ebook img

On the physical structure of IRC+10216. Ground-based and Herschel observations of CO and CCH PDF

0.77 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 On the physical structure of IRC+10216. Ground-based and Herschel observations of CO and CCH

Astronomy&Astrophysicsmanuscriptno.17635 c ESO2012 (cid:13) January10,2012 On the physical structure of IRC+10216 ⋆ Ground-based and Herschel observations of CO and C H 2 E.DeBeck1,R.Lombaert1,M.Agu´ndez2,3,F.Daniel2,L.Decin1,4,J.Cernicharo2,H.S.P.Mu¨ller5,M.Min6,P. Royer1,B.Vandenbussche1,A.deKoter4,6,L.B.F.M.Waters7,4,M.A.T.Groenewegen8,M.J.Barlow9,M. Gue´lin10,13,C.Kahane11,J.C.Pearson12,P.Encrenaz13,R.Szczerba14,andM.R.Schmidt14 1 InstituteforAstronomy,DepartmentofPhysicsandAstronomy,KULeuven,Celestijnenlaan200D,3001Heverlee,Belgium e-mail:[email protected] 2 2 CAB.INTA-CSIC.CrtaTorrejo´nkm4.28850Torrejo´ndeArdoz.Madrid.Spain 1 3 LUTH,ObservatoiredeParis-Meudon,5PlaceJulesJanssen,92190Meudon,France 0 4 AstronomicalInstitute“AntonPannekoek”,UniversityofAmsterdam,SciencePark904,1098XHAmsterdam,TheNetherlands 2 5 I.PhysikalischesInstitut,Universita¨tzuKo¨ln,Zu¨lpicherStr.77,50937Ko¨ln,Germany 6 AstronomicalInstituteUtrecht,UniversityofUtrecht,POBox8000,NL-3508TAUtrecht,TheNetherlands n 7 SRONNetherlandsInstituteforSpaceResearch,Sorbonnelaan2,3584CAUtrecht,TheNetherlands a 8 RoyalObservatoryofBelgium,Ringlaan3,1180Brussels,Belgium J 9 DepartmentofPhysicsandAstronomy,UniversityCollegeLondon,GowerStreet,LondonWC1E6BT,UK 9 10 InstitutdeRadioastronomieMillime´trique(IRAM),300ruedelaPiscine,38406Saint-Martin-dHe`res,France 11 LAOG,ObservatoiredeGrenoble,UMR5571-CNRS,Universite´JosephFourier,Grenoble,France ] 12 JetPropulsionLaboratory,Caltech,Pasadena,CA91109,USA R 13 LERMA,CNRSUMR8112,ObservatoiredeParisandE´coleNormaleSupe´rieure,24RueLhomond,75231ParisCedex05,France S 14 N.CopernicusAstronomicalCenter,Rabianska8,87-100Torun,Poland . h Received—;accepted— p - o Abstract r t Context. Thecarbon-richasymptoticgiantbranchstarIRC+10216undergoesstrongmassloss,andquasi-periodicenhancements s a ofthedensityofthecircumstellarmatterhavepreviouslybeenreported.Thestar’scircumstellarenvironmentisawell-studied,and [ complexastrochemicallaboratory,withmanymolecularspeciesprovedtobepresent.COisubiquitousinthecircumstellarenvelope, whileemissionfromtheethynyl(C2H)radicalisdetectedinaspatiallyconfinedshellaroundIRC+10216.Asreportedinthisarticle, 1 werecentlydetectedunexpectedlystrongemissionfromthe N = 4 3, 6 5, 7 6, 8 7,and9 8transitionsofC Hwiththe v IRAM30mtelescopeandwithHerschel/HIFI,challengingtheavaila−bleche−micala−ndphy−sicalmode−ls. 2 0 Aims. WeaimtoconstrainthephysicalpropertiesofthecircumstellarenvelopeofIRC+10216,includingtheeffectofepisodicmass 5 lossontheobservedemissionlines.Inparticular,weaimtodeterminetheexcitationregionandconditionsofC H,inordertoexplain 2 8 therecentdetections,andtoreconcilethesewithinterferometricmapsoftheN =1 0transitionofC H. 1 Methods. Usingradiative-transfermodelling,weprovideaphysicaldescriptionof−thecircumstellare2nvelopeofIRC+10216,con- . strainedbythespectral-energydistributionandasampleof20high-resolutionand29low-resolutionCOlines—todate,thelargest 1 modelledrangeofCOlinestowardsanevolvedstar.Wefurtherpresentthemostdetailedradiative-transferanalysisofC Hthathas 0 2 beendonesofar. 2 1 Results.Assuming a distance of 150pc to IRC+10216, the spectral-energy distribution is modelled with a stellar luminosity of 11300L andadust-mass-lossrateof4.0 10 8M yr 1.Basedontheanalysisofthe20high-frequency-resolution COobserva- : − − v tions,an⊙averagegas-mass-lossrateforthel×ast1000y⊙earsof1.5 10 5M yr 1isderived.Thisresultsinagas-to-dust-massratioof − − i 375,typicalforthistypeofstar.Thekinetictemperaturethrough×outthec⊙ircumstellarenvelopeischaracterisedbythreepowerlaws: X T (r) r 0.58forradiir 9stellarradii,T (r) r 0.40forradii9 r 65stellarradii,andT (r) r 1.20forradiir 65stellar r rakdinii.T∝his−modelsuccessf≤ullydescribesallk4in9ob∝serv−edCOlines.W≤eal≤soshowtheeffectofdekinnsity∝enh−ancementsint≥hewindof a IRC+10216ontheC H-abundanceprofile,andthecloseagreementwefindofthemodelpredictionswithinterferometricmapsof 2 theC HN =1 0transitionandwiththerotationallinesobservedwiththeIRAM30mtelescopeandHerschel/HIFI.Wereporton theim2portanceo−fradiativepumpingtothevibrationallyexcitedlevelsofC Handthesignificanteffectthispumpingmechanismhas 2 ontheexcitationofalllevelsoftheC H-molecule. 2 Keywords.Stars:individual:IRC+10216—stars:massloss—stars:carbon—astrochemistry—stars:AGBandpost-AGB— radiativetransfer 1. Introduction Thecarbon-richMira-typestarIRC+10216(CWLeo)islocated ⋆ Herschel is an ESA space observatory with science instruments at the tip of the asymptoticgiantbranch(AGB), whereit loses providedbyEuropean-ledPrincipalInvestigatorconsortiaandwithim- massat a highrate ( 1 4 10−5M yr−1; Crosas&Menten, portantparticipationfromNASA. 1997; Groenewegen∼et−al.,×1998; ⊙Cernicharoetal., 2000; 1 E.DeBecketal.:COandCCHaroundIRC+10216 DeBecketal., 2010). Located at a distance of 120 250pc scan coversthe frequencyranges 480 1250GHz and 1410 − − − (Loupetal.,1993;Crosas&Menten,1997;Groenewegenetal., 1910GHz, with a spectral resolution of 1.1MHz. All spectra 1998; Cernicharoetal., 2000), it is the most nearby C-type were measured in dual beam switch (DBS; deGraauwetal., AGB star. Additionaly, since its very dense circumstellar 2010)modewitha3 chopthrow.Thistechniqueallowsoneto ′ envelope (CSE) harbours a rich molecular chemistry, it has correctforanyoff-sourcesignalsinthespectrum,andtoobtain been deemed the prime carbon-rich AGB astrochemical astablebaseline. laboratory. More than 70 molecular species have already Since HIFI is a double sideband (DSB) heterodyne instru- been detected (e.g. Cernicharoetal., 2000; Heetal., 2008; ment, the measured spectra contain lines pertaining to both Tenenbaumetal., 2010), of which many are carbon chains, the upper and the lower sideband. The observation and data- e.g. cyanopolyynesHC N (n=1, 3, 5, 7, 9, 11) and C N (n=1, reductionstrategiesdisentanglethese sidebandswith veryhigh n n 3, 5). Furthermore, several anions have been identified, e.g. accuracy, producing a final single-sideband (SSB) spectrum C H (n= 4,6,8;Cernicharoetal.,2007;Remijanetal.,2007; without ripples, ghost features, and any other instrumental ef- n − Kawaguchietal., 2007), and C N (Thaddeusetal., 2008), fects,coveringthespectralrangesmentionedabove. 3 − C5N− (Cernicharoetal., 2008), and CN− (Agu´ndezetal., ThetwoorthogonalreceiversofHIFI(horizontalH,andver- 2010). Detections towards IRC+10216 of acetylenic chain ticalV)wereusedsimultaneouslytoacquiredataforthewhole radicals (CnH), for n=2 up to n=8, have been reported by e.g. spectralscan,exceptforband2a2.Sincewedonotaimtostudy Gue´linetal. (1978, C4H), Cernicharoetal. (1986a,b, 1987b, polarisationoftheemission,weaveragedthespectrafromboth C5H), Cernicharoetal. (1987a) and Gue´linetal. (1987, C6H), polarisations, reducing the noise in the final product. This ap- Gue´linetal. (1997, C7H), and Cernicharo&Gue´lin (1996, proachisjustifiedsincenosignificantdifferencesbetweentheH C8H). andVspectraareseenforthelinesunderstudy. The smallest CnH radical, ethynyl (CCH, or C2H), was Adetaileddescriptionoftheobservations,andofthedatare- first detected by Tuckeretal. (1974) in its N = 1 0 tran- ductionofthislargesurveyisgivenbyCernicharoetal.(2010b) − sition in the interstellar medium (ISM) and in the envelope and Cernicharoetal. (2011b, in prep.). All C H data in this 2 aroundIRC+10216.It was shownto be one ofthe mostabun- paper are presented in the antenna temperature (T ) scale3. A∗ dant ISM molecules. The formation of C2H in the envelopeof Roelfsemaetal.(2012,submitted)describethecalibrationofthe IRC+10216 is attributed mainly to photodissociation of C2H2 instrumentandmentionuncertaintiesintheintensityoftheorder (acetylene),oneofthemostprominentmoleculesincarbon-rich of10%. AGBstars.Fonfr´ıaetal.(2008)modelledC2H2 emissioninthe ForthecurrentstudyweconcentrateonobservationsofCO mid-infrared,whichsamplesthedust-formationregioninthein- and C H. The ten CO transitions covered in the survey range 2 ner CSE. The lack of a permanent dipole moment in the lin- from J = 5 4upto J = 11 10andfrom J = 14 13upto earlysymmetricC2H2-moleculeimpliestheabsenceofpurero- J =16 15.−TheC2Hrotation−altransitionsN =6 5,−7 6,and tational transitions which typically trace the outer parts of the 8 7ar−ecoveredbythesurveysinbands1a(480− 56−0GHz), envelope.C2H,ontheotherhand,hasprominentrotationallines 1b−(560 640GHz),and2a(640 720GHz),respe−ctively.The thatprobethechemicalandphysicalconditionslinkedtoC2H2 N = 9 −8transitionisdetectedin−band2b(720 800GHz)of inthesecoldouterlayersoftheCSE.Ithasbeenestablishedthat thesurv−ey,butwithverylowS/N;higher-Ntransi−tionswerenot C2H emission arises from a shell of radicals situated at ∼15′′ detected.AsummaryofthepresentedHIFIdataofC2Hisgiven fromthecentralstar(Gue´linetal.,1993).Observationswiththe inTable1,thespectraareshowninFig.1. IRAM30mtelescopeandHerschel/HIFIshowstrongemission inseveralhigh-NrotationaltransitionsofC H,somethingwhich 2 isunexpectedandchallengesourunderstandingofthismolecule 2.2.Herschel/SPIRE inIRC+10216. In the framework of the Herschel guaranteed time key We presentanddiscussthe high-sensitivity,high-resolution programme “Mass loss of Evolved StarS” (MESS; data obtained with the instruments on board Herschel Groenewegenetal., 2011) the SPIRE Fourier Transform (Pilbrattetal., 2010): HIFI (Heterodyne Instrument for the Spectrometer (FTS; Griffinetal., 2010) was used to obtain Far Infrared; deGraauwetal., 2010), SPIRE (Spectral and IRC+10216’s spectrum on 19 November 2009 (OD 189). Photometric Imaging Receiver; Griffinetal., 2010), and PACS The SPIRE FTS measures the Fourier transform of the source (Photodetector Array Camera & Spectrometer;Poglitschetal., spectrum across two wavelength bands simultaneously. The 2010)andwiththeIRAM30mTelescope1inSect.2.Thephys- short wavelength band (SSW) covers the range 194 313µm, icalmodelforIRC+10216’scircumstellarenvelopeispresented − while the long wavelength band (SLW) covers the range and discussed in Section 3. The C H molecule, and our treat- 2 303 671µm. The total spectrum covers the 446 1575GHz mentofitisdescribedinSection4.Asummaryofourfindings − − frequency range, with a final spectral resolution of 2.1GHz. isprovidedinSection5. Thequalityoftheacquireddatapermitsthedetectionoflinesas weakas1 2Jy.Forthetechnicalbackgroundandadescription − 2. Observations of the data-reduction process we refer to the discussions by Cernicharoetal. (2010a) and Decinetal. (2010b). These 2.1.Herschel/HIFI:SpectralScan authors report uncertainties on the SPIRE FTS absolute fluxes The HIFI data of IRC+10216 presented in this paper are part oftheorderof15 20%forSSWdata,20 30%forSLWdata − − ofaspectrallinesurveycarriedoutwithHIFI’swidebandspec- below 500µm, and up to 50% for SLW data beyond 500µm. trometer(WBS; deGraauwetal.,2010)inMay2010,on3con- secutive operational days (ODs) of the Herschel mission. The 2 Because of an incompleteness, only the horizontal receiver was usedduringobservationsinband2a.Supplementaryobservationswill 1 BasedonobservationscarriedoutwiththeIRAM30mTelescope. beexecutedlaterinthemission. IRAM is supported by INSU/CNRS (France), MPG (Germany) and 3 Theintensityinmain-beamtemperatureT isobtainedviaT = MB MB IGN(Spain). T /η ,whereη isthemain-beamefficiency,listedinTable1. A∗ MB MB 2 E.DeBecketal.:COandCCHaroundIRC+10216 CCH(N = 1−0) IRAM 30m CCH(N = 2−1) IRAM 30m 2.0 J=3/2−1/2 J=1/2−1/2 5 J=5/2−3/2 J=3/2−1/2 J=3/2−3/2 1.5 4 CH, v=1 AlCl 4 7 *T (K)A 1.0 C6H AlCl *T (K)A 23 0.5 1 0.0 0 87.25 87.30 87.35 87.40 87.45 87.50 174.65 174.70 174.75 174.80 174.85 ν (GHz) ν (GHz) CCH(N = 3−2) IRAM 30m CCH(N = 4−3) IRAM 30m 4 J=7/2−5/2 J=5/2−3/2 J=5/2−5/2 J=9/2−7/2 J=7/2−5/2 J=7/2−7/2 6 3 *T (K)A 4 SiN *T (K)A 2 2 1 0 0 261.9 262.0 262.1 262.2 349.3 349.4 349.5 349.6 349.7 ν (GHz) ν (GHz) CCH(N = 6−5) HIFI 1a CCH(N = 7−6) HIFI 1b 0.6 J=13/2−11/2 J=11/2−9/2 J=15/2−13/2 J=13/2−11/2 0.3 0.5 0.4 0.2 K) K) *T (A 00..23 SiC2 *T (A 0.1 0.1 0.0 0.0 523.95 524.00 524.05 611.25 611.30 611.35 611.40 ν (GHz) ν (GHz) CCH(N = 8−7) HIFI 2a CCH(N = 9−8) HIFI 2b 0.10 J=17/2−15/2 J=15/2−13/2 J=19/2−17/2 J=17/2−15/2 0.10 0.08 0.06 K) 0.05 K) 0.04 *T (A *T (A 0.02 0.00 0.00 SiC2 −0.02 30SiS SiC2 −0.04 −0.05 698.5 698.6 698.7 785.6 785.7 785.8 785.9 786.0 ν (GHz) ν (GHz) Figure1.C HN =1 0,2 1,3 2,4 3,6 5,7 6,8 7,and9 8rotationaltransitionsinIRC+10216’senvelopeasobserved 2 − − − − − − − − withtheIRAM30mtelescopeandHerschel/HIFI.Thefinestructurecomponentsarelabelledwiththerespective J-transitionsin thetopofeachpanel,whilethehyperfinecomponentsareindicatedwithverticaldottedlines.Forthesakeofclarity,weomittedthe componentsthataretooweaktobedetected.SeeSect.4fordetailsonthespectroscopicstructureofC H.Additionallyobserved 2 featuresrelatedtoothermoleculesarealsoidentified. We refer to Table 2 for a summary of the presented data, and 2.3.Herschel/PACS show an instructive comparison between the HIFI and SPIRE data of the C H lines N = 6 5 up to N = 9 8 in Fig. 2. It 2 − − Decinetal. (2010b) presented PACS data of IRC+10216, isclearfromtheseplotsthatthelowerresolutionoftheSPIRE also obtained in the framework of the MESS programme spectrumcausesmanylineblendsinthespectraofAGBstars. (Groenewegenetal., 2011). The full data set consists of SED scans in the wavelength range 52 220µm, obtained at differ- − ent spatial pointings on November 12, 2009 (OD182). We re- reduced this data set, taking into account not only the central spaxelsofthedetector,butallspaxelscontainingacontribution 3 E.DeBecketal.:COandCCHaroundIRC+10216 80 2 C C S N 60 Si N Si H x (Jy) 40 HCN CS, SiS 13HC CCH, SiS HCN CS SiS, 13CO u Fl 20 N C H 0 480 500 520 540 80 O S 17C 34C ux (Jy) 4600 HO2 S, SiO, 29SiS NH3 CS SiS 13HCN SiO CCH SiS HCN , HCN, Fl Si Cl 20 H N O C C H 0 560 580 600 620 0.3 CCH HIFI 1b U SiS,ν=2 30SiO SiC SiS,ν=1 0.2 2 * (K)TA 0.1 SiC2 AlCl HCN U U U 0.0 609 610 611 612 613 614 80 C N N O C 60 H Si H ux (Jy) 40 SiS, SiS, 13CO SiS CS , CCH, HCN SiS HCN Fl C2 20 N Si N C C H H 0 640 660 680 700 80 2 N H S C C2 C O C Si 60 S, Si S S S, Si S CO 13, H O S, Si S, C CN, Flux (Jy) 2400 Si C Si 30HO, Si2 Si 13 SiS Si C 1713CO, C H 0 720 740 760 780 ν (GHz) Figure2. SPIRE data in the range 475 795GHz, containingthe C H-linesdetected with HIFI. The shaded strips in the panels 2 − indicatethespectralrangesoftheHIFIdatashowninthepanelsofFig.1.Thelabelsmarktheprincipalcontributorstothemain spectralfeatures.Acomparisonofemissionintherange609 614GHz,asobtainedwithSPIRE andwith HIFI,isshowninthe − insetpanel. 4 E.DeBecketal.:COandCCHaroundIRC+10216 Table1.SummaryoftheIRAM30mandHIFIdetectionsofC H.Columnsare,respectively,theinstrumentorbandwithwhichwe 2 detectedtheC Hemission,themain-beamefficiencyη ,thehalf-powerbeamwidthHPBW,theintegrationtime,thenoiselevel 2 MB inthespectra,thefrequencyresolution∆ν,thevelocityresolution∆3,thedetectedC Hrotationaltransitions,thefrequencyrange 2 inwhichweobservedthelines,andthefrequency-integratedintensityI = T dν. ν∗,A A∗ R Instrument η HPBW Int.Time Noise ∆ν ∆3 CCH Freq.Range I MB ν∗,A orBand ( ) (min) (mK) (MHz) (kms 1) transition (MHz) (KMHz) ′′ − IRAM30m A100/B100 0.82 28.2 614 2.5 1.0 3.4 N=1 0 J=3/2−1/2 87278.2 87339.6 15.90 J=1/2−1/2 87392.8−87453.6 7.58 C150/D150 0.68 14.1 215 10 1.0 1.7 N=2−1 − J=5/2−3/2 174621.1 174682.5 58.73 J=3/2−1/2 174708.2−174744.9 38.09 J=3/2−3/2 174796.2−174863.7 5.90 E3 0.57 9.4 161 10 1.0 1.1 N=3−2 − J=7/2−5/2 261974.3 262023.4 110.74 J=5/2−3/2 262048.8−262084.4 60.63 J=5/2−5/2 262191.3−262270.0 6.33 E3 0.35 7.0 79 5.3 2.0 1.7 N=4−3 − J=9/2−7/2 349311.5 349368.3 66.98 J=7/2−5/2 349368.3−349421.6 60.15 J=7/2−7/2 349575.1−349671.8 2.35 − − Herschel/HIFI 1a 0.75 40.5 28 9.9 1.5 0.86 N=6 5 J=13/2 −11/2 523941.9 524000.4 18.60 J=11/2− 9/2 524000.4−524061.9 18.74 J=11/2 −11/2 − 1b 0.75 34.7 13 7.4 0.50 0.25 N=7−6 − − J=15/2 −13/2 611237.1 611301.6 10.09 J=13/2−11/2 611301.6−611364.1 9.44 J=13/2−13/2 − 2a( ) 0.75 30.4 12 13.7 4.5 1.8 N=8−7 − − † J=17/2 −15/2 698487.5 698619.0 4.21 J=15/2−13/2 698619.0−698738.0 4.47 J=15/2−15/2 − 2b 0.75 27.0 15 10 6.0 2.2 N=9−8 − − J=19/2 −17/2 785750.0 785875.0( ) 0.20( ) ‡ ‡ J=17/2−15/2 785875.0−785900.0( ) 0.20( ) ‡ ‡ J=17/2−17/2 − − − − Notes. ( )Onlythehorizontalpolarisationwasobtained.( )Duetothelowsignal-to-noiseratioofthisdataset,thesenumbersaretobeinterpreted † ‡ withcaution. totheflux,i.e.20outof25spaxelsintotal.Theherepresented sitions are shown in Fig. 1, and details on the observationsare dataset,therefore,reflectsthetotalfluxemittedbytheobserved listedinTable1. regions,withinthe approximationthatthere isnoloss between the spaxels. The estimated uncertainty on the line fluxes is of Table 2.SummarisedinformationontheSPIRE spectrumcon- the order of 30%. Due to instrumental effects, only the range tainingtheN =7 6transitionofC H.∆νisthefrequencyreso- 5126C−O1tr9a0nµsimtioonfstJhe=P1A4CS1s3peucptrtuomJi=su4s2abl4e1.,Tchoivserrainnggeenheorlgdys lution,∆3istheve−locityresolution,2andσisthermsnoise.Fν,A∗ levelsfrom 350cm 1 u−pto 3450cm 1.−Asisthecaseforthe isthefrequency-integratedfluxinthegivenfrequencyrange. − − ∼ ∼ SPIREdata,lineblendsarepresentinthePACSspectrumdueto thelowspectralresolutionof0.08 0.7GHz. Frequencyrange 446 1575GHz Integrationtime 2664s − ∆ν − 2.1GHz ∆3 201kms 1 − F 16915JyMHz σ 850mJy ν,A∗ 2.4.IRAM30m:linesurveys We combine the Herschel data with data obtained with the IRAM30mtelescopeatPicoVeleta.TheC H N = 1 0tran- 2 3. Theenvelopemodel:dust&CO − sition was observed by Kahaneetal. (1988); N = 2 1 was − observedbyCernicharoetal.(2000).N = 3 2andN = 4 3 Weassumedadistanced = 150pctoIRC+10216,ingoodcor- − − wereobservedbetweenJanuaryandApril2010,usingtheEMIR respondence with literature values (Groenewegenetal., 1998; receiversasdescribedindetailby(Kahaneetal.,2011,inprep.). Men’shchikovetal., 2001; Scho¨ieretal., 2007, and references The data-reduction process of these different data sets is de- therein).Thesecondassumptioninourmodelsisthattheeffec- scribed in detail in the listed papers. The observed C H tran- tive temperatureT = 2330K, followingthe modelof a large 2 eff 5 E.DeBecketal.:COandCCHaroundIRC+10216 setofmid-IRlinesofC H andHCN,presentedbyFonfr´ıaetal. Table 3. Dust composition in the CSE of IRC+10216, used 2 2 (2008). We determined the luminosity L , the dust-mass-loss to produce the fit to the SED in Fig. 3. The columns list the ⋆ rate M˙ , and the dust composition from a fit to the spec- dustspecies,theassumedshapes,therespectivemassfractions, dust tral energy distribution (SED; Sect. 3.1). The kinetic tempera- and the references to the optical constants of the differentdust tureprofileandthegas-mass-lossrate M˙ weredeterminedfrom species. a CO-line emission model (Sect. 3.2). The obtained envelope model will serve as the basis for the C H-modelling presented 2 Dustspecies Shape Massfraction References in Sect. 4. We point out that all modelling is performedin the (%) radialdimensiononly,i.e.in1D,assumingsphericalsymmetry Amorphouscarbon DHS 53 Preibischetal.(1993) throughouttheCSE. Siliconcarbide CDE 25 Pitmanetal.(2008) Magnesiumsulfide CDE 22 Begemannetal.(1994) 3.1.Dust 6 The dust modelling was performed using MCMax (Minetal., ϕ=0.00 2009),aMonteCarlodustradiativetransfercode.ThebestSED 5 ϕϕ==00..2540 (ISO) 104 fittotheISOSWSandLWSdata,showninFig.3,isbasedona 102 swteitlhlatrhleumreisnuolstsityofLM⋆=e1n1’s3h0c0hLik⊙o.vTehtiasli.s(i2n0g0o1o)4d,ccoornrseisdpeornindgenthcee Jy) 4 (Jy)Fν 1100−20 gdiivffeesreancseteilnlaardroapdtieudsdRi⋆staonfce2.0Tmhiellci-oamrcbsiencaotniodns,oifnLv⋆erayndgoToeffd 4 (10Fν 23 1100−−64 1 10 102 103 agreement with the values reported by e.g. Ridgway&Keady λ (µm) (1988,19mas)andMonnieretal.(2000,22mas). 1 IRC+10216 is a Mira-type pulsator, with a pe- riod of 649days (LeBertre, 1992). Following Eq. 1 0 of Men’shchikovetal., we find that L varies between 10 100 1000 L 15800L at maximum light (at⋆phase ϕ=0), and λ (µm) ⋆,ϕ=0 L ≈ 6250L⊙ at minimum light (ϕ=0.5). Fig. 3 shows the SfiEg⋆,uDϕr=-e0v,.5ai≈triiasbviliistiyb⊙lceotrhraetspthoendspinrgeatdootnhethLe⋆p-hvoatroiambeiltirtiyc.pForionmtscthane 1F0ig−u8rMe3.ySr−E1D, amndodtehlesdfourstIRcoCm+p1o0s2it1io6n, uasnidngprMo˙pduesrtti=es 4li.s0te×d beaccountedforbythemodelscoveringthefullL -range. in Tabl⊙e 3 and spatial distribution and properties shown in ⋆ Fig. 4. The SED model is constrained by the ISO SWS and TheadopteddustcompositionisgiveninTable3.Themain LWS data (full black); photometric points (black diamonds) constituents are amorphouscarbon (aC), silicon carbide (SiC), are taken from Ramstedtetal. (2008) and Ladjaletal. (2010). andmagnesiumsulfide(MgS),withmassfractionsof53%,25%, Dash-dotted grey: model at maximum light, with 15800L , and 22%, respectively. The aC grains are assumed to follow a dashed grey: best fit to the ISO data, at phase ϕ=0.24, u⊙s- distributionofhollowspheres(DHS; Minetal.,2003),withsize ing11300L ,dash-triple-dottedgrey:modelatminimumlight, 0.01µmandafillingfactorof0.8.ThepopulationoftheSiCand with 6250L⊙. The inset is identical, but plotted in log log MgSdustgrainsisrepresentedbyacontinuousdistributionofel- scale. ⊙ − lipsoids(CDE; Bohren&Huffman,1983),wheretheellipsoids allhavethevolumeofaspherewithradiusa =0.1µm.CDEand d DHSarebelievedtogivea morerealistic approximationofthe 3.2.CO characteristicsofcircumstellardustgrainsthana populationof spherical grains (Mie-particles). Assuming DHS for the domi- To constrain the gas kinetic temperature T (r) and the gas kin nant aC grains was found to provide the best general shape of densityρ (r)throughouttheenvelope,wemodelledtheemis- gas theSED.Alldustspeciesareassumedtobeinthermalcontact. sion of 20 rotational transitions of 12CO (Sect. 3.2), measured The absorption and emission between 7µm and 10µm with ground-based telescopes and Herschel/HIFI. The gas ra- and around 14µm that is not fitted in our SED model diative transfer is treated with the non-local thermal equilib- can be explained by molecular bands of e.g. HCN and rium(NLTE)codeGASTRoNOoM(Decinetal.,2006,2010c).To C2H2(Gonza´lez-Alfonso&Cernicharo,1999;Cernicharoetal., ensure consistency between the gas and dust radiative transfer 1999). models,wecombineMCMaxandGASTRoNOoM,bypassingonthe Based on an average of the specific densities of the dust dustproperties(e.g.densityandopacities)fromthemodelpre- components ρs = 2.41gcm−3, we find a dust mass-loss rate sentedinSect.3.1andFig.4tothegasmodelling.Thegeneral of 4.0 10−8M yr−1, with the dust density ρdust(r), dusttem- methodbehindthiswillbedescribedindetailbyLombaertetal. peratur×eTdust(r)⊙and Qext/ad, with Qext the totalextinctioneffi- (2012,inprep.). ciency,showninFig.4.Theinnerradiusofthedustyenvelope, Thelargedatasetofhigh-spectral-resolutionrotationaltran- i.e. the dust condensation radius of the first dust species to be sitions of CO consists of 10 lines observed from the ground formed, is determined at Rinner=2.7R⋆ by taking into account and 10 lines observed with HIFI, which are listed in Table 4. pressure-dependentcondensationtemperaturesfor the different Sincethecalibrationofground-baseddataisattimesuncertain dustspecies(Kamaetal.,2009).Thesedustquantitiesareused (Skinneretal., 1999), large data sets of lines that are observed asinputforthegasradiativetransfer,modelledinSect.3.2. simultaneously and with the same telescope and/or instrument are of great value. Also, the observational uncertainties of the 4 The extensive modelling of Men’shchikovetal. was based on a HIFIdataaresignificantlylowerthanthoseofthedataobtained light-curveanalysisandSEDmodelling,andholdsaquoteduncertainty with ground-basedtelescopes (20 40%). Since the HIFI data − ontheluminosityof20%. ofCOmakeuphalfoftheavailablehigh-resolutionlinesinour 6 E.DeBecketal.:COandCCHaroundIRC+10216 To reproduce the CO lines within the observational un- R R inner outer,dust certainties, we used L =11300L , and a temperature profile5 2.8 R 5000 R ⋆ * * that is a combination of three p⊙ower laws: T (r) r 0.58 kin − −3m) 10−1 8 ((aa)) Tfor (rr)≤ r9R1.⋆2,atTlkainrg(re)rr∝adiri,−b0.a4sefodro1n0thRe⋆w≤orkrby≤Fo6n5f∝Rr´ı⋆a,etanald. g c 10−20 (2k0in08)∝and−Decinetal. (2010a). The minimum temperature in (st 10−2 2 theenvelopeissetto5K.TheradialprofilesofTkin(r)and3(r) ρdu 10−2 4 a6re s1h0ow4nininthFeigin.n5e.rUwsiinndg,awefraficntidonthaaltaabugnadsa-nmcaessC-lOo/sHs2raotef − K) 1000 ((bb)) CM˙o×mofbi1n.i5ng×t1h0e−r5eMsu⊙ltsyrf−ro1mreSpreocdt.u3ce.1s wthiethCtOhoslienefrsovmertyhewCelOl. ( ust 100 model,we find a gas-to-dust-massratio of 375,in the rangeof Td typical values for AGB stars (102 103; e.g. Ramstedtetal., − 10 2008). The predicted CO line profiles are shown and compared 1 101 102 103 104 to the observations in Fig. 6. All lines are reproduced very r (R) * well in terms of integrated intensity, considering the re- ) 108 spective observational uncertainties. The fraction of the pre- −1m 106 ((cc)) dicted and observed velocity-integrated main-beam intensities c (d 104 IMB,model/IMB,data variesintherange74−145%fortheground- a based data set, and in the range 89 109% for the HIFI data /ext 102 set. The shapes of all observed lines−are also well reproduced, Q 100 withtheexceptionoftheIRAMlines(possiblyduetoexcitation 0.1 1 10 100 1000 and/orresolutioneffects). λ (µm) Fig.7showsthecomparisonbetweentheobservedCOtran- sitions J = 14 13 up to J = 42 41 in the PACS spectrum − − and the modelled line profiles, convolved to the PACS resolu- Figure4. Spatialdistributionand propertiesof the dustfor our best SED fit to the ISO data at ϕ=0.24 in Fig. 3: (a) dust den- tion. Considering the uncertainty on the absolute data calibra- sity ρ (r),(b)dusttemperatureT (r),and(c) Q /a .The tionandthepresenceoflineblendswithe.g.HCN,C2H2,SiO, dust dust ext d SiS,orH Oinseveraloftheshownspectralranges(Decinetal., dashedlinesindicatetheinnerandouterradiiofthedustyenve- 2 2010b),theoverallfitoftheCOlinesisverygood,withexcep- lope. tionofthelinesJ =14 13, 15 14,and16 15.Thissignificant − − − differencebetweenthemodelpredictionsandthedatacould,to date,notbeexplained.BoththeHIFIdataandthehigher-JPACS 1000 (K)n 100 −0.58 −0.40 lineSskairnenreerperotdaul.c(e1d9v9e9r)ynwoteeldlbthyethpeosmsiobdleelp.resenceofagradi- Tki 10 −1.20 entintheturbulentvelocity3turb,withdecreasingvaluesforin- creasingradialdistancefromthestar.Wefindnoclearevidence −1s) 1250 10 100 1000 10000 isnhoowunr dinataFisge.t6thuastetsh3is de=cr1e.a5skemins31tu,rbanisdprerepsreondtu.cTehsethmeolidneel m turb − k 10 shapesandintensitiestoahighdegree.InFig.8wecomparethe (exp 5 predictionsofasetofCO-linesusingdifferentvaluesfor3turbin v 0 ordertoassesstheinfluenceofthisparameter. 10 100 1000 10000 Assuming a lower value of 0.5kms−1 or 1.0kms−1 yields r (R) an incomplete reproduction of the emission profile between * 44kms 1 and 41kms 1 for all lines observed with HIFI. − − −Assuming a high−er value of 2kms 1 leads to the overpredic- Figure5. Overview of the gas kinetic temperature and expan- − tion of the emission in this 3-rangefor all observedlines, with sionvelocityusedintheradiativetransfermodelcalculatedwith GASTRoNOoM. In the top panel, we indicated the exponents α the strongest effect for the lowest-J lines. The influence of the from the T (r) rα-power laws used to describe the kinetic differentvaluesofthe turbulentvelocityonthe emissionin the kin ∝ 3-rangeof 41to 12kms 1isnegligible. temperature,andtheradialrangestheyapplyto.SeeSect.3.2. − − − 4. C H 2 sample, these will serve as the starting point for the gas mod- 4.1.Radicalspectroscopy elling. C H( C C H)isalinearmoleculewithanopenshellcon- 2 The adopted CO laboratory data are based on the work of figurat•ion,i.e.itisaradical,withanelectronic2Σ+ground-state Goorvitch&Chackerian (1994) and Winnewisseretal. (1997), configuration (Mu¨lleretal., 2000). Spin-orbit coupling causes and are summarised in the Cologne Database for Molecular fine structure (FS), while electron-nucleus interaction results Spectroscopy (CDMS; Mu¨lleretal., 2005). Rotational levels in hyperfine structure (HFS). Therefore, to fully describe C H 2 J = 0 up to J = 60are taken intoaccountforboththe ground in its vibrationalground-state,we need the following coupling vibrational state and the first excited vibrational state. CO-H 2 collisionalrateswereadoptedfromLarssonetal.(2002). 5 ThecentralstarisassumedtobeablackbodyatatemperatureT . eff 7 E.DeBecketal.:COandCCHaroundIRC+10216 12 NRAO 0.74 1124 SEST 1.00 20 IRAM 1.01 25 IRAM 0.76 60 IRAM 0.96 10 1−0 1−0 1−0 20 1−0 2−1 10 K) 8 8 15 15 40 (TMB 46 46 10 10 20 2 2 5 5 0 0 0 0 0 80 40 CSO 50 JCM T IRAM 40 CSO CSO 0.92 1.24 100 1.11 1.45 1.29 30 3−2 40 3−2 80 3−2 30 4−3 60 6−5 (K)B 20 30 60 20 40 TM 20 40 20 10 10 10 20 0 0 0 0 0 15 HIFI HIFI HIFI 20 HIFI 20 HIFI 5−4 0.89 15 6−5 0.90 15 7−6 0.98 8−7 0.96 9−8 1.00 15 15 (K)B 10 10 10 10 10 M T 5 5 5 5 5 0 0 0 0 0 25 25 20 1H0IF−I9 1. 00 20 1H1IF−I1 0 1. 09 20 1H4IF−I1 3 1. 04 20 1H5IF−I1 4 1. 12 20 1H6IF−I1 5 1. 07 K) 15 15 15 15 15 (MB 10 10 10 10 10 T 5 5 5 5 5 0 0 0 0 0 −60 −40 −20 0 −60 −40 −20 0 −60 −40 −20 0 −60 −40 −20 0 −60 −40 −20 0 v (km s−1) v (km s−1) v (km s−1) v (km s−1) v (km s−1) Figure6.Comparisonoftheobserved12COemissionlines(blackhistogram),andthelineprofilespredictedbytheGASTRoNOoM- model(red dashedline)with parametersas listed in Table 5. Theline transitions J (J 1) andthe telescopeswith which they − − wereobservedareindicatedintheupperleftcornerofeverypanel.ThefactorI /I isgivenintheupperrightcornerof MB,model MB,data everypanel. scheme: C H has one bending mode ν , and two stretching modes, 2 2 ν andν .Therotationalground-stateofthebendingmode(ν ) 1 3 2 J = N+S (1) is situated 530K above the ground state. The C C stretch ∼ F = J +I, (2) (ν3)at 2650Kisstrong.TheC Hstretch(ν1)at 4700Kis weak,a∼ndisresonantwiththefirstexcited2Πelectro∼nicstateat where N istherotationalangularmomentum,notincludingthe 5750KofC2H.Foreachofthesevibrationalstates—ground- ∼ electron or quadrupolar angular momentum (S and I, respec- state,ν2,ν3,andν1—weconsider20rotationallevels,i.e.rang- tively), J isthetotalrotationalangularmomentum,andFisthe ingfromN =0uptoN =19. nuclearspinangularmomentum. The equilibrium dipole moments of the ground and The strong ∆J = ∆N FS components have been detected first excited electronic states have been calculated as µ = upto N = 9 8.Relativetothestrongcomponents,theweaker 0.769 and 3.004Debye, respectively (Woon, 1995). The − ∆J =0componentsdecreaserapidlyinintensitywithincreasing dipole moments of the fundamental vibrational states in the N, andtheyhavebeendetectedupto N = 4 3.TheHFS has ground electronic state are also assumed to be 0.769Debye. − alsobeen(partially)resolvedfortransitionsuptoN =4 3. This is usually a good assumption as shown in the − Spectroscopic properties of the ground vibrational state of case of HCN by Deleon&Muenter (1984). The band C H have most recently been determined by Padovanietal. dipole moments for rovibrational transitions were calcu- 2 (2009)andarethebasisfortheentryintheCDMS. lated from infrared intensities published by Tarroni&Carter Our treatment of the radiative transfer (Sect. 4.3) does not (2004): µ(ν2=1 0)=0.110Debye, µ(ν3=1 0)=0.178Debye, → → dealwiththeFS andHFSofC2H,andislimitedtothepredic- and µ(ν1=1→0)=0.050Debye. The influence of the vibra- tionofrotationallinesN N with∆N = 1,accordingtoa1Σ tionally excited states on transitions in the vibrational ground ′ approximation. Therefore→, all levels treated are described with statewillbediscussedinSect.4.3. only one quantum number N. The final line profiles are calcu- A more completetreatmentof C H will likely haveto take 2 lated by splitting the total predicted intensity of the line over into account, e.g., overtones of the ν -state. These have non- 2 thedifferent(hyper)finecomponents,dependingonthe relative negligible intrinsic strengths (Tarroni&Carter, 2004) because strengthofthesecomponents(CDMS),andpreservingthetotal of anharmonicity and vibronic coupling with the first excited intensity.Notethat,strictlyspeaking,thissplittingisonlyvalid electronic state. Spectroscopic data for several of these are al- underLTEconditions,butthat,duetoourapproachofC Hasa ready available in the CDMS. A multitude of high-lyingstates 2 1Σ-molecule,wewillapplythisschemethroughoutthispaper. mayalsohavetobeconsideredastheyhaveratherlargeintrin- 8 E.DeBecketal.:COandCCHaroundIRC+10216 800 600 400 200 0 −200 1000 800 700 650 550 −400 14−13 15−14 16−15 17−16 18−17 102000300040005000600000 475 375 325 300 250 800 600 −50 −50 −50 −50 −50 400 1607 1612 1617 1721.0 1726.5 1732.0 1835.0 1841.5 1848.0 1949 1956 1963 2063 2071 2079 200 0 −200 450 500 350 350 350 −400 19−18 20−19 21−20 22−21 23−22 102000300040005000600000 200 225 150 150 150 800 600 −50 −50 −50 −50 −50 400 2176 2185 2194 2290.0 2299.5 2309.0 2404 2414 2424 2517 2528 2539 2631 2642 2654 200 0 −200 350 350 750 750 650 −400 24−23 25−24 26−25 27−26 28−27 102000300040005000600000 150 150 350 350 300 ) 800 y 600 (J −50 −50 −50 −50 −50 400 x 2744 2756 2769 2857 2870 2883 2980 2984 2988 3093 3098 3103 3206.0 3211.5 3217.0 200 u 0 Fl −200 650 500 450 300 300 −400 29−28 30−29 31−30 32−31 33−32 102000300040005000600000 300 225 200 125 125 800 600 −50 −50 −50 −50 −50 400 3319 3325 3331 3432.0 3438.5 3445.0 3545 3552 3559 3657.0 3664.5 3672.0 3769.0 3777.5 3786.0 200 0 −200 250 250 200 150 200 −400 34−33 35−34 36−35 37−36 38−37 102000300040005000600000 100 100 75 50 75 800 600 −50 −50 −50 −50 −50 400 3882.0 3890.5 3899.0 3994 4003 4012 4106.0 4115.5 4125.0 4218 4228 4238 4329 4340 4351 200 0 −200 250 100 100 100 −400 39−38 40−39 41−40 42−41 102000300040005000600000 100 25 25 25 −50 −50 −50 −50 4441 4452 4464 4552 4564 4576 4663 4676 4688 4774 4787 4800 ν (GHz) Figure7.ComparisonofCO-emissionlinesmeasuredwithPACS(blackhistogram)andthepredictedlineprofiles(reddottedline). ThetransitionsJ (J 1)arelabelledinthetopleftcornerofeachsubpanel. − − sic intensities because of the vibronic interaction between the 4.2.Observationaldiagnostics groundandthefirstexcitedelectronicstates. The C H-emission doublets due to the molecule’s fine struc- 2 ture(Sect.4.1)areclearlypresentinallhigh-resolutionspectra showninFig1.ForN =1 0,2 1,and3 2,thehyperfinestruc- − − − 9 E.DeBecketal.:COandCCHaroundIRC+10216 20 SEST 10 CSO 10 CSO CSO 4 1−0 8 3−2 8 4−3 15 6−5 3 6 10 K) 6 (MB 2 4 4 5 T 1 2 2 0 0 0 0 −2 −5 −1 4 HI FI 5 HI FI 5 HI FI 5 HI FI 5−4 4 8−7 4 11−10 4 16−15 3 3 3 3 K) (B 2 2 2 2 M T 1 1 1 1 0 0 0 0 −1 −55 −50 −45 −40 −35 −55 −50 −45 −40 −35 −55 −50 −45 −40 −35 −55 −50 −45 −40 −35 v (km s−1) v (km s−1) v (km s−1) v (km s−1) Figure8.ZoomonthebluewingofaselectionoftheCOlineprofiles,withthetransitionsJ (J 1)labelledinthetopleftcorner of each panel.We show the comparisonbetween the observedlines and predictedlines wit−h ass−umed valuesof 3 =0.5kms 1 turb − (fullgrey),1.0kms 1(dottedgrey),theadoptedvalueof1.5kms 1(dashedgrey),and2.0kms 1(dash-dottedgrey). − − − ture is clearlydetected in our observations.Thedouble-peaked lineprofilesaretypicalforspatiallyresolvedopticallythinemis- 15 sion (Olofsson,in Habing&Olofsson, 2003), andindicate that µν)2 N = 3.84 (0.09) x 10+15 cm− 2 the emitting material has reached the full expansion velocity6, πS3 14 Trot = 23.4 (1.3) K ip.oe.si3ti∞on=of14th.5ekemmsit−t1in.gThshiselilssi(n∼a1c6c′′o;rdGanuce´elinweitthalt.h,e19o9b9se)ravnedd /8I,corrv 13 thevTeolodceitryivperoifinfleordmeraitvioendfoonrththiseernevgeilmopeei(nSewcth.i3ch.2)t.he lines 3610 12 k a1r9e83e)x,cuitseidn,gwe construct a rotational diagram (Schloerbetal., og(3 11 1−0 2−1 3−2 4−3 6−5 7−6 8−7 l 0 50 100 150 N exp Eu = 3k×1036 I3,corr, (3) Eu/k (K) Z(T ) × −kT ! 8π3 νSµ2 rot rot whereNisthecolumndensity,Zthetemperaturedependentpar- Figure9. Rotational diagram of the measured C2H transitions, titionfunction,E theenergyoftheupperlevelofthetransition based on Eq. 3. Transitions — in the 1Σ approximation— are u incm 1,ktheBoltzmannconstantinunitsofcm 1K 1,T the labelled in grey. The numbersgiven in parenthesesare the un- − − − rot rotationaltemperatureinK,µthedipolemomentinDebye,and certaintiesonthederivedcolumndensityandrotationaltemper- ν the frequency of the transitions in Hz. I3,corr is the velocity- ature. integratedantennatemperature, 1 3LSR+3 The unique rotational temperature points to excitation of the I3,corr = fbffηMB Z3LSR−3∞∞TA∗d3 (4) liinnaescoinnfionneedsianrgelaethreagticmane,biencahcacroarcdtaenricseedtowthitehaobnuengdaasntceemppeeark- correctedforthebeamfillingfactor(Kramer,1997) ature(Gue´linetal.,1999). 2 1 exp √ln2 θ /θ , θ <θ fbff =1,− (cid:18)−h × source beami (cid:19) θssoouurrccee ≥θbbeeaamm (5) 44..33..1R.aCdoiallitsivioentaralnrastfeesrmodelling andforthebeamefficiencyη (seeTable1).Wehaveassumed MB Forrotationaltransitionswithinthevibrationalgroundstateand auniformemissionsourcewithasizeof32 indiameterforall ′′ within the ν = 1 bending state and the ν = 1 and ν = 1 C Htransitions,basedontheinterferometricobservationsofthe 2 1 3 2 stretching states we adopted the recently published collisional N =1 0emissionbyGue´linetal.(1999). Fro−m Fig. 9 we see that the C2H emission is characterised rsaiotensaolfraHteCsNfo−rHC2Hby,aDnudmthoeucshimelileatraml.o(2le0c1u0l)a.rTmhaeslsaacnkdofsiczoelloif- by a unique rotational temperature T = 23.3( 1.3)K, and a 2 source-averagedcolumndensity N =ro3t.84( 0.09±) 1015cm 2. HCNandC2Hinspirethisassumption.Thecollisionalratesare ± × − givenfor25differenttemperaturesbetween5and500K,hence 6 We assume that the half width at zero level of the CO emission coveringtherangeoftemperaturesrelevantfortheC2Haround lines,i.e.14.5kms−1,isindicativeforthehighestvelocitiesreachedby IRC+10216.Since ν3 and ν1 are at ∼2650K and ∼4700K, re- thegasparticlesintheCSE. spectively, collisional pumping to these vibrationally excited 10

See more

The list of books you might like

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