ebook img

Ordered magnetic fields around radio galaxies: evidence for interaction with the environment PDF

0.8 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 Ordered magnetic fields around radio galaxies: evidence for interaction with the environment

Mon.Not.R.Astron.Soc.000,1–21(2002) Printed11January2011 (MNLATEXstylefilev2.2) Ordered magnetic fields around radio galaxies: evidence for interaction with the environment 1 D. Guidetti⋆1,2,3, R.A. Laing 1, A.H. Bridle 4, P. Parma 2, L. Gregorini 3,2 1 0 1EuropeanSouthernObservatory,Karl-Schwarzschild-Straße2,D-85748Garching-bei-Mu¨nchen,Germany 2 2INAF–IstitutodiRadioastronomia,viaGobetti101,I-40129Bologna,Italy 3DipartimentodiAstronomia,Univ.Bologna,viaRanzani1,I–40127Bologna,Italy n 4NationalRadioAstronomyObservatory,EdgemontRoad,Charlottesville,VA22903-2475,U.S.A. a J 0 1 Accepted.Received;inoriginalform ] O ABSTRACT C WepresentdetailedimagingofFaradayrotationanddepolarizationfortheradiogalaxies . h 0206+35,3C270,3C353andM84,basedonVeryLargeArrayobservationsatmultiplefre- p quenciesintherange1365to8440MHz.Allofthesourcesshowhighlyanisotropicbanded - o rotation measure (RM) structureswith contoursof constant RM perpendicularto the major r axesoftheirradiolobes.AllexceptM84alsohaveregionsinwhichtheRMfluctuationshave t s lower amplitude and appear isotropic. We give a comprehensive description of the banded a RMphenomenonandpresentaninitialattempttointerpretitasaconsequenceofinteractions [ between the sources and their surroundings. We show that the material responsible for the 1 Faradayrotation is in frontof the radio emission and thatthe bandsare likely to be caused v bymagnetizedplasmawhichhasbeencompressedbytheexpandingradiolobes.Wepresent 7 a simple model for the compression of a uniformly-magnetizedexternalmedium and show 0 thatRMbandsofapproximatelytherightamplitudecanbeproduced,butforonlyforspecial 8 initialconditions.Atwo-dimensionalmagneticstructureinwhichthefieldlinesareafamily 1 ofellipsesdrapedaroundthe leadingedgeofthelobecanproduceRM bandsin thecorrect . 1 orientationforanysourceorientation.We also reportthefirstdetectionsofrimsofhighde- 0 polarization at the edges of the inner radio lobes of M84 and 3C270. These are spatially 1 coincidentwithshellsofenhancedX-raysurfacebrightness,inwhichboththefieldstrength 1 and the thermal gas density are likely to be increased by compression. The fields must be : tangledonsmallscales. v i X Keywords: –galaxies:magneticfields–radiocontinuum:galaxies–(galaxies:)intergalactic medium–X-rays:galaxies:clusters r a 1 INTRODUCTION by linearly polarized radiation travelling through a magnetized medium,andcanbedescribedbythetwofollowingrelations: Thedetectionofdiffusesynchrotronemission(radiohalos)onMpc scales in an increasing number of galaxy clusters provides good ∆Ψ[rad] =Ψ(λ)[rad] − Ψ0[rad] =λ2[m2]RM[radm−2], (1) evidence for a distributed magnetic field of µGauss strength in with the hot intracluster medium (ICM; see e.g. Ferrarietal. 2008 for areview).ImagingofFaradayrotationoflinearly-polarizedradio L[kpc] emission from embedded and background sources confirms that RM[radm−2] =812Z ne[cm−3]Bz[µG]dz[kpc], (2) therearefieldsassociatedwiththermalplasmaalonglinesofsight 0 through the clusters (e.g. Carilli&Taylor 2002). Observations of whereΨ(λ)andΨ aretheE-vectorpositionangleoflinearlypo- 0 Faradayrotationcanalsobemadeforradiogalaxiesinsparseren- larizedradiationobservedatwavelengthλandtheintrinsicangle, vironments,allowingthestudyofmagneticfieldsinenvironments respectively,n istheelectrongasdensity,B isthemagneticfield e z too sparse for radio halos to be detected (e.g. Laingetal. 2008; alongtheline-of-sight (B ),andListheintegrationpath. RMis k Guidettietal.2010). therotationmeasure. The Faraday effect (Faraday 1846) is the rotation suffered Observations of Faraday rotation variations across extended radiogalaxiesallowustoderiveinformationabouttheintegralof thedensity-weightedline-of-sightfieldcomponent.Thehot(T ≃ ⋆ E-mail:[email protected] 107 −108K)plasmaemitsintheX-rayenergybandviathermal 2 D. Guidettiet al. bremsstrahlung. Whenhigh quality X-ray datafor aradio-source andinSection3webrieflysummarizethetechniquesusedtoanal- environmentisavailable,itispossibletoinferthegasdensitydis- yse depolarization and two-dimensional variations of RM. Sec- tributionandthereforetoseparateitfromthatofthemagneticfield, tions4and5presenttheRManddepolarizationimagesonwhich subjecttosomeassumptionsabouttherelationoffieldstrengthand ouranalysisisbasedandcorrelationsbetweenthetwoquantities.In density. Section6,weevaluatetheRMstructurefunctionsinregionswhere Most of the RM images of radio galaxies published so far thefluctuationsappear tobeisotropicandderivethepower spec- show patchy structures with no clear preferred direction, consis- tra. A simple model of the source-environment interaction which tent with isotropic foreground fluctuations over a range of linear characterises the effects of compression of a magnetised IGM is scales ranging from tens of kpc to <100pc (e.g. Govonietal. describedinSection7.ThiscanproduceRMbands,butonlyun- ∼ 2006;Guidettietal.2008;Laingetal.2008;Guidettietal.2010). derimplausiblespecialinitialconditions.Empirical“draped”field Numerical modelling has demonstrated that thistype of complex configurationswhichareabletoreproducethebandedRMdistri- RM structure can be accurately reproduced if the magnetic field butionsareinvestigatedinSection8.InSection9,wespeculateon is randomly variable with fluctuations on a wide range of spatial correlationsbetweenradiosourcemorphology,andRManisotropy, scales, and isspread throughout the whole group or cluster envi- discussotherexamplesfromtheliterature,consider theeffectsof ronment(e.g.Murgiaetal.2004;Govonietal.2006;Guidettietal. anisotropicforegroundFaradayscreenonthedetectabilityofRM 2008; Laingetal. 2008; Guidettietal. 2010; Vaccaetal. 2010). bandsandbrieflydiscusspossibleasymmetriesintheamplitudeof Theseauthorsusedforwardmodelling,togetherwithestimatorsof theRMbandsbetweentheapproachingandrecedinglobes.Finally, the spatial statistics of the RM distributions (structure and auto- Section10summarizesourconclusions. correlationfunctionsoramulti-scalestatistic)toestimatethefield Throughout thispaper weassumeaΛCDMcosmology with strength,itsrelationtothegasdensityanditspowerspectrum.The H =71kms−1Mpc−1,Ω =0.3,andΩ =0.7. 0 m Λ techniqueofBayesianmaximumlikelihoodhasalsobeenusedfor thispurpose(Enßlin&Vogt2005;Kuchar&Enßlin 2009). Inordertoderivethethree-dimensionalmagneticfieldpower 2 THESAMPLE spectrum,alloftheseauthorshadtoassumestatisticalisotropyfor thefield,sinceonlythecomponentofthemagneticfieldalongthe High quality radio and X-ray data are available for all of the line-of-sight contributes to the observed RM. This assumption is sources.InthisSectionwesummarizethoseoftheirobservational consistentwiththeabsenceofapreferreddirectioninmostofthe propertieswhicharerelevanttoourRMstudy.Alistofthesources RMimages. and their general parameters is given in Table1, while Table2 Incontrast,thepresentpaperreportsonanisotropicRMstruc- showstheX-rayparameterstakenfromtheliteratureandequipar- tures,observedinlobedradiogalaxieslocatedindifferentenviron- titionparametersderivedfromourradioobservations. ments,rangingfromasmallgrouptooneoftherichestclustersof ThesourceswereobservedwiththeVLAatseveralfrequen- galaxies.TheRMimagesofradiogalaxiespresentedinthispaper cies,infullpolarizationmodeandwithmultipleconfigurationsso showclearlyanisotropic“banded”patternsoverpartoralloftheir that the radio structure is well sampled. The VLA observations, areas.Insomesources,thesebandedpatternscoexistwithregions datareductionanddetaileddescriptionsoftheradiostructuresare ofisotropicrandomvariations.Themagneticfieldresponsiblefor givenfor0206+35andM84byLaingetal.(2011),for3C270by theseRMpatternsmust,therefore,haveapreferreddirection. Laing,Guidetti&Bridle(2011),andfor3C353bySwain(1996). OnesourcewhoseRMstructureisdominatedbybandsisal- All of the radio maps show a core, two sided jets and a double- ready known: M84 (Laing&Bridle 1987). In addition, there is lobedstructurewithsharpbrightnessgradientsattheleadingedges some evidence for RM bands in sources which also show strong ofbothlobes. Thesynchrotron minimumpressures areallsignif- irregularfluctuations,suchasCygnusA(Carilli&Taylor2002).It icantly lower than the thermal pressures of the external medium ispossible,however,thatsomeoftheclaimedbandscouldbedue (Table2). toimperfectsamplingofanisotropicRMdistributionwithlarge- AllofthesourceshavebeenobservedinthesoftX-rayband scalepower,andwereturntothisquestioninSection9.1. bymorethanonesatellite,allowingthedetectionofmultiplecom- Thepresent paper presentsnew RMimagesofthreesources ponentsoncluster/groupandsub-galacticscales.TheX-raymor- whichshowspectacularbandedstructures,togetherwithimproved phologiesarecharacterizedbyacompactsourcesurroundedbyex- dataforM84.Theenvironmentsofallfoursourcesarewellchar- tendedemissionwithlowsurfacebrightness.Theformerincludesa acterizedbymodernX-rayobservations,andwegivethefirstcom- non-thermalcontribution,fromthecoreandtheinnerregionsofthe prehensivedescriptionofthebandedRMphenomenon.Wepresent radiojetsand,inthecaseof0206+35and3C270,athermalcom- aninitialattempttointerpretthephenomenonasaconsequenceof ponent which is well fitted by a small core radius β model. The source-environment interactions and to understand the difference lattercomponentisassociatedwiththediffuseintra-grouporintra- betweenitandthemoreusualirregularRMstructure. clustermedium.Parametersforallofthethermalcomponents,de- TheRMimagesreportedinthispaper arederivedfromnew rivedfromX-rayobservations,arelistedinTable2.Becauseofthe orpreviouslyunpublishedarchiveVeryLargeArray(VLA)1 data irregularmorphologyofthehotgassurrounding3C353andM84, for the nearby radio galaxies 0206+35, M84 (Laingetal. 2011), ithasnotbeenpossibletofitβmodelstotheirX-rayradialsurface 3C270(Laing,Guidetti&Bridle2011)and3C353(Swain,private brightnessprofiles. communication;seeSwain1996). Thepaperisorganizedasfollows.InSection2theradioand 2.1 0206+35(4C35.03) X-ray properties of the sources under investigation are presented 0206+35 is an extended Fanaroff-Riley Class I (FRI; Fanaroff&Riley 1974) radio source whose optical counter- 1 TheVeryLargeArrayisafacilityoftheNationalScienceFoundation, part,UGC11651,isaD-galaxy,amemberofadumb-bellsystem operatedundercooperativeagreementbyAssociatedUniversities,Inc. atthecentreofagroupofgalaxies.Ataresolutionof1.2arcsecthe Faradayrotationmeasurebandsacrossradiogalaxies 3 Figure1. X-rayimagesoverlaidwithradiocontoursforallsources:(a)0206+35:1385.1MHzVLAradiomapwith1.2arcsecFWHM;thecontoursarespaced byfactorof2between0.06and15mJybeam−1.TheROSATPSPCimage(Worrall&Birkinshaw2000),issmoothedwithaGaussianofσ = 30arcsec. (b) M84: 1413.0MHz VLA radio map with 4.5arcsec FWHM; the contours are spaced by factor of 2 between 1 and 128mJy beam−1. TheChandra (Finoguenov&Jones 2001) image is a wavelet reconstruction on angular scales from 4 up to 32arcsec. (c) 3C270: 1365.0MHz VLA radio map with 5.0arcsecFWHM;thecontoursarespacedbyfactorof2between0.45and58mJybeam−1.TheXMM-Newtonimage(Croston,Hardcastle&Birkinshaw 2005)issmoothedwithaGaussianofσ=26arcsec.(d)3C353:1385.0MHzVLAradiomapwith1.3arcsecFWHM;thecontoursarespacedbyfactorof 2between0.35and22mJybeam−1.TheXMM-Newtonimage(Goodgeretal.2008)issmoothedwithaGaussianofσ=30arcsec.TheChandraimageof 3C270isdisplayedinlogarithmicscale. Table1.Generalopticalandradioproperties:Col.1:sourcename;Cols.2&3:position;Col.4:redshift;Col.5:conversionfromangulartospatialscalewith theadoptedcosmology;Col.6:Fanaroff-Rileyclass;Col.7:thelargestangularsizeoftheradiosource;Col.8:radiopowerat1.4GHz;Col.9:angletothe lineofsightofthejetaxis;Col.10:radiospectralindex;Col.11:environmentofthegalaxy;Col.12:reference. source RA DEC z kpc/arcsec FRclass LAS logP1.4 θ env. ref. [J2000] [J2000] [arcsec] [WHz−1] [degree] 0206+35(4C35.03) 020938.6 +354750 0.0377 0.739 I 90 24.8 40 group 1 3C353 172029.1 -005847 0.0304 0.601 II 186 26.3 90 poorcluster 2 3C270 121923.2 +054931 0.0075 0.151 I 580 24.4 90 group 3 M84 122503.7 +125313 0.0036 0.072 I 150 23.2 60 richcluster 3 Referencesfortheenvironmentalclassification:(1)Milleretal.(2002);(2)deVaucoleurs(1991);(3)Trageretal.(2000). 4 D. Guidettiet al. Table2.X-rayandradioequipartitionparametersforallthesources.Col.1:sourcename;Col.2:X-rayenergyband;Col.3:averagethermaltemperature; Cols.4,5,6and7,8,9best-fittingcoreradii,centraldensitiesandβparametersfortheouterandinnerβmodels,respectively;Col.10:averagethermalpressure atthemidpointoftheradiolobes;Cols.11&12:minimumsynchrotronpressureandcorrespondingmagneticfield;Col.13:referencesfortheX-raymodels. source band kT rcxout n0out βout rcxin n0in βin P0 Pmin BPmin ref. [keV] [keV] [kpc] [cm−3] [kpc] [cm−3] [dynecm−2] [dynecm−2] µG 0206+35 0.2-2.5 1.3+1.3 22.2 2.4×10−3 0.35 0.85 0.42 0.70 9.6×10−12 4.31×10−13 5.70 1,2 −0.3 3C353 ” 4.33+0.25 1.66×10−12 11.2 3 −0.24 3C270 0.3-7.0 1.45+0.23 36.8 7.7×10−3 0.30 1.1 0.34 0.64 5.75×10−12 1.64×10−13 3.71 4 −0.01 M84 0.6-7.0 0.6+0.05 5.28±0.08 0.42 1.40±0.03 1.70×10−11 1.07×10−12 9.00 5 −0.05 References:(1)Worrall&Birkinshaw(2000);(2)Worrall,Birkinshaw&Hardcastle(2001);(3)Iwasawaetal.(2000);(4)Crostonetal.(2008);(5)(Finoguenov&Jones 2001). radioemissionshowsacore,withsmoothtwo-sidedjetsalignedin matefortheinclinationofthejetsis≈90◦(Swain,Bridle&Baum theNW-SEdirection and surrounded by adiffuseand symmetric 1998).Theeasternjetisslightlybrighterandendsinawell-defined halo. Laing&Bridle (2011b) have estimated that the jets are hot spot. The radio lobes have markedly different morphologies: inclined by ≈40◦ withrespect tothelineof sight, withthemain theeasternlobeisroundwithsharpedges,whilethewesternlobeis (approaching)jetintheNWdirection. elongatedwithanirregularshape.Thelocationofthesourcewithin 0206+35 has been observed with both the ROSAT theclusterisofparticularinterestforthisworkandmightaccount PSPC and HRI instruments (Worrall&Birkinshaw for the different shapes of the lobes. Fig.1(d) shows the XMM- 1994, 2000; Trussonietal. 1997) and with Chandra Newton image overlaid on the radio contours. The image shows (Worrall,Birkinshaw&Hardcastle 2001). The X-ray emission onlytheNWpartofthecluster,butitisclearthattheradiosource consistsofacompactsourcesurroundedbyagalacticatmosphere liesontheedgeoftheX-rayemittinggasdistribution.sothatthe which merges into the much more extended intra-group gas. The roundeasternlobeisencounteringahigherexternaldensityandis radius of the extended halo observed by the ROSAT PSPC is probablyalsobehindalargercolumnofFaraday-rotatingmaterial ≈2.5arcmin (Fig. 1a). The ROSAT and ChandraX-ray surface (Iwasawaetal.2000,Goodgeretal.2008).Inparticular,theimage brightnessprofilesarewellfitbythecombinationofβmodelswith publishedbyGoodgeretal.(2008)showsthatthegasdensitygra- two different core radii and a power-law component (Hardcastle, dientpersistsonlargerscales. privatecommunication;Table2). 2.2 3C270 2.4 M84 3C270isaradiosourceclassifiedasFRIinmostoftheliterature, M84 is a giant elliptical galaxy located in the Virgo Cluster at althoughinfact,thetwolobeshavedifferentFRclassificationsat about400kpcfromthecore.Opticalemission-lineimagingshows lowresolution(Laing,Guidetti&Bridle2011).Theopticalcoun- adiskofionizedgasaroundthenucleus,withamaximumdetected terpartisthegiantellipticalgalaxyNGC4261,locatedatthecentre extent of 20×7arcsec2 (Hansen,Norgaard-Nielsen&Jorgensen of a nearby group. The radio source has a symmetrical structure 1985;Baumetal.1988;Boweretal.1997,2000).Theradioemis- with a bright core and twin jets, extending E-W and completely sionofM84(3C272.1)hasanangularextensionofabout3arcmin surroundedbylobes.Thelowjet/counter-jetratioindicatesthatthe (≃ 11kpc) and shows an unresolved core in the nucleus of the jets are close to the plane of the sky, with the Western side ap- galaxy,tworesolvedjetsandapairofwidelobes(Laing&Bridle proaching(Laing,Guidetti&Bridle2011). 1987;Laingetal.2011).Theinclinationtotheline-of-sightofthe The XMM-Newton image (Fig.1c) shows a disturbed dis- innerjetaxisis∼60◦,withthenorthernjetapproaching,butthere tribution with regions of low surface-brightness (cavities) at is a noticeable bend in the counter-jet very close to the nucleus, the positions of both radio lobes. A recent Chandra ob- which complicates modelling (Laing&Bridle 2011b). After this servation (Worralletal. 2010) shows “wedges” of low X- bend, the jets remain straight for ≈40arcsec, then both of them ray surface brightness surrounding the inner jets (see also bend eastwards by ∼90◦ and fade into the radio emission of the Croston,Hardcastle&Birkinshaw 2005, Finoguenovetal. 2006, lobes. Jethaetal.2007,Crostonetal.2008).Theoverall surface bright- ThemorphologyoftheX-rayemissionhasaH-shapemadeup ness profile is accurately reproduced by a point source con- of shells of compressed gas surrounding cavities coincident with volved with the Chandra point spread function plus a double both the radio lobes (Finoguenovetal. 2008; Finoguenovetal. β model (Crostonetal. 2008, projb model). Crostonetal. 2008 2006; Finoguenov&Jones 2001). This shape, together with the found no evidence for a temperature gradient in thehot gas. The fact that the initial bending of the radio jets has the same direc- group is characterized by high temperature and low luminosity tionandisquitesymmetrical,suggestsacombinationofinteraction (Finoguenovetal.2006),whichtakentogetherprovideaveryhigh withtheradioplasmaandmotionofthegalaxywithinthecluster levelofentropy. Thismightbeafurthersignofalargedegreeof (Finoguenov&Jones 2001).TheratiobetweentheX-raysurface impactoftheAGNontheenvironment. brightnessoftheshellsofthecompressedgasandtheirsurround- ingsis≈3andisalmostconstantaroundthesource.Theshellsare regionsofenhancedpressureanddensityandlowentropy:theam- 2.3 3C353 plitudeofthedensityenhancements(afactorof≈3)suggeststhat 3C353isanextendedFRIIradiosourceidentifiedwithaD-galaxy theyareproducedbyweakshockwaves(MachnumberM∼1.3) embedded at theperipheryof aclusterof galaxies.Thebest esti- drivenbytheexpandinglobes(Finoguenovetal.2006). Faradayrotationmeasurebandsacrossradiogalaxies 5 Figure2.RMimagesforallsources:(a)0206+35;(b)M84;(c)3C270;(d)3C353.Theangularresolutionandthelinearscaleofeachmapareshowninthe individualpanels. 3 ANALYSISOFRMANDDEPOLARIZATIONIMAGES Sincek ∝ |∇RM|2,Eq.3clearlyillustratesthathigherRMgra- dientsacrossthebeamgeneratehigherkvaluesandinturnhigher Forafully-resolvedforegroundFaradayscreen,theλ2 relationof depolarization. Eq.2holdsexactlyandthereisnochangeofdegreeofpolarization, p, with wavelength. Even in the presence of a small gradient of Our observational analysis is based on the following proce- RMacrossthebeam,λ2rotationisobservedoverawiderangeof dure.WefirstproducedRMandBurnlawkimagesattwodiffer- polarization angle. In this case, the emission tends to depolarize ent angular resolutions for each source and searched for regions withincreasingwavelength,followingtheBurnlaw(Burn1966): withhighk orcorrelatedRMandk values, whichcouldindicate p(λ)=p(0)exp(−kλ4), (3) thepresenceofinternalFaradayrotationand/orstrongRMgradi- entsacrossthebeam.Inregionswithlowkwherethevariationsof where p(0) is the intrinsic value of the degree of polarization RMareplausiblyisotropicandrandom,wethenusedthestructure andk=2|∇RM|2σ2,withFWHM = 2σ(2ln2)1/2 (Burn1966; function(definedinEq.4)toderivethepowerspectrumoftheRM Tribble1991;Laingetal.2008). fluctuations.Finally,toinvestigatethedepolarizationintheareasof 6 D. Guidettiet al. Figure3.BurnlawkimagesforallsourcesatthesameangularresolutionsasfortheRMimages:(a)0206+35at1.2arcsecFWHM;(b)M84at4.5arcsec FWHM;(c)3C270at5arcsecFWHM;(d)3C353at1.3arcsecFWHM.Thecorrespondingintegrateddepolarization(DP;Section5)isindicatedonthetop rightangleofeachpanel.Thecolourscaleisthesameforalldisplays. isotropicRM,andhencethemagneticfieldpoweronsmallscales, WeassumeRMpowerspectraoftheform: wemadenumericalsimulationsoftheBurnlawkusingthemodel Cˆ(f ) = C f−q f ≤f powerspectrumwithdifferentminimumscalesandcomparedthe ⊥ 0 ⊥ ⊥ max resultswiththedata. = 0 f⊥ >fmax Thestructurefunctionisdefinedby wheref isascalar spatialfrequency andfittheobserved struc- ⊥ turefunction(includingtheeffectoftheobservingbeam)usingthe S(r⊥)=<[RM(r⊥+r′⊥)−RM(r′⊥)]2 > (4) Hankel-transform method described by Laingetal. (2008) to de- rivetheamplitude,C andtheslope,q.ToconstraintheRMstruc- 0 (Simonetti,Cordes&Spangler 1984; Minter&Spangler 1996) tureonscalessmallerthanthebeamwidth,weestimatedthemini- wherer andr′ arevectorsintheplaneoftheskyandhiisan mumscaleofthebestfittedfieldpowerspectrum,Λ =1/f , ⊥ ⊥ min max averageoverr′ . whichpredictsameanvalueofkconsistentwiththeobservedone. ⊥ Faradayrotationmeasurebandsacrossradiogalaxies 7 quencytoderivetheresidualsathighresolution.Then,wefitthe Table3.Frequencies,bandwidthsandangularresolutionsusedintheRM andBurnlawkimagesdiscussedinSects.4and5,respectively. residualswithoutallowinganynπambiguitiesandaddedtheresult- ingRM’stothevaluesdeterminedatlowresolution.Thisprocedure source ν ∆ν beam allowedustoobtainanRMmapof0206+35freeofsignificantde- [MHz] [MHz] [arcsec] viationsfromλ2 rotationand fullyconsistent withthe1.2-arcsec measurements. 0206+35 1385.1 25 1.2 Wehaveverifiedthatthepolarizationanglesaccuratelyfollow 1464.9 25 therelation∆Ψ ∝ λ2overthefullrangeofpositionangleessen- 4885.1 50 tiallyeverywhereexceptforsmallareasaroundtheoptically-thick 3C353 1385.0 12.5 1.3 1665.0 12.5 cores:representativeplotsofΨagainstλ2for0206+35areshown 4866.3 12.5 inFig.4.Thelackofdeviationsfromλ2rotationinalloftheradio 8439.9 12.5 galaxies is fully consistent with our assumption that the Faraday 3C270 1365.0 25 1.65 rotatingmediumismostlyexternaltothesources. 1412.0 12.5 TheRMmapsareshown inFig.2. Thetypical rmserror on 4860.1 100 thefit is≈2radm−2. Nocorrection for theGalacticcontribution 1365.0 25 5.0 hasbeenapplied. 1412.0 12.5 All of the RM maps show two-dimensional patterns, RM 1646.0 25 bands, acrossthelobeswithcharacteristicwidthsrangingfrom3 4860.1 100 to12kpc.Multiplebandsparalleltoeachotherareobservedinthe M84 1413.0 25 1.65 4885.1 50 westernlobeof0206+35,theeasternlobeof3C353andthesouth- 1385.1 50 4.5 ernlobeofM84. 1413.0 25 Inallcases,theiso-RMcontoursarestraightandperpendicu- 1464.9 50 lartothemajoraxesofthelobestoaverygoodapproximation:the 4885.1 50 very straight and well-defined bands in the eastern lobes of both 0206+35 (Fig.2a) and 3C353 (Fig.2d) are particularly striking. TheentireareaofM84appearstobecoveredbyabandedstructure, Inthispaper,weareprimarilyinterestedinestimatingtheRM whileinthecentralpartsof0206+35and3C270andthewestern powerspectrumoverlimitedareas,andwemadenoattempttode- lobe of 3C353, regions of isotropic and random RM fluctuations terminetheouterscaleoffluctuations. arealsopresent. Theuseofthestructurefunctiontogether withtheBurnlaw WealsoderivedprofilesofhRMialongtheradioaxisofeach k represents a powerful technique to investigate the RM power source, averaging over boxes a few beamwidths long (parallel to spectrum over a wide range of spatial scales (Laingetal. 2008; the axes), but extended perpendicular to them to cover the entire Guidettietal.2010).Thetwoquantitiesarecomplementary,inthat width of the source. The boxes are all large enough to contain thestructurefunctionallowsustodeterminethepowerspectrumof many independent points. The profiles are shown in Fig.5. For thefluctuationsonscaleslargerthanthebeamwidth,whiletheBurn each radio galaxy, we also plot an estimate of the Galactic con- lawkconstrainsfluctuationsofRMbelowtheresolutionlimit. tributiontotheRMderivedfromaweightedmeanoftheintegrated RM’s for non-cluster radio sources within a surrounding area of 10deg2 (Simard-Normandinetal. 1981). In all cases, both posi- tiveandnegativefluctuationswithrespecttotheGalacticvalueare 4 ROTATIONMEASUREIMAGES present. The RM images and associated rms errors were produced by In 0206+35 (Fig.2a), the largest-amplitude bands are in the weighted least-squares fitting to the observed polarization angles outerpartsofthelobes,withapossiblelow-levelbandjusttothe Ψ(λ)asafunctionofλ2 (Eq.1)atthreeorfourfrequencies(Ta- NWofthecore.Themostprominentband(withthemostnegative ble3,seealsoLaingetal.2011andLaing,Guidetti&Bridle2011) RMvalues)isintheeastern(receding)lobe,about15kpcfromthe usingtheRMtaskintheAIPSpackage. core (Fig.5a). Its amplitude with respect to the Galactic value is Each RM map was calculated only at pixels with rms about40radm−2.Thisbandmustbeassociatedwithastrongor- polarization-angle uncertainties <10◦ at all frequencies. We re- deredmagneticfieldcomponentalongthelineofsight.Ifcorrected feronlytothelower-resolutionRMandk imagesfor3C270and fortheGalacticcontribution,thetwoadjacentbandsintheeastern M84(Table3),asthey show moreof thefaint, extended regions lobewouldhaveRMwithoppositesignsandthefieldcomponent ofthesesourcesandarefullyconsistentwiththehigher-resolution alongthelineofsightmustthereforereverse. versions.TheRMimageofM84isconsistentwiththatshownby M84 (Fig.2b) displays an ordered RM pattern across the Laing&Bridle(1987),butisderivedfromfour-frequencydataand whole source, with two wide bands of opposite sign having the hasahighersignal-to-noiseratio. highestabsoluteRMvalues.Thereisalsoanabruptchangeofsign In the fainter regions of 0206+35 (for which only three fre- acrosstheradiocore(seealsoLaing&Bridle1987).Thenegative quencies are available and the signal-to-noise ratio is relatively bandinthenorthernlobe(associatedwiththeapproachingjet)has low), the RM task occasionally failedto determine the nπ ambi- alargeramplitudewithrespecttotheGalacticvaluethanthecor- guitiesinpositionanglecorrectly.Inordertoremovetheseanoma- responding(positive)featureinthesouthernlobe(Fig.5c). lies,wefirstproducedalower-resolution,buthighsignal-to-noise 3C270(Fig.2c)showstwolargebands:oneonthefrontend RMimagebyconvolvingthe1.2arcsecRMmaptoabeamwidthof of the eastern lobe, the other in the middle of the western lobe. 5arcsecFWHM.Fromthismapwederivedthepolarization-angle The bands have opposite signs and contain the extreme positive rotationsateachofthethreefrequenciesandsubtractedthemfrom andnegativevaluesoftheobservedRM.Thepeakpositivevalueis the observed 1.2arcsec polarization angle maps at the same fre- withintheeasternbandattheextremeendofthelobe(Fig.5e). 8 D. Guidettiet al. The RM structure of 3C353 (Fig.2d) is highly asymmetric. thesouthernjet(Fig.3b).Thereisnocorrespondingfeatureinthe The eastern lobe shows a strong pattern, made up of four bands, RMimage(Fig.2b).Thedepolarizationislikelytobeassociated withverystraightiso-RMcontourswhicharealmostexactlyper- with one of the shells of compressed gas visible in the Chandra pendiculartothesourceaxis.Asin0206+35,adjacentbandshave image(Fig.1b),implyingsignificantmagnetizationwithinhomo- RMwithopposite signsoncecorrectedfortheGalacticcontribu- geneousfieldand/ordensitystructureonscalesmuchsmallerthan tion(Fig.5g).Incontrast,theRMdistributioninthewesternlobe thebeamwidth,apparentlyindependentofthelarger-scalefieldre- showsnosignofanybandedstructure,andisconsistentwithran- sponsiblefortheRMbands.Thispictureissupportedbythegood domfluctuationssuperimposedonanalmostlinearprofile.Itseems spatialcoincidenceofthehighkregionwithashellofcompressed verylikelythatthedifferencesinRMmorphologyandaxialratio gas,asillustratedintheoverlayofthe4.5arcsecBurnlawkimage arebothrelatedtotheexternaldensitygradient(Fig.1d). onthecontoursoftheChandradata(Fig.6(a)).Coolergasassoci- In Table 4 the relevant geometrical features (size, distance atedwiththeemission-linediskmightalsoberesponsible,butthere fromtheradiocore,hRMi)fortheRMbandsarelisted. isnoevidenceforspatialcoincidencebetweenenhanced depolar- izationand Hαemission(Hansen,Norgaard-Nielsen&Jorgensen 1985). Despite the complex morphology of the X-ray emission around M84, its k profile is very symmetrical, with the highest 5 DEPOLARIZATION valuesatthecentre(Fig.5(d)). In this section, we use “depolarization” in its conventional sense 3C270 also shows areas of very strong depolarization tomean“decreaseofdegreeofpolarizationwithincreasingwave- (k ∼550rad2m−4,correspondingtoDP=0.35)closetothecore length” and define DP = p /p . Using the Faraday and surrounding the inner and northern parts of both the radio 1.4GHz 4.9GHz code (Murgiaetal. 2004), we produced images of Burn law k lobes.AsforM84,theareasofhighk arecoincidentwithridges by weighted least-squares fitting to lnp(λ) as a function of λ4 intheX-rayemissionwhichformtheboundariesofthecavitysur- (Eq.3). Onlydatawithsignal-to-noise ratio>4inP ateach fre- roundingthelobes(Fig.6(b)).TheinnerpartsofthisX-raystruc- quencywereincludedinthefits.TheBurnlawkimageswerepro- ture are described in more detail by Worralletal. (2010), whose duced with the same angular resolutions as the RM images. The recenthigh-resolutionChandraimageclearlyreveals“wedges”of 1.65arcsec resolution Burn law k maps for M84 and 3C270 are low brightness surrounding the radio jets. As in M84, the most consistentwiththelow-resolutionones,butaddnoadditionaldetail likely explanation is that a shell of denser gas immediately sur- andarequitenoisy.Thiscouldleadtosignificantlybiasedestimates roundingtheradiolobesismagnetized,withsignificantfluctuations forthemeanvaluesofkoverlargeareas(Laingetal.2008).There- of field strength and density on scales smaller than our 5-arcsec fore,asfortheRMmaps,weusedonlytheBurnlawkimagesat beam, uncorrelated with the RM bands. The k profile of 3C270 lowresolutionforthesetwosources. (Fig.5(f))isvery symmetrical, suggesting that the magnetic-field TheBurnlawkmapsareshowninFig.3.Allofthesources anddensitydistributionsarealsosymmetricalandconsistentwith showlowaveragevaluesofk(i.e.slightdepolarization),suggest- anorientationclosetotheplaneofthesky.Thelargestvaluesofk ing littleRM power on small scales. Withthe possible exception areobservedinthecentre,coincidentwiththefeaturesnotedear- of the narrow filaments of high k in the eastern lobe of 3C353 lierandwiththebulkoftheX-rayemission(thehighk valuesin (which might result from partially resolved RM gradients at the thetwooutermostbinshavelowsignal-to-noiseandarenotsignif- band edges), none of the images show any obvious structure re- icant). latedtotheRMbands.Foreachsource,wehavealsocomparedthe In the Burn law k image of 3C353, there is evidence for a RMandBurnlawkvaluesderivedbyaveragingovermanysmall straightandknottyregionofhighdepolarization≈20kpclongand boxes covering theemission, and we findno correlationbetween extending westwards from the core. This region does not appear them. toberelatedeithertothejetsortoanyother radiofeature.Asin Wealsoderivedprofilesofk(Fig.5b,d,fandh)withthesame M84and3C270,theRMappearsquitesmoothovertheareashow- setsofboxesasfortheRMprofilesinthesameFigure.Thesecon- inghighdepolarization,againsuggestingthattherearetwoscales firmthatthevaluesofkmeasuredinthecentresoftheRMbands ofstructure,onemuchsmallerthanthebeam,butproducingzero arealwayslow,butthatthereislittleevidenceforanydetailedcor- meanRMandtheotherverywellresolved.In3C353,thereisas relation. yetnoevidenceforhotorcoolionizedgasassociatedwiththeen- The signal-to-noise ratiofor 0206+35 isrelativelylow com- hanceddepolarization(contaminationfromtheverybrightnuclear paredwiththatoftheotherthreesources, particularlyat4.9GHz X-rayemissionaffectsanareaof1arcminradiusaroundthecore; (we need to use a small beam to resolve the bands), and this is Iwasawaetal.2000,Goodgeretal.2008). reflectedinthehighproportionofblankedpixelsonthek image. Thekprofileof3C353(Fig.5(h))showsamarkedasymme- Themostobviousfeatureofthisimage(Fig.3a),anapparentdiffer- try,withmuchhighervaluesintheEast.Thisisinthesamesenseas enceinmeankbetweenthehigh-brightnessjets(lessdepolarized) thedifferenceofRMfluctuationamplitudes(Fig.5(g))andisalso andthesurroundingemission,islikelytobeanartefactcausedby consistentwiththeeasternlobebeingembeddedinhigher-density our blanking strategy: pointswhere thepolarized signal islow at gas.Therelativelyhighvaluesofkwithin20kpcofthenucleusin 4.9GHz are blanked preferentially, so the remainder show artifi- theWesternlobearedueprimarilytothediscreteregionidentified ciallyhighpolarizationatthisfrequency.Forthesamereason,the earlier. apparentminimuminkatthecentreofthedeep,negativeRMband (Fig.5aandb)isprobablynotsignificant.Theaveragedvaluesof kfor0206+35arealreadyverylow,however,andarelikelytobe 6 ROTATIONMEASURESTRUCTUREFUNCTIONS slightlyoverestimated,soresidualRMfluctuationsonscalesbelow the1.2-arcsecbeamwidthmustbeverysmall. We calculated RM structure function for discrete regions of the M84 shows onelocalised areaof verystrong depolarization sourceswheretheRMfluctuationsappeartobeisotropicandran- (k ∼500rad2m−4, corresponding to DP = 0.38) at the base of domandforwhichweexpectthespatialvariationsofforeground Faradayrotationmeasurebandsacrossradiogalaxies 9 Table4.PropertiesoftheRMbands:Col.1sourcename;Cols.2&3:overallhRMiandσRM;Col.4:GalactichRMi;Col.5:hRMiforeachband;Col. 6:distanceofthebandmidpointfromtheradiocore(positivedistancesareinthewesterndirectionforallsources,exceptforM84,wheretheyareinthe northerndirection);Col.7:widthoftheband;Col.8:maximumbandamplitude. source hRMi σRM RMG bandhRMi dc width A [radm−2] [radm−2] [kpc] [kpc] [radm−2] 0206+35(4C35.03) −77 23 −72 −140 -15 10 40 −60 -27 4 34 22 6 51 8 4 3C353 −56 24 −69 122 -12 5 50 102 -19 4 -40 -23 4 100 -26 4 3C270 14 10 12 −8 20 12 10 32 37 11 M84 −2 15 2 −27 1 3 10 22 -6 6 thermalgasdensity,rmsmagneticfieldstrengthandpathlengthto effectsoflarge-scalevariationsinpathlengthorfieldstrength(cf. bereasonablysmall.Theseare:theinner26arcsecofthereceding Guidettietal.2010). (Eastern)lobeof0206+35,theinner100arcsecof3C270andthe InordertoconstrainRMstructureonspatialscalesbelowthe inner 40arcsec of thewesternlobeof 3C353. Theselectedareas beamwidth, we estimated the depolarization as described in Sec- of0206+35and3C270arebothwithinthecoreradiiofthelarger- tion3.ThefittedkvaluesarelistedinTable5.Westressthatthese scalebetamodelsthatdescribethegroup-scaleX-rayemissionand values refer only to areas with isotropic fluctuations, and cannot thegalaxy-scalecomponentsaretoosmalltoaffecttheRMstatis- usefullybecomparedwiththeintegrateddepolarizationsquotedin ticssignificantly(Table2).In3C353,theselectedareawaschosen inFig.2. to be small compared with the scale of X-ray variations seen in For M84, using the Burn law k analysis and assuming that Fig.1(d).Inallthreecases,theforegroundfluctuationsshouldbe variation of Faraday rotation across the 1.65-arcsec beam causes fairlyhomogeneous. TherearenosuitableregionsinM84,which theresidualdepolarization,wefindthatΛ <0.1kpcforanyrea- min∼ isentirelycoveredbythebandedRMpattern. sonableRMpowerspectrum. The structure functions, corrected for uncorrelated random noisebysubtracting2σ2 (Simonettietal.1984),areshownin noise Fig.7.Alloftheobservedstructurefunctionscorrespondtopower 7 ROTATION-MEASUREBANDSFROMCOMPRESSION spectra of approximately power-law form over all or most of the It is clear from the fact that the observed RM bands are perpen- rangeofspatialfrequencieswesample.Weinitiallyassumedthat dicular to the lobe axes that they must be associated with an in- thepowerspectrumwasdescribedbyEq.5withnohigh-frequency teraction between an expanding radio source and the gas imme- cut-off (f → ∞) and made least-squares fitsto the structure max diatelysurrounding it.Oneinevitablemechanismisenhancement functions,weightedbyerrorsderivedfrommultiplerealizationsof offieldanddensitybytheshockorcompressionwavesurrounding thepowerspectrumontheobservinggrid,asdescribedindetailby thesource.2TheimplicationofthepresenceofcavitiesintheX-ray Laingetal.(2008)andGuidettietal.(2010). gasdistributioncoincidentwiththeradiolobesisthatthesources Thebest-fittingslopesq andamplitudesC aregiveninTa- areinteractingstronglywiththethermalgas,displacingratherthan 0 ble5. All of the fitted power spectra are quite flat and have low mixing withit (seeMcNamara&Nulsen 2007 for areview). For amplitudes,implyingthatthereislittlepowerintheisotropicand the sources in the present paper, the X-ray observations of M84 randomcomponentofrotationmeasure.Indeed,theamplitudesof (Finoguenovetal. 2008,Fig.1b)and3C270(Crostonetal.2008, thelargest-scaleRMfluctuationssampledinthisanalysisisafew Fig.1c)showcavitiesandarcsofenhancedbrightness,correspond- timeslessthanthatofthebands(seeTables4and5).Thissuggests ing to shells of compressed gas bounded by weak shocks. The thatthefieldresponsibleforthebandsisstrongeraswellasmore strengthofanypre-existingfieldintheIGM,whichwillbefrozen orderedthanthatresponsiblefortheisotropicfluctuations. intothegas,willalsobeenhanced intheshells.Wethereforeex- pectasignificantenhancementinRM.Amoreextremeexampleof The structure functions for 0206+35 and 3C353 rise mono- thiseffectwilloccuriftheexpansionoftheradiosourceishighly tonically, indicating that the outer scale for the random fluctua- supersonic,inwhichcasetherewillbeastrongbow-shockaheadof tions must be larger than the maximum separations we sample. For 3C270, the structure function levels out at r ≈ 100arcsec ⊥ (15kpc; Fig. 7d). This could be the outer scale of the field fluc- 2 Analternativemechanismisthegenerationofnon-linearsurfacewaves tuations, but a better understanding of the geometry and external (Bicknell,Cameron&Gingold 1990).Itisunlikely thatthis canproduce densitydistributionwouldbeneededbeforewecouldruleoutthe large-scalebands,forthereasonsgiveninSection9.5. 10 D. Guidettiet al. Table5.Powerspectrumparametersfortheindividualsub-regions.Col.1:sourcename;Col.2:angularresolution;Col.3:slopeq:Col.4:amplitudeC0; Col.5:minimumscale;Col.6:amplitudeofthelargescaleisotropiccomponent;Col.7:observedmeank;Col.8:predictedmeank.Thepowerspectrumhas notbeencomputedforM84(seeSection6). Source FWHM q logC0 Λmin Aiso kobs ksyn [arcsec] [kpc] [radm−2] [rad2m−4] [rad2m−4] 0206+35 1.2 2.1 0.77 2 10 37 40 3C270 1.65 2.7 0.90 0.3 5 30 26 5.0 2.7 0.90 0.3 5 71 64 3C353 1.3 3.1 0.99 0.1 10 38 33 M84 1.65 <0.1 25 4.5 <0.1 43 0206+35 Figure4.PlotsofE-vectorpositionangleΨagainstλ2atrepresentativepointsofthe1.2arcsecRMmapof0206+35.FitstotherelationΨ(λ)=Ψ0+RMλ2 areshown.ThevaluesofRMaregivenintheindividualpanels. thelobe,behindwhichboththedensityandthefieldbecomemuch lackof detection of strong shocks in theX-raydatafor the other higher.Regardlessofthestrengthoftheshock,thefieldismodified sources. sothatonlythecomponent intheplaneoftheshockisamplified Inthissection,weinvestigatehowtheRMcouldbeaffected andthepost-shockfieldtendsbecomeorderedparalleltotheshock bycompression.Weconsideradeliberatelyoversimplifiedpicture surface. in which the radio source expands into an IGM with an initially uniform magnetic field, B. This is the most favourable situation Theevidencesofarsuggeststhatshocksaroundradiosources forthegenerationoflarge-scale,anisotropicRMstructures:inre- of both FR classes are generally weak (e.g. Formanetal. 2005, ality,thepre-existingfieldislikelytobehighlydisordered,oreven Wilsonetal.2006,Nulsenetal.2005).Thereareonlytwoexam- isotropic,becauseofturbulenceinthethermalgas.Westressthat ples inwhich highly supersonic expansion has been inferred: the wehavenottriedtogenerateaself-consistentmodelforthemag- southern lobe of CentaurusA (M ≈ 8; Kraftetal. 2003) and neticfieldandthermaldensity,butrathertoillustratethegeneric NGC3801(M≈4;Crostonetal.2007).Thereisnoevidencethat effectsofcompressionontheRMstructure. the sources described in the present paper are significantly over- pressured compared with the surrounding IGM (indeed, the syn- Inthismodeltheradiolobeisanellipsoidwithitsmajoraxis chrotronminimumpressure issystematicallylowerthanthether- alongthejetandissurroundedbyasphericalshellofcompressed malpressureoftheIGM;Table 2).Thesidewaysexpansionofthe material.Thisshelliscentredatthemid-pointofthelobe(Fig.8) lobesisthereforeunlikelytobehighlysupersonic.TheshockMach and has a stand-off distance equal to 1/3 of the lobe semi-major numberestimatedforallthesourcesfromrampressurebalancein axis at the leading edge (the radius of the spherical compression theforwarddirectionisalso≈1.3.Thisestimateisconsistentwith isthereforeequalto4/3ofthelobesemi-majoraxis).Inthecom- thatforM84madebyFinoguenovetal.(2006)andalsowiththe pressedregion,thethermaldensityandthemagneticfieldcompo-

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.