The exceptionally powerful TeV gamma-ray emitters in the Large Magellanic Cloud

1.3 MB·English

by  H. E. S. S. Collaboration

The exceptionally powerful TeV γ-ray emitters in the Large
Magellanic Cloud
TheH.E.S.S.Collaboration
Abstract
hasastellarmassofabout4%oftheMilkyWay(6,7).
5 Locatedatadistanceof≈50kpc(8), itisanirregular
1 TheLargeMagellanicCloud,asatellitegalaxyofthe galaxy seen almost face-on (9). Consequently, source
0 Milky Way, has been observed with the High En- confusion is much less of a problem than for the inner
2 ergyStereoscopicSystem(H.E.S.S.)aboveanenergy MilkyWay,andthereislessuncertaintyinthedistances

n of 100 billion electron volts for a deep exposure of of the sources. The LMC stands out among nearby
a 210 hours. Three sources of different types were galaxies for its high star formation rate per unit mass,
J
detected: the pulsar wind nebula of the most en- whichisaboutafactoroffivehigherthanintheMilky
6
ergetic pulsar known N157B, the radio-loud super- Way (10,11), and contains the best example of a local
2
nova remnant N132D and the largest non-thermal starburst,theTarantulaNebula. TheLMCalsoharbors
] X-rayshell–thesuperbubble30DorC.Theunique numerous massive stellar clusters and SNRs. Among
E objectSN1987Ais,surprisingly,notdetected,which theSNRsisauniquesource,SN1987A,theremnantof
H constrains the theoretical framework of particle ac- thenearestsupernovaobservedinmoderntimes(12).
. celerationinveryyoungsupernovaremnants. These High-energyγ-rayemissionfromtheLMCwasde-
h
p detections reveal the most energetic tip of a γ-ray tectedbyEGRET(13)and,morerecently,bytheFermi
- sourcepopulationinanexternalgalaxy,andprovide LargeAreaTelescopes(LAT)(14),whichrevealeddif-
o
via 30 Dor C the unambiguous detection of γ-ray fuseemissionwithanextensionofseveraldegreesindi-
r
t emissionfromasuperbubble. ameter,tracingmassivestarformingregions.VHEγ-ray
s
a telescopes,likeH.E.S.S.,besidesprovidinginformation
[ onmuchhigherenergyCRs,haveanangularresolution
Introduction

1 of a few arcminutes, which is substantially better than
v Fermi-LAT’s resolution at γ-ray energies < 10 GeV.
The origin of cosmic rays (CRs), the very high (VHE,
8 The good angular resolution allows H.E.S.S. to iden-
7 ∼> 1011 eV), and ultra high (∼> 1018 eV) energy par- tify individual sources in the LMC. As we will detail
5 ticles that bombard Earth, has puzzled us for over a
below,adeepH.E.S.S.observationrevealedthreelumi-
6 century. Much progress has been made during the last
0 nous sources in the LMC: the superbubble 30 Dor C,
decade due to the advent of VHE γ-ray telescopes.
. the energetic PWN N157B, and the radio-loud SNR
1 Thesetelescopesdetect∼ 1011−1014 eVγ-raysfrom
N132D. Of these sources, only N157B was detected
0
atomicnuclei(hadronicCRs)collisionswithlocalgas,
5 previously in a 47-hours exposure (15). The observa-
or from ultra-relativistic electrons/positrons (leptonic
1 tionsextendthescopeofVHEγ-rayastronomybypro-
v: CRs), which produce γ-ray emission by upscattering viding examples of sources from a population outside
low-energybackgroundphotons(1).Indeed,asurveyof
i the Milky Way. N157B and N132D belong to known
X theinnerpartoftheMilkyWaywithH.E.S.S.,anarray
γ-raysourceclasses,butbothhavedistinguishingchar-
r of imaging atmospheric Cherenkov telescopes (2), re- acteristics,N157Bbeingpoweredbythemostenergetic
a vealedapopulationofsupernovaremnants(SNRs)and
youngpulsar,andN132DbeingoneoftheoldestVHE
pulsarwindnebulae(PWNe)emittingγ-rayswithener-
γ-rayemittingSNRs. Thesuperbubble30DorC,how-
giesinexcessof100GeV(3).
ever,providesanunambiguousdetectionofasuperbub-
HerewereportonVHEγ-raysourcesdetectedout-
bleinVHEγ-rays. Conspicuouslyabsentfromourlist
side the Milky Way, namely in the Large Magellanic
ofthreesourcesisSN1987A,despitepredictionsthatit
Cloud (LMC). This satellite galaxy of the Milky Way
1
Figure1: SkymapsoftheLargeMagellanicCloud. a)OpticalimageoftheentireLMC(4). Theboxesdenote
theregionsofinterestdiscussedinthispaper. Coloursdenotelevelsof3,5,10and20σstatisticalsignificanceof
theγ-raysignal. b)VHEγ-rayemissionintheregionaroundN157B.Thegreenlinesrepresentcontoursof5,10
and15σstatisticalsignificanceoftheγ-raysignal. c)XMM-NewtonX-rayfluximageoftheregionof30DorC.
The superimposed cyan lines represent contours of 68%, 95% and 99% confidence level of the position of the
γ-ray source. Diamonds denote the positions of the star clusters of the LH 90 association. See supplementary
materialfordetailsontheX-rayanalysis. d)VHEγ-rayemissionintheregionaroundN132D.Thegreenlines
representcontoursof3,4and5σ statisticalsignificance. Thebackgroundoftheγ-rayemission(inpanelsband
d)wasobtainedusingtheringbackgroundmethod(5). Theresultingexcessskymapissmoothedtotheangular
resolutionoftheinstrument.
2
shouldbeabrightγ-raysource(16,17). d3N/(dEdtdA) = Φ (E/1TeV)−Γ (Fig. 2). The
0
best-fit spectral indices and integral γ-ray luminosities
aresummarizedinTable1.
H.E.S.S. Observations
Evenwithadeepexposureof210hours,significant
emission from SN1987A is not detected, and we de-
Wereportonadeep,210-hoursH.E.S.S.exposure,tar-
rive an upper limit on the integral γ-ray flux of F (>
γ
geted at the region of the Tarantula nebula — corre- 1TeV) < 5.6 × 10−14phcm−2s−1 at a 99% confi-
sponding to 30 Doradus (30 Dor) — the largest star-
dencelevel.
forming region in the Local Group of galaxies. We re-
constructed γ-ray showers with an image-fitting analy-
sis (18) and cross-checked with a multivariate analysis Discussion of individual sources
based on image parameterization (19,20), with consis-
tentresults. Inbothanalyses,acutontheuncertaintyof The three VHE emitters belong to different source
the reconstructed γ-ray direction indicated an angular classesandtheirenergyoutputexceedsoratleastequals
resolutionof≈0.05◦. thatoftheirmostpowerfulrelativesintheMilkyWay.
Fig. 1a shows an optical image of the LMC over-
laidwithTeVgamma-raypoint-sourcesignificancecon-
30 Dor C
tours. Inthisdataset,613γrayswithastatisticalsignif-
icanceof33σaredetectedfromthePWNN157B.Fig-
Thesuperbubble30DorCstandsoutinX-raysasitcon-
ure 1b provides a close-up view of the γ-ray emission
from N157B. The diameter of N157B of 100(cid:48)(cid:48) (21) is tains,inthewesternpart,anX-raysynchrotron-emitting
shell with a radius of 47pc, which makes it the largest
oftheorderoftheH.E.S.S.angularresolution. Further
known X-ray synchrotron shell (24–26). X-ray syn-
significantγ-rayemissionisdetectedtotheSouth-West
chrotronemission,whichindicatesthepresenceofVHE
ofN157B.
electrons, is usually associated with 100−2000 year-
A likelihood fit of a model of two γ-ray sources to
old SNRs with radii smaller than 25 pc. In addition,
theon-sourceandbackgroundskymapsestablishesthe
detectionofasecondsourceatanangulardistanceof9(cid:48) the X-ray synchrotron luminosity of 30 Dor C is ten
times that of the archetypal young SNR SN1006 (24).
(corresponding to 130pc at a distance of 50kpc) from
The 30 Dor C shell also emits radio and optical radia-
N157B. The model consisting of two sources is pre-
tion(27),andappearstohavebeenproducedbythestel-
ferred by 8.8σ over the model of one single source.
larwindsandsupernovaeintheOBassociationLH90
Fig. 1c shows an X-ray image with overlaid contours
(NGC2044)(28).
of confidence of the source position. The position
of the second source (RA = 5h35m(55 ± 5)s, Dec = The measured H.E.S.S. flux of 30 Dor C corre-
−69◦11(cid:48)(10 ± 20)(cid:48)(cid:48), equinox J2000, 1σ errors) coin- spondstoa1−10TeVγ-rayluminosityof(0.9±0.2)×
1035ergs−1,withthebest-fitpositionoftheγ-rayemis-
cides with the superbubble 30 Dor C, the first such
sionlyinginbetweenthesixidentifiedsub-clusters(28).
sourcedetectedinVHEγ-rays,andthusrepresentingan
The TeV emission can be explained by the production
additional source class in this energy regime. A γ-ray
ofneutralpionsduetocollisionsofhadronicCRswith
signal around the energetic pulsar PSRJ0540−6919 is
thebackgroundplasma. Alternatively,theso-calledlep-
notdetected,despitethepresenceofanX-rayluminous
tonic emission scenario may apply, in which case the
PWN(22). Afluxupperlimit(99%confidencelevel)is
derivedatF (>1TeV)<4.8×10−14phcm−2s−1. TeV emission is the result of Compton upscattering of
γ
low-energyphotonstoγ-rayenergies,bythesamepop-
Along with the clear detection of N157B and
ulationofelectronsthatisresponsiblefortheX-raysyn-
30DorC,evidenceforVHEγ-rayemissionisobserved
chrotronradiation(1).
from the prominent SNR N132D (Fig. 1d). The emis-
For the hadronic scenario, a combination of en-
sion peaks at a significance of about 5σ above a back-
ergy in CRs (assumed to be protons) and den-
groundwhichisestimatedfromaringaroundeachsky
sity of hydrogen atoms, n , of W = (0.7 −
bin. AtthenominalpositionoftheSNR43γ rayswith H pp
25)×1052(n /1cm−3)−1 erg is required (see S1.3).
astatisticalsignificanceof4.7σarerecorded. H
30 Dor C probably experienced ∼5 supernova explo-
The γ-ray spectra of all three objects are well
sions (25), which likely provided ∼5×1050erg in CR
described by a power law in energy, Φ(E) =
3
1)
-eV10-12 N 157B -12 30 Dor C -12 N 132D
T
1
-s10-13 -13 -13
2
-m
c
x (10-14 -14 -14
u
Fl
10-15 -15 -15
10-16 -16 -16
10-17 1 10 -17 1 10 -17 1 10
χ)
∆ 5 5 5
s (
al 0 0 0
u
d
si -5 -5 -5
e
R
1 10 1 10 1 10
Energy (TeV)
Figure 2: Gamma ray spectra of N157B, 30 Dor C and N132D. Note that the spectral points of 30 Dor C are
notcorrectedforspill-overemissionfromN157B(seeonlinesupplement). Datapointshave1σerrorbars,upper
limitsareatthe99%confidencelevel. Thebottompanelsshowtheresidualsofthedatapointscomparedtothe
best-fitmodel.
Source N157B 30DorC N132D
H.E.S.S.Identifier HESSJ0537−691 HESSJ0535−691 HESSJ0525−696
ExposureTime 181h 183h 148h
γ rays 613 74 43
Significance 33.0σ 8.8σ 4.7σ
PhotonIndexΓ 2.8±0.1 2.6±0.2 2.4±0.3
Φ(1TeV)[10−12cm−2s−1TeV−1] 1.3±0.1 0.16±0.04 0.13±0.05
L (1−10TeV)[1035ergs−1] 6.8±0.3 0.9±0.2 0.9±0.2
γ
Table1: Statisticsandspectralparametersofthethreesources. Theexposuretimeiscorrectedfortheacceptance
differencesduetodifferentoffsetsfromthecameracentre. ThesignificancesofN157BandN132Darestatistical
significances of the γ-ray emission obtained by using formula 17 of (23). The background was estimated from
regions with similar offsets from the camera center as the on-source region. The significance of 30 Dor C is
the significance by which a two-source morphology (N157B and 30 Dor C) is preferred over a single-source
morphology(N157Bonly). Theγ-raycountandthefluxof30DorCarecorrectedforspill-overemissionfrom
N157B(seeonlinesupplement). ΓisthephotonindexandΦ(1TeV)thedifferentialfluxat1TeVofapowerlaw
fittotheenergyspectrum. TheluminosityL iscalculatedforanassumeddistanceof50kpc(8). Thelistederrors
γ
arestatistical,1σerrors. Systematicerrorsareestimatedtobe±0.3forΓand±30%forΦ(1TeV)(15).
4
energy. Hence,theaveragegasdensityshouldbenH ∼> γ-rayandX-rayobservationssuggestactiveparticleac-
20cm−3, which is higher than the density estimate of celeration by a very large, fast expanding shell. This
n ≈0.1−0.4cm−3basedontheX-raythermalemis- may provide the right conditions for accelerating some
H
sion in the southwest (24,29). However, locations of protonstoenergiesexceeding3×1015eV,whichisthe
highdensitiesmaybepresent,iftheX-raythermalemis- maximumenergydetectedforGalacticCRs. Theseob-
sioncomesfromsmallerradiithanthedenseoutershell, servations,therefore,lendsupporttotheviewexpressed
orifcool,dense,clumpedgassurvivedinsidetheother- in (32,39,40) that superbubbles may provide the right
wiserarifiedinteriorofthebubble(30). conditions for particle acceleration to very high ener-
This hadronic scenario puts constraints on the CR gies, because they are thought to contain very turbu-
diffusioncoefficient, becausethediffusionlengthscale lentmagneticfieldsandtheyarelargeenoughtocontain
should be smaller than the radius of the shell: l = VHEparticlesforuptomillionsofyear.
√ diff
2Dt ∼< 47pc for CRs around 10TeV. Therefore, IntheMilkyWay,themostcloselyrelatedobjectto
D(10TeV)∼<3.3×1026(t/106yr)−1cm2s−1,which, 30DorCisthestellarclusterWesterlund1(41),which,
givenanageforthesuperbubbleofafewmillionyears, however,hasacompletelydifferentX-raymorphology.
givesamuchsmallerdiffusioncoefficientthanthetyp- Moreimportantly,itisnotclearwhethertheγ-raysorig-
icalGalacticdiffusioncoefficientofD(10TeV) ∼> 5× inatefromtheclusterwinditself,aPWNorfromthenu-
1029cm2s−1 (31). This small diffusion coefficient re- meroussupernovaethatexplodedinsideWesterlund1in
quires magnetic-field amplification combined with tur- therecentpast. Sincealargefractionofsupernovaeare
bulentmagneticfields,ashypothesizedby(32). thought to go off in superbubbles, this first unambigu-
X-raysynchrotronemissionfrom30DorCrequires ous detection of VHE γ-rays from a superbubble may
largeshockvelocities,vshock ∼> 3000kms−1 (35). As- havebroadimplicationsforthecircumstancesinwhich
suming that this shock originates from an explosion alargefractionofCRsareaccelerated.
centered at the superbubble, we obtain a rough esti-
mate of the age of the western X-ray shell of t =
N157B
0.4R/v ≈ 6000 yr, assuming a Sedov expansion
shock
model(R=2.8×108(Et2/n )1/5 cm). SincetheOB
H
The source HESSJ0537−691 is coincident with
associationismucholder,thisagemostlikelyrefersto
the PWN N157B, which surrounds the pulsar
a recent supernova explosion, whose remnant evolves
PSRJ0537−6910. PWNe are nebulae of ultra-
in the rarified medium of the superbubble. The Sedov
relativistic particles driven by highly-magnetized,
expansion model then gives us a very low estimate for
the density of n ≈ 5×10−4cm−3 for an explosion fast-rotating neutron stars that convert a considerable
H
energy of E = 1051erg. Although this is very low, it amount of their spin-down energy into a particle
wind. The archetypal Crab nebula is one of the
canoccurundercertainconditions(36). Thismodelfor
brightest sources of non-thermal radiation in the sky
theX-raysynchrotronshellcanevenbereconciledwith
and powered by the pulsar with the highest spin-
the hadronic model of the TeV emission, if the rarified
down energy known in the Milky Way (42). With
medium also contains dense clumps. For the leptonic
comparably extreme rotational energy loss rates,
scenariofortheTeVemission,thebroadspectralenergy
N157B(E˙ = 4.9×1038ergs−1)andtheCrabnebula
distribution(SED,Fig.3)requiresanenergyinacceler-
atedelectronsof∼ 4×1048erg,andaveragemagnetic (E˙ = 4.6 × 1038ergs−1) appear to be twins. The
study of N157B thus provides the unique opportunity
field strength of 15µG, low compared to most young
to compare two extreme PWNe, and to disentangle
SNRs(37),butafactorthreetofourhigherthantheav-
object-specificandgenericproperties.
eragemagneticfieldintheLMC(38).
Givenapopulationofultra-relativisticelectronsand
Althoughatthisstagewecannotruleouteitherthe
positrons forming the PWN, the X-ray luminosity is
leptonicorthehadronicscenario,theH.E.S.S.observa-
determined by the strength of the magnetic field and
tions reveal that the conditions inside the superbubble
theγ-rayluminositybytheintensityofradiationfields
must be extreme: the hadronic scenario requires loca-
whichserve astargetsforthe inverseComptonupscat-
tionswithhighdensitiesandahighdegreeofmagnetic
tering. If the radiation fields are known, the magnetic
turbulence, whereas the leptonic scenario requires the
fieldcanbeinferredfromthecombinationofX-rayand
stellar cluster to be extremely rarified. Moreover, the
γ-ray measurements. N157B is likely associated with
5
-1-2m) s 10-11 H.E.S.S. - 30 Dor C Fermi-LAT -1-2m) s 10-11 H.E.S.S. - N 132D ASCA Fermi-LAT
c c
g g
N/dE (er 10-12 Suzaku H.E.S.S. N/dE (er 10-12 Chandra H.E.S.S.
d d
2 E 10-13 2 E 10-13 Australia Telescope
10-14 10-14
10-15 10-15
10-6 10-4 10-2 1 102 104 106 108 1010 1012 10141015 10-6 10-4 10-2 1 102 104 106 108 1010 1012 10141015
Energy (eV) Energy (eV)
Figure 3: Spectral energy distribution of 30 Dor C and N132D. The 30 Dor C X-ray data are from (26). For
N132D,radiodataisfrom(33),andX-raylimitsarefrom(34)andfromre-analysedChandradata. Bothleptonic
(dashed lines) and hadronic (solid lines) models are shown. For further details on Fermi-LAT data and spectral
modelling,seeS1.2andS1.3.
the LH 99 star cluster (21,43,44), and therefore em- downenergyofN157Bisalsohiddenandisnotcarried
beddedinstronginfraredradiationfields(seeS1.5). In byultra-relativisticparticlesormagneticfields. There-
this environment, the magnetic field in the PWN must mainderoftheavailablerotationalenergyislikelytobe
be rather weak, not larger than 45µG, in order to ex- fed into electrons with energies ≤400GeV that radiate
plain the multiwavelength data (Fig. 4). When con- at lower photon energies, adiabatic expansion, and/or
sidering the region from which the hard X-ray emis- particlesescapingintotheinterstellarmediumviadiffu-
sion is coming, the total energy in the magnetic field siveescape(e.g.,(46)). ItthereforeappearsthatN157B
is W = 1.4 × 1047erg – an order of magnitude is such a bright γ-ray emitter because of the enhanced
B,tot
smallerthantheenergyin>400GeVelectrons.Thede- radiation fields, despite the fact that it is apparently a
rivedmaximummagneticfieldisalsomuchlowerthan much less efficient particle accelerator than the Crab
that inferred for the Crab nebula (∼124µG (45)), and nebula.
suggests at least a factor ∼7 lower magnetic pressure.
AsmostoftheelectronsthatradiateintheChandra,X-
N132D
rayandH.E.S.S.,TeVdomainshaveveryshortlifetimes
(≤300years),theenergyinultra-relativisticparticlesin
Inadditiontothetwounambiguouslydetectedsources,
N157Bcanbeinferredindependentlyofthespin-down
we find strong evidence for a third source at the posi-
evolutionofthepulsar. ForthemodelshowninFig.4,a
tionofthecore-collapseSNRN132D.N132DisaSNR
constantfractionof11%ofthecurrentspin-downpower
with strong thermal X-ray emission, which has been
of N157B needs to be injected into the nebula in the
usedtoestimateapre-shockdensityofn ≈2.6cm−3
form of relativistic electrons (compared with 50% for H
(34),ahighexplosionenergyof∼6×1051erg(34),and
the Crab nebula under the same model). This fraction
anageof∼6000yr,basedonaSedovmodel. SuchX-
convertedintoX-rayandTeVemissionisratherinsen-
raybrightSNRsarepredictedtobeγ-rayemitters(48).
sitivetothespectralindexofinjectedelectronsandthe
N132D is also luminous in the radio (33) and infrared
spin-evolution or braking index of the pulsar and only
bands (49). N132D is often compared to the brightest
reliesontheassociationofN157BwithLH99(seeS1.5
radiosourceCasA,which, likeN132D,isanoxygen-
formoreinformation).
richSNR.N132Dhasahigherinfraredluminosity(49),
In this high-radiation field scenario, the situation
but its radio luminosity is 50% that of Cas A. This is
fortheCrabnebulaisverydifferentfromN157B.Not
stillremarkablegiventhatN132Dhasakinematicage
only is the best-fit electron spectrum of N157B harder
of∼ 2500yr(50),whereasCasAis∼ 330yroldand
(Γ = 2.0 vs. 2.35), exhibiting a lower cut-off energy
e
declinesinluminositybyabout0.8%peryear. Thera-
(E = 100TeV vs. 3.5PeV), but much of the spin-
c
6
-1) Chandra
g s 1037 Crab
y (er Fermi-LAT H.E.S.S.
sit
mino 1036
u N 157B
L
45m G
35
10
34
10
33
10
-3 -1 3 5 7 9 11 13 15
10 10 10 10 10 10 10 10 10 10
Energy (eV)
Figure 4: Intrinsic SED of N157B (black) and the Crab nebula (grey). The model shown assumes the same
injection parameters as derived for the Crab nebula (E = 400GeV, E = 3.5PeV, Γ = 2.35 (47)). The
min c e
magnetic field required to explain the Chandra data of N157B in the highest possible radiation fields is 45µG.
A significantly better fit to the Chandra data is obtained with Γ = 2.0, but requires a much lower cut-off of
e
Ec ∼<100TeV.
dio properties have been used to infer a magnetic field believed to interact with a wind-blown cavity wall,
strength of ∼ 40 µG (33). The discrepancy between to possibly have similar ages and sizes (50,51), and
the age estimate based on the X-ray emission and the to have transferred a large fraction of their explosion
kinematicagemayindicatethatthesupernovaexploded energiesintoCRs.
within a bubble created by the progenitor star’s wind ThebrightradiosynchrotronluminosityofN132D
beforeencounteringthehighdensitymaterialitnowin- and the tentative claim of X-ray synchrotron emission
teractswith. fromthissource(52)alsoraisesthepossibilitythatthe
The γ-ray flux measured by H.E.S.S. trans- γ-ray emission is caused by inverse Compton scatter-
lates to a 1 − 10 TeV γ-ray luminosity of ing of low-energy photons. In and around N132D the
(0.9 ± 0.2) × 1035(d/50kpc)2erg/s. Assuming radiation energy density is dominated by the bright in-
that the γ-ray emission is caused by neutral-pion fraredfluxfromdustinsidetheSNR,andcanberoughly
production, this luminosity implies an energy of estimatedtobeatleastu ≈ 1.0eVcm−3. Thislep-
rad
1052(n /1cm−3)−1(d/50kpc)−2erg in relativistic tonic scenario requires that the average magnetic-field
H
protons. Ahadronicoriginoftheγ-rayemission,there- strength needs to be ∼ 20 µG, somewhat lower, but
fore, implies either a large CR-energy fraction of 17% still consistent with the equipartition value (see S1.3).
of the explosion energy, for an estimated post-shock However, this leptonic scenario critically depends on
density of n ≈ 10 cm−3 (34), or the gas density is whetherthe4-6keVX-raycontinuumemissionindeed
H
higher than the x-ray-based estimates. The latter is containsasignificantsynchrotroncomponent.
plausible given that N132D appears to interact with Whatever the emission mechanism for the γ-ray
dense, shocked interstellar clouds, seen in the optical emission from N132D, it is an exciting new γ-ray-
andtheinfraredbands(49). Itisinterestingtocompare emitting SNR, because its age lies in the gap between
N132D to the most luminous Galactic SNR detected young (< 2000yr) TeV-emitting SNRs, and old (∼>
at TeV energies, HESS J1640−465: both SNRs are 10000yr) TeV-quiet SNRs. The latter can be bright
7
pion-decay sources, but their spectra appear to be cut havebeenfound(56). Thusonefindsaconservativeup-
offabove∼10GeV.N132Dprovides,therefore,anin- perlimit,Wpp ∼<1.4×1048f−1erg,where0<f <1is
dication of how long SNRs contain CRs with energies the fraction of accelerated particles that are interacting
inexcessof1013eV. with the dense regions. This upper limit on the energy
of accelerated CR particles corresponds to 0.15f−1%
oftheexplosionenergyof1051erg.
SN1987A
Assuming a spherically-symmetric distribution of
accelerated particles, one can estimate f ∼ 0.2 with
SN1987A, the only naked-eye SN event since the Ke-
thegeometryoftheequatorialringfoundin(57). This
plerSN(AD1604),hasbeenextensivelyobservedatall translates to Wpp ∼< 9 × 1048 erg, implying that less
wavelengthsfromtheradiotothesoftγ-rayband,pro-
than 1% of the explosion energy is carried by acceler-
viding invaluable insights into the evolution of a core-
atedCRnuclei. Thisfractionisrathersmallcompared
collapseSNRinitsearlystage(53).
to typical values of ∼ 10% for young SNRs (of ages
Ithasbeensuggestedthatevenintheearlystagesof
∼ 1000 − 2000 years), but is not unreasonable for a
the SNR development, the shock wave, which is heat-
veryyoungobjectlikeSN1987A.
ing the dense circumstellar medium (CSM) structured
by stellar winds of the progenitor star, should have led
toefficientaccelerationofVHEnuclearCRs,accompa- Summary
niedbystrongmagneticfieldamplificationthroughCR-
induced instabilities (16,54). In collisions of the CRs With the deep H.E.S.S. observations of the LMC, we
withCSMparticles,γ-raysareproduced. Estimatesfor have detected three luminous examples of CR sources
theγ-rayflux(16,17)stronglydependonthemagnetic inanexternalgalaxy. Thesesourcesdetectedinγ rays
fieldtopologyandonthepropertiesofthenon-uniform include a superbubble and counterparts to the most lu-
CSM(55),makingfluxestimatesuncertainbyatleasta minous sources in the Milky Way. N157B provides a
factorof2(16). counterparttotheCrabNebula,butitselectronacceler-
Based on a nonlinear kinetic theory of CR accel- ationefficiencyisfivetimeslessthanfortheCrabneb-
eration, successfully applied to several young Galac- ula, and its magnetic field pressure is seven times less.
tic SNRs, the volume-integrated γ-ray flux at TeV en- N132Dhasbeenlongregardedanolderversionofthe
ergies, F (> 1TeV), from SN1987A was predicted brightest Galactic radio SNR Cas A, and is one of the
γ
to be rising in time, and to have reached a level of most radio-luminous SNRs known. N132D is also re-
≈ 2.5×10−13phcm−2s−1 intheyear2010(16). An markableinthatitisoneoftheoldestVHEγ-rayemit-
analysiswithdifferentassumptionsonCSMproperties tingSNRs. Withthethreedetectedsources,weincrease
and a more phenomenological approach to CR accel- ourunderstandingofthevarietyofVHEγ-raysources,
eration resulted in a predicted flux of ∼ 8 × 10−14 which will likely require observations with the future
phcm−2s−1 intheyear2013(17). TheH.E.S.S.upper Cherenkov Telescope Array (58), which should be an
limitF (>1TeV)<5×10−14phcm−2s−1 ata99% orderofmagnitudemoresensitivethanH.E.S.S.
γ
confidence level, obtained from observations made be-
tween2003and2012,beingbelowtheaforementioned
References and Notes
predictionsandafactorof3belowsimilarestimatesfor
theyear2005,thereforeplacesconstraintsonthemod-
1. J.A.Hinton,W.Hofmann,ARA&A47,523(2009).
elsdespitetheiruncertainties.
2. (H.E.S.S. collaboration) Aharonian, F. et al., A&A
The H.E.S.S. upper limit on the γ-ray flux trans-
457,899(2006).
lates into an upper limit for the γ-ray luminosity of 3. (H.E.S.S. collaboration) Aharonian, F. et al., Sci-
Lγ(> 1TeV) < 2.2×1034erg/s, which can be used ence307,1938(2005).
to derive an approximate upper limit on the energy of 4. A.Mellinger,PASP121,1180(2009).
5. D. Berge, S. Funk, J. Hinton, A&A 466, 1219
theacceleratedparticles,W ,foragivenaveragetarget
pp
(2007).
density. Multi-wavelengthstudiesofSN1987Asuggest
6. P.J.McMillan,MNRAS414,2446(2011).
that the shock at the current epoch has reached and is
7. R.P.vanderMarel,TheLocalGroupasanAstro-
interactingwiththeso-calledequatorialring,forwhich
physical Laboratory, M. Livio, T. M. Brown, eds.
gasdensitiesrangingfrom103 cm−3 to3×104 cm−3
(2006),pp.47–71.
8
8. G.Pietrzyn´ski,etal.,Nature495,76(2013). 43. Y.-H. Chu, R. C. Kennicutt, Jr., R. A. Schommer,
9. S.Kim,etal.,ApJ503,674(1998). J.Laff,AJ103,1545(1992).
10. J.Harris,D.Zaritsky,AJ138,1243(2009). 44. Q.D.Wang,E.V.Gotthelf,ApJ494,623(1998).
11. T. P. Robitaille, B. A. Whitney, ApJ 710, L11 45. M. Meyer, D. Horns, H.-S. Zechlin, A&A 523, A2
(2010). (2010).
12. W. D. Arnett, J. N. Bahcall, R. P. Kirshner, S. E. 46. J.A.Hinton,S.Funk,R.D.Parsons,S.Ohm,ApJ
Woosley,ARA&A27,629(1989). 743,L7(2011).
13. P.Sreekumar,etal.,ApJ400,L67(1992). 47. N. Bucciantini, J. Arons, E. Amato, MNRAS 410,
14. A.A.Abdo,etal.,A&A512,A7(2010).
381(2011).
15. (H.E.S.S. collaboration) Abramowski, A. et al. ,
48. B.Katz,E.Waxman,J.CosmologyAstropart.Phys.
A&A545,L2(2012).
1,18(2008).
16. E.G.Berezhko,L.T.Ksenofontov,H.J.Vo¨lk,ApJ
49. A. Tappe, J. Rho, W. T. Reach, ApJ 653, 267
732,58(2011).
(2006).
17. V.V.Dwarkadas,MNRAS434,3368(2013).
50. F.Vogt,M.A.Dopita,Ap&SS331,521(2011).
18. M. de Naurois, L. Rolland, Astroparticle Physics
51. (H.E.S.S. collaboration) Abramowski, A. et al.,
32,231(2009).
MNRAS439,2828(2014).
19. S. Ohm, C. van Eldik, K. Egberts, Astroparticle
52. X.Xiao,Y.Chen,AdvancesinSpaceResearch41,
Physics31,383(2009).
416(2008).
20. C.-C. Lu, H.E.S.S. collaboration, 33rd Interna-
53. S. Immler, K. Weiler, R. McCray, eds., SUPER-
tional Cosmic Ray Conference, International Cos-
NOVA1987A:20YEARSAFTER:Supernovaeand
micRayConference(2013).
Gamma-Ray Bursters, vol. 937 of American Insti-
21. Y.Chen,etal.,ApJ651,237(2006).
tuteofPhysicsConferenceSeries(2007).
22. P.Kaaret,etal.,ApJ546,1159(2001).
54. A. R. Bell, K. M. Schure, B. Reville, G. Giacinti,
23. T.Li,Y.Ma,ApJ272,317(1983).
MNRAS431,415(2013).
24. A.Bamba,M.Ueno,H.Nakajima,K.Koyama,ApJ
55. R. A. Chevalier, V. V. Dwarkadas, ApJ 452, L45
602,257(2004).
(1995).
25. D.A.Smith,Q.D.Wang,ApJ611,881(2004).
56. S.Mattila,etal.,ApJ717,1140(2010).
26. H. Yamaguchi, A. Bamba, K. Koyama, PASJ 61,
57. C.-Y. Ng, et al., The Astrophysical JournalLetters
175(2009).
728,L15(2011).
27. D.S.Mathewson,etal.,ApJS58,197(1985).
58. M. Actis, et al., Experimental Astronomy 32, 193
28. M.C.Lortet,G.Testor,A&A139,330(1984).
29. P. J. Kavanagh, et al., A&Ain press (avail- (2011).
able http://arxiv.org/abs/1409.6547) 59. W.B.Atwood,etal.,ApJ697,1071(2009).
60. P.L.Nolan,etal.,ApJS199,31(2012).
(2014).
61. T.Kamae,N.Karlsson,T.Mizuno,T.Abe,T.Koi,
30. S. Gabici, F. A. Aharonian, MNRAS 445, L70
ApJ647,692(2006).
(2014).
62. G.W.Pratt,M.Arnaud,A&A408,1(2003).
31. A.W.Strong,I.V.Moskalenko,V.S.Ptuskin,An-
63. E.R.Micelotta,B.R.Brandl,F.P.Israel,A&A500,
nual Review of Nuclear and Particle Science 57,
807(2009).
285(2007).
64. A. M. Atoyan, F. A. Aharonian, MNRAS 278, 525
32. A.M.Bykov,SpaceSci.Rev.99,317(2001).
(1996).
33. J.R.Dickel,D.K.Milne,AJ109,200(1995).
34. J.P.Hughes,I.Hayashi,K.Koyama,ApJ 505,732
(1998). Acknowledgements
35. F. A. Aharonian, A. M. Atoyan, A&A 351, 330
(1999).
ThesupportoftheNamibianauthoritiesandoftheUni-
36. M.S.Oey,G.Garc´ıa-Segura,ApJ613,302(2004).
37. E.A.Helder,etal.,SpaceSci.Rev.173,369(2012). versity of Namibia in facilitating the construction and
38. B.M.Gaensler,etal.,Science307,1610(2005). operation of H.E.S.S. is gratefully acknowledged, as is
39. E.Parizot,A.Marcowith,E.vanderSwaluw,A.M. the support by the German Ministry for Education and
Bykov,V.Tatischeff,A&A424,747(2004). Research(BMBF),theMaxPlanckSociety,theGerman
40. G.Ferrand,A.Marcowith,A&A510,A101(2010). Research Foundation (DFG), the French Ministry for
41. A.Abramowski,etal.,A&A537,A114(2012).
Research,theCNRS-IN2P3andtheAstroparticleInter-
42. R. Bu¨hler, R. Blandford, Reports on Progress in
disciplinaryProgrammeoftheCNRS,theU.K.Science
Physics77,066901(2014).
and Technology Facilities Council (STFC), the IPNP
9
of the Charles University, the Czech Science Founda- 1, K. Katarzyn´ski 41, U. Katz 40, S. Kaufmann 27,
tion, the Polish Ministry of Science and Higher Edu- B.Khe´lifi33,M.Kieffer20,S.Klepser39,D.Klochkov
cation, the South African Department of Science and 21, W. Kluz´niak 36, D. Kolitzus 15, Nu. Komin∗ 26,
TechnologyandNationalResearchFoundation, andby K. Kosack 23, S. Krakau 13, F. Krayzel 38, P.P. Kru¨ger
theUniversityofNamibia. Weappreciatetheexcellent 18, H. Laffon 30, G. Lamanna 38, J. Lefaucheur 33,
work of the technical support staff in Berlin, Durham, V.Lefranc23,A.Lemie`re33,M.Lemoine-Goumard30,
Hamburg, Heidelberg, Palaiseau, Paris, Saclay, and in J.-P.Lenain20,T.Lohse6,A.Lopatin40,C.-C.Lu∗ 2,
Namibiaintheconstructionandoperationoftheequip- V.Marandon2,A.Marcowith22,R.Marx2,G.Maurin
ment. The H.E.S.S. collaboration will make public the 38, N. Maxted 32, M. Mayer∗ 37, T.J.L. McComb
smoothedexcessskymapandtheassociatedcorrelated 8, J. Me´hault 30,31, P.J. Meintjes 42, U. Menzler
significance map as shown in Figure 1, together with 13, M. Meyer 28, A.M.W. Mitchell 2, R. Moderski 36,
the source spectral points, on the HESS website on M.Mohamed27,K.Mora˚28,E.Moulin23,T.Murach6,
the link to this publication: http://www.mpi-hd. M.deNaurois16,J.Niemiec25,S.J.Nolan8,L.Oakes
mpg.de/hfm/HESS/pages/publications. We 6, H. Odaka 2, S. Ohm∗ 39, B. Opitz 1, M. Ostrowski
wouldliketothankBoazKatz,EliWaxmanandRanny 10, I. Oya 6, M. Panter 2, R.D. Parsons 2, M. Paz Ar-
Budnik for their external proposal supporting observa- ribas 6, N.W. Pekeur 18, G. Pelletier 34, J. Perez 15,
tionsoftheSNRN132Dbasedontheirworkonγ-ray P.-O. Petrucci 34, B. Peyaud 23, S. Pita 33, H. Poon
emissionfromshell-typeSNRs(48). 2, G. Pu¨hlhofer 21, M. Punch 33, A. Quirrenbach 27,
S.Raab40,I.Reichardt33,A.Reimer15,O.Reimer15,
M. Renaud 22, R. de los Reyes 2, F. Rieger 2, L. Rob
The H.E.S.S. Collaboration
43, C. Romoli 3, S. Rosier-Lees 38, G. Rowell 32,
B.Rudak36, C.B.Rulten19, V.Sahakian5,4, D.Salek
A. Abramowski 1, F. Aharonian 2,3,4, F. Ait Benkhali 44, D.A. Sanchez 38, A. Santangelo 21, R. Schlickeiser
2, A.G. Akhperjanian 5,4, E.O. Angu¨ner 6,, M. Backes 13, F. Schu¨ssler 23, A. Schulz 39, U. Schwanke 6,
7, S. Balenderan 8, A. Balzer 9, A. Barnacka 10,11, S. Schwarzburg 21, S. Schwemmer 27, H. Sol 19,
Y. Becherini 12, J. Becker Tjus 13, D. Berge 14, F. Spanier 18, G. Spengler 28, F. Spies 1, Ł. Stawarz
S. Bernhard 15, K. Bernlo¨hr 2,6, E. Birsin 6, J. Biteau 10, R. Steenkamp 7, C. Stegmann 37,39, F. Stinzing 40,
16,17, M. Bo¨ttcher 18, C. Boisson 19, J. Bolmont 20, K. Stycz 39, I. Sushch 6,18, J.-P. Tavernet 20, T. Tav-
P. Bordas 21, J. Bregeon 22, F. Brun 23, P. Brun 23, ernier 33, A.M. Taylor 3, R. Terrier 33, M. Tluczykont
M.Bryan9,T.Bulik24,S.Carrigan2,S.Casanova25,2, 1,C.Trichard38,K.Valerius40,C.vanEldik40,B.van
P.M.Chadwick8,N.Chakraborty2,R.Chalme-Calvet Soelen 42, G. Vasileiadis 22, J. Veh 40, C. Venter 18,
20,R.C.G.Chaves22,M.Chre´tien20,S.Colafrancesco A. Viana 2, P. Vincent 20, J. Vink∗ 9, H.J. Vo¨lk 2,
26, G. Cologna 27, J. Conrad 28,29, C. Couturier 20, F.Volpe2,M.Vorster18,T.Vuillaume34,S.J.Wagner
Y. Cui 21, M. Dalton 30,31, I.D. Davids 18,7, B. De- 27, P. Wagner 6, R.M. Wagner 28, M. Ward 8, M. Wei-
grange 16, C. Deil 2, P. deWilt 32, A. Djannati-Ata¨ı dinger 13, Q. Weitzel 2, R. White 35, A. Wierzcholska
33, W. Domainko 2, A. Donath 2, L.O’C. Drury 3, 25, P. Willmann 40, A. Wo¨rnlein 40, D. Wouters 23,
G. Dubus 34, K. Dutson 35, J. Dyks 36, M. Dyrda 25, R.Yang2,V.Zabalza2,35,D.Zaborov16,M.Zacharias
T. Edwards 2, K. Egberts 37, P. Eger 2, P. Espigat 33, 27,A.A.Zdziarski36,A.Zech19,H.-S.Zechlin1
C. Farnier 28, S. Fegan 16, F. Feinstein 22, M.V. Fer-
nandes 1, D. Fernandez 22, A. Fiasson 38, G. Fontaine 1 Universita¨t Hamburg, Institut fu¨r Experimental-
16,A.Fo¨rster2,M.Fu¨ßling37,S.Gabici33,M.Gajdus
physik, Luruper Chaussee 149, D 22761 Hamburg,
6, Y.A. Gallant 22, T. Garrigoux 20, G. Giavitto 39, Germany, 2 Max-Planck-Institut fu¨r Kernphysik, P.O.
B. Giebels 16, J.F. Glicenstein 23, D. Gottschall 21, Box 103980, D 69029 Heidelberg, Germany, 3 Dublin
M.-H.Grondin2,27, M.Grudzin´ska24, D.Hadasch15,
Institute for Advanced Studies, 31 Fitzwilliam Place,
S. Ha¨ffner 40, J. Hahn 2, J. Harris 8, G. Heinzelmann Dublin 2, Ireland, 4 National Academy of Sciences
1, G. Henri 34, G. Hermann 2, O. Hervet 19, A. Hillert
of the Republic of Armenia, Marshall Baghramian
2, J.A. Hinton 35, W. Hofmann 2, P. Hofverberg 2,
Avenue, 24, 0019 Yerevan, Republic of Armenia ,
M. Holler 37, D. Horns 1, A. Ivascenko 18, A. Ja- 5 Yerevan Physics Institute, 2 Alikhanian Brothers
cholkowska 20, C. Jahn 40, M. Jamrozy 10, M. Janiak St., 375036 Yerevan, Armenia, 6 Institut fu¨r Physik,
36, F. Jankowsky 27, I. Jung 40, M.A. Kastendieck
10