransson, C., & Chevalier, R. A. 1987, ApJL, 322, L15,doi: 10.1086/185028Fruscione, A., McDowell, J. C., Allen, G. E., et al. 2006, inSociety of Photo-Optical Instrumentation Engineers(SPIE) Conference Series, Vol. 6270, Society ofPhoto-Optical Instrumentation Engineers (SPIE)Conference Series, ed. D. R. Silva & R. E. Doxsey,62701V, doi: 10.1117/12.671760Harrison, F. A., Craig, W. W., Christensen, F. E., et al.2013, ApJ, 770, 103, doi: 10.1088/0004-637X/770/2/103Johnston, S., Smith, D. A., Karastergiou, A., & Kramer, M.2020, MNRAS, 497, 1957, doi: 10.1093/mnras/staa2110 Kaastra, J. S., & Bleeker, J. A. M. 2016, A&A, 587, A151,doi: 10.1051/0004-6361/201527395Kargaltsev, O., & Pavlov, G. G. 2008, in AmericanInstitute of Physics Conference Series, Vol. 983, 40 Yearsof Pulsars: Millisecond Pulsars, Magnetars and More, ed.C. Bassa, Z. Wang, A. Cumming, & V. M. Kaspi,171–185, doi: 10.1063/1.2900138Kaspi, V. M., Roberts, M. S. E., & Harding, A. K. 2006,Isolated neutron stars, Vol. 39, 279–339Long, K. S., Blair, W. P., Winkler, P. F., et al. 2010, ApJS,187, 495, doi: 10.1088/0067-0049/187/2/495Magnier, E. A., Primini, F. A., Prins, S., van Paradijs, J.,& Lewin, W. H. G. 1997, ApJ, 490, 649.http://cdsads.u-strasbg.fr/cgi-bin/nph-bib query?bibcode=1997ApJ...490..649M&db key=ASTMcCray, R. 1993, ARA&A, 31, 175,doi: 10.1146/annurev.aa.31.090193.001135McCray, R., & Fransson, C. 2016, ARA&A, 54, 19,doi: 10.1146/annurev-astro-082615-105405Miceli, M., Orlando, S., Burrows, D. N., et al. 2019, NatureAstronomy, 3, 236, doi: 10.1038/s41550-018-0677-8Ono, M., Nagataki, S., Ferrand, G., et al. 2020, ApJ, 888,111, doi: 10.3847/1538-4357/ab5dbaOrlando, S., Miceli, M., Pumo, M. L., & Bocchino, F. 2015,ApJ, 810, 168, doi: 10.1088/0004-637X/810/2/168Orlando, S., Miceli, M., Petruk, O., et al. 2019, A&A, 622,A73, doi: 10.1051/0004-6361/201834487 Greco et al.
Orlando, S., Ono, M., Nagataki, S., et al. 2020, A&A, 636,A22, doi: 10.1051/0004-6361/201936718Page, D., Beznogov, M. V., Garibay, I., et al. 2020, arXive-prints, arXiv:2004.06078.https://arxiv.org/abs/2004.06078Panagia, N. 1999, in New Views of the Magellanic Clouds,ed. Y. H. Chu, N. Suntzeff, J. Hesser, & D. Bohlender,Vol. 190, 549Parizot, E., Marcowith, A., Ballet, J., & Gallant, Y. A.2006, A&A, 453, 387, doi: 10.1051/0004-6361:20064985Park, S., Zhekov, S. A., Burrows, D. N., et al. 2006, ApJ,646, 1001, doi: 10.1086/505023Parthasarathy, A., Johnston, S., Shannon, R. M., et al.2020, MNRAS, 494, 2012, doi: 10.1093/mnras/staa882Reynolds, S. P. 1998, ApJ, 493, 375.http://cdsads.u-strasbg.fr/cgi-bin/nph-bib query?bibcode=1998ApJ...493..375R&db key=ASTReynolds, S. P., Zoglauer, A., Boggs, S. E., Harrison, F., &NuSTAR Team. 2015, in American Astronomical SocietyMeeting Abstracts, Vol. 225, American AstronomicalSociety Meeting Abstracts ndication of a pulsar wind nebula in SN 1987A A. DATA REDUCTION
Chandra data were reprocessed with the CIAO v4.12.2software, using CALDB 4.9.2. We reduced the datathrough the task chandra repro and we extracted theACIS-S spectra by using the tool specextract which alsoprovided the corresponding arf and rmf files.
NuSTAR data were reprocessed with the standardpipelines provided by the
NuSTAR data analysis soft-ware NuSTARDAS by using nupipeline and nuproducts .Details of the observations are reported in Table 3.Spectral analysis has been performed with XSPEC(v12.11.1, (Arnaud 1996)) in the 0 . − − Chandra and
NuSTAR data, respec-tively. All spectra were rebinned adopting the opti-mal binning procedure described in Kaastra & Bleeker(2016), and the background spectrum to be subtractedwas extracted from a nearby region immediately out-side of the source. We verified that our results are notaffected by the choice of the background regions. B. X-RAY ABSORPTION FROM COLD EJECTAWe used the 3D MHD simulation of SN87A by Or20 toestimate the absorption pattern of the cold ejecta. Thesimulation reproduces most of the features observed inthe remnant of SN87A in various spectral bands, andlinks the SNR with the properties of the asymmetricparent SN explosion (Ono et al. 2020) and with the na-ture of its progenitor star (Urushibata et al. 2018). Themodel provides all relevant physical quantities in eachcell of the 3D spatial domain (MHD variables, plasmacomposition of the CSM and of the ejecta vs time). Themost notable quantities for our purposes are: the elec-tron temperature, the ion density for many chemicalspecies ( H, He, He, C, N, O, Ne, Mg, Si, S, Ar, Ca, Ti, Cr, Fe, Fe, Ni, and thedecay products, including Fe), and the ionization age.The model also predicts that the putative NS relic ofthe supernova explosion has received a kick towards theobserver in the north with a lower limit to the kick veloc-ity of ≈
300 km/s, resulting from a highly asymmetricexplosion (Ono et al. 2020, Or20).For the present study, we assumed a slightly higherkick velocity of 500 km/s. We checked that the resultsdo not change significantly for values ranging between300 and 700 km/s. Given the kick velocity and the direc-tion of motion of the NS as predicted by the model (Ono https://heasarc.gsfc.nasa.gov/docs/nustar/analysis/nustar swguide.pdf et al. 2020, Or20), and orienting the modeled remnantas it is observed in the plane of the sky (Or20), we estab-lished the position of the NS in the 3D spatial domain ofthe simulation for each year analyzed in this work. Con-sidering that the extension of the putative radio PWNis of the order of < ≈
180 AU, we can consider the centralsource as point-like in our procedure.We reconstructed the absorption pattern encounteredby the synchrotron X-ray emission of the putative PWNthrough the subsequent absorbing layers along the line ofsight. The physical effect responsible for the absorptionis the photo-electric effect, since the material surround-ing the considered source is cold (
T <
100 K). We ex-tracted values of temperature, column density and abun-dances associated to each absorbing layer of the 3D do-main of the model and we included these parameters inthe spectral analysis through the XSPEC photo-electricabsorption model vphabs .For the years considered in this work, the absorptiondue to cold ejecta is comparable with an equivalent Hcolumn density higher than 10 cm − . This indicatesthat potential signature of a PWN emission must besearched in the high energy part ( (cid:38)
10 keV) of the X-ray spectra, less affected by absorption. C. SYNTHETIC LYNX SPECTRUMWe produced a synthetic LYNX observation ofSN87A, as predicted by our MHD model for year 2037.We expect the thermal emission to stay almost con-stant in the next 15-20 years, though ejecta contribu-tion and/or interaction of the remnant with other inho-mogeneities beyond the ring may affect our predictions.Fig. 4 shows the synthetic spectrum extracted from acircular region with radius R = 0 . (cid:48)(cid:48) , well within thebright ring of SN87A (central panel of Fig. 1), assum-ing an exposure time of 300 ks. If we consider the emis-sion of the PWN absorbed by the ejecta pattern thatthe model predicts for 2037, we notice that the resultingcomponent would have a flux higher than that associ-ated with thermal emission above 4 keV, thus becomingdetectable in the soft x-ray band (see Fig. 4.). Furtherdetails about the future detectability of the PWN andon the thermal emission of the putative NS will be de-scribed in a forthcoming paper (Greco et al., in prep.).0 Greco et al.
Table 3.
Summary of the main characteristics of the analyzed observations.Telescope OBS ID PI Date (yr/month/day) Exposure time (ks)13735 Burrows 2012/03/28 4814417 Burrows 2012/04/01 27
Chandra
NuSTAR − − − . . r m a li z ed c oun t s s − k e V − Energy (keV)Synthetic Lynx spectrum − PWN in 2037