ebook img

Deep Observation of the Giant Radio Lobes of Centaurus A with the Fermi Large Area Telescope PDF

0.96 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Deep Observation of the Giant Radio Lobes of Centaurus A with the Fermi Large Area Telescope

Astronomy&Astrophysicsmanuscriptno.cena˙v6 (cid:13)c ESO2012 January6,2012 Deep Observation of the Giant Radio Lobes of Centaurus A with the Fermi Large Area Telescope Rui-zhiYang1,2,NarekSahakyan3,4,EmmadeOnaWilhelmi1,FelixAharonian1,5,andFrankRieger1 1 Max-Planck-Institutfu¨rKernphysik,P.O.Box103980,69029Heidelberg,Germany 2 Key Laboratory of Dark Matter and Space Astronomy, Purple Mountain Observatory, Chinese Academy of Sciences, Nanjing, 210008,China 3 UniversityofRomeSapienzaandICRANet,Dip.Fisica,p.leA.Moro2,00185-Rome,Italy 2 4 InstituteforPhysicalResearch,NASofArmenia,Ashtarak-2,0203,Armenia 1 5 DublinInstituteforAdvancedStudies,31FitzwilliamPlace,Dublin2,Ireland 0 2 Preprintonlineversion:January6,2012 n a ABSTRACT J Thedetection of high energy (HE)g -ray emissionup to ∼3GeV from thegiant lobes of the radiogalaxy Centaurus A hasbeen 5 recentlyreportedbytheFermi-LATCollaborationbasedontenmonthsof all-skysurveyobservations. Adatasetmorethanthree timeslargerisusedheretostudythemorphologyandphotonspectrumofthelobeswithhigherstatistics.Thelargerdatasetresults ] E inthedetectionofHEg -rayemission(upto∼6GeV)fromthelobeswithasignificanceofmorethan10and20s fortheNorthand H theSouthlobe,respectively.Basedonadetailedspatialanalysisandcomparisonwiththeassociatedradiolobes,wereportevidence forasubstantialextensionoftheHEg -rayemissionbeyondtheWMAPradioimageinthecaseoftheNorthernlobeofCenA.We h. reconstructthespectralenergydistribution(SED)ofthelobesusingradio(WMAP)andFermi-LATdatafromthesameintegration p region.Theimplicationsarediscussedinthecontextofhadronicandleptonicscenarios. - o Keywords.Gammarays:galaxies r t s a 1. Introduction [ The bright, nearby radio galaxy Centaurus A (Cen A; NGC 5128) has been extensively studied from radio to very-high-energy 1 (VHE)g -rays(e.g.,seeIsrael1998,Steinle2010forreviews).Itsuniqueproximity(d∼3.7Mpc;Ferrareseetal.2007)andpeculiar v morphologyallowforadetailedinvestigationofthe non-thermalaccelerationandradiationprocessesoccurringbothinitsactive 7 nucleusanditsrelativisticoutflows.AtradiofrequenciesCenArevealsgiantstructures,theso-called”lobes”,withatotalangular 1 sizeof∼10◦(Shain1958,Burnsetal.1983),correspondingtoaphysicalextensionof∼600kpc(d/3.7Mpc). 2 1 At high-energy(HE;200 MeV<E<100 GeV) Fermi-LAT hasrecently detected g -rayemission fromboth the core (i.e., within . ∼0.1◦)andthegiantradiolobesofCenA(Abdoetal.2010a,Abdoetal.2010b):Ananalysisoftheavailable10-monthsdataset 1 revealsapoint-likeemissionregioncoincidentwiththepositionoftheradiocoreofCenA,andtwolargeextendedemissionregions 0 detectedwithasignificanceof5and8s fortheNorthernandtheSouthernlobe,respectively.TheHEemissionfromthecoreextends 2 upto∼10GeVanditwelldescribedbyapower-lawfunctionwithphotonindex∼2.7.Itcanbesuccessfullyinterpretedasorigi- 1 natingfromsynchrotronself-Compton(SSC)processesintheinnermostpartoftherelativisticjet.However,asimpleextrapolation : v oftheHEcorespectrumtotheTeVregimetendstounder-predicttheTeVfluxobservedbyH.E.S.S.(Aharonianetal.2009),afact Xi thatmayindicateanadditionalcontributionrelatedtoe.g.non-thermalmagnetosphericprocessesemergingatthehighestenergies (seeRieger2011forreview).TheextendedHEemissionregions,ontheotherhand,seemmorphologicallycorrelatedwiththegiant r a radiolobesandcontributemorethanone-halftothetotalHEsourceemission.Itisagainspectrallywelldescribedbyapower-law functionextendingupto2or3GeVwithphotonindicesofG ∼2.6. If the extended HE emission indeed arises due to inverse-Compton up-scattering of CMB and EBL (Extragalactic Background Light)photons,thenthiscouldoffersauniquepossibilitytospatiallymaptheunderlyingrelativisticelectrondistribution.Thede- tectionofGeVg -raysfromtheradiolobesimpliesmagneticfieldstrengths ∼< 1µG(e.g.,Abdoetal.2010a).Thisestimatecanbe obtainedquitestraightforwardlyfromthecomparisonofradioandg -rays,assumingthattheseradiationcomponentsareproduced inthesameregionbythesamepopulationofelectronsthroughsynchrotronandinverse-Comptonprocesses.Ingeneral,however, theradioandtheg -rayregiondonotneedtocoincide.Whiletheradioluminositydependsontheproductoftherelativisticelectrons densityN andthe magnetic-fieldsquareB2, the inverse-Comptong -ray luminosityonly dependson N . Thisimpliesthatg -rays e e cangiveusmodel-independentinformationaboutboththeenergyandthespatialdistributionofelectrons,whiletheradioimageof synchrotronradiationstronglydependsonthemagneticfield.Asaconsequence,theg -rayimagecanbelargerthantheradioimage if the magneticfield dropsatthe peripheryofthe regionoccupiedbyelectrons.Thisprovidesoneofthe motivationfora deeper studyoftheextendedHE(lobe)emissionregioninCenA.Tothisendweanalyze3yrofdata,increasingtheavailableobservation timebymorethanafactorthreewithrespecttothepreviouslyreportedresults.Thelargerdatasetallowsforadetailedinvestigation of the spectrum andmorphologyof the lobeswith better statistics, especially above1 GeV, where the spectralshape may reflect 1 Rui-zhiYangetal.:DeepObservationoftheGiantRadioLobesofCentaurusAwiththeFermiLargeAreaTelescope cooling effects and/or maximum energy constraints on the parent population of particles generating the HE g -ray emission. We alsore-analyzeradiodatafromWMAP(Pageetal.2003)forthesameregionfromwheretheHEemissionisevaluatedfrom,and discusstheimplicationsoftheresultantspectralenergydistribution(SED)fordifferentemissionscenarios. Thepaperisstructuredasfollows.InSec.2thespectralandspatialHEanalysisresultsaredescribed,whereastheanalysisofthe WMAPresultsforthelobesarepresentedinSec.3.ImplicationsforleptonicandhadronicemissionmodelsarediscussedinSec.4 andconclusionsarepresentedinSec.5. 2. Fermi-LATDataAnalysis TheLargeAreaTelescope(LAT)onboardoftheFermig -RaySpaceTelescope,operatingsinceAugust4,2008,candetectg −ray photonswithenergiesintherangebetween100MeVandafew100GeV.DetailsabouttheLATinstrumentcanbefoundinAtwood et al. (2009). Here we analyze the field of view (FoV) of Cen A, which includes the bright core and the giant radio lobes. We selecteddataobtainedfromthebeginningoftheoperationuntilNovember14,2011,amountingto∼3yrofdata(MET239557417– 342956687).WeusedthestandardLATanalysissoftware(v9r23p1)1.Inordertoavoidsystematicerrorsduetopoordetermination of the effective area at low energies, we selected only events with energies above 200 MeV. The region-of-interest (ROI) was selectedtobearectangularregionofsize14◦×14◦ centeredonthepositionofCenA(RA=201◦21′54′′,DEC=−43◦1′9′′).To reducethe effectof Earth albedobackgrounds,time intervalswhen the Earth was appreciablyin the FoV (specifically, when the centeroftheFoVwasmorethan52◦ fromzenith)aswellastimeintervalswhenpartsoftheROIwereobservedatzenithangles >105◦ were also excludedfromthe analysis.The spectralanalysiswas performedbased onthe P7v6versionofthe post-launch instrumentresponsefunctions(IRFs).WemodeledtheGalacticbackgroundcomponentusingtheLATstandarddiffusebackground modelgal 2yearp7v6 v0 and we left the overallnormalizationand index as free parameters.We also used iso p7v6sourceas the isotropicg -raybackground. TheresultantFermi-LATcountsmapforthe3yrdatasetisshowninFig.1(a).The(green)crossesshowthepositionofthepoint- like sources from the 2FGL catalog (Abdoetal.2011) within the ROI. The core of Cen A is clearly seen with a test statistic of TS>800,correspondingtoadetectionsignificanceof28s .ExtendedemissiontotheNorthandSouthofCenAisdetectedwith significancesofTS>100(10s )andTS>400(20s ),respectively. 2.1. SpatialAnalysis Eventswith energiesbetween 200 MeV and 30 GeV were selected. The residualimage after subtractingthe diffuse background and point-like sources including the core of Cen A is shown in Fig. 1(b). The fluxes and spectral indices of 11 other point-like sourcesgeneratedfromthe2FGLcatalogwithintheROIarealsoleftasfreeparametersintheanalysis.The2FGLcatalogsource positionsareshowninFig.1(a),where2FGLJ1324.0-4330eaccountsforthelobes(bothNorthandSouth).Anewpoint-likesource (2FGLJ1335.3-4058),locatedatRA=203◦49′30′′,DEC=−40◦34′48′′accountsforsomeresidualemissionfromtheNorthlobe, althoughno knownsource at other wavelengthsis foundto be associated. We treat it as partof the North lobe here.The core of CenAismodeledasapoint-likesource.Thenthefollowingstepswereperformed: (1)Toevaluatethetotal(extended)HEg -rayemissionwefirstusedatemplatebasedontheresidualmap(T1;correspondingtothe bluecontoursinFig.2).TheTSvaluesfortheSouthandtheNorthlobeinthistemplateare411and155,respectively.Theresidual mapwasalsocomparedwithradio(WMAP,22GHz)lobecontours(greencontoursoverlaidonFig.1(b)).WefindthattheSouth lobeoftheHEg -rayimageissimilartotheSouthlobeoftheradioone,whereastheHEemissionintheNorthextendsbeyondthe radiolobeemissionregion. (2)Inordertounderstandthisfeaturebetter,were-fittheexcessusinganadditionaltemplate(T2;redcontoursinFig.3)generated fromtheradio(WMAP)image.ThetwotemplatesareshowninFig.2,andthecorrespondingresidualmapsareshowninFig.1. WhilethereissomeresidualemissiontotheNorthofCenAfortemplateT2,thisresidualemissionisobviouslyabsentfortemplate T1.ThequalitativefeaturesofthedifferentresidualmapsareconfirmedbythecorrespondingTSvalues,whicharelistedinTable1. Accordingly,theHESouthlobeseemreasonablywellinagreementwiththeradioSouthlobe,whereasfortheNorth,thetemplate generatedfromtheradiolobe(T2)fitstheHEexcesssubstantiallyworsethanT1(110vs155). Table1:TSvalueforthetwotemplatesused. Model NorthLobe SouthLobe T1 155 411 T2 110 406 (3) To further investigate a possible extension (or contribution of a backgroundsource) of the North lobe, we evaluated the projectionofarectangularregionontheexcessimage(inwhiteinFig.1(b)).Fig.3showstheprojectionfortheNorthandSouth regionsforboth,theFermi-LATimage(inblack)andtheradioone(inred).TheSouthprojectionfortheradiomapiswell-fitted bya singleGaussiancenteredat∼0.05(0 isdefinedasthe centeroftherectangleonRA=201◦21′54′′,DEC=−43◦1′9′′) with an extension of s =0.99◦. For the Fermi-LAT map the Gaussian is centered at ∼0.5 and has s =1.01◦, compatible with the radiomapprojection.Onthecontrary,theNorthprojectionfortheFermi-LATmaphasaGaussianprofilewiths =1.68◦, while forthe radiomaps is0.97◦. Theextensionin thenorthprojectionsfortheFermi-LATmapindicatesthattheg -rayNorthlobeis 1 http://fermi.gsfc.nasa.gov/ssc 2 Rui-zhiYangetal.:DeepObservationoftheGiantRadioLobesofCentaurusAwiththeFermiLargeAreaTelescope moreextendedthantheradiooneorthatan(otherwiseunknown)sourceinthebackgroundmaybecontributingtothetotalemission. 2.2. SpectralAnalysis Our morphological analysis indicates some incongruity between the morphology of the radio lobe and g -ray lobe in the North. Hence, in order to model the g -ray lobe as self-consistently as possible, we used the template generated with the residual map (T1).Integratingthe whole g -rayemission observed,we then derivedthe totalflux and index in the 100MeV to 30 GeV energy range. For the North lobe the integral HE flux is (0.93±0.09)×10−7ph cm−2s−1 and the photon index is 2.24±0.08, while for the South lobe we find (1.4±0.2)×10−7ph cm−2s−1 and 2.57±0.07, respectively. The core region of Cen A has a flux (1.4±0.2)×10−7ph cm−2s−1 while the photon index is 2.7±0.1. The results are summarized in Table 2, where the subscripts 3aand10mrefertothe3-yeardata(analyzedhere)andthe10-monthdata(reportedinAbdoetal.2010a),respectively.We find thatthefluxandphotonindicesintheT2templatesaresimilartothe10-monthdata.Ontheotherhand,theanalysisusingtheT1 templateresultsinaharderspectrumfortheNorthlobe. Toderivethespectralenergydistribution(SED)wedividedtheenergyrangeintologarithmicallyspacedbandsandapplygtlikein eachofthesebands.Onlytheenergybinsforwhichasignalisdetectedwithasignificanceofatleast2s areconsidered,whilean upperlimitiscalculatedforthosebelow.Asaresult,thereare7binsintheSEDfortheSouthlobe.TheSEDisshowninFig.4. Toclarifytheoriginoftheg -rayemission,weevaluatedthespectrumindifferentpartsofeachlobe.Tothisend,wedividedeach lobeintotwoparts(seeFig.5)andusedgtliketoevaluatethespectrum.IntheSouthlobethephotonindexis2.8±0.2neartheCen Acoreand2.3±0.1farawayfromthecore.FortheNorthlobethevaluesare2.2±0.2forbothparts.Theenergyspectrumofthe Southlobeseemstobehardening(DG =-0.5±0.3)farfromthecoreofCenAwhereasinNorthlobethereisnoobviouschange S inthespectrum.Unfortunately,thestatisticsisstillnotenoughtoclaimahardeinngonthespectrumoftheSouthlobe.Thespectra forthefourregionsareshowninFig.6andFig.7. Table2:Fluxesandspectraofthelobes SourceName F 3a(T1) G 3a(T1) F 10m G 10m F 3a(T2) G 3a(T2) SouthLobe 1.43±0.15 2.57±0.07 1.09±0.24 2.60±0.15 1.40±0.15 2.56±0.08 NorthLobe 0.93±0.09 2.24±0.08 0.77±0.20 2.52±0.16 0.64±0.15 2.56±0.08 F istheintegralflux(100MeVto30GeV)inunitsof10−7ph·cm−2s−1andG isthephotonindex.Thesubscripts”3a”and”10m” referto the3-yeardataanalyzedhereandto the 10-monthresults(basedona WMAP template)reportedinAbdoetal. (2010a), respectively. 3. WMAPdata InSec.2wefoundevidencethattheHEg -rayemissionregionsdonotcoincidewellwiththeradiolobes,especiallyfortheNorth lobe.Tocorrectlycomparetheg -rayemissionwiththeradioone,ananalysisoftheradiodataforthesameregionisrequired,rather thansimplyfortheNorthradiolobeitself.WeperformedthisanalysisfollowingthemethoddescribedinHardcastleetal.(2009). SevenyearsofWMAPdatawere analyzed(Komatsuetal.2011). TheWMAPmapsin allfivebandsareconvolvedwitha 0.83◦ Gaussiantogetsimilarresolutionforallbands.Theinternallinearcombination(ILC)CosmicMicrowaveBackground(CMB)map was treated as backgroundand subtractedfrom all maps. The intensity maps(in mK) were convertedto flux maps (in Jy/beam) (Pageetal.2003)andintegratedintheregiondefinedbytheNorthg -raylobetoobtainthetotalflux.Fluxvaluesalmosttwiceas largeareobtainediftheg -rayregionisused.Inordertocross-checkouranalysismethod,wealsoderivedthefluxoftheNorthradio lobeforthe sameregionasusedin Hardcastleetal. (2009).AllresultsaresummarizedinTab.4.OurresultsfortheNorthradio lobeiscompatiblewithinerrorswiththeoneobtainedbyHardcastleetal.(2009)(showninbracketsinTab.5).NotethatintheV Table3:RadiofluxfortheNorthlobe(in1011Jy·Hz)fromWMAPdata. band Northg -rayLobe NorthradioLobe K(22.5GHz) 5.34±0.37 2.71±0.19(2.61±0.25) Ka(33GHz) 5.24±0.41 2.55±0.22(2.50±0.24) Q(41GHz) 5.53±0.56 2.94±0.29(2.58±0.25) V(61GHz) 7.65±1.19 <4.13 W(94GHz) <25.6 <9.43 andtheWbandthesignalislessthanthe5s ,soweonlygiveupperlimitshere. 3 Rui-zhiYangetal.:DeepObservationoftheGiantRadioLobesofCentaurusAwiththeFermiLargeAreaTelescope 13 2 0 12 2FGL J-36.00123F5G9L. 9J-13374476.7-3752 2FGL J1259.8-3749 11 36.000 11..68 - 2FGDeclinationL J1420-40.000-44.000F7G.L5 -J41235257F2G.L6 -2J42F14FG31GL33L 5 J.J1313-324225042F.522.FG68FF0GL-GG-L 4LL4 J3 3J10JJ31301103033e07006.34.5..9-53-4--434400632025823 5789...823 Declination -40.00044.000 0011....792488 22FFGGLL JJ11332286..54--44772289 - 2FGL J1446.8-4701 -48.000 2FGL J1305.8-4925 34..57 -48.000 00..3599 212.000 206.000 200.000 194.000 2.3 212.000 206.000 200.000 194.000 0.19 Right ascension Right ascension 1.2 0 (a)LATcountsmapofthe14◦×14◦ROI. (b)Excessmapafterbackgroundsubtraction. 2 2 0 1.8 0 1.8 0 0 0 0 36. 1.6 36. 1.6 - - 0 1.4 0 1.4 0 0 0 0 on -40. 1.2 on -40. 1.2 Declinati 44.000 01.8 Declinati 44.000 01.8 - - 0 0.6 0 0.6 0 0 0 0 48. 0.4 48. 0.4 - - 212.000 206.000 200.000 194.000 0.2 212.000 206.000 200.000 194.000 0.2 Right ascension Right ascension 0 0 (c)ResidualmapfortemplateT1. (d)ResidualmapfortemplateT2. Fig.1:ThedifferentmapsfortheCenAregion:(a)LATcountsmapofthe14◦×14◦ regionofinterest(ROI)aroundtheposition ofCenA.ThecountsmapissmoothedwithaGaussianofkernel0.8◦.Thegreencrossesmarkthepositionofthe2FGLpoint-like sources.(b)Excessmapaftersubtractionofdiffusebackground,point-likesourcesandCenAcore.ThecontoursareWMAPradio lobecontours,whilethewhiteboxesrepresenttheprojectionregionsdiscussedinSec.2.1.(c)ResidualmapusingtemplateT1for thelobes.(d)ResidualmapusingtheradiotemplateT2forthelobes. 4. OntheOriginoftheNon-ThermalLobeEmission UsingtheWMAPandFermi-LATresultsreportedhere,wecancharacterizethespectralenergydistributions(SEDs)fortheNorth and the South lobe, respectively. While the radio emission is usually taken to be caused by electron synchrotron emission, the originof the HE g -ray emissioncouldin principlebe related to bothleptonic (inverse-Comptonscattering)or hadronic(e.g.,pp- interaction)processes.Inthefollowingwediscusspossibleconstraintsfortheunderlyingradiationmechanismasimposedbythe observedSEDs. 4.1. Inverse-Comptonoriginofg -rays Both theHE g -rayandthe radioemissioncouldbe accountedforina leptonicscenario.Inthe simplestversion,a single popula- tion ofelectronsN(g ,t) is used to modelthe SED throughsynchrotronandinverse-Comptonemission, with particleacceleration beingimplicitlytreatedbyaneffectiveinjectiontermQ=Q(g ,t).Thelaterallowsustodisentangleaccelerationby,e.g.,multiple shocksorstochasticprocesses(e.g.,O’Sullivanetal.2009)fromemission,andenablesastraightforwardinterpretation.Thekinetic equationdescribingtheenergeticandtemporalevolutionoftheradiatingelectronscanthenbewrittenas ¶ N ¶ N = (PN)− +Q, (1) ¶ t g¶ t esc whereP=P(g )=−dg isthe(time-independent)energylossrateandt isthecharacteristicescapetime.Fornegligibleescape(as dt esc appropriateheregiventhelargesizeoftheg -rayemittingregion)andquasi-stationaryinjectionQ(g ,t)=Q(g ),thesolutionofthe kineticequationbecomes N(g ,t)= 1 g0Q(g )dg , (2) P(g )Zg 4 Rui-zhiYangetal.:DeepObservationoftheGiantRadioLobesofCentaurusAwiththeFermiLargeAreaTelescope where g 0 is found by solving the characteristic equation for a given epoch t, t = gg0 Pd(gg) (e.g., Atoyan&Aharonian1999). If synchrotronorinverse-Compton(Thomson)losses(P(g )=ag 2)providethedominanRtlosschannel,theng =g /(1−ag t),sothat 0 attheenergyg = 1 thestationarypower-lawelectroninjectionspectrumsteepensbyafactorof1(coolingbreak)duetoradiative br at losses,naturallygeneratingabrokenpower-law. We use the above particle distribution described in eq. (2), for a representation of the observed lobe SEDs. The magnetic field strength B, the maximum electron energyg and the epoch time t are left as free parametersto modelthe data. Klein-Nishina max (KN)effectsontheinverse-Compton-scatteredHEspectrumaretakenintoaccount. Fig.8showstheSEDresultsobtainedfortheNorthandSouthlobes.TheHEpartofbothspectracanbedescribedbyapower-law withphotonindexG g ≃2.2and2.6fortheNorthandtheSouthlobe,respectively.Atlowenergies,theSouthlobespectrumshows asynchrotronpeakatabout5GHz,whiletheNorthoneiswelldescribedbyapower-lawwithanindex>2.Notethatifonewould use a simple power-lawelectron injection spectrumQ(g )(cid:181)g −a , evolvingin time with a cooling break,to describe the HE g -ray spectrum,apowerindexa =3.2wouldberequiredfortheSouthlobe.Yet,assumingthatthesameelectronpopulationisresponsible forboththeradio-synchrotronandHEinverse-Comptonemission,suchavaluewouldbeinconflictwiththeresultsobtainedfrom theWMAPdataanalysis,indicatinganelectronpopulationwithpower-lawindexa ≃2basedonthedetectedsynchrotronemission. As it turnsout, however,this issue could be accommodatedby consideringa more naturalspectral inputshape, e.g., an electron injectionspectrumwithanexponentialcut-off g Q(g )=Q g −a exp − , (3) 0 (cid:18) g (cid:19) max wheretheconstantQ canbeobtainedfromthenormalizationtotheinjectionpowerL=m c2 Q(g )g dg . 0 e The lifetime of the giant radio lobes is somewhat uncertain. Dynamical arguments suggestRa minimum age of several 106 yr, whilesynchrotronspectralagingargumentsindicateanageofsome107yr(e.g.,Alvarezetal.2000,Hardcastleetal.2009).Inthe followingwethusdiscusstheSEDimplicationsforaminimumandamaximumepochtimetoft =106yrandt =8×107yr. min max For the South lobe, the radio data suggest a break frequency n =5 GHz above which the spectrum drops down. The break in br the synchrotron spectrum is related to the break in the electron spectrum via n =1.3g 2B Hz. In principle, a change in the br 1µG spectralshapeoftheelectronpopulationcouldbeduetocoolingeffectsor/andtheexistenceofamaximumenergyfortheelectron population.Foraminimumepochtimet =106 yr,coolingwouldaffectthesynchrotronspectrumatfrequency≈7.5B THz, min 1µG much higher than inferred from the radio data. Therefore, to obtain a break at 5 GHz in the South lobe, a maximum energy in theelectronpopulation(g ),lowerthang definedbyt =t wouldbeneeded.Ontheotherhand,foramaximumepochtime max br min t =8×107 yr, the power-lawspectralindex changesat frequency≃1B GHz, providinga satisfactory agreementwith the max 1µG radiodata.Inthiscasethemaximumelectronenergyisobtainedfromtheradiodataabovethebreakfrequency5GHz.Resultsfor theconsideredminimumandmaximumepochtime,andforafixedpower-lawelectronindexa =2areillustratedinFigs.8and9. NotethatforB≤3µG,theenergylossrateP(g )isdominatedbytheICchannelonly,sothattheresultsofthecalculationsarequite robust. Figure 8 shows a representation of the SED for the North and the South lobe, respectively, using the parameters t =106 yr min andg =9.5×104.ThedashedlineshowstheHEcontributionproducedbyinverse-ComptonscatteringofCosmicMicrowave max Background (CMB) photons by relativistic electrons within the lobes. In this case the resulting g -ray flux can only describe the first two data points and then drops rapidly. Thus, in order to be able to account for the observed HE spectrum, Extragalactic BackgroundLight(EBL)photonsneedtobeincludedinadditiontoCMBphotons(seedot-dashedlineinFigure8).Upscattering of infrared-to-optical EBL photons was already required in the stationary leptonic model reported in the original Fermi paper (Abdoetal.2010a).InourapproachweadoptthemodelbyFranceschinietal.(2008)toevaluatethisEBLcontribution.Thesolid lineinFig.8representsthetotal(CMB+EBL)inverse-Comptoncontribution.Themaximumtotalenergyofelectronsinbothlobes isfoundto be≃2×1058 erg.Dividingthisbythe epochtime 106 yr,wouldimplya meankineticpowerofthe jetsinflatingthe lobesof≃6×1044erg/s,roughlyanorderofmagnitudesmallerthantheEddingtonpowerinferredfortheblackholemassinCen A,yetwellabovetheestimatedpowerofthekpc-scalejetinthecurrentepochofjetactivity(Crostonetal.2009).Thiswouldimply thatthejetwasmorepowerfulinthepast.However,therequirementonthejetpowercanbesignificantlyreducedifoneassumes anageofthelobes≫106yr. Figure 9 shows a representation of the SED for the maximum epoch time t =8×107 yr, with a maximum electron Lorentz max factor g =2.5×106 and 1.5×106 for the North lobe and the South lobe, respectively.Note that in this case the contribution max byinverse-ComptonscatteringofCMBphotonsaloneissufficienttoaccountfortheobservedHEspectrum(seethesolidlinein Figure9).Theinverse-ComptoncontributionofEBLphotonsonlybecomesimportantathigherenergies(seethedot-dashedline inFig.9).Ontheotherhand,foranepochtimet exceedingt =8×107yr,thehighenergypartoftheSEDwouldnolongerbe max consistentwiththedata(seethedashedlineinFig.9fort=108yr).Thiscouldbeinterpretedasadditionalevidenceforafiniteage <108yrofthelobes.Themaximumtotalenergyofelectronsinbothlobesisfoundtobe≈6×1057erg,requiringonlyarelatively modestmeankineticjetpowerof2×1042erg/s. 4.2. Hadronicg -rays? Once protons get efficiently injected, they are likely to remain energetic since the cooling time for pp-interactions is t ≈ pp 1015(n/1cm−3)−1 s. High-energy protons interacting with the ambient low density plasma, can then produce daughter mesons andthosep 0 componentdecaysintotwog -rays.Thedatareportedhereallowustoderiveanupperlimitontheenergeticprotons 5 Rui-zhiYangetal.:DeepObservationoftheGiantRadioLobesofCentaurusAwiththeFermiLargeAreaTelescope containedinthelobesofCenA.Asbefore,weuseapower-lawprotondistributionwithanexponentialcut-off,i.e., g N(g )=N g −a exp − p p 0 p (cid:18) g (cid:19) max where the constant N can be expressed in terms of the total proton energyW =m c2 g N(g )dg . Current estimates for the 0 p p p p p thermalplasmadensityinthegiantradiolobesofCenAsuggestavalueintherangen≃(1R0−5−10−4)cm−3(e.g.,Isobeetal.2001, Feainetal.2009). We use n=10−4 cm−3 for the modelrepresentationshown in dotted line in Fig. 9. In both lobes, the power- law index of the proton populationis a =2.1, and the high-energycut-off is E ≃55 GeV. The maximum total energyW is max p proportionaltothegasnumberdensityn,sothatW ≃1061(n/10−4cm−3)−1 erg,obtainedhere,shouldbeconsideredasanupper p limit.Inprinciple,protonscouldbeaccumulatedoverthewholeevolutionarytimescaleofthelobes.Forlongtimescale≥109yr,an averageinjectionpower≤3×1044 erg/sandameancosmic-raydiffusioncoefficientofD∼R2/t∼< 3×1030(R/100kpc)2 cm2/s wouldbeneeded. 5. DiscussionandConclusion Results based on an detailed analysis of 3 yr of Fermi-LAT data on the giant radio lobes of Cen A are described in this paper. We haveshownthatthedetectionoftheHElobeswithasignificancemorethantwice ashighasreportedbefore(i.e.,withmore than 10 and 20s forthe Northernand the Southernlobe, respectively)allows for a better determinationof their spectralfeatures andmorphology.AcomparisonoftheFermi-LATdatawithWMAPdataindicatesthattheHEg -rayemissionregionsdonotfully coincidewiththeradiolobes.Thereisofcoursenoapriorireasonforthemtocoincide.Theresultsreportedhereparticularlysupport thecase of asubstantialHEg -rayextensionbeyondthe WMAPradioimageinthe case oftheNorthernlobeofCen A. We have reconstructedtheSEDbasedondatafromthesameemissionregion.Asatisfactoryrepresentationispossibleinaleptonicscenario withradiativecoolingtakenintoaccountself-consistentlyandinjectiondescribedbyasinglepower-lawwithexponentialcut-off. Theresultsimplyafiniteage<108yrofthelobesandameanmagneticfieldstrengthB∼< 1µG.Whileforlobelifetimesoftheorder of8×107yr,inverse-Comptonup-scatteringofCMBphotonsalonewouldbesufficienttoaccountfortheobservedHEspectrum, up-scattering of EBL photons is needed in the case of shorter lobe lifetimes. In a leptonic framework,the HE emission directly traces (via inverse-Comptonscattering) the underlying relativistic electron distribution and thereby provides a spatial diagnostic tool.Theradioemission,arisingfromsynchrotronradiation,ontheotherhandalsotracesthemagneticfielddistribution.Together, theHE g -rayandtheradioemissionthusofferimportantinsightsintothe physicalconditionsofthe source.Thefactthatthe HE emissionseemsextendedbeyondtheradioimagecouldthenbeinterpretedasduetoachangeinthemagneticfieldcharacterizing theregion.Thiswouldimplythataquasi-homogeneousSEDmodelforthelobescanonlyserveasafirst-orderapproximationand thatmoredetailedscenariosneedtobeconstructedtofullydescribethedata.Thisalsoappliestotheneedofincorporatingelecton re-accelerationself-consistently.ExtendedHEemissioncouldinprinciplealsoberelatedtoacontributionfromhadronicprocesses. Thecoolingtimescalesforprotonsappearmuchmorefavourable.Ontheotherhand,boththespectralshapeofthelobesandthe required energetics seem to disfavor pp-interaction processes as sole contributor. One of the insights emerging from the present paperistheneedforamoredetailedtheoreticalSEDapproachinordertobeabletotakefulladvantageofthecurrentobservational capabilities. References Abdo,A.A.etal.(Fermi-LAT)2010a,Science,328,725 Abdo,A.A.etal.(Fermi-LAT)2010b,ApJ,719,1433 Abdo,A.A.etal.(Fermi-LAT)2011,arXiv:1108.1435[astro-ph.HE]. Aharonian,F.etal.(H.E.S.S.Collaboration)2009,ApJ,695,L40 Alvarez,H.etal.2000,A&A,355,863 Atoyan,A.M.,andAharonian,F.A.1999MNRAS,302,253 Atwood,W.B.etal.(Fermi-LATCollaboration)2009,ApJ,697,1071 Burns,J.O.etal.1983,ApJ,273,128 Croston,J.H.etal.2009,MNRAS,395,1999 Feain,I.J.etal.2009,ApJ,707,114 Ferrarese,L.etal.2007,ApJ,654,186 FranceschiniA.,RodighieroG.,VaccariM.2008,Astron.Astrophys.487,837 Hardcastle,M.J.,Cheung,C.C.,Feain,I.J.,Stawarz,L.2009,MNRAS,393,104 Isobe,N.etal.2001,ASPC,250,394 Israel,F.P.1998,A&ARv8,237 O’Sullivan,S.,Reville,B.,Taylor,A.2009,MNRAS,400,248 Page,L.etal.2003,ApJS,148,39 Komatsu,E.etal.(WMAPCollaboration)2011ApJS,192,18 Rieger,F.M.2011,IJMPD20,1547 Shain,C.A.1958,Aust.J.Phys.11,517 Steinle,H.2010,PASA,27,431 http://fermi.gsfc.nasa.gov/ssc/data/analysis/documentation/Cicerone/ http://fermi.gsfc.nasa.gov/ssc/data/analysis/documentation/Pass7 usage.html 6 Rui-zhiYangetal.:DeepObservationoftheGiantRadioLobesofCentaurusAwiththeFermiLargeAreaTelescope 0 0 0 6. 3 - 0 0 0 0. on -4 ati n cli 00 e 0 D 4. 4 - 0 0 0 8. 4 - 212.000 206.000 200.000 194.000 Right ascension Fig.2:Thetwotemplatesusedintheanalysis.ThebluecontourscorrespondtoT1andtheredtoT2. be) 500be) be)350 350be) o o o o mi LAT - South L450000 340000WMAP - South L mi LAT - North L223050000 223050000WMAP - North L mber of Events (Fer123000000 0120000Number of Events ( mber of Events (Fer11055000 051100500Number of Events ( Nu Nu 0 -50 -4 -2 0 2 4 -6 -4 -2 0 2 4 6 q q (a)ProjectionoftheSouthlobe (b)ProjectionoftheNorthlobe Fig.3:TheprojectionoftherectangularregionshowninFig.1(b)forbothlobes.ThecurvesareGaussianfits,withWMAPdatain redandFermi-LATdatainblack. 7 Rui-zhiYangetal.:DeepObservationoftheGiantRadioLobesofCentaurusAwiththeFermiLargeAreaTelescope 1e-10 ) 2 - m c 1 - s 1e-11 g r e ( x u l F ) h ut 1e-12 o S 2 / h t r 1.5 o N ( 1 o i t a 0.5 R x u 0 L 100 1000 10000 F Energy(MeV) Fig.4: SED for template T1. Squares and crosses are for the North lobe and the South lobe, respectively.The ratio of the fluxes (North/South)areshowninthebottompanel. 0 0 0 6. 3 - 0 0 0 0. on -4 ati n cli 00 e 0 D 4. 4 - 0 0 0 8. 4 - 212.000 206.000 200.000 194.000 Right ascension Fig.5: The two lobes are both divided into two parts (near core of Cen A core and further away) for spectral comparison. The contoursindifferentcolorsrepresentthedifferentparts. 8 Rui-zhiYangetal.:DeepObservationoftheGiantRadioLobesofCentaurusAwiththeFermiLargeAreaTelescope 1e-10 1e-10 -2m) 1e-11 -2m) 1e-11 -1Flux(erg s c 1e-12 -1Flux(erg s c 1e-12 1e-13 1e-13 100 1000 100 1000 Energy(MeV) Energy(MeV) Fig.6:TheSEDsforthecorrespondingsubregionsintheNorth.TheleftpanelistheSEDofthesubregionfarfromthecore,while therightpanelreferstothesubregionnearthecore.ThecolorsarethesameasusedinthedefinitioninFig.5. 1e-10 1e-10 -2m) 1e-11 -2m) 1e-11 -1Flux(erg s c 1e-12 -1Flux(erg s c 1e-12 1e-13 1e-13 100 1000 100 1000 Energy(MeV) Energy(MeV) Fig.7:SameasFig.6butfortheSouthernsubregions.TheleftpanelistheSEDofthesubregionfarfromthecore,whiletheright panelreferstothesubregionnearthecore. 13.5 13.5 13.0 13.0 -2LL -2LL m m -1gsc 12.5 -1gsc 12.5 -2310er 12.0 -2310er 12.0 ´ ´ ΥHFΥ 11.5 ΥHFΥ 11.5 Hg10 Hg10 lo 11.0 lo 11.0 10.5 10.5 10 15 20 25 10 15 20 25 log HΥHHzLL log HΥHHzLL 10 10 (a)Southlobe (b)g -rayexcess(lobe)regionintheNorth Fig.8:Synchrotronandinverse-Comptonfluxesfort=106 yr.TheradiodatafortheSouthlobearefromHardcastleetal.(2009) (sum of Region 4 and Region 5 in their Table 1), while the radio data for the North region are from the WMAP analysis in this paper.ThemeanmagneticfieldvalueBusedfortheNorthandtheSouthlobeis0.38µGand0.45µG,respectively.Thedot-dashed linereferstotheICcontributionduetoEBLupscattering. 9 Rui-zhiYangetal.:DeepObservationoftheGiantRadioLobesofCentaurusAwiththeFermiLargeAreaTelescope 13.0 13.0 -2LLm 12.5 12.5 -1gsc 12.0 -´2310er 12.0 HFΥ 11.5 ΥHg10 11.5 lo 11.0 11.0 10.5 10 15 20 25 log10HΥHHzLL 10.5 10 15 20 25 (a)Southlobe (b)g -rayexcess(lobe)regionintheNorth Fig.9: Synchrotronand inverse-Comptonfluxes fort =8×107 yr. The mean magnetic field value B for the South lobe and the g -ray excess region in North is 1.17µG and 0.91µG respectively. The dot-dashed line refers to the IC contribution due to EBL upscattering.Thedashedline(a)showstheresultfort=108yr.Thepossibleg -rayfluxexpectedfrompp-interactionsforathermal gasdensityn=10−4cm−3arealsoshown(dottedline). 10

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.