A&A603,A125(2017) Astronomy DOI:10.1051/0004-6361/201630014 & (cid:13)c ESO2017 Astrophysics The two-component giant radio halo in the galaxy cluster Abell2142 T.Venturi1,M.Rossetti2,G.Brunetti1,D.Farnsworth3,4,F.Gastaldello2,S.Giacintucci5,6,D.V.Lal7,L.Rudnick3, T.W.Shimwell8,D.Eckert9,S.Molendi2,andM.Owers10,11 1 INAF–IstitutodiRadioastronomia,viaGobetti101,40129Bologna,Italy e-mail:[email protected] 2 INAF–IASF-Milano,viaBassini15,20133Milano,Italy 3 MinnesotaInstituteforAstrophysics,SchoolofPhysicsandAstronomy,UniversityofMinnesota,116ChurchStreetSE, Minneapolis,MN,55455,USA 4 Cray,Inc.,380JacksonStreet,Suite210,St.Paul,MN55101,USA 5 NavalResearchLaboratory,Washington,DC20375,USA 6 DepartmentofAstronomy,UniversityofMaryland,CollegePark,MD,20742-2421,USA 7 NationalCentreforRadioAstrophysics,TIFR,PostBag3,Ganeshkhind,411007Pune,India 8 LeidenObservatory,LeidenUniversity,POBox9513,2300RALeiden,TheNetherlands 9 DepartmentofAstronomy,UniversityofGeneva,ch.d’Écogia16,1290Versoix,Switzerland 10 AustralianAstronomicalObservatory,POBox915,NorthRyde,NSW1670,Australia 11 DepartmentofPhysicsandAstronomy,MaquarieUniversity,NSW2109,Australia Received4November2016/Accepted10March2017 ABSTRACT Aims.WereportonaspectralstudyatradiofrequenciesofthegiantradiohaloinA2142(z=0.0909),whichweperformedtoexplore itsnatureandorigin.TheopticalandX-raypropertiesoftheclustersuggestthatA2142isnotamajormergerandthepresenceofa giantradiohaloissomewhatsurprising. Methods. We performed deep radio observations of A2142 with the Giant Metrewave Radio Telescope (GMRT) at 608 MHz, 322MHz,and234MHzandwiththeVeryLargeArray(VLA)inthe1–2GHzband.Weobtainedhigh-qualityimagesatallfrequen- ciesinawiderangeofresolutions,fromthegalaxyscale,i.e.∼5(cid:48)(cid:48),upto∼60(cid:48)(cid:48)toimagethediffusecluster–scaleemission.Theradio haloiswelldetectedatallfrequenciesandextendsouttothemostdistantcoldfrontinA2142,about1Mpcawayfromthecluster centre.Westudiedthespectralindexintworegions:thecentralpartofthehalo,wheretheX–rayemissionpeaksandthetwobrightest dominantgalaxiesarelocated;andasecondregion,knownastheridge(inthedirectionofthemostdistantsouth–easterncoldfront), selectedtofollowthebrightpartofthehaloandX-rayemission.WecomplementedourdeepobservationswithapreliminaryLOw FrequencyARray(LOFAR)imageat118MHzandwiththere-analysisofarchivalVLAdataat1.4GHz. Results.Thetwocomponentsoftheradiohaloshowdifferentobservationalproperties.Thecentralbrightestparthashighersurface brightessandaspectrumwhosesteepnessissimilartothoseoftheknownradiohalos,i.e.α1.78GHz =1.33±0.08.Theridge,which 118MHz fadesintothelargerscaleemission,isbroaderinsizeandhasconsiderablylowersurfacebrightessandamoderatelysteeperspectrum, i.e.α1.78GHz ∼ 1.5.WeproposethatthebrightestpartoftheradiohaloispoweredbythecentralsloshinginA2142,inaprocess 118MHz similartowhathasbeensuggestedformini-halos,orbysecondaryelectronsgeneratedbyhadroniccollisionsintheICM.Onthe otherhand,thesteeperridgemayprobeparticlere-accelerationbyturbulencegeneratedeitherbystirringthegasandmagneticfields onalargerscaleorbylessenergeticmechanisms,suchascontinuousinfallofgalaxygroupsoranoff-axis(minor)merger. Keywords. galaxies:clusters:general–galaxies:clusters:individual:A2142–radiocontinuum:general 1. Introduction historyandprovideinvaluableinformationconcerningtheirdy- namical state (see Markevitch & Vikhlinin 2007). At the same time,thedeepradioimagingachievedbelow1GHzbytheGi- Cluster mergers are the most energetic events in the Universe. Withtotalenergyoutputsoforder1063−1064 erg,mergersarea antMetrewaveRadioTelescope(GMRT),andmorerecentlyby LOFAR,issheddinganewlightonthepropertiesofnon-thermal naturalwaytoaccountformassassembly:galaxyclustersform componentsingalaxyclusters. asaconsequenceofmergertreestoreachandexceedmassesof order1015 M .Thegravitationalenergyreleasedintothecluster Radio halos represent the most spectacular non-thermal ef- (cid:12) volumeduringmergersdeeplyaffectsthedynamicsofthegalax- fectsofclustermergers.Radiohalosarediffusesynchrotronra- iesandthepropertiesofthethermalintraclustermedium(ICM) diosourcesthatcoverthewholeclustervolume,i.e.upandbe- andnon-thermalrelativisticparticleandmagneticfieldemission yondMpcsize.Thesehaloshavesteepspectra,forwhichtypical (Brunetti & Jones 2014). The impressive quality of the X–ray valuesareα ∼ 1.2–1.3,forS ∝ ν−α,whereS isthefluxdensity ChandraandXMM–Newtonimages(e.g.Markevitchetal.2000; andαisthespectralindexofthesynchrotronradiospectrum,and Rossettietal.2013)showsavarietyoffeaturesinthehotICM, extremely low surface brightness of a fraction of µJy arcsec−2 suchascoldfrontsandshocks,whichtracetheclusterformation (see the Feretti et al. 2012, for a recent observational review). ArticlepublishedbyEDPSciences A125,page1of17 A&A603,A125(2017) Radiohalosarenotubiquitousingalaxyclusters;only30–40% attheunprecedenteddistanceof∼1Mpcfromtheclustercentre ofmassiveclustersintheUniverse(M ≥ 1015 M )hostaradio was detected with XMM–Newton and studied by Rossetti et al. (cid:12) halo.Inthosecases,the1.4GHzradiopoweroftheMpc-scale (2013),whoproposedthatlarge-scalesloshingwastherespon- halocorrelateswithboththeclusterX-rayluminosityandmass siblemechanismforitsorigin;thiseitherresultedfromthelong- (e.g.Liangetal.2000;Brunettietal.2007;Basu2012;Cassano termevolutionofcentralsloshingcommoninmanyrelaxedclus- etal.2013;Cucitietal.2015).Forthoseclusterswithoutaradio tersorfromamergerofintermediatestrength. halo, upper limits to the radio power derived from the GMRT Diffuse radio emission on the scale of few hundred kpc radiohalosurveyareordersofmagnitudebelowthecorrelation around the two brightest cluster galaxies (BCGs) was detected (Venturi et al. 2007; and 2008; Cassano et al. 2013; Kale et al. byGiovannini&Feretti(2000).Amorerecentstudyperformed 2013and2015a). withtheGreenBankTelescope(GBT)at1410MHzshowsthat Theoriginofgiantradiohalosisstilladebatedissue,how- thesizeofthisdiffuseradioemissionis∼2Mpc,extendingeven ever the connection between radio halos and cluster mergers is beyondthemostdistantcoldfront.Themajoraxisofthisgiant quantitativelysupportedbythedistributionofclusterswithand radiohaloisalignedinthesamesouth-eastdirectionwherethe without radio halos as a function of a number of indicators of large-scale cold front is located (Farnsworth et al. 2013). The the X-ray substructure: giant radio halos are always found in GBT image shows that its surface brightness is very low, i.e. unrelaxed clusters (Buote 2001; Cassano et al. 2010). On the ∼0.2 µJyarcsec−2, but the poor angular resolution does not al- other hand, a strong correlation is found between the cool-core lowadetailedcomparisonwiththeX-rayimages.Toovercome strengthinrelaxedclustersandthepresenceofradiomini-halos; the resolution limitations of the GBT image and to study the thatis,diffusecluster-scalesourcesthataresmallerinsizethan spectralpropertiesofthisexceptionalradiohaloforcomparison giant radio halos (of the order of up to few hundreds of kpc), withtheX-rayemissionandopticalinformation,weundertooka which always encompass the radio emission from the cluster studywiththeGMRTat608MHz,322MHzand234MHz,and dominant galaxy, but whose origin is not closely related to the with the Karl Jansky Very Large Array (VLA) in the 1–2 GHz currentcycleofactivity(e.g.Bravietal.2016;Giacintuccietal. band.WecomplementedouranalysiswithLOwFrequencyAR- 2014;Kaleetal.2015b;Mittaletal.2009). ray(LOFAR)dataat118MHzandwitharchivalVLAobserva- A possible explanation for the origin of giant radio halos tionsat1.4GHz. is the re-acceleration of in situ relativistic particles by turbu- In this paper we present the results of our work. The pa- lence injected into the ICM during cluster merger events (the perisorganizedasfollows:inSect.2wedescribetheobserva- so-called re-acceleration model; see Brunetti et al. 2001). Sec- tionsanddataanalysis;theimagesarepresentedinSect.3;the ondary(hadronic)modelsfortheoriginoftherelativisticparti- spectral study is presented in Sect. 4; the results are discussed cles(e.g.Dennison1980;Blasi&Colafrancesco1999)arecur- in Sect. 5; and the summary and future prospects are given in rentlylessfavouredbecauseofthelackofdetectionofpredicted Sect.6.InAppendixAwecomplementtheinformationandre- γ-ray emission from galaxy clusters by the Fermi satellite, and port on the radio emission associated with the cluster galaxies. becauseofthediscoveryofradiohaloswithultra-steepspectra Throughout the paper we use the convention S ∝ ν−α. We as- (α ≥ 1.5), whose relativistic energy would exceed the thermal sumeastandardΛCDMcosmologywithH =70kms−1Mpc−1, o energy under the secondary model assumptions (see the proto- Ω =0.3,Ω =0.7.Attheclusterredshift(z=0.0909)thiscor- M v type case of A521; Brunetti et al. 2008). Mixed hadronic and respondstoascaleof1.705kpc/(cid:48)(cid:48) andtoaluminositydistance re-accelerationmodelshavebeenproposedandsomelevelofra- D =418.6Mpc. L dioemissionisexpectedin“radiooff-stateclusters”(e.g.Brown etal.2011;Brunetti&Lazarian2011;Zandaneletal.2014). Despite the statistical connection between radio halos and 2. Observationsanddatareduction mergers, a few radio halos have been found in less disturbed 2.1. ObservationswiththeGMRT systems. An example of these outliers is the giant radio halo in the strong cool-core cluster CL1821+643 (Bonafede et al. We observed A2142 with the GMRT (Pune, India) in March 2014). A study of the X-ray morphological parameters of this 2013 at 608, 322, and 234 MHz. The logs of the observations cluster shows that it shares the same properties of galaxy clus- aregiveninTable1. ters hosting a radio halo, suggesting that it may be a case of a Thedatawerecollectedinspectral-linemodeatallfrequen- clustermergerinwhichthecoolcorehasbeenpreserved(Kale cies, i.e. 256 channels at 322 and 608 MHz, and 128 channels & Parekh 2016). More recently, a giant radio halo has been re- at 234 MHz, with a spectral resolution of 125 kHz/channel at portedinA2261andA2390,neitherofwhichisamajormerger 322MHzand608MHz,and65kHz/channelat234MHz.The (Sommer et al. 2016). Current models predict that radio halos raw data were first processed with the software flagcal (Prasad canalsobegeneratedinlessdisturbedsystems,althoughwitha & Chengalur 2013; Chengalur 2013) to remove RFI and ap- probabilitythatissignificantlylowerthaninthecaseofmassive ply bandpass calibration, then further editing, self-calibration, majormergers(Cassanoetal.2006;andBrunetti&Jones2014, andimagingwereperformedusingtheNRAOAstronomicalIm- forareview).Suchoutliersarethusveryimportant,astheymay ageProcessingSystem(AIPS)package.Thesources3C286and provide important constraints on the origin of radio halos, and 3C48wereusedasprimary(amplitude)calibrators.Inorderto probe a piece of the theoretical framework that is still poorly findacompromisebetweenthesizeofthedatasetandtheneedto explored. minimizebandwidthsmearingeffectswithintheprimarybeam, ThegalaxyclusterA2142(z=0.0909)isanotherchallenge after bandpass calibration the central channels in each individ- toourunderstandingoftheformationofgiantradiohalos. ualdatasetwereaveragedto30,39,and26channelsof∼1MHz A2142 is massive (M ∼ 8.8×1014 M , Cuciti et al. 2015) eachat608MHz,322MHz,and234MHz,respectively. (cid:12) and is located at the centre of a supercluster (Einasto et al. Ateachfrequencyweperformedmulti-facetimaging,cover- 2015; Gramann et al. 2015) with ongoing accretion groups. inganareaof∼2.5◦×2.5◦at234MHz,∼1.8◦×1.8◦at322MHz It is the first object where cold fronts were discovered by and∼1.4◦×1.5◦ at608MHz,respectively.ThefieldofA2142 Chandra (Markevitch et al. 2000). Recently, another cold front isextremelycrowded,asisclearfromFig.1,andincludesmany A125,page2of17 T.Venturietal.:Thetwo–componentgiantradiohaloinA2142 Table1.Observations. Array ProjectID Obs.Date ν ∆ν Time FWHM rmsa MHz MHz h (cid:48)(cid:48)×(cid:48)(cid:48), ◦ mJyb−1 GMRT 23_017 23-03-13 234 16 10 12.7×11.0,68 0.18 23_017 28-03-13 322 32 10 9.8×7.5,58 0.13 23_017 22-03-13 608 32 10 5.4×4.5,59 0.03 VLAb VLA11B-156 09-10-11 1500 1000 1.5 c c Notes. (a) Valuemeasuredfarfromthefieldcentre;(b) theVLAobservationsconsistofthreepointingswiththesameset-upandexposuretime (seeSect.2.2).Herewegivethetotaltimeoverthethreepointings;(c)seeSect.2.2.andTable2. strong sources, which had to be properly self-calibrated and Table2.ParametersofthefullresolutionVLAimages. cleanedtoreachthetargetedrms(seeTable1). At each frequency we produced final images, over a wide ν ∆ν FWHM rms range of resolutions and with different tapering and weight- MHz MHz (cid:48)(cid:48)×(cid:48)(cid:48) mJyb−1 ing schemes, to account for the complexity of the radio emis- 1380 250 40(cid:48)(cid:48)×37(cid:48)(cid:48) 0.15 sion in the field. In particular, full resolution images were used 1780 200 32(cid:48)(cid:48)×29(cid:48)(cid:48) 0.07 to subtract the strongest radio sources at distances larger than ∼0.8◦−1.5◦ (depending on the frequency) from the field cen- tre,andthentaperingandrobustweightingwereusedtoimage the diffuse radio sources and the radio halo. Images with mul- noise and less complete u−v coverage. Standard data flagging andreductiontechniqueswereperformedwithCASA,usingthe tiple resolutions were produced with resolutions ranging from 5.5(cid:48)(cid:48) × 4.5(cid:48)(cid:48) to 39.1(cid:48)(cid:48) × 36.6(cid:48)(cid:48) at 608 MHz, 9.8(cid:48)(cid:48) × 7.5(cid:48)(cid:48) to VLAcalibratorsourcesJ1331+3030(3C286)andJ1609+2641 53.4(cid:48)(cid:48)×43.7(cid:48)(cid:48)at322MHz,and12.7(cid:48)(cid:48)×11.0(cid:48)(cid:48)to45.19(cid:48)(cid:48)×44.80(cid:48)(cid:48) for flux and phase calibration, respectively. After editing for RFI, roughly 45% of the total bandwidth remained, yielding at234MHz. ∼450MHzoversevencleanspectralwindows.Wecreatedtwo Finally,toimagetheradiohalowefirstproducedhighreso- images, one using 250 MHz bandwidth around 1.38 GHz and lutionimagesusingtheu−vspacings>2kλ.Thenweusedthese one using 200 MHz around 1.78 GHz. The standard phase tosubtractthecleancomponentsofthediscretesourcesfromall and amplitude calibration were successful enough that self- of the u−v data, and imaged the residual emission with a taper calibrationdidnotproduceasignificantimprovement,soitwas and natural weighting (ROBUST=+2 in AIPS) to a resolution not used in the image presented here. Residual amplitude cali- of∼50(cid:48)(cid:48)−60(cid:48)(cid:48) atallfrequencies.Allimageswereprimarybeam brationerrorsareestimatedtobewithin3%. correctedusingthetaskPBCORinAIPS.Theshortestbaselines Weusedthemulti-frequencymulti-scalecleantaskinCASA inourdatasetsare∼0.2kλ,∼0.1kλ,and∼0.07kλ,respectively (Rau&Cornwell2011)bothtodeconvolveandcreateamosaic at 608 MHz, 322 MHz, and 234 MHz. The largest detectable featuresarehence17(cid:48),32(cid:48)and44(cid:48),respectively. fromthethreepointingsforeachofthe1.38and1.78GHzmaps. Correction for primary beam attenuation was performed using Thefinalrmsvaluesforthefullresolutionimagesaregiven theCASAtaskimpbcor. inTable1.Atlowerresolutionweobtainedrms∼0.05mJyb−1 Toisolatethediffuseclusteremissionwesubtractedthecon- at 608 MHz (13.18(cid:48)(cid:48) × 11.06(cid:48)(cid:48)), ∼0.25 mJy b−1 at 322 MHz tribution from radio galaxies, as follows. We used the C con- (14.80(cid:48)(cid:48) × 13.06(cid:48)(cid:48)), and ∼0.3 mJy b−1 at 234 MHz (19.46(cid:48)(cid:48) × figuration 1.6 GHz maps (resolution of 11(cid:48)(cid:48), rms sensitivity of 17.92(cid:48)(cid:48)).Weestimatethatresidualcalibrationerrorsateachfre- 90 µJy b−1) to create masks of the location and extents of ra- quency are within 4–5% at 608 MHz and of the order of 10% dio galaxies; these data were not calibrated accurately enough at 234 MHz and 322 MHz. We point out that the difference in to directly subtract from the D configuration data. With these the flux density between the Baars et al. (1977) scale adopted masks, we then performed an interactive single-scale clean of here and the Scaife & Heald (2012) scale, suggested for obser- the1.38GHzand1.78GHzfullresolution∼40(cid:48)(cid:48)Dconfiguration vations at ν ≤ 500 MHz, is of the order of few percent, hence images until the radio galaxies were no longer visible. Model withinourfinalestimatederrors.Figure1showsthefullfieldof u−vdatasetsrepresentingtheradiogalaxyemissionwerecreated viewat234MHzandhighlightsthesizeofthefieldsimagedat fromthesecleancomponentsandsubtractedfromtheoriginalD 322MHzand608MHz. configurationu−vdata.Theresidualu−vdatasetswerethenim- agedwithmulti-scalecleanandcorrectedfortheprimarybeam, 2.2. JanskyVLAobservations convolving the final maps to 60(cid:48)(cid:48) with an rms of ∼180 µJy b−1 (∼140 µJy b−1) at field centre for 1.38 (1.78) GHz, to increase A2142 was observed with the Karl G. Jansky VLA in D and thesignal-to-noiseratioofthediffuseemission.Thelargestde- C configurations at 1–2 GHz as part of NRAO observing pro- tectableangularsizeoftheseobservationsis<∼1000(cid:48)(cid:48)and<∼820(cid:48)(cid:48), gramme VLA11B-156. Three pointings were acquired to re- at1.38GHzand1.78GHz,respectively. cover the full extent of the radio halo detected with the GBT (Farnsworth et al. 2013) with 28 min of integration time per pointing. Observations were made in spectral line mode with 3. Radioemissionfromclustergalaxies 16 spectral windows, each 64 MHz wide, spread across the full 1–2 GHz band. For technical reasons due to the recent TheinnerportionofthefieldofviewoftheGMRTobservations, VLA upgrade, only two seconds of every five second integra- i.e.A2142itself,isshowninFigs.2and3,wherethecontours tion on source were recorded, which resulted in higher thermal of the 234 MHz and 608 MHz radio emission are overlaid on A125,page3of17 A&A603,A125(2017) 0 . 0 0 : 0 3 608 MHz field 0 . 0 0 : 0 0 : 8 2 0 . 0 0 n 0: o 3 i t a n 0 li . c 0 e 0 D : 0 0 : 7 2 0 . 0 0 : 0 3 322 MHz field 04:00.0 02:00.0 16:00:00.0 58:00.0 56:00.0 15:54:00.0 Right ascension Fig.1.FieldofA2142atGMRT234MHz.Thelowresolutionimageisrestoredwithabeamof45.2(cid:48)(cid:48)×40.8(cid:48)(cid:48),PA–37.9◦.Thecontinuousand dashedblackboxesshowthesizeofthe322MHzand608MHzfields,respectively. theredplateoftheDigitizedSkySurveyDSS–2andtheXMM– CFHT (Canadian French Hawaii Telescope) MegaCam g-band Newtonimage,respectively. image. ThecentralregionofA2142isdominatedbythepresenceof two extended FRI radio galaxies (Fanaroff & Riley 1974) with 3.1. Radiogalaxiesattheclustercentre head-tailmorphology(Sect.3.1)andbydiffuseemissioncoinci- dentwiththebrightestpartoftheX-rayemissionfromtheintra- ThemoststrikingradiogalaxiesatthecentreofA2142aretwo clustermedium,whichweclassifyasagiantradiohalo(Sect.4). long tailed sources labelled T1 and T2 in Fig. 2. A zoom on Beyond these striking features, many radio sources in the field eachofthemisshownintheupperleftandupperrightpanelsof areassociatedwithclustergalaxies. Fig.A.1. TableA.1reportsthefullsampleofradiosourceswithopti- ThesourceT1istheradiogalaxyB21556+27(Collaetal. calcounterpartattheredshiftofA2142inthe608MHzfieldof 1972, Owen et al. 1993), associated with a m = 17.5 clus- g view (Fig. 1). The list is based on the full resolution 608 MHz ter galaxy (z = 0.0955). A compact counterpart is visible in imagewithadetectionlimitS =0.25mJy(i.e.5σ)prior the XMM-Newton image. The head of this radio galaxy is co- 608MHz totheprimarybeamcorrection.Forthisreason,theradiosource incident with the north-western cold front in the cluster. The catalogueisnotcompleteinradiopower,whosedetectionlimit length of the tail is ∼610 kpc. Figure A.1 clearly shows that increases away from the cluster centre. Flux density values at the long and straight tail has small amplitude wiggles. The 234MHzhavealsobeenreportedinthetable.Theradiogalax- imaging process at all frequencies reveals that it is embedded ies presented in Sect. 3.1 are shown in Fig. A.1, where the full in the diffuse emission of the radio halo. Its radio power is resolution 234 MHz and 608 MHz contours are overlaid on a logP = 24.80 W Hz−1, which is high for this class of 608MHz A125,page4of17 T.Venturietal.:Thetwo–componentgiantradiohaloinA2142 0 . 0 W1 0 : T2 5 2 G 0 . 0 0 : 0 2 : 7 2 W2 T1 0 n . 0 o 0 i at 5: n 1 i l c e D 0 . 0 0 : 0 1 C4 0 C1 C2 C3 . 0 0 : 5 0 59:00.0 40.0 20.0 15:58:00.0 57:40.0 Right ascension Fig.2.RadioemissionofA2142overlaidontheredplateofDSS–2.BlackcontoursshowalowresolutionGMRT234MHzimagerestoredwith abeamof45.2(cid:48)(cid:48)×40.8(cid:48)(cid:48),PA–37.9◦(sameasFig.1);contourlevelsare±3,6,12mJy/b;andthermsintheimageis∼0.9mJyb−1farfromthefield centre(negativecontoursarewhitedashed).Redcontoursshowthe608MHzfullresolutionimagerestoredwithabeamof5.2(cid:48)(cid:48)×4.5(cid:48)(cid:48),PA52.6◦; contourlevelsare±0.1,0.2,0.4,0.8,1.6,6.4,25.6,102.4mJyb−1;andthermsintheimageis∼35µJyb−1 farfromthefieldcentre.Negative contoursareshowninwhite. objects.Wemeasureα608MHz = 0.73±0.15,whichsteepensto Two wide-angle tail (WAT) sources are also present, and 234MHz α1400MHz =1.04±0.111. labelled W1 and W2 in Fig. 2. A zoom on each of them is 608MHz shown in the left and right central panels of Fig. A.1, respec- The source T2 is associated with the galaxy 2MASX tively. Interestingly, neither of these is located at the cluster J15582091+2720010 (m = 16.6, z = 0.0873 from the g redshift. The source W1 has a very faint optical counterpart NASA/IPACExtragalaxticDatabase;NED),locatednorthofthe (m = 23 from NED) with z = 0.574; its flux density is cluster centre and outside the brightest part of the X–ray emis- g phot S = 37.99 mJy. The radio peak of W2 coincides with sion,asisclearfromFig.3.Itslengthis∼370kpcanditsradio 608MHz anX-raysourceandhasaveryfaintopticalcounterpart.Despite power is logP = 24.51 W Hz−1. From Table A.1 it is 608MHz thepresenceofthreeverynearbyclustergalaxies,anassociation clear that T2 has a very steep integrated spectrum. If we com- with any of these seems unlikely. Considering that wide-angle plementourGMRTfluxdensitymeasurementswiththearchival tails are tracers of galaxy clusters (e.g. Giacintucci & Venturi 1.4GHzdata,weobtainα608MHz =1.21±0.14,whichsteepens 234MHz 2009;Maoetal.2009)andthatbothW1andW2areassociated toα1400MHz =1.92±0.11. 608MHz withveryfaintobjects,atleastanotherclusteralongthelineof sightandbehindA2142mustbepresent. TheradioemissionlabelledG(bottomleftpanelofFig.A.1) 1 TheVLAC+Dconfiguration1.4GHzobservationsareare-analysis waspresentedinEckertetal.(2014)at608MHz.Itisaremark- ofprojectAG344.Weproducedimageswithangularresolution24.1(cid:48)(cid:48)× 21.9(cid:48)(cid:48) and ∼38.6(cid:48)(cid:48) ×34.9(cid:48)(cid:48) with rms of the order of ∼15 µJy b−1 and able blend of discrete radio sources associated with a group at ∼20µJyb−1,respectively.Individualsourcesubtractionwasperformed z ∼ 0.094 located north-east of the cluster centre. The south- followingthesameproceduredescribedinSect.2.1toobtainimagesof erntipoftheradioemissioniscoincidentwiththebrighttipof thediffuseemission. thelong(∼800kpc)X-raytailvisibleinFig.3,whichhasbeen A125,page5of17 A&A603,A125(2017) Southern X-ray tip 0 . 0 0 : 5 2 0 . 0 0 : 0 2 : 7 2 n o ati 0.0 1 Mpc n 0 cli 5: e 1 D 0 . 0 0 : 0 1 0 . 0 0 : 5 0 59:00.0 30.0 15:58:00.0 57:30.0 Right ascension Fig.3.RadioemissionofA2142overlaidontheX-rayimagefrom XMM-Newton.Magentacontoursshowa234MHzlowresolutionimage restoredwithabeamof60(cid:48)(cid:48)×60(cid:48)(cid:48).Contoursaredrawnat1.5,3,6,12,24,48mJyb−1.Negativescontoursareshowninblue(–3mJyb−1).The whitecontoursshowthe608MHzimage(samecontoursandresolutionasinFig.2). interpretedasthesignatureoftheinfallofthegroupintoA2142 and T2 (GMRT J155820+272000), the two tailed radio galax- (Eckertetal.2014and2017).Weassociatethesouthernpeakof iesattheclustercentre,allradiosourcesareeitherpoint–likeor thisemissionwiththebrightest(m =16.1)galaxyinthegroup barelyextendedatthefullresolutionofthe608MHzimagewith g (TableA.1),eventhoughtheoverlayshowninFig.A.1suggests radio powers in the range 21.85 ≤ logP608MHz <∼ 23 W Hz−1. thatitcouldbeablendofemissionfrommoregalaxies. ThesevaluessuggestthattheradiogalaxypopulationofA2142 ThethreecompactradiogalaxieslabelledC1,C2,andC3in mayincludebothstarburstgalaxiesandfaintradioactivenuclei. Fig.2formanotherinterestinggroup.Theyarealignedandasso- Forthelatter,theradiopowersarewellbelowtheFRI/FRIIdivi- ciatedwithclustergalaxiesofsimilaropticalmagnitude(inthe sionandaretypicalofthefaintFR0radiogalaxies,whoseradio rangem =17.4–17.7fromNED).Afourthclusterradiogalaxy, morphologylacksextendedemissionintheformofradiolobes g C4,islocatedjustwestofthistriplet,asseeninthebottomright (Baldietal.2015,andreferencestherein). panelofFig.A.1. The radio emitting galaxies in A2142 span a range of red- The high resolution images at 608 MHz reveal that some shifts from ∼0.079 to >∼0.12 (see Table A.1); this is consistent clustergalaxiesembeddedinthegianthalohaveassociatedradio withthepresenceofmultiplegroups,ashighlightedinthespec- emission.ThemostluminousBCG(m = 16.2andz = 0.0904) troscopic analysis performed by Owers et al. (2011) and in the g hosts a faint radio source (see upper left panel of Fig. A.1). structureanalysisshowninEinastoetal.(2015).InFig.A.2we Moreover, a fainter cluster galaxy (m = 18.8, z = 0.0806) showthelocationoftheclusterradiogalaxieslistedinTableA.1 g justnorth-westofthemostluminousBCGandtwomorecluster withcolourcodestohighlightthedifferentredshifts. galaxies located along the X-ray elongation from the BCGs to The wider field of view of the 234 MHz (Fig. 1) includes theC1–C2–C3groupshowradioemission. manymoreradiosourcesassociatedwithclustergalaxies.Start- ingfromthevisualinspectiononDSS–2andafterconsultation 3.2. OverallradiopropertiesofthegalaxiesinA2142 of the NED database we found 71 counterparts in the redishift rangez∼0.07−0.13.Thesearemostlylocatedsouthandeastof A total of 42 radio sources have an optical counterpart. These theclustercentre,followingtheoverallelongationofthecluster numbers refer to the region covered by the 608 MHz observa- subgroups (Owers et al. 2011) and of the supercluster (Einasto tions.WithexceptionsmadeforT1(GMRTJ155814+271619) et al. 2015). A full characterization of the population of radio A125,page6of17 T.Venturietal.:Thetwo–componentgiantradiohaloinA2142 00.0 0.0 18: 8:0 1 Declination 14:00.0 Declination 14:00.0 0 27:10:00. 7:10:00.0 2 59:00.0 40.0 20.0 15:58:00.0 50.0 40.0 30.0 20.0 10.0 15:58:00.0 Right ascension Right ascension Fig.4.Leftpanel:radiohaloinA2142overlaidontheXMM–Newtonemission.The234MHzimageisshowninblack(negativecontoursin yellow).Contourlevelsaredrawnat±2.5,4,8mJyb−1,θ = 44.86(cid:48)(cid:48) ×40.96(cid:48)(cid:48),PA−39◦.The608MHzemissionisshowninwhite(negative contoursshowningreen).Contourlevelsaredrawnat±0.6,1.2,2.4mJyb−1,θ=50(cid:48)(cid:48)×50(cid:48)(cid:48).Bothimageswereobtainedaftersubtractionofthe discretesourcesfromtheu−vdata.Rightpanel:VLAcontoursat1377MHz(θ =60(cid:48)(cid:48)×60(cid:48)(cid:48),contoursaredrawnat±0.2,0.4,0.8,1.6mJyb−1) overlaidonthe322MHzGMRTimage.Negativecontoursaredrawnasdashedlines.Thegreenarrowhighlightsthenorth-westernextension(see Sect.4.1). sourcesinA2142isbeyondthescopeofthispaperandwillbe presentedinafuturework. 0 0 7: 1 4. Radiohalo 00 6: 1 4.1. Morphology The diffuse extended emission in A2142 is well visible in ation 5:00 n 1 Figs. 1–3, and this emission is best highlighted in the low res- cli e 0 olutionimagesshownintheleftandrightpanelsofFig.4(ob- D 4:0 tained after subtraction of the individual radio galaxies at each 7:1 2 frequency;seeSects.2.1and2.2),andinFig.5.Ourimagesare 0 suggestive of a multi-component cluster-scale emission, which 0 3: alltogetherwerefertoastheradiohalo.Forourstudyweiden- 1 tifytworegions,whichareshownintheleftpanelofFig.6and 100 kpc 0 0 are named H1 and H2. The operational definition of these two 2: 1 regionsisgiveninSect.4.2. 30 25 15:58:20 15 10 05 The region H1 is the brightest part of the halo, and this re- Right ascension gion is best highlighted in Fig. 5, which shows an intermediate Fig.5.Zoomonthecentralpartoftheradiohalo.TheGMRT608MHz resolution608MHzimageoverlaidontheChandraX-rayemis- image(discreteradiosourcesnotsubtracted)attheresolutionof16.3(cid:48)(cid:48)× sion.TheradioandX-raypeaksarecoincident.Thisregionwas 14.2(cid:48)(cid:48),PA−84.6◦,isoverlaidonChandra.Contoursaredrawnat±0.18, formerly classified as mini–halo (Giovannini & Feretti 2000). 0.36, 0.72, 1.44, 2.88 mJy b−1; positive contours are shown in black, Figure5clearlyshowsthatitisconfinedbytheinnermostcold negative contours white). The green arrows show the location of the front, whose position is indicated by the green arrows. Even innerSEcoldfront. though the north-western boundary is more difficult to define, owing to the presence of the radio emission from T1, H1 does not seem to extend beyond the north-western cold front. This brightness distribution of this component is very different from regionhasthesameextentandboundariesatallfrequencies. H1.ItisconsiderablymoreextendedthanH1andlacksacentral The left panel of Fig. 4 shows the entire extended radio peak. The difference between H1 and H2 is confirmed by our emission in A2142 at 234 MHz and 608 MHz overlaid on the analysis of the radio brightness profile across the two regions X-ray emission detected by XMM-Newton, while the emission shownintheleftpanelofFig.6(seeSect.4.2fordetails).The at 1.38 GHz is given as contours in the right panel, overlaid regionH1showsaregularprofilewithaprominentpeak,follow- with the 322 MHz image. At all frequencies, the radio halo is ingcloselythebrightnessdistributionofthehotgas.Theprofile elongated in the same north-west to south-east direction of the of H2 is less regular, with a flat shoulder and a steeper profile X-rayemissionfromtheintraclustergas,andcoversthebright- farfromthecore,andshowsnoclearcorrelationwiththeX-ray est X-ray ridge of emission out to the most distant cold front. profile. The north-western extension visible at 1.38 GHz (right Its largest angular size is ∼10(cid:48), i.e. ∼1 Mpc at the cluster red- panelofFig.4)isnotdetectedatlowerfrequenciesanditismost shift. We define as H2 the ridge-like emission extending from likelyduetoincompletesubtractionofT1. H1 towards the most distant old front in the south-east direc- Finally, our VLA datasets detect further extended emission tion.InspectionofbothpanelsofFig.4suggeststhatthesurface surroundingH1andH2(seetheleftpanelofFig.6),whichwe A125,page7of17 A&A603,A125(2017) Fig.6.Leftpanel:tworegionsH1andH2usedfortheevaluationoftheintegratedspectrum(seeSect.4.2)shownonthe1.38GHzVLAimageat theresolutionof60(cid:48)(cid:48).Rightpanel:sliceofthesurfacebrightnessofthe1.38GHzimagealongthedirection−33◦(seeSect.4.2),sothatlengths alongtheslicecanbecalculatedas(Dec2-Dec1)/cos(33d).Theverticalgreydottedlineshowsthelocationofthesouth-eastinnermostcoldfront, whiletheXMMX-raybrightnessprofileconvolvedwiththesameresolution(60(cid:48)(cid:48))isshownasablackdottedline. interpretastheremainsofthe2Mpcscaleemissionimagedwith Finally,weinspectedthe74MHzimageintheVLSS–Redux theGBT(Farnsworthetal.2013),whosefullextentisunrecov- (VLA Low-Frequency Sky Survey) to check for the presence eredinallourimages.Thisclearlyshowsthelimitationsofinter- of diffuse radio emission at this frequency and found no clear ferometricobservations,whoselackofzerospacingsisasevere emissionabovethenoiselevel. limit in imaging very low brightness extended and complex ra- TodeterminethespectralindicesforH1andH2,wecarried dioemission(asisthecaseoftheradiohaloinA2142)atleast out the following steps: (a) we defined the H1 and H2 bound- outtoz∼0.1. aries; (b) we adopted a procedure that was insensitive to varia- tions in the detailed structures at different frequencies and that In the next sections we refer to H1 as the central emission, H2astheridge,andthelargerscaleemissionforthemorediffuse removed the varying amounts of flux from the 2 Mpc compo- nentstomeasurethefluxdensityofeachcomponent;and(c)we emissiondetectedwiththeVLAandGBT. evaluatedtheu−vcoveragetoensurereliablefluxestimates.We describeeachofthesebelow. 4.2. Radiospectrum a) H1andH2definition–TheextentsofH1andH2wereini- tiallyestimatedbyeyefromgreyscaleimagesofthebestim- ThecomplexityoftheradioemissioninA2142atallfrequencies ages. To look at this more quantitatively, we calculated the doesnotallowreliableimagingofthespectralindexdistribution averagefluxateachpositionalongthemajoraxisofthera- throughouttheradiohalo.Toovercomethisdifficultyandobtain dioemissioninastrip160(cid:48)(cid:48) widealongtheminoraxis.The someinformationaboutpossiblechangesofthespectralproper- sliceofthesurfacebrightness,obtainedusingtheVLA1.38 ties,wederivedtheintegratedspectrumoftheradiohaloinH1 GHz image convolved to 60(cid:48)(cid:48), is shown in the right panel and H2 (see the left panel of Fig. 6, where the two regions are ofFig.6.BecauseH1andH2overlap,thedivisionbetween overlaidontheVLAimage). themissomewhatarbitrary.ItwasfirstchosenaswhereH1 To complement and further extend the frequency cov- beginstosignificantlyaffecttheshapeoftheH2profileand erage of the GMRT and VLA observations we re-analysed thencomparedtothepositionofthecoldfront.TheH1,H2 1.4 GHz archival data in the C and D configuration (see division is at the southern base of the cold front (see Fig. 5 Sect. 3.1) and used a preliminary image obtained with LOFAR and right panel of Fig. 6). The north-west boundary of H1 at118MHz.TheLOFARhba_dual_inner118-190MHzdata isapproximatelyatitshalf-powerlevel,asisthesouth-east were recorded on April 19 2014 (project ID LC1_017). A sub- boundary of H2. Because of the variations in the detailed setofthetargetdatasetintheband118–124MHzwascalibrated structure at the different frequencies, the widths of the H1 andimagedusingthestandarddirectionindependentcalibration and H2 regions were chosen to be the FWHM of the emis- procedure(seee.g.Shimwelletal.2017).Theresultingimage, sionaftersmoothingby160(cid:48)(cid:48) alongthemajoraxis.ForH1, made from the visibilities from baselines shorter than 7kλ, has we used the width of the bright core, rather than the lower a sensitivity of 3 mJy b−1, an angular resolution of ∼50(cid:48)(cid:48) and brightness emission. The H2 profile is somewhat asymmet- the peak flux density measurements of compact sources are in ric,sothelarger(SW)widthwasused.Theseproceduresled agreement with the 150MHz TGSS ADR (Intema et al. 2016) tothelengthandwidthoftheboxesas160(cid:48)(cid:48)×136(cid:48)(cid:48)(6square surveytowithin30%.Afulldirectiondependentcalibrationof arcmin)forH1and216(cid:48)(cid:48)×160(cid:48)(cid:48)(9.6squarearcmin)forH2, the dataset (see e.g. van Weeren et al. 2016a,b; Williams et al. orientedat−33deg. 2016;Hardcastleetal.2016;andShimwelletal.2016)willbe b) Flux density measurements – All images were first con- appliedinthefuturetocorrectforionosphericdisturbancesand volved to a resolution of 60(cid:48)(cid:48), except for the VLSSr, which improvetheimagefidelity,sensitivityandresolution,butthisis has a beam size of 75(cid:48)(cid:48). The total flux densities in the H1 beyondthescopeofthispaper. and H2 boxes were then measured as follows. For each A125,page8of17 T.Venturietal.:Thetwo–componentgiantradiohaloinA2142 Fig.7.Integratedspectrumoftheradiohalo.Leftpanel:plotofH1.Theupperlimitat74MHzisshownasaredtriangle.Rightpanel:plotofH2. Filledbluecircleshighlightthemeasurementswithgoodu−vcoverageatshortspacings;theremainingmeasurementsareshownasfilledgreen dotsseeSect.4.2).Inbothpanelstheweightedlinearfitsareshownwiththesamecolourcodeandtheupperlimitat74MHzisshownasared triangle. component,wecalculatedtherunningsumofthefluxwithin frequency,canleadtoelevatedbackgroundlevelsaroundH1 aboxfixedtothesamelengthandwidthasH1andH2,re- andH2.Thedeterminationofabackgroundlevelinthe1D spectively.Theboxwasslidalongthelineperpendicularto cuts,asdiscussedabove,removesanysucheffects.Thetotal themajoraxisoveralengthof1120(cid:48)(cid:48)crossingthemajoraxis spectrum of H1 and H2 between 118 MHz and 1.778 GHz (andH1andH2,respectively)nearthemiddle.Wethenmea- is shown in the left and right panel of Fig. 7, respectively. sured both the off-source background level in each running We point out that H2, whose largest extent along its ma- sum,alongwiththepeakfluxabovethebackground.These joraxis(216(cid:48)(cid:48))iswellwithinthenominallargestdetectable peak fluxes are reported in Table 3. The background level structure, is not sufficiently well sampled in our interfero- includesanysmallcontributionsfromthemuchlargerscale metric images owing to to its surface brightness, which is (20(cid:48) component detected in our single dish measurements), considerably lower than H1. In particular, the inner portion aswellasanyinstrumentaleffectsintheinterferometerim- of the u−v coverage (within 1 kλ) is much better sampled ages. We then calculated the residual rms scatter in the off- at 118 MHz (LOFAR), 322 MHz (GMRT), 1.38 GH, and sourcebackgroundofeachrunningsumandreporttheseas 1.78 GHz (VLA-D). For this reason we refer to those im- the errors for the peak fluxes. Such errors therefore reflect agesastothewell-sampled(weareusingthisorgoodcov- random noise and instrumental artifacts but do not reflect erage)images.Bycontrast,theu−vcoverageofthe1.4GHz any systematic effects such as those due to u−v coverage, archival VLA and the 608 MHz and 234 MHz GMRT data which we discuss below. It is important to note that the H1 ispoorerontheshortestspacingsandsomeofthefluxden- andH2fluxesarenotthe“totalflux”fromeachcomponent, sityfromH2ismissing.Thisproblemiswellknownandits since they each have broad wings beyond their box widths. effecthasbeenquantifiedinearlierworksongiantradioha- ItisnotclearwhetherthesewingsareextensionsofH1and los (see Brunetti et al. 2008; and Venturi et al. 2008). The H2 or part of the larger scale emission. However, since the flux density measurements for H2 are plotted separately in boxes were kept the same at all frequencies, they provide a therightpanelofFig.7. well-defined sample of the H1 and H2 emission. The right panel of Fig. 6 shows the surface brightness profile along A weighted mean square fit provides α1.78GHz = 1.33 ± 0.08 theridgeorientedat−33deg.Thecleardifferenceinsurface for H1 and α1.78GHz = 1.42 ± 0.08 for11H82M.HTzhe difference is brightness and extent between H1 and H2 separately moti- 118MHz onlymarginal.However,itbecomesmoresignificantifwetake vatedourstudyofthetotalspectrumforthesetworegions. thedifferentu−vcoverageintoaccount.Afitofthetwosetsof c) Evaluationoftheu−vcoverage–Despitethenominallargest measurementsforH2separately(filledgreenandfilledbluedots detectableangularsizeofourobservationsateachfrequency in the right panel of Fig. 7), provides α1.78GHz = 1.55±0.08. (seeSects.2.1and2.2,forcompletenessnotethatthelargest 118MHz Thefluxdensityupperlimitsat74MHzareconsistentwiththe angularsizedetectablebyLOFARat118MHzis∼1◦),be- spectral index both for H1 and H2. The future analysis of the causeoftheirverylowsurfacebrightnessnoneofourinter- LOFARdatawillallowustoobtainbetterconstraints. ferometer data sample well the largest scale (20(cid:48)) emission seenwiththesingledish.Wenotethatthetotalfluxdensity detectedwiththeVLAat1377MHzis23±2mJy,whichin- 5. Originoftheradiohalo cludes H1, H2, and the emission beyond these two regions until they fade into the background, while the 1410 MHz The observations presented in this work confirm that in many flux density measured in the GBT image is ∼55 mJy (see respectsA2142isacasestudyofgalaxyclusters.Ourmostim- Farnsworth et al. 2013). However, the partial sampling of portant results is the finding that the radio halo consists of two thisverydiffusecomponent,whichvariesfromfrequencyto regionswithdifferentmorphologicalandspectralproperties. A125,page9of17 A&A603,A125(2017) Table3.Radiohalofluxdensity. theturbulenteddiesof∼100kms−1 onascaleofabout50kpc. Theserequirementsareindeedconsistentwiththosemeasuredin Region Frequency Fluxdensity simulationsofgassloshingandthosenecessaryforturbulentre- MHz mJy accelerationtomaintain(ortore-accelerate)radioemittingelec- trons in these regions (ZuHone et al. 2013); in addition, values H1 1778 4.4±0.5 ofthesamemagnituderangehavebeenderivedforthePerseus 1465 7.0±0.5 cluster(Hitomicollaboration2016). 1377 7.9±0.4 608 16.5±1.4 322 56.2±4.1 5.2. Theridge 234 69.3±4.9 118 275.0±94.1 TherightpanelofFig.6clearlyshowsthatthesurfacebrightness 74 <700 and extent of the radio emission in the ridge (H2) are very dif- H2 1778 3.3±0.5 ferent from those in the core region (H1). The clear separation 1465 2.2±0.4 between these two components suggests that they might origi- 1377 4.9±0.4 natefromdifferentmechanismsorthattheytracedifferentevo- 608 6.8±1.4 lutionary stages of the same phenomenon. On the scale of the 322 40.5±4.1 ridgetheclusterappearsunrelaxedwithaclearelongationinthe 234 34.5±4.9 south-eastern direction. The radio emission follows the spatial 118 260.0±94.1 distributionoftheX-rayemittinggasandextendsuptothecold 74 <700 frontlocatedat1Mpcsouth-eastofthecore.Onthesescalesthe clustermayhavebeenperturbedbyanumberofprocessesthat inducegassloshingandgeneratetheexternalcoldfront.Thede- tectionofabridgeoflow-entropygasbetweenthecentralregion Ouranalysis,whichspreadsoverafrequencyrangeofmore oftheclusterandthemostdistantcoldfrontisinsupportofthis thanoneorderofmagnitude(118–1.78GHz),suggeststhatthe picture(Rossettietal.2013).Itislikelythatthegasandmagnetic regionofemissionextendingsouth-eastoftheclustercore(H2; fieldsinthisregionhavebeenmoderatelyperturbedandstirred. seeleftpanelofFig.6)hasalowersurfacebrightness,isbroader We can speculate that this induces turbulence that cascades on insize,andhasamoderatelysteeperspectrumthanthebrighter, smaller scales and damps into particle re-acceleration and fast morecompactregioninthecore(H1).Inthissectionwediscuss the possible origin of such differences in the framework of the magetic reconnection, which are two interconnected processes ifincompressibleturbulenceisconsidered(Brunetti&Lazarian cluster dynamics, as inferred from its broadband properties. In 2016). Fig.8weshowalltheobservationalinformationthatisrelevant tothediscussion. Thelowluminosityoftheridgeandofthe2Mpcscaleemis- sion,andthemoderatelysteeperspectrumoftheridge,suggest thattheenergybudgetthatbecomesavailabletothenon-thermal 5.1. Thecentralemission components in H2 is smaller than that of classical giant radio halos,thatindeedaregeneratedduringmajormergerevents.An The radio emission observed in the core of A2142 (H1) is alternative possibility is that the radio emission is very old and bounded by the two inner cold fronts that are detected on the marks the switched off phase of the radio halo, when merger 100–200kpcscale,asclearfromFigs.5and8.Thismaysuggest turbulence is dissipated at later times. Both cases are possible that this emission traces the dissipation of the energy produced for massive clusters that show intermediate properties between by the sloshing of the low-entropy gas oscillating between the mergingandrelaxedsystems(e.g.Cassanoetal.2006;Brunetti twoinnercoldfronts.Thisscenarioissimilartowhathasbeen et al. 2009; Donnert et al. 2013). This is indeed the case of suggestedfortheoriginofmini–halosincool-coreclusters(e.g. A2142. As a matter of fact, this radio halo is considerably un- Mazzotta&Giacintucci2008;ZuHoneetal.2013).TwoBCGs derluminous compared to the correlation between radio power are present in A2142 (see Figs. 8 and A.1) and the sloshing in andclustermass(Cassanoetal.2013;Cucitietal.2017),even thecoremaybeduetotheperturbationinducedbyagaslessmi- considering the GBT measurement in Farnsworth et al. (2013, nor merger with the group associated with the secondary BCG seeSects.4.2and6). (Owersetal.2011).ThesetwoBCGsmayalsoplayaroleinthe originoftheradioemission.Atpresentonlythemostluminous Differentscenarioscanexplaintheoriginoftheperturbations BCG hosts a radio source (Sect. 3.1, Table A.1 and Fig. A.1), inducedintheICM.Apossibilityisthatthethreecoldfrontsde- but it is likely that over the last Gyr or so both BCGs have re- tected in X-rays (see their location in Fig. 8) are generated by leasedrelativisticparticlesandmagneticfieldsinthecoreregion. a single event (Rossetti et al. 2013). In this case, H1 and H2 Duringthistimescalerelativisticelectronscouldbeadvectedby mayprobetheevolutionofthisphenomenonondifferentscales turbulentmotionsinthesloshinggas.Theveryweakcorrelation and/ortimesandmaytracedifferentlevelsofperturbationsand between the radio power of the BCG and that of the mini-halo magneticfieldstrengththatarepresentintheICM.Thesymme- foundforasizeablesampleofmini-haloclusters(Govonietal. tryofthethreecoldfrontsandthewaytheyencompassH1and 2009;Giacintuccietal.2014)isindeedsuggestiveofthefactthat H2isinsupportofthisscenario. thecentralAGNactivitymaynotbepoweringtheradioemission Anotherpossibilityisthattheridgetracesaturbulentregion directly,butitisthemostlikelysourceofseedelectronsforre- that is generated by the continuous accretion of subhalos along accelerationintheICM. theS–Edirection.Thishypothesismaybesupportedbythefact Inordertocoveradistanceoftheorderof100kpcin1Gyr, that optical data show the presence of several groups of galax- thespatialdiffusioncoefficientduetoturbulenttransportshould iesthattracealarge-scalefilamentintheS–Edirection(Owers be of the order of D ∼ δV L ∼ 2 × 1030 cm2s−1 (Brunetti etal.2011;Einastoetal.2015),andgroupaccretioninA2142 L & Jones 2014, and references therein), implying a velocity of iscaughtinaction(Eckertetal.2014). A125,page10of17