Mon.Not.R.Astron.Soc.000,??–??(2011) Printed4January2012 (MNLATEXstylefilev2.2) Dust in Historical Galactic Type Ia Supernova Remnants with ⋆ Herschel 2 H.L. Gomez1, C.J.R.Clark1, T. Nozawa2, O. Krause3, E.L.Gomez1,4, M. Matsuura5, 1 0 M.J.Barlow5, M.-A. Besel3, L. Dunne6, W.K.Gear1, P. Hargrave1, Th. Henning3, 2 n R.J. Ivison 7,8, B. Sibthorpe7, B.M.Swinyard5,9, R. Wesson10 a J 1SchoolofPhysics&Astronomy,CardiffUniversity,TheParade,Cardiff,CF243AA,UK 3 2InstituteforthePhysicsandMathematicsoftheUniverse,UniversityofTokyo,Kashiwa,Chiba277-8583,Japan 3Max-Planck-Institutfu¨rAstronomie,Ko¨nigstuhl17,69117Heidelberg,Germany ] 4LascumbresObsveratoryGlobalTelescopeNetwork,6740CortonaDriveSuite102,Goleta,CA93117 A 5Dept.ofPhysicsandAstronomy,UniversityCollegeLondon,GowerStreet,LondonWC1E6BT,UK G 6SchoolofPhysics&Astronomy,UniversityofNottingham,UniversityPark,NottinghamNG72RD,UK 7UKAstronomyTechnologyCentre,RoyalObservatoryEdinburgh,BlackfordHill,EdinburghEH93HJ,UK . h 8InstituteforAstronomy,UniversityofEdinburgh,BlackfordHill,EdinburghEH93HJ,UK p 9SpaceScienceandTechnologyDepartment,RutherfordAppletonLaboratory,Oxfordshire,OX110QX,UK - 10EuropeanSouthernObservatory,AlonsodeCordova3107,Casilla19001,Vitacura,Santiago19,Chile o r t s a [ ABSTRACT 2 Theoriginofinterstellardustin galaxiesispoorlyunderstood,particularlytherelativecon- v tributions from supernovae and the cool stellar winds of low-intermediate mass stars. Re- 7 2 cently,largemassesofnewly-formeddusthavebeendiscoveredintheejectaofcore-collapse 6 supernovae. Here, we present Herschel PACS and SPIRE photometry at 70-500µm of the 6 historical, young supernova remnants: Kepler and Tycho; both thought to be the remnants . of Type Ia explosion events. We detect a warm dust component in Kepler’s remnant with 11 Td = 82+−46K and mass ∼ (3.1+−00..86)×10−3M⊙; this is spatially coincident with thermal X-rayemissionandopticalknotsandfilaments,consistentwiththewarmdustoriginatingin 1 1 the circumstellarmaterialsweptupbythe primaryblastwave ofthe remnant.Similarlyfor v: Tycho’sremnant,wedetectwarmdustat90+−57Kwithmass(8.6+−21..38)×10−3M⊙.Comparing thespatialdistributionofthewarmdustwithX-raysfromtheejectaandswept-upmedium, i X andHαemissionarisingfromthepost-shockedge,weshowthatthewarmdustissweptup interstellarmaterial.We findnoevidenceof a cool(25-50K) componentofdustwith mass r a ≥0.07M⊙asobservedincore-collapseremnantsofmassivestars.Neitherthewarmorcold dust components detected here are spatially coincident with supernova ejecta material. We compare the lack of observed supernovadust with a theoreticalmodel of dust formation in TypeIaremnantswhichpredictsdustmassesof88(17)×10−3M⊙forejectaexpandinginto ambient surrounding densities of 1(5)cm−3. The model predicts that silicon- and carbon- rich dust grains will encounter, at most, the interior edge of the observed dust emission at ∼400yearsconfirmingthatthemajorityofthewarmdustoriginatesfromsweptupcircum- stellarorinterstellargrains(forKeplerandTychorespectively).Thelackofcolddustgrains intheejectasuggeststhatTypeIaremnantsdonotproducesubstantialquantitiesofiron-rich dust grains and has importantconsequencesfor the ‘missing’ iron mass observed in ejecta. Finally, although,we cannotcompletely rule out a small mass of freshly-formedsupernova dust,theHerschelobservationsconfirmthatsignificantlylessdustformsintheejectaofType Iasupernovaethanintheremnantsofcore-collapseexplosions. Keywords: Supernovae:individual:Kepler,Tycho–ISM:submillimetredust:ISM–Galax- ies:abundances–submillimetre ⋆ HerschelisanESAspaceobservatorywithscienceinstrumentsprovided byEuropean-ledPrincipalInvestigatorconsortiaandwithimportantpartic- ipationfromNASA. 2 H. Gomezet al. 1 INTRODUCTION Wilson & Batrla 2005). Recent observations with the Balloon- borne Large Aperture Submillimeter Telescope, BLAST and the Dustingalaxiesisthought tobeproducedbyinthestellarwinds HerschelSpaceObservatoryshoweditwasdifficulttodistinguish of both low-intermediate mass (LIM) Asymptotic Giant Branch between colddust inthe remnant and cold dust fromintervening (AGB)stars(e.g.Gehrz1989; Whittet2003; Sargentetal.2010) interstellarclouds usingphotometry alone (Sibthorpeet al.2010; and,toanunknownextent,bymassivestars(Ho¨fner2009;Gomez Barlowetal.2010).Thesestudiesrevealedanewcoolcomponent etal.2010;Andrewsetal.2011;Gall,Hjorth&Anderson2011a). of dust in Cas A with mass 0.08M⊙ at ∼ 35K, yet even if all Massive stars may also produce dust in the expanding gas ejecta ofthedustsurvivedthepassagethroughtheshock,thedustmass whentheyexplodeassupernovae(e.g.Claytonetal.2001;Todini isaboutoneorderofmagnitudelowerthannecessarytosolvethe &Ferrara2001).IntheMilkyWay(seeZhukovska&Gail2008), dustbudgetproblem. themajor dust source ispresumed tobeLIMstars, but whenac- Apossiblemethodtodistinguishbetweensupernovadustand countingfordustdestructiontimescalesandtheobserveddustmass contaminationfrominterstellarmaterialalongthelineofsightwas intheGalaxy,thedustinjectionratefromstarsrequiredisanorder proposedbyDunneetal.(2009) wheretheytracedthealignment ofmagnitudehigherthanobserved.Analternativesourceofdustis ofthedusttowardsCasAwiththeSNRmagneticfieldusingpo- requiredtomakeupthedustbudget(e.g.Jones2001).Thisshortfall larimetry at 850µm. They find that a significant fraction of the inthedustmassestimatedfromthedustinjectionratesfromLIM submmemissiontowardsCasAisstronglyalignedwiththemag- starsisalsoobservedintheinterstellarmedium(ISM)oftheLarge netic field of the remnant. The dust mass inferred from the po- MagellanicCloud(Matsuuraetal.2009),ingalaxiesouttoz=0.4 larimetry study of Cas A agrees with the dust masses estimated (Dunneetal.2011) andindustyhigh-redshiftgalaxies(seeMor- fromHerschel photometry observations of the core-collapse SNe gan&Edmunds2003;Dwek,Galliano&Jones 2007).Although SN1987A (Matsuuraetal.2011); intheformer0.2−1.0M⊙ of recentworkbyJones&Nuth(2011)suggest thatthedust-budget dustwasdetecteddependingonthecomposition.Thelimitonthe problemisnotasproblematicaspreviousworkssuggestgiventhe ejectedheavyelementmassinthesesourcesplacesafurthercon- uncertaintiesinmasslossratesandtheefficienciesofdustforma- straintonthemaximumphysicaldustmass,finding0.2 < M < d tioninstellarwinds, itisclear that significant dust production in 0.5M⊙. SNejectawould alleviatethisdust budgetary problem. Theaver- Theevidencefordustformationincore-collapseSNRsisac- agedustyieldperSNrequiredtoexplaindustygalaxiesatbothlow cumulating,thoughitisnotyetunderstoodwhethertheamountof andhighredshiftsisoforder0.5−1M⊙(Michałowski,Watson& SNdustcreatedisdependentontheprogenitormass,themetallic- Hjorth2010;Gall,Anderson&Hjorth2011b). ityorthetypeofexplosion.Neitherisitclearhowmuchdustwill AlthoughSNehavelongbeenproposedasasourceofdustin survive,thoughtheoreticalstudiessuggestasignificantfractionof theISM(Dwek&Scalo1980;Claytonetal.2001),observational dustgrainsformedinejectawillbedestroyedbeforeinjectioninto evidencehasbeenscarceuntilrecently.Woodenetal.(1993)first the ISM (Jones 2001; Bianchi & Schneider 2007) particularly in detected emission from 10−4M⊙ of dust in SN1987A using the theejectaofenvelope-poorcore-collapseSNeorTypeIaremnants Kuiper Airborne Observatory. Subsequent Mid-far-infrared (FIR) (Nozawaetal.2010;Nozawaetal.2011).Asyet,thereisnoclear observationsofGalacticandextra-galacticcore-collapseremnants evidenceforfreshlyformeddustintheremnantsofTypeIaexplo- withtheSpitzer SpaceTelescopedetect small quantitiesof warm sions,thoughtheoreticalmodelsofdustformationintheremnants dust(∼(1−50)×10−3M⊙)inyoung(t<1000days-Sugerman ofTypeIaeventssuggeststhatupto0.2M⊙dustcouldforminthe etal.2006;Meikleetal.2007,2011;Kotaketal.2009;Andrewset ejecta(Nozawaetal.2011). al.2010,2011;Fabbrietal2011;Szalaietal.2011)andoldrem- Historicalremnantsareuniqueprobesforunderstandingdust nants(Williams,Chu&Gruendl2006;Rhoetal.2008).Theevi- productionandindeed,destructionfollowingcorecollapseTypeII dencefordustproductioninTypeIaremnantsisscarce(Borkowski supernovaeaswellaslowmassbinaryTypeIevents.IntheHer- et al. 2006; Seoket al. 2008): mid-IR observations of the Kepler schelera,wearefinallybeginningtoprobethecontributionofSNe SNRwiththeInfraredSpaceObservatory(ISO;Douvion,Lagage tothedustbudget,yetgiventhesomewhatlargeuncertaintiesinour & Pantin 2001) found thermal emission from warm dust (100K) understandingofthedustyieldfromstellarsourcesinothergalax- withmass10−4M⊙.Morerecently,Ishiharaetal.(2010)detected iesandininterpretingstatisticaldescriptionsofthedustcontentin 10−3M⊙ofswept-up(warm)dustintheTychoSNRwithAKARI, galaxiesfromwide-areasurveys(e.g.Dunneetal.2011),itisim- withatentativesuggestionthat10−4M⊙ofdustseeninthenorth- portanttoinvestigatedustproductioninthosenearest,youngestand east of the remnant may have been formed inthe ejecta. Outside resolvedgalacticremnants.Onecanalsoarguethatunderstanding of our Galaxy, Borkowski et al.(2006) obtained mid-IRobserva- dustformationinsupernovaeisimportantregardlessoftheexplo- tions of four Ia remnants in the Large Magellanic Cloud, finding sionmechanism:dustformationfollowingType-IISNemightsug- < 10−2M⊙ of swept-up interstellar dust. The dust masses from gestthatType-IaSNewouldalsobelikelydustproducers(Clayton themid-farIRstudiesareordersofmagnitudelowerthanneeded etal.1997;Travaglioetal.1999).Alternatively,detectingdustfrom tosolvethedustbudgetissues. circumstellarmaterialsweptupbytheblastwaveofaTypeIacan Inordertoinvestigateifsupernovaremnants(SNRs)contain beausefulprobeoftheprogenitorandexplosionmechanism. colddust missedbyprevious infrared(IR)observations, theSub- The resolution, wavelength coverage and sensitivity of Her- millimetre Common User Bolometer Array (SCUBA) was used schel allows us to investigate the origin and quantity of dust in to observe Cassiopeia A (Cas A), a remnant of a Galactic core- historicalsupernovaremnants.HerewepresentIRandsubmmob- collapseSNe(Krauseetal.2008a).Thecolddustseenat450and servations (§2) of the Tycho and Kepler remnants. Wedetermine 850µmwasinterpretedasemissionfromdustassociatedwiththe thedustmassinandtowardstheremnantsin§3and§4andcom- remnantduetothespatialcorrelationwiththeX-ray,radioandsub- paretheIRemissionwithmultiwavelengthtracersin§3.1and§4.1. millimetre(submm)emission(Dunneetal.2003).Subsequently,a Theoriginofthedustisdiscussedin§5andweusethetheoretical significantfractionofthesubmmemissionwasshowntooriginate modelofNozawaetal.(2011)tocomparetheexpecteddustmasses frommolecularcloudsalongthelineofsight(Krauseetal.2004; withtheobservations.Theresultsaresummarisedin§6. Dust intheKeplerand TychoRemnants 3 2 OBSERVATIONSANDDATAREDUCTION (Bonnareletal.2000).ForKepler’sremnantweusedobservations inthe filters658, 660 and 502nm where the emission originates 2.1 HerschelPhotometry fromHα+[NII],[NII]and[OIII]respectively(e.g.Sankritetal. TheKeplerandTychosupernovaremnantswereobservedwiththe 2008). Badpixelsweremasked by hand. ForTycho, weuse data Herschel (Pilbratt et al. 2010) Photodetector Array Camera and observed at 656nm (Lee et al. 2010), tracing Hα in the eastern Spectrometer(PACS;Poglitschetal.2010)andSpectralandPho- limboftheremnant. tometricImagingReceiver(SPIRE;Griffinetal.2010)at70,100, Spitzer Multiband Imaging Photometer (MIPS; Rieke et al. 160,250,350and500µmaspartoftheMassLossfromEvolved 2004)Level2calibrateddatawereobtainedviathearchive3 (Ke- StarS survey (MESS: Groenewegen et al. 2011). Kepler was ob- pler: PI W.Blair;Tycho: PI J.Rho) at 24 and 70µm. For 100µm servedon10thSeptand11thMarch2010; Tychoon20thMarch data,weusetheImprovedReprocessingoftheIRASSurvey(IRIS; 2010 and 18th Jan 2010 for PACS and SPIRE respectively. The Milville-Deschenes&Lagache2005) whichhavebeencalibrated PACSdatawereobtainedin‘scan-map’modewithspeed20′′/sin- with respect to data taken with the Diffuse Infrared Background cludingapairoforthogonalcross-scansoveranarea∼22′×22′. Experiment(DIRBE).WidefieldInfraredSurveyExplorer(WISE; TheSPIREmaps wereobserved in‘LargeMap’ mode withscan Wrightetal.2010)datawerealsoobtainedat22µm. lengthof30′ over∼ 32′×32′;across-scanlengthisalsotaken, Theradiodata at18cmfor Tychowas obtained through the witharepetitionfactorofthree.Thedatawereprocessedfollowing Very Large Array (VLA) archive. The Kepler data at 6cm was thedetaileddescriptiongiveninGroenewegenetal.(2011). kindlyprovidedbyT.DeLaney(DeLaneyetal.2002).TheChan- The PACSmaps werereduced withthe Herschel Interactive draX-raydataforKepler (PIS.Reynolds; Reynolds etal.2008) ProcessingEnvironment(HIPE;Ott2010)applyingalllow-levelre- and Tycho (PI J.Hughes; see Badenes et al. 2006; Katsuda et al. ductionsteps(includingdeglitching)toLevel1.TheSCANAMOR- 2010) were available online as calibrated files extracted into soft PHOSsoftware (Roussel 2011) wasthenused remove effectsdue andhardregionsoftheX-rayspectrum4.ForKepler’sremnant,the tothethermaldriftsandtheuncorrelated1/f noiseoftheindivid- X-ray data used here range from 0.30− 0.72, 0.72 −0.90 and ualbolometers.FinalLevel2mapcreationthroughdataprojection 1.7−2.1keV,andforTycho,wehave0.95−1.26,1.63−2.26 onto a changeable spatial grid was performed with SCANAMOR- and4.1−6.1keV.Theserangesshowlineemissionfromsilicon PHOS.SincetheHIPEreductionwasbasedonthecalibrationfile andironorginatingfromtheejectaaswellassoftandhard(with versionv5,interimcalibrationfactorsof1.12,1.15and1.17were thermalandnon-thermal contributions) X-raycontinuum orginat- appliedtothePACSdataattherespectivewavelengthstobeconsis- ingfromtheblastwave. tentwiththecurrentcalibrationfilev6.TheFullWidthHalfMax- To compare the submm data with the molecular emission imum(FWHM)at70,100and160µmis6,8and12′′respectively. towards and surrounding the remnants, we use high resolution ThefluxcalibrationuncertaintyforPACSiscurrentlyestimatedas 12CO(J = 2−1) and13CO(J = 2−1) observations for Ke- 10percentforthe70and100µmbandsand20percentat160µm pler (seeGomezet al.2009 forfull details).Wealsoinclude our (Poglitschetal.2010).ColourcorrectionsforPACSareexpected SCUBAimageofKeplerat850µm(Morganetal.2003;Gomezet tobesmallcomparedtothecalibrationerrors.Tocorrectforcolour al.2009).Lowresolutioncarbonmonoxide12CO(J =1−0)maps withdustspectrumν1.0andtemperaturesof100K,wedividedby forTycho’sremnantweretakenfromtheCanadianGalacticPlane 1.00and1.031at70and100µm.The20Kinterstellardustcompo- Survey(CGPS;Tayloretal.2003).Thedatawereintegratedover nentdominatesthe160µmimage,andassuchthecorrectionfactor −68 to −53kms−1 to compare the remnant with molecular gas is0.957. thoughttobeinteractingwiththesupernovaremnant(see§3.1). ForSPIRE,thestandardphotometerpipeline(HIPEv.5.0)was used (Griffinet al. 2010) with an additional iterativebaseline re- moval step (e.g. Bendo et al. 2010). The SPIRE maps were cre- 2.3 ZeroLevels ated with the standard pipeline NA¨IVE mapper. We multiply the Due to the difficulty in finding a suitable baseline with which to 350µmdataproductby1.0067tobeinlinewithmostrecentcali- subtractthebackgroundintheHerschelSPIREmaps,wecorrelate brationpipeline(equivalenttoHIPEv7reduction).Sincethebeam thepixelintensityfromHerschelwiththearchivalIRASIRISmaps areas increase as a function of pixel scale2, the FWHM is 18.1, (calibratedwithrespecttoDIRBE)toestimatethezerolevel(see 24.9 and 36.4′′ for pixel sizes of 6, 10 and 14′′, at 250, 350 and Fig. 1, Table 1). The Herschel data were convolved to the same 500µmrespectively. TheSPIREcalibrationmethodsandaccura- angularresolutionastheIRIS100-µmmap(258′′)andregridded cies are outlined by Swinyard et al. (2010) and are estimated to to 90′′ pixels. A least squares fit analysis was performed on the beapprox10percent.Thepipelineproduces monochromaticflux Herschel vs. IRIS data (whilst masking out the remnant) to de- densities for point sources with νSν ∝ ν1.0 but at the longer termine the flux offsets; this sets the zero level of the PACS and wavelengths,colourcorrectionsbecomesignificant.Wemultiplied SPIREdataassumingthattheIRISdataisatruerepresentationof by thecolour corrections 0.999,1.0045,1.0302 as defined inthe thesky.Afurther scalingfactor needstobeaccounted fordueto Herschel SPIREmanual2, appropriate for β = 1.5 and extended the expected colour correction between the bands when compar- sources. ing to the 100µm band (the so-called ‘gain’ factor). These were estimatedbyinterpolatingamodifiedblackbodycurveν2B(ν,T ) d (after a preliminary correction of the zero levels). The zero level 2.2 MultiwavelengthDatasets of the PACS and SPIRE maps are negative except for the PACS Optical data for the Tycho and Kepler remnants were obtained 70µmmapofKepler.Theleastsquaresfitwasthenrepeated(see from the Hubble Space Telescope (HST) archive via ALADIN e.g. Bracco et al. 2011). The errors on the fitting were estimated 1 http://herschel.esac.esa.int/twiki/pub/Public/PacsCalibrationWeb/ 3 http://irsa.ipac.caltech.edu/data/SPITZER/docs/spitzerdataarchives/ 2 http://herschel.esac.esa.int/Docs/SPIRE 4 http://chandra.harvard.edu/pwarmo/openFITS/ 4 H. Gomezet al. Wavelength Zerolevel(MJy/sr) (µm) Kepler Tycho 70 4.7±0.4 −0.9±0.1 100 −3.1±0.2 −2.5±0.3 160 −4.8±0.3 −2.6±0.2 250 −37.4±3.8 −30.8±4.8 350 −30.5±2.0 −20.4±2.1 500 −25.2±1.4 −7.7±1.3 Table1.ZerolevelonthePACSandSPIREmapscalculated usingleast squaresfittoHerscheldataandIRASIRISdata.Errorsarethe1σestimated frombootstrappingthedata1000timesandrefittingeachset. Figure1.TheHerschelSPIREintensitiesfortheKeplerobservationsver- susIRASIRIS100-µminMJy/sr.TheSPIREmapsareconvolvedtothe sameresolutionastheIRISmapsandregriddedontothesamepixelscale. Figure2.Three-colourFIR-submmimageofKepler’sSNRwithPACSat Theredlineistheleastsquaresfunction,thegreycirclesaretheHerschel 70(blue),100(green)and160µm(red).Theimagehasbeensmoothed. datadividedbythegainfactor(duetothecolourcorrectionbetweenIRIS Theareashownis16.5′×16.6′. at100µmandtheSPIREwavelengths).Thedataaremaskedatthelocation oftheremnant.Thegreyshadedregionsindicatetheresultsfromthe1000 bootstrapfitstothesamedataset. themostlikelyclassificationusingbothironabundanceconstraints andthelackofaneutronstar.ThedeepX-rayimagesalsoprovide evidenceoftheblast-waveinteractingwithdensecircumstellarma- using the confidence intervals from the bootstrap method, where terial(CSM).Thisisunexpected for TypeIaevents although ob- 1000fitsweremadetothebootstappeddata,Table1.Thiserroris servationsofCSMarebecomingmorewidespread;thepresenceof addedinquadraturealongwiththecalibrationerrorslistedabove suchadenseCSMsurroundingKeplerledReynoldstosuggestthe andintroduces(atmost)anerrorof∼ 17%(wellwithinthe20% thermonuclear explosion of a single massive star (M∗ ∼ 8M⊙) calibrationerrorusedforPACSat160µm);notethatthisdoesnot asan alternative to thecanonical White-Dwarf (WD) binary sce- accountforanysystematicerrors. narioof Iaexplosions. Whether Kepler was theresult of a single progenitor or WD-AGB-binarysystem explosion, the mechanism was likely to be a deflagration explosion (Yang et al. 2009; see Woosley,Taam&Weaver1986forareviewonexplosionmodels). 3 KEPLER’SSUPERNOVAREMNANT Inthisscenario, theheating of theejectabyradioactive elements TheKeplersupernova event wasfirstseenbyJohannes Keplerin ishigher than in core-collapse remnants, the ejecta densities will 1604andtheremnanthasashell-likestructure∼3arcminindiam- alsodecreaserapidly,thoughtheejectaiswellmixedsincethere- eter.Estimatesofitsdistance,usingHIabsorptionfeatures,range verseshockprogressesquicklythroughthelayersinadeflagration from3.9to6kpc(Reynoso&Goss1999;Sankritetal.2008).Its explosion(see§5.1). classification has been controversial (see the excellent reviews in The warm dust emission detected in Kepler with ISO (Dou- Blair et al. 2007; Reynolds et al. 2008; Sankrit et al. 2008) with vionetal.2001),followsthewell-knownnorth-southshellasym- evidencepointingtowardsbothType-Ia–thethermonuclearexplo- metryseenalsoatoptical,X-rayandradiowavelengths.Thecor- sionofalow-massaccretingstarinabinarysystem–orType-Ib– relation of the dust with the optical Hα suggested the dust was thecorecollapseofamassivestar.ThedeepestX-rayanalysiswith shock-heatedcircumstellardust,sweptupbytheshockwaveofthe Chandra(Reynoldsetal.2008),however,suggeststhataType-Iais blast. Blair et al. (2007) used Spitzer to revise the dust mass to Dust intheKeplerand TychoRemnants 5 5.4×10−4M⊙. They also found that emission at 70µm arises inKepler:continuumX-raysfromtheblastwavesweepingupsur- onlyinthebrightestknotsoftheshellseenat24µm.The160µm rounding circumstellar and interstellar material, X-ray line emis- Spitzerdatawaspoor-qualityduetoitslowresolutionandstrong sion fromthe warmejecta, and continuum emission arising from background gradient. Morgan et al. (2003) originally interpreted thenon-thermalsynchrotronemissioninthenorth,southandeast- theirlongerwavelength450and850-µmSCUBAdataasevidence ernlimbs.Thewarm dust emissionintheFIR isconfined within for0.3–1M⊙ofcolddustassociatedwiththeremnant.Theycom- theradiusoftheouterblast-waveandisseenonlyatthebrightest pared the emission from interstellar material towards Kepler us- points in the optical, radio and X-ray (Fig. 4 - as noted in Blair ingsubmmcontinuumimagingandspectroscopicobservationsof etal.2007)withpeaksat70µmseenwherethesoftandhardther- atomic and molecular gas, via HI, 12CO and 13CO and detected malX-raysoverlap.Therefore,thewarmdustisassociatedwiththe weakCOemissionfromdiffuse,opticallythingasatthelocations densest gasbehind theforward shock rather thantheinner ejecta of some of the submm clumps. The contribution to the submm material; thisconfirms the dust originates fromthe outermost re- emissionfromforegroundmolecularandatomiccloudsusingthese gionsoftheremnanti.e.fromthesweptupshockedcircumstellar COtracerswasfoundtobenegligible,buttheseemissionlinesmay and/orinterstellarmedium(§5.2). notbetracingallofthegasintheISM. ThethreecolourHerschelPACSimageofKeplerisdisplayed 3.2 Thespectralenergydistribution inFig.2.Theremnantisclearlyseeninblueandfollowsthesame asymmetric structure as seen in the X-ray and radio (§3.1): the TheIR-submmspectralenergydistribution(SED)towardsKepler’s emissionisbrightestinthenorthalonganarc-likeshellandsome SNR isshown inFig. 5. Thefluxes estimated from Spitzer, Her- emissionisseentowardsthecentreoftheremnant.Thelowlevel schelandSCUBAinthisworkwithinanapertureofradius120′′ redstructuresseenacrossthemap(outsideoftheremnant)origi- centredontheremnant,areplottedwiththeirerrors(estimatedfrom natefromdustyinterstellarcloudswithtemperatures∼13−20K. thecalibrationerror,theerrorinthezerolevelestimatedin§2.3and fromthevariationsinthepixelRMSwithintheaperture).Allthe data were smoothed to the resolution of the SPIRE 500µm data 3.1 Themultiwavelengthview usingaGaussiankernel.Theapertureissomewhatlargerthanthe InFig.3,wepresentamultiwavelengthviewcentredonKepler’s SNR boundary defined by the edge of the radio (103′′) but was SNR,comparingtheHerschelPACSandSPIREphotometricdata chosen to encompass the SNR emission in the images smoothed with synchrotron emission from the remnant at 6cm, unrelated tothe500µmbeam.TheSpitzerfluxeswerecolourcorrectedus- molecular gas(tracedbytheCOemission) and850µm emission ingalgorithmsdetailedintheMIPShandbook5.Wealsocompare fromcolddust.Thecompleteshell-likestructureofSNmaterialis ourfluxeswithpreviouslypublishedfluxesintheliterature(Arendt clearlyseenatradiowavelengths,tracingtheforwardandreverse 1989,Braun1987,Douvionetal.2001,Morganetal.2003,Blairet shocks (DeLaney et al. 2002). In Fig. 4 we compare the optical al.2007)althoughthesewereobtainedusingapertureswithradius and X-ray emission with the warm dust from Herschel-PACS at rangingfrom100−120′′andthereforedifferfromthefluxesinTa- 70µm.Thewell-knownnorth-southasymmetryseenatoptical,ra- ble2accordingly.Notealsothatwemeasurethefluxencompassed dio and X-ray wavelengths with the enhanced northern emission intheaperturearoundKepler,andthereforeincludesemissionfrom (thought to originate from the motion of Kepler through the sur- theforegroundcloud. rounding interstellarmaterial)isalsoseeninthePACSdata.The The contribution to the IR/submm fluxes from synchrotron ‘arm’structureacrossthemiddle(clearlyseenintheradio,Fig.3) emissionisobtainedbyscalingtheVLAradiomapofKeplerus- originatesfromaprojectioneffectofmaterialinthefrontandback ingaconstant spectral index, Sν ∝ να wherethemean valueof oftheshell(Blairetal.2007).HerscheldetectsfaintIRemission αis−0.71butrangesfrom-0.6−-0.85(DeLaneyetal.2002).The arisingfromthiscomponentat70µm.Thefaint‘ear’-likefeatures synchrotronemissionat500µmis0.20±0.06Jy(∼ 15percent seenalongtheoutermostmidsectionoftheexpanding shellatra- of thetotal flux). Wefitatwo-component modified blackbody to dio and X-ray wavelengths (Figs. 3 & 4) are not seen in the in- thedustSEDafterthesynchrotroncomponenthasbeensubtracted; frared.Atthelonger(160µm)PACSandtheSPIREwavelengths, themodelisthesumoftwomodifiedPlanckfunctionseachwitha itisdifficulttodisentangleanySNRemissionfromthelarge-scale characteristictemperature,TwandTc(Eq.1): interstellarstructuresextendingacrosstheregion;themapsarein- Sν =Nw×νβB(ν,Tw)+Nc×νβB(ν,Tc) (1) creasinglydominated by unrelateddust cloudsat temperaturesof ∼16−20K.Thesoutherncloudofdustwhichisoutsidethearea whereNw andNc representtherelativemassesinthewarm oftheblastwaveandejectaregionsisthesamecloudorginallyde- andcoldcomponent, B(ν,T)isthePlanckfunction andβ isthe tectedinmolecularemissionbyGomezetal.(2009)at5kms−1. dustemissivityindex.ThemodelwasfittedtotheSED(constrained The cold dust cloud in the north-east is coincident with the CO bythe12µmflux)andtheresultingparameters(Nw,Nc,Tw,Tc, structureat10kms−1(see§5.3formorediscussion). β) which gave the minimal chi-squared were found. We also ap- Opticalemissionarisingfromdense,knottyregionsinthecir- pliedabootstrapanalysistoourSED-fittingtodeterminetheerrors cumstellar medium overrun by the blast wave, is traced in [NII] ontheSEDmodel.Thephotometrymeasurementswereperturbed andHα.Interestingly,the70µmemissioncorrelateswiththeHα accordingtotheerrorsateachwavelength,thenewfluxeswerefit- +[NII]features(Fig.4)inboththenorthernlimb(particularlyin ted in the same way, and the SED parameters recorded, with the thenorth-westregion)andacrosstheprojectedshellalongthemid- procedurerepeated1000times. dle.SincetheHαemissionoriginatesfromcollisional heatingof ThetotaldustmassiscalculatedusingEq.2: trhoeuntdhiinngremgiaotenribael)h,inthdetshpeaStiNalashgorecekmfreonnttw(ii.eth.ftrhoemPAswCeSpet-mupisssiuorn- Md = S500D2 × Nw + Nc (2) κ500 (cid:20)B(500µm,Tw) B(500µm,Tc)(cid:21) and the ionised features suggest this dust component arises from theswept up medium. Thisisfurther supported when comparing the FIR with X-ray emission. There are three X-ray components 5 http://irsa.ipac.caltech.edu/data/SPITZER/docs/mips/ 6 H. Gomezet al. Figure3.Multiwavelength montageofKepler’s supernovaremnantanditsenvirons.Top:PACSandmiddle:SPIREbandscentredatRA = 262.671◦, Dec=−21.4914◦(J2000.0;yellowcross);theregionshowis7.2′×7.2′.Thelocationoftheforwardshock(radius103′′)isindicatedwiththewhitecircle (DeLaneyetal.2002).Alsoshownarebottom:theSCUBA850µm(Morganetal.2003);theCO(J =2−1)mapintegratedfrom−200<v<200kms−1 (Gomezetal.2009)andthe6cmVLAmap(DeLaneyetal.2002). IRFluxesTowardsKeplerSNR(Jy) 24µm 70µm 100µm 160µm 250µm 350µm 500µm 850µm Published 9.5±1.0a 5.6±1.4b 2.9±1.1b <0.9a .. .. 3.0±0.2c,d 1.0±0.2c Thiswork 9.5±1.0e 12.3±2.7 11.2±2.4 16.5±2.9 13.0±2.8 5.8±1.2 2.7±0.6 0.7±0.1 Nonthermal 0.02 0.05 0.06 0.09 0.12 0.16 0.20 0.29 Table2.FluxesinJymeasuredforKepler’sremnantwithinthe120′′aperture.Notethatthepreviouspublishedintegratedfluxesusedifferentaperturesizes tothiswork.References:a-Blairetal.(2007);b-averagevaluesfromDouvionetal.(2001),Arendt(1989)andSaken,Fesen&Shull(1992)IRASdata;c- Gomezetal.(2009);d-450µm.e-MeasuredonSpitzerMIPSarchivedata. whereκ500 isthedustabsorptioncoefficientandDisthedistance followBlairetal.(2007)inadoptingadistanceof4kpc.Todeter- (wechoosethe500µmfluxsincethisislesssensitivetothetemper- minethedustmass,weuseκ500 = 0.1m2kg−1 withκ ∝ λ−β, ature).Thedistancetotheremnantisuncertain,withmeasurements appropriate for grains inthe MilkyWay and β ∼ 1−2(Draine ofHIabsorptionplacingKeplerat∼ 5kpc(seeReynoso&Goss &Lee1984). Thebest-fitdust masswithintheaperture obtained 1999 for a full discussion). Sankrit et al. (2008) have revised the fromtheSEDmodel(Eqs.1&2)isthereforeMw = 0.0035M⊙ measurement to 3.9+−11..94kpc using HST proper motion data. We andMc =2.2M⊙,withtemperaturesTw =84K,Tc =19Kand Dust intheKeplerand TychoRemnants 7 Figure4.ComparisonofthewarmdustemissiontracedbyPACS70-µmdata(whitecontours)withopticalandX-ray.Left:ThreecolourimageofKepler’s supernovaremnantwiththeHSTat658nm(red),660nm(green)and502nm(blue)indicatingemissionatHα+[NII],[NII]and[OIII]respectively. The opticalemissionmostlyarisesfromshockedcircumstellar/interstellarmaterialinteractingwiththeprimaryblastwave.Thenorthcomponentindicatedwith thefiducialisthoughttobeejectamaterial(Sanrkitetal.2008)whichhasoverruntheprimaryshockboundaryduetothecontactdiscontinuityintheblastwave. Right:ThreecolourX-rayimageusingChandraarchivedatashowingsoft(0.3−0.72keV:red),(0.72−0.9keV:green)andhardX-rays(1.7−2.1keV: blue).TheimagesarecenteredonαJ2000=262.663◦andδJ2000 =−21.485◦andare3′×3′across. β = 1.5. The temperature of the warm component is consistent withcollisional heating of dust intheshocked gaswithtempera- tureandelectrondensityof1.2keVandne ∼ 20cm−3 (Bouchet etal.2006,theirFig.15). The median parameters and confidence intervals from the bootstrap method are listed in Table 4. The dust mass depends strongly on thechoice of κ, for example, typical interstellar dust grains are described well with κ = 6.7m2kg−1 at 70µm (e.g. Draine&Li2001).Inthiscase,thewarmdustmasswouldbere- ducedtoMw = 1×10−3M⊙.Wenotethattheuncertaintyinκ couldbeanorderofmagnitude. TemperaturemapsofthedustemissiontowardsKeplerwere createdusingtheconvolvedIR-submmmapsfrom24-500µm,and fittinga two-temperature modified blackbody SED to each pixel. Weextractedthewarmandcoldcomponentsforeachpixel,shown inFig.6(a)and(b).Inordertoaccountforthenon-thermalcompo- Figure5.TheIR-radioSEDtowardsKepler’s SNR.Thepreviouslypub- nent,thefluxduetosynchrotronradiationwassubtractedfromthe lished fluxes from the literature (Table 2) are shown with errors as ver- tical bars (Spitzer = purple triangles; IRAS = blue diamonds; SCUBA = dustmapsonapixel-by-pixelbasisbeforethegreybodyfitswere orange squares). The solid black line is the sum of the two-temperature made. The synchrotron contribution in each pixel was estimated thermal dustcomponents fittedtotheIR(Tw = 86KandTc = 20K) usingaspectralindexmapcreatedfromthe6cmand23cmVLA andsubmmfluxes(dot-dashedanddashedlinesrespectively).Thedot-dot- observationsconvolvedtothesamebeam(seealsoDeLaneyetal. dot-dashedlineshowsthenon-thermalSEDexpectedfromthesynchrotron, 2002). The somewhat flatterspectral index inthe northern bright Sν ∝ν−0.71scaledbythe6cmfluxes.TheredlineshowsthetotalSED ring(α∼−0.68comparedtotheaveragevalueof-0.71)resulted fromthenon-thermalandthermalcomponents.Notethatthepreviouspub- in removing most of the cold dust in this region, though there is lishedintegratedfluxesusedifferentaperturesizestothiswork. residualemissionassociatedwiththiscomponent withintheSNR shell (Fig. 6(b)). This cold dust feature appears tobe coincident withasimilarstructureseeninthewarmdust component butthe colddustisfurtheroutfromthecentreby∼ 10′′ comparedtothe 8 H. Gomezet al. Kepler, the interstellar dust structures are associated with atomic ratherthanmoleculargas. If the cold dust inthe north of the remnant islocated at the distanceoftheremnant,thelocationoftheemissionissuggestive ofdustsweptupbytheblastwave,thoughthemassofdustinthe northern shell is rather substantial at ∼ 1.0M⊙ for T ∼ 20K. Elsewhere, the cold dust component does not spatially coincide with any of the SNR tracers including X-ray, radio and optical emissionandisunlikelytoberelatedtotheSNitself.Wediscuss theoriginofthecolddustfurtherin§5.3. Finally,wealsotestedforthepresenceofcooldustwithtem- peraturesintherange30-40KasfoundinCasA(Sibthorpeetal. 2010;Barlowetal.2010):thefluxcontributionofthewarmcom- ponent(visibleat24µm)wassubtractedfromthe70and100µm maps,however,wefindnoevidenceofacooldustcomponent. 4 TYCHO’SSUPERNOVAREMNANT (a) TheTychosupernovaeventwasobservedbyTychoBrahein1572. Itliesatadistanceof1.5−3.8kpc(Reynosoetal.1997;Krause et al. 2008b) and has a thin shell-like morphology with diameter ∼ 8′ across. Here,weusethedistance giveninD = 3.8kpcas this is consistent with the historical records of the explosion, the non-detectionoftheremnantandthedistancetotheproposedbi- narycompanion(Luetal.2011)(seeKrauseetal.2008bformore details). The outer blast wave extends to an average distance of 251′′ from thecentre, withthe contact discontinuity between the reverse shock material and swept up ISMat angular distances of 183,241and260′′ dependingonthepositionangle(Warrenetal. 2005).Tycho’slightechospectrumhasrevealedittohavebeena TypeIaevent(Krauseetal.2008b);itistheremnantofacanoni- calsingle-degeneratebinaryexplosion(Luetal.2011) asevident fromX-rayejecta abundances (Decourchelle et al.2001). Thisis furthersupportedbythelackofsimilarityinTycho’sopticallight curve withtypical sub-luminous (van den Bergh1993) and over- luminous(Ruiz-Lapuente2004)TypeIaexplosions(Krauseetal. 2008b). Badenesetal.(2006) performeddetailedcomparison be- tweentheX-rayspectraandtheresultsfromhydrodynamicalmod- elingofthesupernovashocks,andfoundthatthedelayeddetona- (b) tionexplosionmodelwasthebestfittothedata.Inthisscenario, theSNRisexpandingintoanambientdensityof∼0.6−3cm−3, Figure6.Dusttemperaturemapsderivedfromfittingamodifiedgreybody withanejectamassof1.3M⊙. to the 24-500µm maps towards Kepler’s supernova remnant showing(a) Many authors have suggested that Tycho is interacting with warmdustand(b)colddust.ThecolourbarshowsthetemperatureinK.The solidblackcircleindicatesthephotometryapertureusedwithradius120′′ molecular clouds (Reynoso & Goss 1999; Lee, Koo & Tatenatsu 2004; Cai, Yang & Lu 2009; Xu, Wang & Miller 2011) partic- andtheblack dashedcircle isthe average radius ofthe outershockfront 103′′. ularly with a large CO structure seen in the north and north-east atvelocitiesbetween-63.5and-61.5kms−1;thisstructureisbe- lievedtoberesponsible for theasymmetryseenintheX-rayand warm dust emission which is spatially coincident with the radio radio (see Fig. 8). Cai et al. (2009) showed that interaction with (Fig.6(a)). theISMoccursalongtheentireboundaryoftheremnantwithin- Thecoldduststructuresinthenortharealsocoincidentwith terstellar structures detected at velocities -69 − -58kms−1; the thestructuresseenat850µmwiththeSCUBAcamera(Gomezet totalgasmassofcloudsinteractingwithTychois103M⊙ (Xuet al.2009,Fig.3)originallyattributedtosupernovadustduetothe al. 2010). Conversley, Tian & Leahy (2011) propose that theCO lackofcorrespondancebetweentheCOgasfrommolecularinter- clouds are not interacting withtheremnant sincethe densities of stellar clouds towards Kepler (Fig. 3) and the 850µm emission. themoleculargascloudsinthenorthare∼200cm−3,comparedto ThelargerHerschelmapclearlyshowslarge-scaleandexendedin- theX-rayejectaparameterswhichsuggesttheremnantisplough- terstellarduststructuresrightacrosstheentireregiontowardsKe- ingintomaterialofdensity0.2−3cm−3(Decourchelleetal.2001, plerandimpliesthatthattheargumentinGomezetal.wasflawed. Badenesetal.2006).Thedifferenceinthedensityestimatedforthe Thiscouldsimplybeduetoanon-constantratioofmoleculargas- surroundingmaterialfromCO/X-rayabundancesisafactorof60 to-dust, or that the12COand 13COlinesarenot effectivetracers (ratherthanthethreeordersofmagnitudequotedinTian&Leahy ofthegasinthisregion.Alternatively,itispossiblethat,towards (2011)).However,thedensityoftheHIcloudtheydetectat-43− Dust intheKeplerand TychoRemnants 9 remnant.InFig.9,wecomparethePACS70µmemissionwitha three-colour X-ray image (0.95-1.26, 1.63-2.26 and 4.1-6.2keV). Theemissionfromwarmdustat70µmisboundedwithinthefor- wardandreverseshocksoftheSNRat251′′and183′′respectively (Warrenetal.2005) andwedonotseeFIRemissiontowardsthe interioroftheremnant.The250µmemission(Fig.8),however,ex- tendsfaroutsidetheshockboundaryandspatiallycorrelateswith thelow-levelstructuresandthepeaksseenintheCOmap(Fig.10). Theredoesn’tappeartobeanyagreementbetweenthepeaksat70 and250µmthoughweseeemissionatbothFIRwavelengthswhere thereisalsoCOemission.ComparingwiththeCOmoleculargas, theinterstellarcloudsseenbyHerschelclosetotheremnantinthe north(labeledNWandNE)andtotheeast,matchstructuresseen at−61and−65kms−1.Inthesoutheastandsouthwest,twofaint COcloudsareseeninthesamerangeofvelocity,bothwithassoci- ated70µmand250µmemission. The correlation between the 70µm emission and the X-rays emittedattheouteredgesoftheSNRisstriking.ThehardX-ray emission(4.1-6.2keV)isfilamentaryinstructurewiththinrimsex- tendingoutsideofthecircularedgebothinthenorthandsouth-east sideoftheremnant.Thepeaksinthewarmdusttracethepeaksin Figure7.Three-colourFIRimageofTycho’sSNRwithPACSdataat70, 100and160µm.Theareashownis23.2′×27.2′. thehardX-rays.TheFIRemissiondoesnotspatiallycoincidewith thesoftX-rays(Fig.9)withpeaksinthesoftX-raylaggingbehind the70µmpeaksby27′′inthenorthwest,thoughthesoftemission 57kms−1 is comparable to the density obtained from X-ray ob- attheedgesoftheouterblastwavedocorrelatewiththeFIR(re- servations and could thereforebe tracing gasinteracting withthe gions‘A’and‘B’inFig.9).Thepeaksintheradioemissionalsolag remnant. Usinga flatgalactic rotationmodel (withstandard con- by∼16′′.Theejectamaterialdoesnotspatiallycorrelatewiththe stants)placesthisHIcloudatd=3.4kpcwhichisconsistentwith IRemission,withthepeakinthenorthwestlaggingbehindtheFIR thedistanceofTycho’sSNR. by23′′.ThecorrelationofwarmdustwiththehardX-rayemission Ishihara et al. (2010) detect warm dust features in the north at theedges of the remnant isslightlysurprising due tothelarge oftheremnantwithAKARIandsuggestthesearelinkedtotheCO fractionofnon-thermalemissionexpectedfromshocksinthisen- structuresinthenortheast.Theyalsofindenhanceddustemission ergyregime.However,dustisseenwheresoftX-rays(fromthermal in thenorthwest region of the shell and since they find no corre- emission due tothe blast wave sweeping up interstellar material) spondingCOfeature,theysuggestthiscomponentcouldoriginate and hard X-rays overlap. Lee et al. (2010) showed that the hard from SN dust. Without the long wavelength data, they were un- X-raysalsocontainasignificantfractionofthermalemission,we abletoconstraintheamountofcolddustarisingfromSNmaterial thereforeconclude that thewarmdust isspatiallycorrelatedwith withintheblastwave. thesweptupmaterial.Thelackof70µmemissionintheSWre- ThethreecolourHerschelPACSimageispresentedinFig.7. gionoftheremnant,wherethehardX-raysarebright,maysuggest AsinKepler,theremnantisclearlyvisibleinblue,andtheinterstel- thattheCO(Fig.10)and250-µmstructureinthisregionorginates larstructuresbecomecleareratthelongerPACSandSPIREwave- fromaforegroundcloudwhichisnotassociatedwiththeremnant; lengths(Fig.8).Cold,dustycloudspeakingat250µmareseento this would mean very littleinterstellar dust is being swept up on theimmediateeastandwestoftheremnantandhavetemperatures thisside(inagreement withthefasterexpansion rateoftheSNR typical of the ISM. Thebright SPIREinterstellar structure to the observedinthisregion).ThehardX-raysinthewest-southwestre- immediatenorth of theremnant inFig. 8, ishighly suggestive of gion are produced by non-thermal synchrotron emission revealed emissionoriginatingfromgasanddustsweptupbytheexpanding by the presence of X-ray ‘stripes’ due to tangled magentic fields SNR.Thestructureissimilarinappearancetothesubmmcavities on small scales (Eriksen et al. 2011). Here then the hard X-rays blownoutbythewindsfromwarmOBstars(e.g.Schneideretal. aredominatedbynon-thermalprocessesandthelackofwarmdust 2010)withfilamentaryfingerspointinginwardstothecentreofthe inthisregionfurthersupportsthehypothesisthatweonlyseedust ionisingradiation(inthiscasetheSNR).Inthehigher resolution in the shocked, swept-up ISM where we find thermal continuum PACSdata (Fig. 7), we seethree small, warm sources embedded X-raysandHαemission-thisishighlightedinFig.11. withintheSPIREclouds(see§4.2forfurtherdiscussion).Thecom- There is no correlation between the FIR and X-rays for the pactbluesourcetothenorth(Fig.7)istheCepheidVariablestar, outermost ejecta material which has broken through the reverse V∗ASCas. shockregionouttotheveryedgeoftheSNR(aslabeledinFig.9). Weconcludethatthereisnoevidencetosuggestthatthewarmdust isassociatedwiththeejecta. 4.1 Themultiwavelengthview Theemissionfromwarmdustat70µminthenorthhastwobright 4.2 Thespectralenergydistribution componentsintheNEandNW(Fig.8);seenalsointheAKARIdata atwavelengthsoutto160µm.Thiscorrespondsspatiallytopeaks The IR-submm SED towards Tycho’s SNR is shown in Fig. 12. in24µmemissionaswellastheX-raycontinuumwhichtracesthe ThefluxesestimatedfromSpitzerandHerschelwithinanaperture forward blastwave. Thelonger wavelength Herschel-SPIREdata ofradius251′′centredontheremnant,areplottedwiththeirerrors do not spatially correlate with radio or X-ray emission from the (Table 3; the zero-level offsets were applied to the maps before 10 H. Gomezet al. Figure8.MultiwavelengthmontageofTycho’ssupernovaremnant.Fromlefttoright:top:12CO(J =1−0)dataintegratedfrom−68<v<−53kms−1), 18cmradiodataandSpitzerMIPS24µm.Middle:HerschelPACSandbottom:SPIRE.TheimagesarecentredatRA=6.3308◦,Dec=64.1372◦(J2000.0; yellowcross)withsize9′×9′.Thelocationoftheforwardshock(radius251′′)isindicatedwiththewhitecircle. IRFluxestowardsTychoSNR(Jy) 12µm 24µm 70µm 100µm 140µm 160µm 250µm 350µm 500µm Published 1.8±0.8a .. 38.6±7.7b,c .. 46.7±14.0b 44.5±13.4b .. .. .. .. 28.3±3.2b,d .. 40.6±8.1b,e .. .. .. .. .. Thiswork .. 20.1±3.0 44.8±10.8 41.1±8.8 .. 59.0±14.1 42.7±8.6 32.8±6.5 17.6±3.5 .. 18.9±3.8f Nonthermal .. 0.65 1.28 1.58 1.98 2.09 2.73 3.34 4.14 Table3.FluxesinJymeasuredtowardsTycho’sremnantwithinthe251′′aperturedefinedinFig.8.Referencesa-Douvionetal.2001;b-Ishiharaetal. 2010;AKARIfluxesmeasuredatwavelengths:c-65µm;d-22.9µm;e-90µm.f MeasuredonWISEdataat22µm.