arXiv subjects

Eli Waxman

Publications and source records attributed to Eli Waxman.

At least 55 records · Page 3Linked to original sources

arXiv2017

Shock breakout theory

The earliest supernova (SN) emission is produced when the optical depth of the plasma lying ahead of the shock, which ejects the envelope, drops below c/v, where v is the shock velocity. This "breakout" may occur when the shock reaches the edge of the star, producing a bright X-ray/UV flash on time scales of seconds to a fraction of an hour, followed by UV/optical "cooling" emission from the expanding cooling envelope on a day time-scale. If the optical depth of circumstellar material (CSM) ejected from the progenitor star prior to the explosion is larger than c/v, the breakout will take place at larger radii, within the CSM, extending its duration to days time scale. The properties of the early, breakout and cooling, emission carry unique signatures of the structure of the progenitor star (e.g. its radius and surface composition) and of its mass-loss history. The recent progress of wide-field transient surveys enable SN detections on a day time scale, and are being used to set unique constraints on the progenitors of SNe of all types. This chapter includes a pedagogical description of SN breakout theory, and a concise overview of what we have learned from observations so far, and of advances in observational capabilities that are required in order to make further significant progress.

astro-ph.HE↗

arXiv2017

UV/Optical emission from the expanding envelopes of type II supernovae

The early part of a supernova (SN) light-curve is dominated by radiation escaping from the expanding shock-heated progenitor envelope. For polytropic Hydrogen envelopes, the properties of the emitted radiation are described by simple analytic expressions and are nearly independent of the polytropic index, $n$. This analytic description holds at early time, $t<$~few days, during which radiation escapes from shells initially lying near the stellar surface. We use numerical solutions to address two issues. First, we show that the analytic description holds at early time also for non-polytropic density profiles. Second, we extend the solutions to later times, when the emission emerges from deep within the envelope and depends on the progenitor's density profile. Examining the late time behavior of polytropic envelopes with a wide range of core to envelope mass and radius ratios, $0.1\le M_{\rm c}/M_{\rm env}\le10$ and $10^{-3}\le R_{\rm c}/R\le10^{-1}$, we find that the effective temperature is well described by the analytic solution also at late time, while the luminosity $L$ is suppressed by a factor, which may be approximated to better than $20[30]\%$ accuracy up to $t=t_{\rm tr}/a$ by $A\exp[-(at/t_{\rm tr})^α]$ with $t_{\rm tr} = 15 (M_{\rm env}/M_\odot)^{3/4}(E/10^{51}{\rm erg})^{-1/4}~\rm d$, $A=0.9[0.8]$, $a=1.7[4.6]$ and $α=0.8[0.7]$ for $n=3/2[3]$. This description holds as long as the opacity is approximately that of a fully ionized gas, i.e. for $T>0.7~\rm eV$, $t<14(R/10^{13.5}{\rm cm})^{0.55}~\rm d$. The suppression of $L$ at $t_{\rm tr}/a$ obtained for standard polytropic envelopes may account for the first optical peak of double-peaked SN light curves, with first peak at a few days for $M_{\rm env}<1M_\odot$.

astro-ph.HE↗

arXiv2016

Constraining High-Energy Cosmic Neutrino Sources: Implications and Prospects

We consider limits on the local ($z=0$) density ($n_0$) of extragalactic neutrino sources set by the nondetection of steady high-energy neutrino sources producing $\gtrsim50$ TeV muon multiplets in the present IceCube data, taking into account the redshift evolution, luminosity function and neutrino spectrum of the sources. We show that the lower limit depends moderately on source spectra and strongly on redshift evolution. We find $n_0\gtrsim{10}^{-8}-{10}^{-7}~{\rm Mpc}^{-3}$ for standard candle sources evolving rapidly, $n_s\propto{(1+z)}^3$, and $n_0\gtrsim{10}^{-6}-{10}^{-5}~{\rm Mpc}^{-3}$ for nonevolving sources. The corresponding upper limits on their neutrino luminosity are $L_{ν_μ}^{\rm eff}\lesssim10^{42}-10^{43}~{\rm erg}~{\rm s}^{-1}$ and $L_{ν_μ}^{\rm eff}\lesssim10^{41}-10^{42}~{\rm erg}~{\rm s}^{-1}$, respectively. Applying these results to a wide range of classes of potential sources, we show that powerful blazar jets associated with active galactic nuclei are unlikely to be the dominant sources. For almost all other steady candidate source classes (including starbursts, radio galaxies, and galaxy clusters and groups), an order of magnitude increase in the detector sensitivity at $\sim0.1-1$ PeV will enable a detection (as point sources) of the few brightest objects. Such an increase, which may be provided by next-generation detectors like IceCube-Gen2 and an upgraded KM3NET, can improve the limit on $n_0$ by more than two orders of magnitude. Future gamma-ray observations (by Fermi, HAWC and CTA) will play a key role in confirming the association of the neutrinos with their sources.

astro-ph.HE↗

arXiv2015

The energy budget of GRBs based on updated prompt \& afterglow observations

We compare the isotropic equivalent 15-2000 keV gamma-ray energy, E_gamma, emitted by a sample of 91 swift Gamma-Ray Bursts (GRBs) with known redshifts, with the isotropic equivalent fireball energy, E_fb, as estimated within the fireball model framework from X-ray afterglow observations of these bursts. The uncertainty in E_gamma, which spans the range of ~10^51 erg to ~10^53.5 erg, is approximately 25% on average, due mainly to the extrapolation from the BAT detector band to the 15-2000 keV band. The uncertainty in E_fb is approximately a factor of 2, due mainly to the X-ray measurements' scatter. We find E_gamma and E_fb to be tightly correlated. The average(std) of η^11hr_gamma is approximately log_10(E_gamma/(3ε _eE^11hr_fb)) are -0.34(0.60), and the upper limit on the intrinsic spread of η_gamma is approximately 0.5 (ε_e is the fraction of shocked plasma energy carried by electrons and E^x hr_fb is inferred from the X-ray flux at x hours). We also find that E_fb inferred from X-ray observations at 3 and 11 hours are similar, with an average(std) of log_10(E^3hr_fb/E^11hr_fb) of 0.04(0.28). The small variance of η_gamma implies that burst-to-burst variations in ε_e and in the efficiency of fireball energy conversion to gamma-rays are small, and suggests that both are of order unity. The small variance of η_gamma and the similarity of E^3hr_fb and E^11hr_fb further imply that ε_e does not vary significantly with shock Lorentz factor, and that for most bursts the modification of fireball energy during the afterglow phase, by processes such as radiative losses or extended duration energy injection, are not significant. Finally, our results imply that if fireballs are indeed jets, then the jet opening angle satisfies θ>0.1 for most cases. [abridged]

astro-ph.HE↗

arXiv2015

The detection rate of early UV emission from supernovae: A dedicated GALEX/PTF survey and calibrated theoretical estimates

The radius and surface composition of an exploding massive star,as well as the explosion energy per unit mass, can be measured using early UV observations of core collapse supernovae (SNe). We present the first results from a simultaneous GALEX/PTF search for early UV emission from SNe. Six Type II SNe and one Type II superluminous SN (SLSN-II) are clearly detected in the GALEX NUV data. We compare our detection rate with theoretical estimates based on early, shock-cooling UV light curves calculated from models that fit existing Swift and GALEX observations well, combined with volumetric SN rates. We find that our observations are in good agreement with calculated rates assuming that red supergiants (RSGs) explode with fiducial radii of 500 solar, explosion energies of 10^51 erg, and ejecta masses of 10 solar masses. Exploding blue supergiants and Wolf-Rayet stars are poorly constrained. We describe how such observations can be used to derive the progenitor radius, surface composition and explosion energy per unit mass of such SN events, and we demonstrate why UV observations are critical for such measurements. We use the fiducial RSG parameters to estimate the detection rate of SNe during the shock-cooling phase (<1d after explosion) for several ground-based surveys (PTF, ZTF, and LSST). We show that the proposed wide-field UV explorer ULTRASAT mission, is expected to find >100 SNe per year (~0.5 SN per deg^2), independent of host galaxy extinction, down to an NUV detection limit of 21.5 mag AB. Our pilot GALEX/PTF project thus convincingly demonstrates that a dedicated, systematic SN survey at the NUV band is a compelling method to study how massive stars end their life.

astro-ph.HE↗

arXiv2013

AMS02 results support the secondary origin of cosmic ray positrons

We show that the recent AMS02 positron fraction measurement is consistent with a secondary origin for positrons, and does not require additional primary sources such as pulsars or dark matter. The measured positron fraction at high energy saturates the previously predicted upper bound for secondary production (Katz et al 2009), obtained by neglecting radiative losses. This coincidence, which will be further tested by upcoming AMS02 data at higher energy, is a compelling indication for a secondary source. Within the secondary model the AMS02 data imply a cosmic ray propagation time in the Galaxy of < Myr and an average traversed interstellar matter density of order 1/cc, comparable to the density of the Milky Way gaseous disk, at a rigidity of 300 GV.

astro-ph.HE↗

arXiv2013

The energy production rate density of cosmic rays in the local universe is $\sim10^{44-45}\rm erg~Mpc^{-3}~yr^{-1}$ at all particle energies

The energy output (per logarithmic interval of particle energies) of Cosmic Rays (CRs) with energies $10{\rm GeV}\lesssim\varepsilon_p\lesssim100{\rm GeV}$ is $\sim 10^{47}\rm erg$ per solar mass of star$-$formation, based on the CR production rate in the Milky Way and in starburst galaxies, implying a generation rate of $\varepsilon_p^2Q\sim 10^{45}\rm erg~Mpc^{-3}~yr^{-1}$ in the local universe. It is only $\sim 10$ times larger than the output, $\varepsilon_p^2 Q=0.5\pm0.2\times 10^{44}\rm erg~Mpc^{-3}~yr^{-1}$, of Ultra High Energy CRs (UHECRs) at energies $10^{10.5}{\rm GeV}<\varepsilon_p<10^{12}\rm GeV$ (obtained assuming they are mostly protons), which in turn is comparable to the lower limit of $\varepsilon_p^2 Q\ge 0.5\times 10^{44}\rm erg~Mpc^{-3}~yr^{-1}$ of high energy CRs with $10^6{\rm GeV}\lesssim\varepsilon_p\lesssim 10^{8}\rm GeV$ implied by the saturation of the Waxman-Bahcall bound by the neutrino excess recently discovered by IceCube. These similarities are consistent with a flat production spectrum, $\varepsilon_p^2 Q\sim \text{const}$ for CRs at all observed energies. If a flat production spectrum is generated by our galaxy, the observed CR flux in the range $10^{6.5}-10^{9.5}\rm GeV$, above the "knee", is suppressed compared to lower energies due to propagation effects rather than acceleration upper limits. As suggested by Parizot and Aublin, the most exciting possibility is that cosmic rays at all energies are emitted from a single type of (unknown) sources, which can not be supernova remnants.

astro-ph.HE↗

arXiv2013

Constraints on the source of ultra-high energy cosmic rays using anisotropy vs chemical composition

The joint analysis of anisotropy signals and chemical composition of ultra-high energy cosmic rays offers strong potential for shedding light on the sources of these particles. Following up on an earlier idea, this paper studies the anisotropies produced by protons of energy >E/Z, assuming that anisotropies at energy >E have been produced by nuclei of charge Z, which share the same magnetic rigidity. We calculate the number of secondary protons produced through photodisintegration of the primary heavy nuclei. Making the extreme assumption that the source does not inject any proton, we find that the source(s) responsible for anisotropies such as reported by the Pierre Auger Observatory should lie closer than ~20-30, 80-100 and 180-200 Mpc if the anisotropy signal is mainly composed of oxygen, silicon and iron nuclei respectively. A violation of this constraint would otherwise result in the secondary protons forming a more significant anisotropy signal at lower energies. Even if the source were located closer than this distance, it would require an extraordinary metallicity >120, 1600, 1100 times solar metallicity in the acceleration zone of the source, for oxygen, silicon and iron respectively, to ensure that the concomitantly injected protons not to produce a more significant low energy anisotropy. This offers interesting prospects for constraining the nature and the source of ultra-high energy cosmic rays with the increase in statistics expected from next generation detectors.

astro-ph.HE↗

arXiv2013

What can we really learn about Magnetic Fields in Galaxy Clusters from Faraday Rotation observations?

We construct a simple and robust approach for deriving constraints on magnetic fields in galaxy clusters from rotation measure (RM) maps. Relaxing the commonly used assumptions of a correlation between the magnetic field strength and the plasma density and of a power-law (in wave number) magnetic field power spectrum, and using an efficient numerical analysis method, we test the consistency of a wide range of magnetic field models with RM maps of 11 extended sources in 5 clusters, for which the data were made available to us. We show that the data reveal no indication for a radial dependence of the average magnetic field strength, and in particular no indication for a correlation between the gas density and the field strength. The RM maps of a considerable fraction of the sources either require or are consistent with the presence of a spatially uniform magnetic field of a relatively small strength, 0.02-0.3 muG, which contributes significantly to the RM. The RM maps of all but one source do not require a power-law magnetic field power spectrum, and most are consistent with a power spectrum dominated by a single wave length. The uncertainties in the magnetic field strengths (and spatial correlation lengths) derived from RM maps exceed an order of magnitude (and often more). These uncertainties imply, in particular, that there is no indication in current RM data for a systematic difference between the magnetic field strengths in radio-halo clusters and in radio-quiet clusters. With the improvement expected in the near future of the quality and quantity of RM data, our analysis method will enable one to derive more accurate constraints on magnetic fields in galaxy clusters.

astro-ph.HE↗

arXiv2013

Non-relativistic radiation mediated shock breakouts: III. Spectral properties of SN shock breakout

The spectrum of radiation emitted following shock breakout from a star's surface with a power-law density profile $ρ\propto x^n$ is investigated. Assuming planar geometry, local Compton equilibrium and bremsstrahlung emission as the dominant photon production mechanism, numerical solutions are obtained for the photon number density and temperature profiles as a function of time, for hydrogen-helium envelopes. The temperature solutions are determined by the breakout shock velocity $v_0$ and the pre-shock breakout density $ρ_0$, and depend weakly on the value of n. Fitting formulas for the peak surface temperature at breakout as a function of $v_0$ and $ρ_0$ are provided, with $T_{peak}\approx 9.44\exp{[12.63(v_0/c)^{1/2}]}$ eV, and the time dependence of the surface temperature is tabulated. The time integrated emitted spectrum is a robust prediction of the model, determined by $\mathcal{T}_{\rm peak}$ and $v_0$ alone and insensitive to details of light travel time or slight deviations from spherical symmetry. Adopting commonly assumed progenitor parameters, breakout luminosities of ~10^45 erg/s and ~10^44 erg/s in the 0.3-10 keV band are expected for BSG and RSG/He-WR progenitors respectively (T_{peak} is well below the band for RSGs, unless their radius is ~10^13 cm). >30 detections of SN1987A-like (BSG) breakouts are expected over the lifetime of ROSAT and XMM-Newton. An absence of such detections would imply that either the typical parameters assumed for BSG progenitors are grossly incorrect or that their envelopes are not hydrostatic. The observed spectrum and duration of XRF 080109/SN2008D are in tension with a non-relativistic breakout from a stellar surface interpretation.

astro-ph.HE↗

arXiv2013

A Search for Correlation of Ultra-High Energy Cosmic Rays with IRAS-PSCz and 2MASS-6dF Galaxies

We study the arrival directions of 69 ultra-high energy cosmic rays (UHECRs) observed at the Pierre Auger Observatory (PAO) with energies exceeding 55 EeV. We investigate whether the UHECRs exhibit the anisotropy signal expected if the primary particles are protons that originate in galaxies in the local universe, or in sources correlated with these galaxies. We cross-correlate the UHECR arrival directions with the positions of IRAS-PSCz and 2MASS-6dF galaxies taking into account particle energy losses during propagation. This is the first time that the 6dF survey is used in a search for the sources of UHECRs and the first time that the PSCz survey is used with the full 69 PAO events. The observed cross-correlation signal is larger for the PAO UHECRs than for 94% (98%) of realisations from an isotropic distribution when cross-correlated with the PSCz (6dF). On the other hand the observed cross-correlation signal is lower than that expected from 85% of realisations, had the UHECRs originated in galaxies in either survey. The observed cross-correlation signal does exceed that expected by 50% of the realisations if the UHECRs are randomly deflected by intervening magnetic fields by 5 degrees or more. We propose a new method of analysing the expected anisotropy signal, by dividing the predicted UHECR source distribution into equal predicted flux radial shells, which can help localise and constrain the properties of UHECR sources. We find that the 69 PAO events are consistent with isotropy in the nearest of three shells we define, whereas there is weak evidence for correlation with the predicted source distribution in the two more distant shells in which the galaxy distribution is less anisotropic.

astro-ph.HE↗

arXiv2013

The early UV/Optical emission from core-collapse supernovae

We derive a simple approximate model describing the early, hours to days, UV/optical supernova emission, which is produced by the expansion of the outer <~0.01 solar mass part of the shock-heated envelope, and precedes the optical emission driven by radioactive decay. Our model includes an approximate description of the time dependence of the opacity (due mainly to recombination), and of the deviation of the emitted spectrum from a black body spectrum. We show that the characteristics of the early UV/O emission constrain the radius of the progenitor star, its envelope composition, and the ratio of the ejecta energy to its mass, E/M. For He envelopes, neglecting the effect of recombination may lead to an over estimate of progenitor radius by more than an order of magnitude. We also show that the relative extinction at different wavelengths may be inferred from the light-curves at these wave-lengths, removing the uncertainty in the estimate of progenitor radius due to reddening (but not the uncertainty in E/M due to uncertainty in absolute extinction). The early UV/O observations of the type Ib SN2008D and of the type IIp SNLS-04D2dc are consistent with our model predictions. For SN2008D we find progenitor radius to be approx. 10^11 cm, and an indication that the He envelope contains a significant C/O fraction.

astro-ph.HE↗

arXiv2012

Imploding ignition waves: I. one dimensional analysis

We show that converging spherical and cylindrical shock waves may ignite a detonation wave in a combustible medium, provided the radius at which the shocks become strong exceeds a critical radius, R_c. An approximate analytic expression for R_c is derived for an ideal gas equation of state and a simple (power-law-Arrhenius) reaction law, and shown to reproduce the results of numerical solutions. For typical acetylene-air experiments we find R_c~0.1 mm (spherical) and R_c~1 mm (cylindrical). We suggest that the deflagration to detonation transition (DDT) observed in these systems may be due to converging shocks produced by the turbulent deflagration flow, which reaches sub (but near) sonic velocities on scales >>R_c. Our suggested mechanism differs from that proposed by Zel'dovich et al., in which a fine-tuned spatial gradient in the chemical induction time is required to be maintained within the turbulent deflagration flow. Our analysis may be readily extended to more complicated equations of state and reaction laws. An order of magnitude estimate of R_c within a white dwarf at the pre-detonation conditions believed to lead to Type Ia supernova explosions is 0.1 km, suggesting that our proposed mechanism may be relevant for DDT initiation in these systems. The relevance of our proposed ignition mechanism to DDT initiation may be tested by both experiments and numerical simulations.

astro-ph.HE↗

arXiv2011

Early emission from type Ia supernovae

A unique feature of deflagration-to-detonation (DDT) white dwarf explosion models of SNe of type Ia is the presence of a strong shock wave propagating through the outer envelope. We consider the early emission expected in such models, which is produced by the expanding shock-heated outer part of the ejecta and precedes the emission driven by radioactive decay. We expand on earlier analyses by considering the modification of the pre-detonation density profile by the weak-shocks generated during the deflagration phase, the time evolution of the opacity, and the deviation of the post-shock equation of state from that obtained for radiation pressure domination. A simple analytic model is presented and shown to provide an acceptable approximation to the results of 1D numerical DDT simulations. Our analysis predicts a thousand second long UV/optical flash with a luminosity of ~1 to 3*1e39 erg/s. Lower luminosity corresponds to faster (turbulent) deflagration velocity. The predicted luminosity of the UV flash is an order of magnitude lower than that of earlier estimates, and is expected to be strongly suppressed at times longer than an hour due to the deviation from pure radiation domination.

astro-ph.HE↗

arXiv2011

PTF11eon/SN2011dh: Discovery of a Type IIb Supernova From a Compact Progenitor in the Nearby Galaxy M51

On May 31, 2011 UT a supernova (SN) exploded in the nearby galaxy M51 (the Whirlpool Galaxy). We discovered this event using small telescopes equipped with CCD cameras, as well as by the Palomar Transient Factory (PTF) survey, and rapidly confirmed it to be a Type II supernova. Our early light curve and spectroscopy indicates that PTF11eon resulted from the explosion of a relatively compact progenitor star as evidenced by the rapid shock-breakout cooling seen in the light curve, the relatively low temperature in early-time spectra and the prompt appearance of low-ionization spectral features. The spectra of PTF11eon are dominated by H lines out to day 10 after explosion, but initial signs of He appear to be present. Assuming that He lines continue to develop in the near future, this SN is likely a member of the cIIb (compact IIb; Chevalier and Soderberg 2010) class, with progenitor radius larger than that of SN 2008ax and smaller than the eIIb (extended IIb) SN 1993J progenitor. Our data imply that the object identified in pre-explosion Hubble Space Telescope images at the SN location is possibly a companion to the progenitor or a blended source, and not the progenitor star itself, as its radius (~10^13 cm) would be highly inconsistent with constraints from our post-explosion photometric and spectroscopic data.

astro-ph.CO↗

arXiv2011

X-rays, gamma-rays and neutrinos from collisoinless shocks in supernova wind breakouts

We show that a collisionless shock necessarily forms during the shock breakout of a supernova (SN) surrounded by an optically thick wind. An intense non-thermal flash of <~ MeV gamma rays, hard X-rays and multi-TeV neutrinos is produced simultaneously with and following the soft X-ray breakout emission, carrying similar or larger energy than the soft emission. The non-thermal flash is detectable by current X-ray telescopes and may be detectable out to 10's of Mpc by km-scale neutrino telescopes.

astro-ph.HE↗

arXiv2011

On The GeV & TeV Detections of the Starburst Galaxies M82 & NGC 253

The GeV and TeV emission from M82 and NGC 253 observed by Fermi, HESS, and VERITAS constrains the physics of cosmic rays (CRs) in these dense starbursts. We argue that the gamma rays are predominantly hadronic in origin, as expected by previous studies. The measured fluxes imply that pionic losses are efficient for CR protons in both galaxies: we show that a fraction F_cal ~ 0.2 - 0.4 of the energy injected in high energy primary CR protons is lost to inelastic proton-proton collisions (pion production) before escape, producing gamma rays, neutrinos, and secondary electrons and positrons. We discuss the factor ~2 uncertainties in this estimate, including supernova rate and leptonic contributions to the GeV-TeV emission. We argue that gamma-ray data on ULIRGs like Arp 220 can test whether M82 and NGC 253 are truly calorimetric, and we present upper limits on Arp 220 from the Fermi data. We show that the observed ratio of the GeV to GHz fluxes of the starbursts suggests that non-synchrotron cooling processes are important for cooling the CR electron/positron population. We briefly reconsider previous predictions in light of the gamma-ray detections, including the starburst contribution to the gamma-ray background and CR energy densities. Finally, as a guide for future studies, we list the brightest star-forming galaxies on the sky and present updated predictions for their gamma-ray and neutrino fluxes.

astro-ph.HE↗

arXiv2011

Non-relativistic radiation mediated shock breakouts: II. Bolometric properties of SN shock breakout

Exact bolometric light curves of supernova shock breakouts are derived based on the universal, non relativistic, planar breakout solutions (Sapir et al. 2011), assuming spherical symmetry, constant Thomson scattering opacity, κ, and angular intensity corresponding to the steady state planar limit. These approximations are accurate for progenitors with a scale height much smaller than the radius. The light curves are insensitive to the density profile and are determined by the progenitor radius R, and the breakout velocity and density, v_0 and ρ_0 respectively, and κ. The total breakout energy, E_BO, and the maximal ejecta velocity, v_max, are shown to be E_BO=8.0πR^2κ^-1cv_0 and v_max=2.0v_0 respectively, to an accuracy of about 10%. The calculated light curves are valid up to the time of transition to spherical expansion, t_sph\approx R/4v_0. Approximate analytic expressions for the light curves are provided for breakouts in which the shock crossing time at breakout, t_0=c/κρ_0v_0^2, is << R/c (valid for R<10^14 cm). Modifications of the flux angular intensity distribution and differences in shock arrival times to the surface, Δt_asym, due to moderately asymmetric explosions, affect the early light curve but do not affect v_max and E_BO. For 4v_0<<c, valid for large (RSG) progenitors, L\propto t^{-4/3} at max(Δt_asym,R/c)< t<t_sph and R may be accurately estimated from R\approx 2*10^13 (L/10^43 erg s^-1)^{2/5}(t/1 hr)^{8/15}.

astro-ph.HE↗