Detection of Gamma-Ray Emission in the Region of the Supernova Remnants G296.5+10.0 and G166.0+4.3
aa r X i v : . [ a s t r o - ph . H E ] J un Detection of Gamma-Ray Emission in the Region of theSupernova Remnants G296.5+10.0 and G166.0+4.3
Miguel Araya
Space Research Centre (CINESPA), Universidad de Costa RicaSan Jos´e 2060, Costa Rica [email protected]
ABSTRACT
52 months of accumulated observations by the Large Area Telescope onboardthe
Fermi Gamma-ray Space Telescope in the region of the supernova remnantsG296.5+10.0 (PKS 1209-51/52) and G166.0+4.3 (VRO 42.05.01) are analyzed. GeVemission is detected coincident with the position of the sources at the ≃ σ and 11 σ levels above the background, respectively, for the best-fit spectral and spatial scenar-ios. The gamma-ray spectrum of the sources can be described with a power-law inenergy. G166.0+4.3 shows a soft GeV spectrum while that of G296.5+10.0 is flat (inthe νF ν representation). The origin of the gamma-ray emission from the sources isexplored. Both leptonic and hadronic mechanisms can account for the high-energyemission from G296.5+10.0, while a leptonic scenario is preferred for G166.0+4.3. Subject headings: gamma-ray: observations; ISM: supernova remnants;ISM:individuals:G296.5+10.0, G166.0+4.3, acceleration of particles, radiation mech-anisms: non-thermal
1. Introduction
Supernova remnants (SNRs) are known to be high-energy sources exhibiting non-thermalphoton spectra from radio to gamma-rays. High-energy photon emission may result from inter-actions of high-energy protons with ambient nuclei resulting in the production of neutral pionsthat decay to gamma-rays. This is generally referred to as the hadronic scenario. Leptonic mech-anisms for the production of gamma-rays in SNRs include inverse Compton up-scattering (IC)of low energy photons by high-energy electrons and non-thermal bremsstrahlung emission (e.g.,Gaisser et al. 1998) from high-energy electrons interacting with ambient particles. 2 –Certainly, high-energy electrons are known to be accelerated in SNRs from their radio and X-ray synchrotron emission; the latter is usually associated to the forward shock of young remnants(e.g., Gotthelf et al. 2001; Berezhko et al. 2002; Hwang et al. 2002; Rho et al. 2002; Long et al. 2003;Vink & Laming 2003; Berezhko & V¨olk 2004). Particles are thought to gain energy through firstorder Fermi acceleration (Bell 1978; Blandford & Eichler 1987), also known as diffusive shockacceleration (DSA) after crossing the shock front of the supernova explosion, which occurs manytimes as the particles move in the presence of magnetic fields that result either from shock compres-sion of the interstellar field or from amplification by cosmic ray instabilities (see Schure et al. 2012,for a recent review).It is thought that a considerable ( − ) fraction of the SNR energy can be transferred toparticles via DSA, which is one of the reasons why they are considered the main source of Galac-tic cosmic rays. However, identifying and separating the emission from leptonic and hadroniccomponents is a challenging task. Some SNRs interacting with molecular clouds, such as W28,W49B, W51C and G8.7-0.1, show gamma-ray emission that seems to favour a hadronic ori-gin (Abdo et al. 2010b; Abdo et al. 2010c; Abdo et al. 2009a; Ajello et al. 2012). Other studiesshow that some young SNRs with a relatively hard GeV photon spectrum are probably leptonic-dominated, such as RX J1713.7-3946 (Abdo et al. 2011) and RX J0852.0-4622 (Tanaka et al. 2011).The softer spectrum of other young sources such as Cas A (Abdo et al. 2010a; Araya & Cui 2010)and Tycho SNR (Giordano et al. 2012) might also favour a hadronic origin. Recent observationsof the SNRs IC 443 and W44 seem to have confirmed the presence of cosmic ray protons (moregenerally, ions) from their characteristic spectrum below ∼ MeV (Ackermann et al. 2013).Observations from the recently-launched
Fermi satellite (Atwood et al. 2009) have contributedto form a more consistent picture of gamma-ray emission from SNRs and particle acceleration inSNR shocks (e.g., Caprioli 2012). Despite the advances in the understanding of SNR properties,important limitations are still common for gamma-ray studies. For example, the broad point spreadfunction (PSF) of the Large Area Telescope (LAT) onboard
Fermi and the high Galactic radiationbackground often present challenges for data analysis. In this paper, a study of
Fermi
LAT ob-servations is carried out for two remnants located outside the Galactic plane, where the expectedGalactic diffuse level is lower.The sources are G296.5+10.0 and G166.0+4.3, shell-type SNRs with radio extensions ′ × ′ and ′ × ′ , respectively. The distance adopted here for G296.5+10.0 is 2.1 kpc (Giacani et al. 2000).In the case of G166.0+4.3, a distance estimate of . ± . kpc (Landecker et al. 1989) is used.G166.0+4.3 is composed of two parts as seen in radio images: a regular shell to the Eastand a region expanding into a low density medium (the extended ‘wing’) to the West. The mor-phology of this remnant possibly results from the shock encountering a density discontinuity(Pineault et al. 1985) in the interstellar medium (ISM). The X-ray emission is only present in the 3 –interior and peaks towards the West wing (Burrows & Guo 1994), where the shock has encountereddenser material (Landecker et al. 1989). HI observations have shown ISM features interacting withthe SNR (Landecker et al. 1989).G296.5+10.0 is a barrel-shaped SNR. The detected X-ray emission is in good correspon-dence with the radio. The morphology of barrel-shaped SNRs might be the result of their in-teractions with the ISM material or the magnetic field in the ISM, or result from the intrinsicproperties of the outburst and later interaction with the ISM (e.g., Kesteven & Caswell 1987). Lo-cated relatively far above the Galactic plane, G296.5+10.0 is possibly surrounded by low-density,uniform ISM. However, HI observations have revealed three clouds that are associated with theSNR (Giacani et al. 2000): a long, broad structure of size ◦ × ′ to the northeast, a cloud alongthe southwestern limb, near the Galactic coordinates ( l, b ) ∼ (296 ◦ . , +9 ◦ . (volume den-sity ∼ cm − ), and the HI cloud across the eastern limb close to the brightest filaments near ( l, b ) ∼ (296 ◦ . , +9 ◦ . (density ∼ cm − ).The structure of this paper is as follows. In Section 2 the gamma-ray data reduction is dis-cussed, paying attention to the morphology and spectral properties of the emission detected in thedirection of the SNRs. In Section 3, the non-thermal spectral energy distributions (SED) are mod-eled with different emission mechanisms. The discussion of results and final remarks are given inSection 4.
2. Observations2.1. Radio data
The radio data points for G296.5+10.0 and G166.0+4.3 are obtained from the literature (Milne & Haynes 1994;Leahy & Tian 2005; Leahy & Tian 2006). The radio spectrum of the sources used here can be ac-counted for by optically-thin synchrotron emission from cosmic ray electrons. The particle spec-trum responsible for the radio emission is a power-law in energy ( ∝ ǫ − s with ǫ the particle energy).From standard results concerning the synchrotron emission by a power-law population of electrons,the radio data imply s ≃ and ≃ . for G296.5+10.0 and G166.0+4.3, respectively. 4 – Fermi
LAT data
Fermi
LAT data taken between 04 August 2008 and 24 January 2013 were analyzed with thestandard software
ScienceTools version v9r27p1 released April 18, 2012. Several selection criteriaare applied to events, including the selection of events with high probability of being gamma-rays(the so-called Pass 7 Source class) and with a reconstructed zenith angle less than 100 ◦ to avoidcontamination from gamma rays from Earth’s limb. Time intervals when the spacecraft is withinthe range of rocking angles used during nominal sky-survey observations (the rocking angle is lessthan 52 ◦ ) are also selected. The spectral analysis is further restricted above 200 MeV to avoiduncertainties in the effective area and broad PSF at lower energies, and below 100 GeV due tolimited statistics. The same data selection criteria were used for regions near SNRs G296.5+10.0and G166.0+4.3.Events within a square region of 14 ◦ × ◦ of the catalogued positions of the SNRs G296.5+10.0and G166.0+4.3, RA (J2000)= 12 h m s , Dec (J2000)= -52 ◦ ′ ′′ and RA (J2000)= 05 h m s ,Dec (J2000)= 42 ◦ ′ ′′ , respectively, are included in the analysis. This is necessary to account forthe large PSF of the LAT. The emission model that is used in the analysis includes the positions andspectral parameter values of the sources within this region that are found in the LAT 2-year SourceCatalog (Nolan et al. 2012). In the case of the region containing G296.5+10.0 nearby extendedsources (known as Cen A and MSH 15-52) are modeled with spatial templates provided within the ScienceTools . The data are binned in sky coordinates with the tool gtbin in square bins of size 0. ◦ Fermi
LAT data are explored by means of a maximumlikelihood analysis using the tool gtlike . The likelihood is defined as the probability of obtainingthe data given an input spatial and spectral model for the sources. The starting point for the fittingprocedure is obtained from the LAT 2-year Source Catalog, as mentioned above. The currentlyreleased instrument response functions ( P SOU RCE V ) are used throughout the analysis, aswell as the latest galactic and extragalactic diffuse background components (as specified in thefiles gal 2yearp7v6 v0.fits and iso p7v6source.txt , respectively). The spectral parameters of thecatalogued sources beyond 7 ◦ from the position of G296.5+10.0 and G166.0+4.3 are kept fixedto the values reported in the catalog. The fit is performed with the optimizer NEWMINUIT untilconvergence is achieved.In the case of the SNR G166.0+4.3, the previously detected point source 2FGL J0526.6+4308is removed from the model. This source is found at the position RA (J2000)= 05 h m s , Dec See http://fermi.gsfc.nasa.gov/ssc . See http://fermi.gsfc.nasa.gov/ssc/data/analysis/scitools/binned likelihood tutorial.html . ◦ ′ ′′ within the remnant’s shell. Therefore, no source at the positions of theSNRs is originally included in this part of the analysis. The resulting model is referred to as thenull hypothesis and it is thus a model of the background.The significance of a source is estimated by means of a test statistic (TS) defined as − log( L /L ), where L and L are the maximum likelihood values for the null hypothesis and fora model including the additional source, respectively (Mattox et al. 1996). In the limit of a largenumber of counts, despite some caveats (Protassov et al. 2002), the detection significance of thesource is roughly given by √ TS, and therefore TS = 25 is usually considered the threshold valuefor detection (corresponding to a significance of σ for one degree of freedom).The best-fit models that result from the maximization procedures are used to generate mapsof the background emission which are smoothed with a boxcar of length ◦ . . When subtractedfrom the corresponding smoothed count maps, the resulting residuals maps can be used to visualizethe disagreement between the observed counts and the model. These background-subtracted mapsare obtained for different energy bands and then divided by the square root of the background,resulting in the signal-to-noise maps that are shown in Figs. 1 and 2 for the two sources.Representative radio contours of G296.5+10.0 and G166.0+4.3 obtained from observations ofthe Green Bank catalog of radio sources (GB6, 4860 MHz, Gregory et al. 1996) and the Wester-bork Northern Sky Survey (WENSS, 325 MHz, Rengelink et al. 1997), respectively, are overlaidonto the signal-to-noise maps. Residual emission coincident with the positions of the sources isapparent. Maximum likelihood analyses are next performed in order to quantify the significanceand test the spatial morphology and spectral parameters of the emission. As shown in Figs. 1 and 2 only events above 0.5 GeV are used for morphology studies, totake advantage of the narrower PSF at higher energies. Several hypotheses for the morphology ofthe residual gamma-ray emission are tested.
G296.5+10.0
Fig. 1 shows the signal-to-noise maps obtained for the region around this SNR with the corre-sponding radio contours of the bipolar shell for two energy bands: 2-7 GeV and 7-100 GeV. In theenergy range above 2 GeV the emission seems to come in part from a region towards the interiorof the remnant, and it becomes more prominent towards the western limb of the SNR at the highestenergies. This excess seems to follow the western radio emission. 6 –In order to evaluate the significance of the emission, a ‘TS map’ is calculated above 500 MeVfor the region near the SNR. The map is obtained by evaluating the TS value of a point source thatis moved in a predefined grid. Based on these results, the gamma-ray morphology is assessed byfitting different spatial models for G296.5+10.0: (a) a spatial template obtained from the radio GB6observation, (b) a uniform disk template and (c) a point-source located at the position of maximumvalue in the TS map. A simple power-law model is assumed for the energy spectrum. For the disk,the TS values with respect to the null hypothesis are evaluated for different locations and sizes.The radius of the disk template is systematically increased in steps of 0 ◦ .1 and the position of itscentroid is changed within the interior of the remnant shell.The TS values obtained for the radio template, uniform disk and point-source hypotheses are28, 36 and 25, respectively, as shown in Table 1. The best-fit disk radius and centroid locationare 0 ◦ .6 ± ◦ . and RA (J2000)= 12 h m , Dec (J2000)= -52 ◦ ′ . The error of the centroid is ◦ . at 95 % confidence level. Although the fit with the disk template has a higher TS with respectto the fit with the radio template (the difference being ∆ TS = 8 ), the corresponding model has3 additional parameters and therefore both can be considered statistically similar descriptions ofthe source morphology. The disk hypothesis is an improvement with respect to the point-source( ∆ TS = 11 , see Table 1), the analysis thus indicates that the source extension is detected at the σ level (Lande et al. 2012). The disk hypothesis will then be adopted for spectral analysis hereand as explained later on the systematic uncertainties of changing the hypothesis for the sourcemorphology will be considered.As an additional test to evaluate the amount of background emission that may have not beenaccounted for in the model, new fits are performed by moving the disk template in different az-imuthal positions outside and around the radio contours. No significant emission is detected out-side/around the shell of the SNR. G166.0+4.3
Fig. 2 shows a correspondence between the gamma-ray signal and the radio contours of the shellof the SNR for two energy bands: 0.5-2 GeV and 2-100 GeV. For the latter energy band, thegamma-ray emission seems to become more significant near the west, where the extended wing islocated.Next, two hypotheses for the gamma-ray emission are tested: (a) one following a spatial tem-plate obtained from the WENSS 325MHz observation, and (b) a point-source hypothesis, locatedat the position of the 2-year LAT Catalog source 2FGL J0526.6+4308. A simple power-law isused to model the energy spectrum in both cases. Table 1 shows the TS values obtained for the twospatial hypotheses with respect to the null hypothesis (no emission from G166.0+4.3). 7 –The TS value for the radio contour hypothesis (91) increases with respect to the TS value forthe point source (83). The radio template scenario for the emission may be considered a slightimprovement in the fit with respect to the point source and has been adopted for the spectralanalysis. Association
There are no known radio pulsars near the position of the SNRs. G296.5+10.0 is known to harbora radio-quiet neutron star left from the supernova explosion (Zavlin et al. 2000), but these compactobjects do not show gamma-ray emission. Moreover, the results shown here support the presenceof an extended source of gamma-ray emission. The CRATES catalog of flat spectrum radio sources(Healey et al. 2007) contains a few sources near the contours of G296.5+10.0. The three closestare located at the positions ( l, b ) ∼ (295 ◦ . , +10 ◦ . , (296 ◦ . , +10 ◦ . , (296 ◦ . , +10 ◦ . .These sources are found towards the north of the shell, on average around ◦ . from the peak ofthe 7-100 GeV signal-to-noise map. In what follows, then, the gamma-ray emission seen in thedirection of G166.0+4.3 and G296.5+10.0 is assumed to be associated with the SNRs. A binned likelihood analysis is performed in the energy band 0.2-100 GeV for both sourcesusing the best-fit spatial templates found in Section 2.2.1. Using 10 energy bins per decade in theexposure map calculation and assuming power-law spectra, the resulting TS values for the sourcesG166.0+4.3 and G296.5+10.0 in this energy band are 136 and 36, corresponding to significancesof roughly σ and σ over the background for 2 and 5 degrees of freedom, respectively.In order to probe for curvature in the spectra, a log-parabolic spectral shape, dN/dE = N ( E/E b ) − ( α + β log ( E/E b ) is tested in the fit for both sources, but no significance of a deviationfrom a power-law spectral distribution can be claimed for either one. The best-fit spectral parame-ters are summarized in Table 2.The spectral indices are very different for both sources, G166.0+4.3 having a much steeperspectrum. No significant emission for this source is detected above 10 GeV. On the other hand,G296.5+10.0 would be among the dimmest gamma-ray emitting SNRs detected so far, with aluminosity of (3 . ± . × ergs/s (1-100 GeV), about three times that of the Cygnus Loop(Katagiri et al. 2011). A similar TS value is obtained for a disk morphology hypothesis for G166.0+4.3. Since this hypothesis introduces3 additional free parameters, it is not used here. σ an upper limit for the flux is derived. Several sources of systematic errors are consideredfor these bins: a) the effect of changing the morphology of the emission as explained in Section2.2.1; b) the uncertainty of the Galactic diffuse emission, which is evaluated by artificially varyingthe best-fit value of the normalization of the Galactic level by ± in each bin (as done by Abdoet al. 2009a); and c) the systematic uncertainty in the effective area, which is energy-dependentand given by at 100 MeV, at 560 MeV and at 10 GeV (Abdo et al. 2009b). Fig. 3shows the broadband SEDs and gamma-ray data points with the resulting statistical and systematicerrors added in quadrature.
3. Emission Model
The electron distributions used here are broken power-laws of the form N e ( ǫ ) = Kǫ − s (cid:18) ǫǫ br (cid:19) ! − δ extending up to a maximum energy ǫ max . The break ( ǫ br ) and maximum energy as well as theparameter δ are varied to fit the shape of the gamma-ray spectra. The proton distributions used tomodel the data are simple power-laws, N p ∝ ǫ − s p , based on the fact that the gamma-ray spectra ofthe sources are also power-laws.Instead of applying a model to follow the evolution of the SNRs and predict the break energiesand magnetic field values, these parameters are derived under the assumption that the nature of thegamma-ray emission is either leptonic or hadronic, in other words, the necessary conditions thatproduce either outcome are explored. It is noted also that a full exploration of the parameters thatreproduce the gamma-ray data is beyond the scope of this work and only a representative set ofparameters is presented for each source.The volume occupied by the particles is the volume of the SNR and the supernova explo-sion energy is set to E SNR = 10 ergs. Gamma-ray emission mechanisms include IC of cos-mic microwave background (CMB) photons, bremsstrahlung emission from electrons and neutralpion decay, which depend on the target material density, denoted as n H (cm − ). For both SNRs,leptonic-dominated and hadronic-dominated scenarios are considered. The gamma-ray flux result-ing from hadron interactions is calculated as in Kamae et al. (2006) 9 – As noted before, the spectral index of electrons is fixed from radio observations to s = 2 . Theremnant is approximated as a sphere of radius 0 ◦ .7 (at a source distance of 2.1 kpc this is equivalentto 26 pc). The gamma-ray emission can be accounted for by IC on the CMB, as shown in Fig. 3, witha ‘softening’ given by δ = 0 . , an electron break energy ǫ br ≃ GeV, a magnetic field of 60 µ G, a total electron energy of . × ergs and ambient density 0.05 cm − or less. Althoughnot shown, a bremsstrahlung-dominated scenario results with a magnetic field B ≃ µ G, totalelectron energy . × ergs and ambient density 1 cm − . The maximum electron energy adoptedis ǫ max = 1 . TeV.
Given the relatively low statistics of the observation and the high systematic errors at sev-eral hundreds of MeV, the gamma-ray index is not well constrained. If a standard power-lawproton distribution is assumed with index s p = 2 , a total cosmic-ray proton energy of . × ( n H / cm − ) − ergs is needed to account for the high-energy SED. The proton energy spec-trum shown in Fig. 4 extends from the pion production threshold to 250 GeV, although Fermi
LATdata are not inconsistent with higher proton energies, as there is no evidence for a spectral cutoff.The magnetic field is µ G in this case and the total electron energy is . × ergs, decreasingwith increasing magnetic field. The spectral index of electrons, again from radio observations, is fixed at s = 1 . . The largerestimated distance and smaller angular extension of G166.0+4.3 compared to G296.5+10.0 yield,however, a similar intrinsic remnant size. The remnant is also approximated as a sphere of radius26 pc. 10 – In the case that the emission is mainly from IC on the CMB, an electron spectral ‘softening’given by the parameter δ = 0 . , particle break energy and maximum energy of 50 GeV and 500GeV, respectively, are necessary. The magnetic field and total electron energy are ≃ µ G and . × ergs, respectively. The value for the density used for calculating the bremsstrahlungemission is n H = 0 . cm − , although a value lower than this is probably more realistic, as X-rayobservations show (Burrows & Guo 1994). The gamma-ray data requires a steep cosmic-ray proton spectrum (the photon spectral indexis 2.7). Fig. 4 shows a scenario where the gamma-ray emission is accounted for by a power-lawproton distribution with index s p = 2 . . The magnetic field is 50 µ G and the total proton energy (2 . +2 . − . ) × ( n H / cm − ) − ergs, considering the uncertainty in the distance ( . ± . kpc,Giacani et al. 2000). The maximum particle energy is not well-constrained since the contributionfrom the highest energy particles is less important for such a steep distribution, the value used inFig. 4 is ≃ GeV. Fig. 4 shows a scenario with n H = 1 cm − , a total proton energy of ≃ of the available SNR energy, a total electron energy of . × ergs and a bremsstrahlung fluxcalculated also with an ambient density n H = 1 cm − . Note that if instead a much lower ambientdensity is adopted as implied by X-ray observations ( n H ∼ . cm − , Burrows & Guo 1994), thegamma-ray observations require an unrealistically large total proton energy.
4. Discussion and Conclusions
Excess gamma-ray emission in the region of the SNRs G166.0+4.3 and G296.5+10.0 is re-vealed by analysis of accumulated observations from the
Fermi
LAT. In the case of G166.0+4.3, ahypothesis for the emission from the radio shell and wing (325 MHz) is slightly preferred over thepreviously detected point source located in the shell. For SNR G296.5+10.0 extended gamma-raymorphology is also preferred over point source emission (see Table 1).The gamma-ray spectra of the sources are very different. They can be described by simplepower-laws with no clear evidence of curvature. The spectrum of G166.0+4.3 is very steep (photonindex . ± . , statistical errors only) while that of G296.5+10.0 (photon index . ± . ,statistical errors only) is somewhat harder than that expected for a standard test-particle DSAspectrum with a power-law index of 2. The broad-band SEDs of both sources can be either of 11 –leptonic or hadronic origin.The hadronic interpretation for the emission from G166.0+4.3 presents, however, some diffi-culties. The steep proton spectrum required to account for the gamma-ray SED is hard to under-stand in the context of standard shock acceleration, and the ambient density required is much higherthan that obtained from X-ray observations (Burrows & Guo 1994). The contradiction could besolved if the SNR interacts with irregular ISM with clumps of high-density material located in alow-density environment, as has been proposed for other SNRs (Castro & Slane 2010). It is be-lieved, based on HI observations of the region, that the SNR has interacted with material in itsenvironment (Landecker et al. 1989). However, it is unclear whether the density of the target ma-terial is enough to account for the flux and even if it was, it would still be hard to explain the steepspectrum.Another possibility is that the emission from G166.0+4.3 is leptonic. As shown in Fig. 3,the IC on the CMB level is enough to account for the Fermi
LAT points with reasonable SNR pa-rameters, and it is compatible with the low ambient density implied by X-ray observations, whichaffects the bremsstrahlung level only. In this scenario, the observed steep gamma-ray spectrum isproduced mainly by the highest-energy synchrotron-emitting electrons in the tail of the distribu-tion, above a particle break of ≃ GeV according to the model. The leptonic emission scenariois compatible with the observed correspondence between the radio contours of the source and thegamma-ray maps, as seen in Fig. 2.From these considerations, the leptonic emission scenario for the SED of G166.0+4.3 is foundto be a more natural explanation. Future observations might allow performing a more detailed anal-ysis considering different electron populations, perhaps disentangling the properties of particles inthe wing and shell, as well as studying spatial variations of spectral parameters.In the case of G296.0+10.0, the hypothesis of a uniform disk, containing most of the ra-dio shell, of radius 0 ◦ .6 ± ◦ . and centroid location RA (J2000)= 12 h m , Dec (J2000)= -52 ◦ ′ (Galactic coordinates (l,b) ∼ (296 ◦ . , +9 ◦ . ) was adopted here. However, the residuals in Fig.1 suggest that the gamma-ray morphology might be more complicated than the scenarios exploredhere would imply.At the highest energies, the peak of the signal-to-noise maps (see Fig. 1) is located to the eastof the western radio contours, where synchrotron emission is also present (Gregory et al. 1996).This is interesting since both the X-ray and radio fluxes are highest in the eastern hemisphere.It is possible, in the context of a hadronic scenario, that at least part of the gamma-ray emissionis produced by high-energy particles that interact with the dense ( n H = 13 cm − ) southwesterncloud seen (Giacani et al. 2000) in contact with the shell, although there is another dense cloudalso interacting with the shell in the east. The SED parameters derived for leptonic and hadronic 12 –scenarios are both physically reasonable.The required mean density for a bremsstrahlung-dominated scenario for G296.5+10.0 ( n H =1 cm − ) might be in conflict with previous observations ( . cm − and . cm − ; Kellett et al. 1987;Matsui et al. 1988, respectively). Roger et al. (1988) argue that this SNR is in the adiabatic (Sedov-Taylor) phase of evolution which is consistent with the remnant size, the observed low ambientdensity and thus an IC origin for the gamma-ray emission. Furthermore, the particle break andmagnetic field required by the gamma-ray data ( ǫ br ≃ GeV and B ≃ µ G) are consistentwith a synchrotron cooling break (see, e.g., Tanaka et al. 2008) and a reasonable SNR age ( ∼ yr, Vasisht et al. 1997). If the emission from this SNR is mainly leptonic, an IC origin seems thenmore plausible than a bremsstrahlung origin.This research has made use of NASA’s Astrophysical Data System and of the SIMBADdatabase, data from the WENSS team (Rengelink et al. 1997) and from the NRAO. Financial sup-port from Universidad de Costa Rica is acknowledged. Valuable observations from the anonymousreferee helped improve the quality of this work substantially. 13 –
REFERENCES
Abdo, A. A., et al., 2011, ApJ, 734, 28Abdo, A. A., et al., 2010a, ApJ, 710, L92Abdo, A. A., et al., 2010b, ApJ, 722, L1303Abdo, A. A., et al., 2010c, ApJ, 718, 348Abdo, A. A., et al. 2009a, ApJ, 706, L1Abdo, A. A., et al. 2009b, Physical Review Letters, 103, 251101Ackermann, M., et al. 2013, Science, 339 (6121), 807Ajello, M., et al. 2012, ApJ, 744, 80Araya, M. & Cui, W. 2010, ApJ, 720, 20Atwood, W. B., et al. (the Fermi LAT Collaboration), 2009, ApJ, 697, 1071Bell, A. R., 1978, Mon. Not. R. astr. Soc., 182, 147Berezhko, E. G., & V¨olk, H. J. 2004, A&A, 419, L27Berezhko, E. G., Ksenofontov, L. T., & V¨olk, H. J. 2002, A&A, 395, 943Blandford, R. D., & Eichler, D., 1987, PHYSICS REPORTS, 154, No. 1, 1-75Burrows, D. N. & Guo, Z. 1994, ApJL, 421, 19Gaisser, T. K., Protheroe, R. J. & Stanev, T. 1998, ApJ, 492, 219Caprioli, D. 2012, JCAP, 07, 038Castro, D. & Slane, P. 2010, ApJ, 717, 372Giacani, E. B., Dubner, G. M., Green, A. J., Goss, W. M. & Gaensler, B. M. 2000, ApJ, 119, 281Giordano, F. et al. 2012, ApJ, 744, L2Gotthelf, E. V., Koralesky, B., Rudnick, L., Jones, T. W., Hwang, U., Petre, R. 2001, ApJ, 552,L39Gregory, P. C., Scott, W. K., Douglas, K., & Condon, J. J. 1996, ApJS, 103, 427 14 –Healey, S. E., Romani, R. W., Taylor, G. B., Sadler, E. M., Ricci, R., Murphy, T., Ulvestad, J. S. &Winn, J. N. 2007, ApJSS, 171, 61Hwang, U., Decourchelle, A., Holt, S. S., Petre, R., 2002, ApJ, 581, 1101Kamae, T., Karlsson, N., Mizuno, T., Abe, T., & Koi, T., 2006, ApJ, 647, 692Katagiri, H., Tibaldo, L., Ballet, J., Giordano, F., Grenier, I. A., Porter, T. A., Roth, M., Tibolla,O., Uchiyama, Y. & Yamazaki, R. 2011, ApJKellett, B. J., Branduardi-Raymont, G., Culhane, J. L., Mason, I. M., Mason, K. O. & Whitehouse,D. R. 1987, MNRAS, 225, 199Kesteven, M. J. & Caswell, J. L. 1987, A&A, 183, 118Lande, J., Ackermann, M., Allafort, A., Ballet, J., Bechtol, K., Burnett, T. H., Cohen-Tanugi, J.,Drlica-Wagner, A., Funk, S., Giordano, F., Grondin, M.-H., Kerr, M. & Lemoine-Goumard,M. 2012, ApJL, 756, 5Landecker, T. L., Pineault, S., Routledge, D. & Vaneldik, J. F. 1989, MNRAS, 237, 277Leahy, D. & Tian, W. 2005 A&A, 440, 929Leahy, D. & Tian, W. 2006 A&A, 451, 991Long, K. S., Reynolds, S. P., Raymond, J. C., Winkler, P. F., Dyer, K. K., & Petre, R. 2003, ApJ,586, 1162Nolan, P. L., et al. 2012, ApJS, 199, 31Matsui, Y., Long, K. S. & Tuohy, I. R. 1988, in
IAU Colloquium 101, Supernova Remnants andthe Interstellar Medium , ed. R. S. Roger and T. L. Landecker (Cambridge: CambridgeUniversity Press), p. 157Mattox, J. R., et al., 1996, ApJ, 461, 396Milne, D. K. & Haynes, R. F. 1994, MNRAS, 270, 106Pineault, S., Pritchet, C. J., Landecker, T. L., Routledge, D. & Vaneldik, J. F. 1985, A&A, 151, 52Protassov, R., van Dyk, D. A., Connors, A., Kashyap, V. L., & Siemiginowska, A. 2002, ApJ, 571,545Rengelink, R. B., Tang, Y., de Bruyn, A. G., Miley, G. K., Bremer, M. N., Roettgering, H. J. A. &Bremer, M. A. R., 1997, A&AS, 124, 259 15 –Rho, J., Dyer, K.K., Borkowski, K.J., Reynolds, S.P., 2002, ApJ, 581, 1116Roger, R. S., Milne, D. K., Kesteven, M. J., Wellington, K. J. & Haynes, R. F. 1988, ApJ, 332, 940Schure, K. M., Bell, A. R., Drury, L. O’C & Bykov, A. M. 2012, 1203.1637v1Tanaka, T., Uchiyama, Y., Aharonian, F. A., Takahashi, T., Bamba, A., Hiraga, J. S., Kataoka, J.,Kishishita, T., Kokubun, M., Mori, K., Nakazawa, K., Petre, R., Tajima, H. & Watanabe,S. 2008, ApJ, 685, 988Tanaka, T., et al., 2011, ApJ, 740, L51Vasisht, G., Kulkarni, S. R., Anderson, S. B. & Hamilton, T. T. 1997, ApJ, 476, L43Vink, J. & Laming, J. M. 2003, ApJ, 584, 758Zavlin, V. E., Pavlov, G. G., Sanwal, D. & Tr¨umper, J. 2000, ApJ, 540, L25
This preprint was prepared with the AAS L A TEX macros v5.2.
16 –Table 1: Test Statistic for Different Spatial Models Compared to the Null Hypothesis (0.5-100GeV) Test Statistic Additional Degrees of FreedomG296.5+10.0Null hypothesis (background only) 0 0Uniform disk a
36 5Point source 25 44860 MHz radio continuum 28 2G166.0+4.3Null hypothesis (background only) 0 0Point source b
83 2325 MHz radio continuum 91 2 a Best fit radius and centroid position 0 ◦ .6 ± ◦ . , RA (J2000)= 12 h m , Dec (J2000)= -52 ◦ ′ ( ± ◦ .2 at 95 % confi-dence level). b Found in the 2-year LAT catalog (Nolan et al. 2012).
Table 2: Best-Fit Spectral Parameters a (0.2-100 GeV)Spectral Index Integrated Photon Flux (cm − s − ) TSG296.5+10.0 . ± .
13 (3 . ± . × − . ± . . ± . × − a Assuming power-law spectra and best-fit spatial template. Only statistical errors are shown.
17 – (a) (b)
Fig. 1.— Signal-to-noise maps in the region of the SNR G296.5+10.0 smoothed with a boxcar oflength 0 ◦ .5 in two energy bands: (a) 2-7 GeV, and (b) 7-100 GeV (see text), and green contours ofa GB6 4860 MHz observation of the SNR. The position of the CRATES sources are representedby (yellow) crosses and Galactic coordinates are shown in degrees. The insets in the bottom leftcorner show the smoothed PSF. 18 – . deg . . (a) (b) Fig. 2.— Signal-to-noise maps in the region of the SNR G166.0+4.3 smoothed with a boxcar oflength 0 ◦ .5 in two energy bands: (a) 0.5-2 GeV and (b) 2-100 GeV. Overlaid are green contoursof the radio WENSS observation (325 MHz) of the SNR. The magenta cross indicates the positionof the source 2FGL J0526.6+4308. Galactic coordinates are shown in degrees and the PSF is onlyshown for the higher energy map. ν F ν ( e r g s / c m / s ) −14 −13 −12 −11 −10 −9 ν (Hz) (a) ν F ν ( e r g s / c m / s ) −14 −13 −12 −11 −10 −9 ν (Hz) (b) Fig. 3.— Leptonic emission model: (a) G296.5+10.0, (b) G166.0+4.3. The emission componentsare: synchrotron (solid line), IC-CMB (dotted line) and non-thermal bremsstrahlung (dashed line,shown only above ∼ × Hz). The dark solid line represents the total gamma-ray emission.Blue squares are obtained from the LAT observation presented in this paper. 19 – ν F ν ( e r g s / c m / s ) −14 −13 −12 −11 −10 −9 ν (Hz) (a) ν F ν ( e r g s / c m / s ) −14 −13 −12 −11 −10 −9 ν (Hz) (b) Fig. 4.— Hadronic emission model: (a) G296.5+10.0, (b) G166.0+4.3. The emission componentsare: synchrotron (solid line), IC-CMB (dotted line), non-thermal bremsstrahlung (dashed line,shown only above ∼ × Hz) and hadronic emission from π0