Toappear in ApJ PreprinttypesetusingLATEXstyleemulateapjv.5/2/11 COLD CO GAS IN THE ENVELOPESOF FU ORIONIS-TYPE YOUNG ERUPTIVE STARS A´. Ko´spa´l1,2, P. A´braha´m1, T. Csengeri3, Th. Henning2, A. Moo´r1, R. Gu¨sten3 Toappear inApJ ABSTRACT FUorsareyoungstellarobjectsexperiencinglargeopticaloutburstsduetohighlyenhancedaccretion 7 fromthecircumstellardiskontothestar. FUorsareoftensurroundedbymassiveenvelopes,whichplay 1 a significant role in the outburst mechanism. Conversely, the subsequent eruptions might gradually 0 clear up the obscuring envelope material and drive the protostar on its way to become a disk-only 2 T Tauri star. Here we present an APEX 12CO and 13CO survey of eight southern and equatorial FUors. We measure the mass of the gaseous material surrounding our targets. We locate the source n of the CO emission and derive physical parameters for the envelopes and outflows, where detected. a OurresultssupporttheevolutionaryscenariowhereFUorsrepresentatransitionphasefromenvelope- J surrounded protostars to classical T Tauri stars. 8 Subject headings: stars: pre-mainsequence— stars: variables: T Tauri— stars: circumstellarmatter 1 ] R 1. INTRODUCTION envelopes comes from modeling broad-band spectral en- ergydistributionsofthedustemission,basedonspatially S FU Orionis-type objects (FUors) constitute a small unresolvedphotometricdatamainlyatinfraredandsub- . group of young stars characterized by large optical- h millimeter wavelengths. The gascomponent, however,is infrared outbursts, attributed to highly enhanced ac- p typically much less studied. With the goal to obtain a cretion (Hartmann & Kenyon 1996). During these out- - generalpicture ofthe moleculargascontent,we perform o bursts,accretionratesfromthe circumstellardisk to the r starareintheorderof10−4M⊙/yr,threeordersofmag- a comprehensive and homogeneous survey of all known t FUors, by measuring millimeter CO lines using single s nitude higher than in quiescence or in normal T Tauri dish telescopes. In this paper we present observations a stars. FUors are natural laboratorieswhere not only en- [ hancedaccretionbutenhancedmasslosscanbe studied. of the envelopes of eight southern and equatorialFUors, andstudythedistributionandkinematicsofthecircum- Most FUors have optical jets, molecular outflows, and 1 stellar gas, including the characterization of the molec- opticallyvisiblering-likestructuresona0.1pcscalethat, v ular outflows where detected. Our data reveal the large 8 in some cases, might be connected to expanding shells variety and trends in the envelope structures predicted 1 thrownoffduringpreviousoutbursts(McMuldroch et al. by the evolutionary models. 0 1993). 5 Circumstellar envelopes are supposed to play a sig- 0 nificant role in the outbursts of FUors, partly by 2. OBSERVATIONS . replenishing the disk material after each outburst Table 1 lists the targets selected for our study 1 (Vorobyov & Basu 2006), partly by triggering the erup- from the list of Audard et al. (2014). We used the 0 7 tions (Bell & Lin 1994). For this reason, envelopes are FLASH+ receiver (Klein et al. 2014) at the APEX tele- 1 not static, but evolve with time. Based on the appear- scope (Gu¨sten et al. 2006) to measure the 12CO(3–2), : ance of the 10µm silicate feature, Quanz et al. (2007) 13CO(3–2),and 12CO(4–3)lines towardsour targets be- v defined two categories of FUors: objects showing the tween 2014 August 23 – 28. APEX is a 12m diame- i feature in absorption are younger, still embedded in a X ter millimeter-wave telescope located on the Llano de circumstellarenvelope;objectsshowingthesilicateband Chajnantor at 5104m altitude in the Chilean Atacama r a inemissionaremoreevolved,withdirectviewonthesur- desert. FLASH+ is a dual-frequency heterodyne re- face layer of the accretion disk. A similar evolutionary ceiver, operating simultaneously in the 345GHz and the sequence was outlined by Green et al. (2006) based on 460GHz atmospheric windows, providing 4GHz band- the amount of far-infraredexcess. These studies suggest width in each sideband. The lower frequency channel that FUors represent a fundamentally important tran- was tuned to 344.2GHz in USB to cover the 13CO(3–2) sition period during early star formation when the em- at 330.588GHz, and the 12CO(3–2) at 345.796GHz, re- bedded protostar clears away its enshrouding envelope spectively. The higher frequency channel was tuned to to become a Class II TTauri star (Sandell & Weintraub the 12CO(4–3)line at 461.041GHz in USB. We used the 2001; Green et al. 2013). XFFTS backends providing a nominal 38kHz spectral Traditionally, a large part of our knowledge on FUor resolution for the 3–2 lines and 76kHz for the 4–3 line. For each target, 90′′×90′′ on-the-fly (OTF) maps were 1Konkoly Observatory, Research Centre for Astronomy and obtained, using a relative reference off position 1000′′ EarthSciences,HungarianAcademyofSciences,Konkoly-Thege away in right ascension. We started each observation Mi2kMlo´saxu´-tP1la5n-1c7k,-I1n1s2t1ituBtufdu¨arpAesstt,rHonuonmgaier,yK¨onigstuhl 17, 69117 by checking in total power mode whether the off posi- tionsareclean. Ifneeded,wemodifiedthe OFFposition Heidelberg,Germany 3Max-Planck-Institut fu¨r Radioastronomie, Auf dem Hu¨gel (1200′′ for V900 Mon and 800′′ for Bran 76), until we 69,53121Bonn,Germany made sure that there is no CO emission at the velocity 2 K´ospa´l et al. Table 1 COobservations ofourtargets. Name Distancea vLSR F(12CO(3–2)) F(12CO(4–3)) F(13CO(3–2)) τ12 τ13 Mtot Outflow? Sifeature (pc) (kms−1) (Jykms−1) (Jykms−1) (Jykms−1) (M⊙) AR6A/6B 800 5.3 3800±7 5190±39 1150±7 57 0.8 1.3b n ? Bran76 1700 17.7 18.8±1 21.4±5.7 4.27±1.39 24 0.4 0.02 n em HBC494 460 4.3 3660±11 4780±50 1070±15 130 1.9 0.4 y abs HBC687 400 17.2 173±6 164±22 37.0±6.7 15 0.2 0.01 n em Haro5aIRS 470 11.2 7940±9 13800±50 2990±8 76 1.1 1.2 y abs OOSer 311 8.1 15500±24 27800±140 3250±27 48 0.7 0.6 ? abs V346Nor 700 −3.0 2490±8 4780±33 383±8 52 0.8 0.3 y abs V900Mon 1100 13.6 199±2 234±11 50.3±2.2 67 1.0 0.1b n em a DistancesarefromAudardetal.(2014)andReipurth&Aspin(1997). b ForAR6A/6BandV900Mon,thereisnodistinctpeakintheCOemissionatthestellarposition,thereforetheCOemission(andmassesgiven here)aremostlikelynotassociatedwiththesesources. of the target. even the 13CO line seems to be double-peaked. Because A first order baseline was removed from the spectra. the line ratios do not indicate extraordinarily high opti- The data were calibrated using a main beam efficiency cal depths, we suspect that in this case several different of 0.73 and 0.60 at 352 and 464GHz, respectively, and velocity components are superimposed along the line of the values were converted to Jy using 41JyK−1 and 48 sight. Theobserveddiversityofthelineprofilesseemsto JyK−1 at 352 and 464GHz, respectively. The rms noise beacharacteristicoftheFUorclass. Evans et al.(1994) levelcalculatedfortheline-freechannelsin1kms−1 bins presented single-dish CO line data for a sample contain- is 0.8Jy for the 12CO(3–2) and 13CO(3–2) lines, and ing both northern and southern FUors. In RNO 1B, 2.3Jy for the 12CO(4–3) line. The telescope’s beam is V1735 Cyg, and V346 Nor they detected self-absorbed 19′.′2, and 15′.′3 at the corresponding frequencies. 12CO, while Z CMa, V1057 Cyg, and V1515 Cyg dis- played narrow, single-peaked 12CO emission. The 13CO 3. RESULTSANDANALYSIS line was single-peakedinallof their sources,similarlyto CO emission for all targeted isotopologues and transi- our results. tions were detected in our maps. In Fig. 1 we show the COline profilesintegratedwithina10000auradiuscen- Integrated emission maps— For all of our targets there teredonthenominalpositionofourtargets,whileFig.2 is some CO emission towards the stellar position, but showsthetotalCOlineintensitymapsintegratedforthe there is also significant confusion from extended emis- velocitychannelswhereatleasta3σsignalwasdetected. sion. Bran76,HBC494,HBC687,Haro5aIRS,OOSer, The flux-weighted average vLSR velocities for the CO andV346Nor,where the CO emissionpeaksat the stel- emission are listed in Table 1. The velocity-integrated lar position, are clearly detected. For AR 6A/6B, and line fluxes for same spatial areas (within 10000au) are V900 Mon, the CO emission peak is offset, so it cannot also given in Table 1. We used the optically thin 13CO be unambiguously associatedwith the star. In any case, lines to convert the observed line fluxes to total gas the masses we calculated should be considered as upper masses assuming local thermodynamic equilibrium, us- limitsfortheenvelopemassesduetoconfusion. Threeof ing 20K temperature, a 13CO/12CO abundance ratio of oursourcesweretargetedbySandell & Weintraub(2001) 69(Wilson 1999) and12CO/H abundance ratio of10−4 in850µmand1.3mmcontinuum. WhileBran76wasun- 2 (Bolatto et al.2013). Wenotethatifweuse50Kinstead detected, for HBC 494 and V346 Nor they give envelope of 20K, the masses would be a factor of 1.06 lower, and masses assuming 50K for the dust temperature. Their if we used 15K instead of 20K, the masses would be a values (0.1M⊙ for HBC 494 and 0.5M⊙ for V346 Nor) factor of 1.29 times higher. are in good agreement with our mass estimates from Line profiles— Figure 1 demonstrates that out of our the CO line fluxes using 50K (0.4M⊙ for HBC 494 and 0.3M⊙ for V346 Nor). sample, Bran76 and HBC 687show the narrowestlines, the FWHM is only about 0.7–0.8kms−1. V900 Mon is Comparison with dust continuum emission— In Fig. 1 we somewhat broader, while the rest of the targets show overplottedwithcontoursthe250µmemissionusingHer- very broad lines, and prominent line wings in the 12CO schel/SPIRE data from the Herschel Science Archive lines. For most of the sources (AR 6A/6B, HBC 494, (proposal IDs: KPGT fmotte 1, KPGT pandre 1, Haro5aIRS,V346Nor,andV900Mon),the13COlineis OT1 maudar01 1). Herschel at this wavelength had a single-peaked,whilethe12COlinesareeitherflat-topped similarbeamsize(18′′)toourAPEXbeamforthe J=3– or show self-absorption. This suggests that 12CO is op- 2 lines (19′.′2). Generally there is a good agreement be- tically thick. The same is true for the two targets with tweenthecontinuumandtheCOlinemaps,althoughthe thenarrowlines,wheretheratioofthe12CO(3–2)tothe continuumpeaksaremoreprominentthantheCOpeaks, 13CO(3–2) line peaks suggest a maximum optical depth and there is less extended emission in continuum than of τ =15–24for the former and τ =0.2–0.4for the lat- in CO. Just like in CO, Bran 76, HBC 494, HBC 687, 12 13 ter. For the rest of the targets, the line peaks indicate Haro 5a IRS, and V346 Nor are clearly detected in con- somewhat larger optical depths, in the 50–130 range for tinuum,OOSerismarginallydetected,whileAR6A/6B τ and in the 0.7–1.9 range for τ . The line profile seem to be sitting in the middle of a cavity. Unfortu- 12 13 of OO Ser is different from the other sources, because nately, V900 Mon was not observed by Herschel. Cold CO in the envelopes of FUors 3 1200 250 800 AR 6A 25 Bran 76 HBC 494 HBC 687 12CO(3-2) 6B 1000 200 20 y) 600 y) y) 800 y) 150 12CO(4-3) J J 15 J J ux ( 400 ux ( 10 ux ( 600 ux ( 100 13CO(3-2) Fl 200 Fl 5 Fl 400 Fl 50 200 0 0 0 -5 0 -5 0 5 10 15 10 15 20 25 -5 0 5 10 10 15 20 25 v (km/s) v (km/s) v (km/s) v (km/s) lsr lsr lsr lsr 4000 6000 120 Haro 5a OO Ser V346 Nor V900 Mon IRS 5000 600 100 3000 Jy) Jy) 4000 Jy) 400 Jy) 80 x ( 2000 x ( 3000 x ( x ( 60 u u u u Fl Fl 2000 Fl 200 Fl 40 1000 1000 20 0 0 0 0 5 10 15 20 0 5 10 15 -10 -5 0 5 5 10 15 20 v (km/s) v (km/s) v (km/s) v (km/s) lsr lsr lsr lsr Figure 1. COlineprofilesofourtargetsobservedwithAPEX.Fluxeswereintegratedwithina10000auradiuscenteredonthenominal positionofourtargets. Thelinewingsmarkedbytheverticallinesindicatepossibleoutflows. AR 6A/6B Bran 76 HBC 494 HBC 687 n) n) n) n) mi 0.5 mi 0.5 mi 0.5 mi 0.5 arc arc arc arc et ( 0 et ( 0 et ( 0 et ( 0 s s s s off off off off C C C C DE -0.5 DE -0.5 DE -0.5 DE -0.5 10000 au 0.5 0 -0.5 0.5 0 -0.5 0.5 0 -0.5 0.5 0 -0.5 RA offset (arcmin) RA offset (arcmin) RA offset (arcmin) RA offset (arcmin) Haro 5a IRS OO Ser V346 Nor V900 Mon n) n) n) n) mi 0.5 mi 0.5 mi 0.5 mi 0.5 arc arc arc arc set ( 0 set ( 0 set ( 0 set ( 0 off off off off C C C C DE -0.5 DE -0.5 DE -0.5 DE -0.5 0.5 0 -0.5 0.5 0 -0.5 0.5 0 -0.5 0.5 0 -0.5 RA offset (arcmin) RA offset (arcmin) RA offset (arcmin) RA offset (arcmin) Figure 2. Integrated CO intensity maps of our targets for the 12CO(3–2) line observed with APEX (grayscale) and 250µm continuum emissionfromHerschel(contours). Plussignsmarkthestellarposition,whilethehatched circleshowstheHerschelbeamsize. Outflows— Some 12CO lines have high-velocity wings, imation, and by correcting for the optically thick emis- indicated by the vertical lines in Fig. 1. We integrated sion using the (1 − e−τ12)/τ correction factor, where 12 the emission for the red-shifted and blue-shifted parts τ was calculated from the 12CO / 13CO line ratio in 12 and plotted the resulting maps with red and blue con- each velocity channel. The outflow masses, momenta, tours in Fig. 3. We detected clear signs for outflows in and energies of the FUors fall into the upper 30% com- HBC 494, Haro 5a IRS, and V346 Nor. OO Ser may pared to the distribution of these values measured by also drive an outflow, but at this spatial resolution, the Dunham et al. (2014) for a sample of 28 outflows driven detection is only tentative due to confusion in the area. bylow-massprotostars. TheoutflowofV346Norwasal- Forthe three unambiguouslydetectedoutflows,we mea- ready detected by Evans et al. (1994) in 12CO(3–2) and sured the masses, momenta, and energies of the blue by Reipurth et al. (1997) in the 12CO(1–0), revealing a and red lobes, following the method and equations pre- similar morphology of the outflowing gas as our data. sented in Dunham et al. (2014). The results, listed in Lee et al.(2002)observedHBC494in12CO(1–0). Their Tab. 2, are calculated both in the optically thin approx- channel maps at ≈15′′ resolution look very similar to 4 K´ospa´l et al. AR 6A/6B Bran 76 HBC 494 HBC 687 n) n) n) n) mi 0.5 mi 0.5 mi 0.5 mi 0.5 arc arc arc arc et ( 0 et ( 0 et ( 0 et ( 0 s s s s off off off off C C C C DE -0.5 DE -0.5 DE -0.5 DE -0.5 10000 au 0.5 0 -0.5 0.5 0 -0.5 0.5 0 -0.5 0.5 0 -0.5 RA offset (arcmin) RA offset (arcmin) RA offset (arcmin) RA offset (arcmin) Haro 5a IRS OO Ser V346 Nor V900 Mon n) n) n) n) mi 0.5 mi 0.5 mi 0.5 mi 0.5 arc arc arc arc set ( 0 set ( 0 set ( 0 set ( 0 off off off off C C C C DE -0.5 DE -0.5 DE -0.5 DE -0.5 0.5 0 -0.5 0.5 0 -0.5 0.5 0 -0.5 0.5 0 -0.5 RA offset (arcmin) RA offset (arcmin) RA offset (arcmin) RA offset (arcmin) Figure 3. IntegratedCOintensitymapsofourtargetsforthe12CO(3–2)lineobservedwithAPEX(grayscale). Theredandbluecontours show redshifted and blueshifted emission integrated in the velocity ranges indicated inFig. 1 (contour levels are at 3, 6, 9, ...σ). Thick black lines mark the directions of the detected bipolar outflows. Plus signs mark the stellar position, while the hatched circle shows the Herschelbeam size. ours. Haro 5a IRS and its surroundings were observed are more evolved. We checked the Cornell Atlas of in 12CO(1–0) and 12CO(3–2) by Takahashi et al. (2006) Spitzer/IRSSources4formid-infraredspectra,andfound and Takahashi et al. (2008). They clearly detected the thatHBH494,Haro5aIRS,OOSer,andV346Norhave CO outflow from Haro 5a IRS, and found an embedded silicateabsorptionfeature,whileBran76,HBC687,and protostellar candidate, MMS 7-NE,which also drives an V900 Mon exhibit emission (see also Tab. 1). Our divi- outflow. This complex spatial and velocity structure of sion based on the CO gas properties of our targets cor- the CO emission is also reflected in our observations. relateswell with the divisionbasedonthe 10µmsilicate feature. Objects possessing massive gas envelopes ex- 4. DISCUSSIONANDCONCLUSIONS hibitsilicateabsorption,whilethosewithlower-massen- Figure 1 and Table 1 reveal a striking variety of en- velopes show silicate emission (Tab. 1). This conclusion velope properties within our sample. One of these suggestsaparallelevolutionofthecircumstellardustand properties is envelope mass. Haro 5a IRS, OO Ser, gas in FUors. V346 Nor, and HBC 494 contain a significant amount When young stellar objects transition from the em- of gas (>0.3M⊙). HBC 678 and Bran 76 are associ- bedded to the disk-only phase, they clear away their en- ated with only 0.01-0.02M⊙ of material. Having no as- velopes to become an optically visible star. According sociated CO peaks, AR 6A/6B and V900 Mon probably tothehypothesisofQuanz et al.(2007),repetitiveFUor also have low-mass envelopes. Interestingly, other enve- outburstsmaydrivethis process,sincea thickobscuring lope parameters suggest an almost identical division of envelope produces an absorption feature, while for an the sample. Envelopeswith higher mass exhibit broader emissionfeature,largeopeningangleforthepolarcavity lines, while the low-mass envelopes have narrower lines in the envelope is needed to provide a clear line-of-sight or remain undetected towards the source. Temperatures to the inner disk (see also Kenyon & Hartmann 1991, calculatedfrom the ratioof the 12CO(4–3)and 12CO(3– Green et al. 2006). FUor outbursts gradually widen the 2) line ratios show that the low-mass envelopes are typ- outflow cavitydue to the enhanced outflow activity dur- ically cold (5-7K), while the higher-mass envelopes are ing the eruptions. warm (>40K). Outflows are only detected from sources During a typical FUor outburst 0.01M⊙ mass is ac- with higher-mass envelopes. By observing a large sam- creted onto the central star (Hartmann 2008). Depend- pleoflow-massprotostars,Jørgensenet al.(2009)found ing on the length of the quiescent periods, a similar that the envelope mass decreases sharply from typically amount of material may be accreted between the out- 1M⊙ inClass0objectsto <0.1M⊙ intheClassI phase. bursts. Thisiscomparabletotheenvelopemassfoundin Placingourtargetsintothisevolutionaryscheme,FUors ourmoreevolvedsubsample. Themassreservoirinthese with higher envelope masses represent a very early evo- systems to replenish the disk after an outburst is very lutionaryphase. TheFUorswithlowenvelopemassesor small. Therefore, these objects are probably very close upperlimitsmaybeclosetotheendoftheClassIphase. to the transition to the disk-only phase and may repre- As we summarized in the Introducion, Quanz et al. sent the links between the ClassI and ClassII phases of (2007) proposed a different way to order FUors into protostellar evolution. an evolutionary sequence: objects showing silicate ab- sorption at 10µm are younger, still embedded in their opaque envelopes, while those showing silicate emission 4 http://cassis.sirtf.com/ Cold CO in the envelopes of FUors 5 Table2 Bell,K.R.,&Lin,D.N.C.1994, ApJ,427,987 Outflowmasses(M),momenta(P)andenergies(E). Bolatto,A.D.,Wolfire,M.,&Leroy,A.K.2013,ARA&A,51, 207 Parameter HBC494 Haro5aIRS V346Nor Dunham,M.M.,Arce,H.G.,Mardones,D.,etal.2014,ApJ, opticallythin 783,29 M(blue)M⊙ 0.004 0.023 0.020 Evans,II,N.J.,Balkum,S.,Levreault, R.M.,Hartmann,L.,& M(red)M⊙ 0.007 0.020 0.054 Kenyon,S.1994,ApJ,424,793 P(blue)M⊙kms−1 0.016 0.039 0.076 Green,J.D.,Hartmann,L.,Calvet,N.,etal.2006,ApJ,648, P(red)M⊙kms−1 0.017 0.042 0.228 1099 E(blue)erg 7.8×1041 8.2×1041 3.8×1042 Green,J.D.,Evans,II,N.J.,K´ospa´l,A´.,etal.2013,ApJ,772, E(red)erg 1.2×1042 1.0×1042 1.3×1043 117 opticallythick Gu¨sten,R.,Nyman,L.˚A.,Schilke,P.,etal.2006,A&A,454,L13 M(blue)M⊙ 0.021 0.311 0.100 Hartmann,L.2008,AccretionProcessesinStarFormation M(red)M⊙ 0.053 0.131 0.092 Hartmann,L.,&Kenyon,S.J.1996,ARA&A,34,207 P(blue)M⊙kms−1 0.061 0.417 0.281 Jørgensen,J.K.,vanDishoeck,E.F.,Visser,R.,etal.2009, P(red)M⊙kms−1 0.166 0.244 0.312 A&A,507,861 E(blue)erg 2.1×1042 6.2×1042 9.3×1042 Kenyon,S.J.,&Hartmann,L.W.1991,ApJ,383,664 E(red)erg 5.6×1042 4.8×1042 1.5×1043 Klein,T.,Ciechanowicz, M.,Leinz,C.,etal.2014,IEEE Transactions onTerahertzScienceandTechnology, 4,588 Lee,C.-F.,Mundy,L.G.,Stone, J.M.,&Ostriker,E.C.2002, ApJ,576,294 This work was supported by the Momentum grant of McMuldroch,S.,Sargent,A.I.,&Blake,G.A.1993,AJ,106, the MTACSFK Lendu¨letDiskResearchGroup,andthe 2477 HungarianResearchFund OTKA grantK101393. T.Cs. Quanz,S.P.,Henning,T.,Bouwman,J.,etal.2007,ApJ,668, 359 acknowledges support from the Deutsche Forschungsge- Reipurth,B.,&Aspin,C.1997,AJ,114,2700 meinschaft,DFG viatheSPP(priorityprogramme)1573 Reipurth,B.,Olberg,M.,Gredel,R.,&Booth,R.S.1997, A&A, ‘Physics of the ISM’. 327,1164 Facilities: APEX. Sandell,G.,&Weintraub,D.A.2001,ApJS,134,115 Takahashi,S.,Saito,M.,Ohashi,N.,etal.2008,ApJ,688,344 Takahashi,S.,Saito,M.,Takakuwa, S.,&Kawabe,R.2006, ApJ, REFERENCES 651,933 Vorobyov,E.I.,&Basu,S.2006, ApJ,650,956 Audard,M.,A´brah´am,P.,Dunham,M.M.,etal.2014, Wilson,T.L.1999,ReportsonProgressinPhysics,62,143 ProtostarsandPlanetsVI,387