Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.arcetri.astro.it/science/starform/preprints/fontani_6.pdf Äàòà èçìåíåíèÿ: Mon Mar 26 19:23:37 2012 Äàòà èíäåêñèðîâàíèÿ: Mon Feb 4 15:09:23 2013 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: cygnus

Mon. Not. R. Astron. Soc. 000, 1­17 (2011)

Printed 26 March 2012

A (MN L TEX style file v2.2)

Dense gas in IRAS 20343+4129: an ultracompact Hii region caught in the act of creating a cavity
´ F. Fontani1 , Aina Palau2, G. Busquet3, A. Isella4, R. Estalella5, A. Sanchez-Monge1, P. Caselli6 and Q. Zhang7

arXiv:1203.5258v1 [astro-ph.GA] 23 Mar 2012

INAF-Osservatorio Astrofisico di Arcetri, L.go E. Fermi 5, Firenze, I-50125, Italy Institut de Ci`ncies de l'Espai (CSIC-IEEC), Campus UAB-Facultat de Ci`ncies,Torre C5-parel l 2, Bel laterra, E-08193, Catalunya, Spain e e 3 INAF-Istituto di Fisica del lo Spazio Interplanetario, Via Fosso del Cavaliere 100, Roma, I-00133, Italy 4 Division of Physics, Mathematics and Astronomy, California Institute of Technology, MC 249-17, Pasadena, CA 91125, USA 5 Departament de Astronomia i Meteorologia (IEEC-UB), Institut de Ci`ncies del Cosmos, Universitat de Barcelona, Marti Franqu`s 1, e e E-08028 Barcelona, Spain 6 School of Physics and Astronomy, University of Leeds, Leeds, LS2 9JT , UK 7 Harvard-Smithsonian Center for Astrophysics, 60 Garden Street MS78, Cambridge, MA 02138, USA
2

1

Accepted date. Received date; in original form date

ABSTRACT

The intermediate- to high-mass star-forming region IRAS 20343+4129 is an excellent laboratory to study the influence of high- and intermediate-mass young stellar objects on nearby starless dense cores, and investigate for possible implications in the clustered star formation process. We present 3 mm observations of continuum and rotational transitions of several molecular species (C2 H, c-C3 H2 , N2 H+ , NH2 D) obtained with the Combined Array for Research in Millimetre-wave Astronomy, as well as 1.3 cm continuum and NH3 observations carried out with the Very Large Array, to reveal the properties of the dense gas. We confirm undoubtedly previous claims of an expanding cavity created by an ultracompact Hii region associated with a young B2 zero-age main sequence (ZAMS) star. The dense gas surrounding the cavity is distributed in a filament that seems squeezed in between the cavity and a collimated outflow associated with an intermediate-mass protostar. We have identified 5 millimeter continuum condensations in the filament. All of them show column densities consistent with potentially being the birthplace of intermediate- to high-mass ob jects. These cores appear different from those observed in low-mass clustered environments in sereval observational aspects (kinematics, temperature, chemical gradients), indicating a strong influence of the most massive and evolved members of the protocluster. We suggest a possible scenario in which the B2 ZAMS star driving the cavity has compressed the surrounding gas, perturbed its properties and induced the star formation in its immediate surroundings. Key words: Stars: formation ­ ISM: individual ob jects: IRAS 20343+4129 ­ ISM: molecules

1

INTRODUCTION

Most of the stars of all masses in the Galaxy form in rich clusters. Despite this, the details of the clustered star formation process are still p oorly understood. Studies of low-mass protoclusters have started to unveil similarities and differences b etween isolated and clustered dense cores (e.g. Andr´ e et al. 2007, Foster et al. 2009, Friesen et al. 2009). Globally



these studies suggest that cluster environment has a relatively smaller influence on the prop erties of the cores (temp erature, mass, velocity disp ersion, chemical abundances of early phase molecules) than is typically assumed (Foster et al. 2009). However, the conclusions describ ed ab ove do not include observations of high-mass star forming regions. Because the phenomena associated with massive star formation have a stronger impact on the environment (massive outflows, UV radiation, expanding Hi i regions), it is plausible that these energetic phenomena have ma jor effects on the

E-mail: fontani@arcetri.astro.it

c 2011 RAS


2

F. Fontani et al.
in Millimeter Astronomy (CARMA) at 3 mm and the Very Large Array (VLA) at 1.3 cm. All selected molecular transitions are commonly used to characterise dense gas: (i) NH3 and N2 H+ are excellent tracers of dense and cold cores b ecause either sp ecies do not suffer from depletion up to 105 cm-3 , and NH3 is extensively used as thermometer in b oth low- and high-mass star forming regions; (ii) NH2 D provides an estimate of the degree of deuteration (with NH3 ). This combination of diagnostic lines was successfully used by Busquet et al. (2010) to identifying pre-protostellar cores in the proto-cluster associated with IRAS 20293+3952; (iii) C2 H and c-C3 H2 are b oth high-density PDR tracers useful to shed light on the interaction among the cold gas and the UV radiation field coming from IRS 1 and IRS 3. C2 H is also a tracer of cold gas (e.g. Beuther et al. 2008, Padovani et al. 2009). In this pap er we concentrate on the gas morphology, temp erature and kinematics of the region adjacent to IRS 1 and IRS 3, and confirm the hyp othesis prop osed by Palau et al. (2007b) that IRS 3 is op ening a cavity in the surrounding dense gas and starless material is b eing accumulated on the cavity walls. In Sect. 2 we describ e the observations. The observational results are presented in Sect. 3, and discussed in Sect. 4. In Sect. 5 we summarise the main findings of the work and give a general conclusion.

surrounding dense material. The study of such interaction is esp ecially imp ortant to quantify the effect of protostellar feedback on the environment and test recent models of highmass star formation including outflows and radiation from the newly b orn stars (e.g. Krumholz et al. 2011, Henneb elle et al. 2011). Interferometric observations of dense gas and dust tracers (N2 H+ , mm continuum, NH3 ) have revealed the presence of pre-stellar core candidates surrounding ultracompact Hi i regions (UC Hi i's) and other massive young stellar ob jects (YSOs) that do show evidences of such an interaction. For example, Fontani et al. (2009) found that in the protocluster associated with IRAS 05345+3157 the kinematics of two pre-stellar core candidates is influenced by the passage of a massive outflow. UV radiation and p owerful outflows affect the chemistry of starless cores in IRAS 20293+3952 (Palau et al. 2007a). On the other hand, a crucial chemical process in pre­stellar cores, i.e. the deuteration of sp ecies like N2 H+ and NH3 , seems to remain as high as in pre-stellar cores isolated and associated with low-mass star forming regions (Fontani et al. 2008, Busquet et al. 2010, Pillai et al. 2011). Therefore, to date it is not clear if and how the presence of massive ob jects affects the prop erties and evolution of the other (pre-)protocluster memb ers. The protocluster associated with IRAS 20343+4129 (hereafter I20343) represents an excellent lab oratory to study this issue. The IRAS source is located to the northeastern side of the Cygnus OB2 association, at 1.4 kp c of distance from the Sun (Sridharan et al. 2002, Rygl et al. 2011), and two bright nebulous stars, IRS 1 (north) and IRS 3 (south), are found inside the IRAS error ellipse when observed at high angular resolution (Kumar et al. 2002). The bright infrared stars are emb edded in a cometary-like cloud whose head, facing the Cygnus OB2 association, is bright at centimeter wavelengths and whose tail, bright in the midinfrared, is extending for ab out 10 ( 4 p c) towards the north-east (Fig. 1). This kind of clouds are also known as bright rimmed clouds. Thanks to interferometric observations of 12 CO and 1.3 mm continuum, Palau et al. (2007b) concluded that IRS 1 is an intermediate-mass Class I YSO driving a molecular outflow in the east-west direction, while IRS 3 is likely a more evolved intermediate/high-mass star. This is further confirmed through mid-infrared photometric and sp ectroscopic observations, from which Campb ell et al. (2008) also estimated a b olometric luminosity of the order of 1000 L for b oth IRS 1 and IRS 3. Furthermore, IRS 3 is at the centre of an UC Hi i-region detected through VLA centimeter continuum emission (Carral et al. 1999), and of a fan-shap ed emission in the 2.12 µm rovibrational line of molecular hydrogen (Kumar et al. 2002). East and west of this fan-shap ed feature, Palau et al. (2007b) detected molecular gas and dust resolved into several millimeter continuum compact sources. Palau et al. (2007b) interpreted these starless condensations as b eing accumulated on the walls of the expanding shock front, but could not derive firm conclusions on their origin and nature. This work aims at b etter understading the nature of the dense cores in I20343, and its relation with the neighb ouring more evolved ob jects. To achieve the goal we p erformed observations of molecular sp ecies obtained at high angular resolution with the Combined Array for Research

2 2.1

OBSERVATIONS AND DATA REDUCTION CARMA

3 mm CARMA observations of I20343 were obtained on 29 Mar 2010 in C- and 01 May 2010 in D-configuration under good weather conditions for observations at 3 mm, characterized by ab out 5 mm of precipitable water and atmospheric noise rms of ab out 300 µm as measured on a baseline of 100 m at the frequency of 225 GHz. The phase centre was the same as in Palau et al. (2007b), namely: R.A. (J2000) = 20h 36m 07s 3 and Dec. (J2000) = 41 39 57. 2. The local . standard of rest velocity of the cloud is assumed to b e 11.5 km s-1 , as determined from single-dish ammonia observations (Sridharan et al. 2002). The primary b eam of the 10 m and 6 m dishes at ab out 85 GHz is 73 and 121 , resp ectively. The single-side-band system temp erature during the observations was b elow 150 K. During C-configuration observations, the correlator provided 4 bands which were configured to observe the continuum, the C2 H, the NH2 D and the N2 H+ lines simultaneously. D-array observations were obtained with the new CARMA correlator, which provides more bands. Two 500 MHz bands were used to observe the continuum and 5 bands set up to observe the C2 H, N2 H+ , NH2 D, CCS and c-C3 H2 line emission. The pass-band was calibrated by observing 1733-130; flux calibration was set by observing MWC349. The estimated uncertainty of the absolute flux calibration is 10% , and it is determined from p eriodic observations of MWC349. Atmospheric and instrumental effects were corrected by observing the nearby quasar 2007+404 every 15 minutes. The tracers observed and the main observational parameters (frequency, synthesised b eam, linear resolution, sp ectral resolution, 1 rms channel noise, largest detectable angular scale) are rep orted in Table 1. The CCS line is the only undetected transition and will b e not discussed in the
c 2011 RAS, MNRAS 000, 1­17


Dense gas in IRAS 20343+4129

3

Figure 1. Large-scale view of the located to the south-west as shown Spitzer image at 5.8 µm, where we the colour scale units are MJy sr-1 with the VLA (D-configuration).

surroundings by the arrow. highlight the . The dashed

of IRAS 20343+4129 as seen by Spitzer at The enlargement panel to the right shows the two infrared sources IRS 1 (saturated in the contours correspond to the 3.6 cm continuum

8.0 µm. The Cygnus OB2 association is region of our interest as it appears in the IRAC image) and IRS 3. In both panels, emission detected by Carral et al. (1999)

following. The continuum was derived by averaging the 500 MHz bands. Visibility data were edited and calibrated with the MIRIAD package. A minor flagging of the data was p erformed using the UVFLAG task, mainly to remove the intervals characterized by the bad atmospheric phase coherence. The channel spacing and the corresp onding 1 rms noise are shown in Table 1. Merging the visibilities obtained in C and D configuration, imaging, deconvolution, and analysis of channel maps and continuum were p erformed using the standard tasks of the GILDAS1 package (e.g. UVMERGE, UVMAP, CLEAN). Images were created applying natural weighting to the visibilites.

for calibration of high frequency data, using the NRAO package AIPS. The absolute flux scale was set by observing the quasar 1331+305 (3C286), for which we adopted a flux of 2.52 Jy at 1.3 cm, and 1.45 Jy at 0.7 cm. The quasar 2015+371, with a b ootstrapp ed flux of 1.39 ± 0.02 Jy at 1.3 cm and 2.1 ± 0.2 Jy at 0.7 cm, was observed regularly to calibrate the gains and phases. Final images were produced with the robust parameter of Briggs (1995) set to 5, corresp onding to natural weighting. At 7 mm we applied a tap er at 60 k with the aim of recovering faint extended emission, but no emission was detected at this wavelength. The VLA was also used to map the (J ,K )=(1,1) and (2,2) inversion transitions of the ammonia molecule on 2001 July 23, with the array in the C configuration. The phase center was set to R.A.(J2000) = 20h 36m 08s 013; . Dec.(J2000)= +41 39 56. 93. The FWHM of the primary b eam at the observed frequency was 110 , and the range of pro jected baselines was 2.59 to 267.20 k. The absolute flux calibration was p erformed by using 3C286, adopting a flux density at 1.3 cm of 2.41 Jy. The phase calibrator was QSO 2013+370, with a 1.3 cm b ootstrapp ed flux density of 2.34±0.04 Jy, and 3C273 was used as the bandpass calibrator. The NH3 (1,1) and NH3 (2,2) lines were observed simultaneously in the 4 IF correlator mode of the VLA (with 2 p olarizations for each line), providing 63 channels with a sp ectral resolution of 0.62 km s-1 across a bandwidth of 3.13 MHz, plus an additional continuum channel containing the central 75% of the total bandwidth. The bandwidth was centered at the systemic velocity VLSR =11.5 km s-1 (Sridharan et al. 2002) for the NH3 (1,1) line, and at VLSR =6.5 km s-1 for the NH3 (2,2) line (to cover the main and one of the satellite comp onents). Data were

2.2

VLA

I20343 was observed with the Very Large Array (VLA2 at 1.3 and 0.7 cm on 2007 Mar 26 using the array in the D configuration3 . The phase center of the observations was R.A. (J2000) = 20h 36m 07. 51; Dec. (J2000) = 41 40 00. 9. The data reduction followed the VLA standard guidelines

1

the GILDAS software is developed at the IRAM and the Observatoire de Grenoble, and is available at http://www.iram.fr/IRAMFR/GILDAS 2 The Very Large Array (VLA) is op erated by the National Radio Astronomy Observatory (NRAO), a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. 3 The VLA continuum observations presented in this work correspond to pro ject AP525. Another pro ject, AP533, was undergone at 3.6, 2 and 0.7 cm, but due to technical problems the data of AP533 pro ject were lost. c 2011 RAS, MNRAS 000, 1­17


4

F. Fontani et al.
Table 1. Observed tracers and basic parameters for the CARMA and VLA observations. The synthesised beam and 1 rms for CARMA observations are based on the combined configurations C+D (unless when differently specified) Instrument/Tracer a CARMA 3 mm continuum NH2 D JKa ,Kc = 11,1 C2 H NJ,F = 13/2,2 - ortho-c-C3 H2 JKa ,Kc CCS NJ = 76 - 65 N2 H+ JF1 = 12 - 01 VL A 7 mm continuum 1.3 cm continuum NH3 (1,1) NH3 (2,2) 43. 22. 23. 23. 3399 4601 6945 7226 4. 3. 4. 4. 67 02 20 25 â â â â 4. 2. 3. 3. 57 98 19 19 0.03 0.02 0.025 0.025 ­ ­ 0.62 0.62 0.0004 0.000057 0.0035 0.0035 18 34 34 34 - 10,1 (F = 1 - 1) 01/2,1 = 21,2 - 10,1 e (F = 3 - 2)f 86. 85. 87. 85. 86. 93. 4197 9263 3169 3389 1814 1737 4. 3. 3. 5. 5. 2. 80 32 35 81 67 04 â â â â â â 4. 2. 2. 4. 4. 1. 36 68 72 54 84 94 0.03 0.02 0.02 0.034 0.035 0.013 ­ 0.07 0.41 0.07 0.07 0.065 0.0004 0.045d 0.02 0.07 0.04 0.07 33 33 33 33 33 16 Frequency b (GHz) Synth. beam ( â ) Linear resolution (pc) v (km s-1 ) 1 rms (Jy beam-1 ) LASc ( )

a For the molecular transitions, in the text we will use the following abbreviations: NH2 D (11,1 - 10,1 ) = NH2 D (1­1); C2 H (13/2,2 - 01/2,1 ) = C2 H (1­0); c-C3 H2 (21,2 - 10,1 ) = c-C3 H2 (2­1); N2 H+ (12 - 01 ) = N2 H+ (1­0); b rest frequency of the transition listed in Col. 1; c largest angular scale (at half power) detectable by the interferometer, estimated from the minimum baseline of the array configuration, and following the appendix in Palau et al. (2010). For the lines observed in C and D configuration, this refers to the merged uv coverage; d sensitivity in the merged C+D channel maps smoothed to a spectral resolution of 0.1 km s-1 ; e observed in D configuration only; f observed in C configuration only.

calibrated using standard procedures of AIPS, and imaging was p erformed by applying a tap er in the uv -data of 50 k and using natural weighting to recover the faint and extended emission. The synthesised b eams and rms noises for all the VLA observations are listed in Table 1.

3 3.1

RESULTS Continuum emission maps

In Fig. 2 the 3 mm (CARMA, C+D configuration, solid contours) and 1.3 cm (VLA, D configuration, dashed contours) continuum emission maps are shown. As reference, we overplot the 1.3 mm continuum emission observed by Palau et al. (2007b, obtained with ab out a factor 1.5 b etter angular resolution) as well as the direction of the lob es of the 12 CO outflow originated by IRS 1 and the (rough) edge of the fan-shap ed H2 emission associated with IRS 3 (Kumar et al. 2002). The 3 mm continuum emission is resolved into 5 main condensations, which we call MMA, MMB, MMC, MMD and MME in order of increasing R.A. The brightest are MME and MMA, east and west of IRS 3, resp ectively. From Fig. 2 we note that MME roughly coincides with a 1.3 mm continuum p eak (Palau et al. 2007b), while the westernmost one, MMA, encompasses 3 p eaks seen at 1.3 mm with SMA. The faintest core, MMC, is detected towards IRS 1. Two more 3 mm condensations, MMD and MMB, are not detected at 1.3 mm. Sp ecifically, MMB is clearly detected close to IRS3, with a shift of only 4 to the northeast.

The most evident differences among the 3 and 1.3 mm continuum maps are that at 3 mm core MME (MM7 in Palau et al. 2007b) is more extended, and the eastern and western cores are connected by a filamentary emission passing through IRS 3 which is undetected at 1.3 mm. Both differences are probably just the consequence of CARMA b eing sensitive to larger structures than the SMA. We estimated that the SMA in the observations of Palau et al. (2007b) was sensitive to structures with FWHM < 9 (using the minimum baseline of the observations and following Palau et al. 2010), while CARMA using C+D configurations was sensitive to structures < 33 (Table 1), allowing CARMA to recover more extended emission.

As shown in Fig. 2, the continuum emission at 1.3 cm is dominated by one strong and compact source with its emission p eak (20h 36m 07.3s , +41 39'52") coincident with IRS 3, and matching well the fan-shap ed H2 emission detected by Kumar et al. 2002. A Gaussian fit to this compact centimeter source yields a p eak intensity of 0.8 mJy/b eam, a flux density of 1.3 mJy, and a deconvolved size of 4.3 â 1.0 , with PA=78 , corresp onding to a size of 6000 AU in the eastwest direction (and an unresolved size in the north-south direction). In addition to the compact 1.3 cm source associated with IRS 3, there is a secondary p eak at around 6 which falls 2 to the south of MMD, and faint emission joining this secondary p eak and the centimeter source in IRS 3, suggesting that the two p eaks of centimeter emission could b e linked.
c 2011 RAS, MNRAS 000, 1­17


Dense gas in IRAS 20343+4129

5

Figure 2. Continuum emission obtained with CARMA at 3 mm (solid contours) and with the VLA at 1.3 cm (dashed contours) towards I20343. The solid contours start from the 3 rms level (1.2â10-3 Jy beam-1 ) and are in steps of 1 rms. The dashed first contour and step correspond to the 3 rms level (1.5â10-5 Jy beam-1 ) and are in steps of 1 rms The grey scale indicates the 1.3 mm continuum emission observed by Palau et al. (2007b) with the SMA (levels = 10% of the maximum, first level = 20% of the maximum). The crosses indicate the two infrared sources idenfitifed by Kumar et al. (2002). Red and blue arrows highlight the direction of the blue- and red-shifted emission detected in 12 CO (2-1) by Palau et al. 2007b . The green arc roughly shows the fan-shaped feature detected in H2 µm associated with IRS 3 (Kumar et al. 2002), interpreted as an expanding cavity by Palau et al. 2007b. The ellipse at bottom-left corresponds to the CARMA synthesised beam (4.8 â 4.4 , at P.A. 92o ). The dashed ellipse at bottom-right is the VLA synthesised beam (3.71 â 3.50 , at P.A. -56.2o ).

3.2 3.2.1

Distribution of the Integrated intensity of molecular line emission Molecular tracers observed with CARMA

The maps of the integrated intensity of the lines observed and detected with CARMA (see Table 1) are shown in Fig. 3. The emission map of each tracer has b een sup erimp osed on the images obtained from the Spitzer Space Telescop e in the four mid-IR IRAC bands (centred at 3.6, 4.5, 5.8 and 8 µm, resp ectively). The location of the near-infrared sources IRS 1 and IRS 3 are also indicated, as well as the direction of the outflow centred on IRS 1 and the edge of the H2 emission associated with IRS 3, as in Fig. 2. We also sup erimp ose the 1.3 cm continuum emission detected in this work (see Fig. 2), which marks clearly the Hi i region associated with IRS 3. The molecular gas seems to b e squeezed in b etween the two dominant mid-IR sources IRS 1 and IRS 3 in all molecules except N2 H+ . Other compact mid-IR sources are present in the region but do not seem to b e associated with any clear molecular counterpart. The diffuse IR nebulosity, esp ecially evident in the 5.8 and 8 µm bands, is probably emission from small dust grains distributed around I20343 b ecoming brighter at longer wavelengths. Some of the difc 2011 RAS, MNRAS 000, 1­17

fuse emission detected at 8 µm may also b e PAHs emission (e.g. Peeters et al. 2002). The morphology of the integrated intensity of c-C3 H2 (2-1) and C2 H (1-0) delineates clearly a cavity around IRS 3 (top panels of Fig. 3), providing a strong supp ort to the hyp othesis prop osed by Palau et al. (2007b ), namely that IRS3 is most likely a more evolved intermediate-mass star creating a cavity. The C2 H emission is more extended than that of cC3 H2 , p erhaps due to the smaller sensitivity that we have in the c-C3 H2 channel maps (see Table 1). Sp ecifically, a narrow filament extended 30 is clearly detected north-east of the field center (Fig. 3, top-left panel), inclined roughly as the tail of the mid-IR cometary shap e (see Fig. 1), suggesting that the two features can have the same origin. The b ottom panels in Fig. 3 show the integrated emission of the two Nitrogen-b earing sp ecies N2 H+ and NH2 D. The emission in N2 H+ consists mainly of one cloud to the east of IRS 3 elongated in the southeast-northwest direction, and extending up to IRS 1, and two smaller clouds, one immediately to the south-west of IRS 3 with no continuum counterpart (called IRS3-SW), and the other associated with MMA. In addition, there is one clump ab out 1 to the west of IRS 3, almost at the b order of the bright rim, which falls on a region with no infrared emission associated. On the other hand, the emission of NH2 D consists mainly of one filament elongated in the east-west direction, passing through IRS 3, and with some emission at IRS3SW. The N2 H+ clump 1 to the west of IRS 3 is detected also in NH2 D, but looks more extended in NH2 D. However, this can b e just an effect of the different angular resolution and different filtering of extended emission, as NH2 D was observed with C+D configuration while N2 H+ was observed in C configuration only (see Table 1). If we put together the two mostly extended molecular tracers, namely C2 H and NH2 D, we can notice a sort of 'snake-like' filament of molecular gas (Fig. 4) extending from the south-western side of I20343, clearly detected in NH2 D, up to the north-eastern corner, in which a long and narrow filament is detected in C2 H. The bulk of the emission is in b etween IRS 1 and IRS 3. This 'snake-like' filament matches very well the 1.2 mm continuum emission detected with MAMBO (Beuther et al. 2002, Fig. 4), and its SWNE inclination follows roughly the 'head-tail' orientation of the mid-IR diffuse cometary emission (Fig. 1), suggesting a p ossible common origin. However, the highest sensitivity region of the CARMA maps, i.e. the field of view of the 10 m dishes ( 73 ), includes only the central region of the filament (see Fig. 4). For this reason, in this work we focus on the centre of I20343, where the interaction b etween the two brightest IR sources and the surrounding molecular gas seems predominant, and plan a large interferometric mosaic which will allow us to unveil the overall distribution of the molecular gas and its relation with all the IR sources. 3.2.2 Ammonia (1,1) and (2,2) inversion transitions

The integrated intensity maps of NH3 (1,1) and (2,2) are presented in Fig. 5. In the (1,1) line, the emission resembles that seen with CARMA in the 3 mm continuum. Four main p eaks are detected, which roughly corresp ond to those detected in the 3 mm continuum. On the other hand, none of the 1.3 mm continuum p eaks identified by Palau et al. (2007b)


6

F. Fontani et al.

Figure 3. Integrated maps of the molecular lines detected with CARMA (Table 1) towards I20343 and superimposed on the images obtained in the Spitzer-IRAC bands (3.6 µm, 4.5 µm, 5.8 µm, 8 µm, colour scale in MJy sr-1 units). In each panel, black contours depict the velocity-averaged emission of the transition labelled at the top-right corner (the corresponding synthesised beam is shown in the bottom-right corner). For all lines, the emission has been averaged over all the velocity channels with signal, except for N2 H+ (bottom-left panel) for which the integrated emission was averaged over the main group of hyperfine components only. For C2 H and NH2 D, first contour and step correspond to the 20% level of the maximum (corresponding roughly to the 3 rms level of the averaged map), while for c-C3 H2 and N2 H+ contours start from the 30% level of the maximum and are in steps of 20%. The white contours represent the VLA 1.3 cm (K-band) continuum image (same contours as in Fig. 2). The position of IRS 1 and IRS 3, the associated outflow and fan-shaped H2 emission are shown as in Fig. 2.

exactly coincides with any of the NH3 (1,1) p eaks, despite the similar angular resolution. The NH3 (1,1) emission reveals two main clouds, one to the east and the other to the west of IRS 3. The eastern cloud includes MMD and MME and app ears elongated in the southeast-northwest direction, similar to the N2 H+ eastern cloud and to the 3 mm continuum emission. The western cloud has a condensation near MMB (near IRS 3) and another condensation near MMA. The overall emission in the western cloud is elongated in the east-west direction. We stress that NH3 (1,1) emission is marginally detected also towards IRS 3 and the cavity driven by it, as for the 3 mm continuum. On the other hand, the millimeter continuum sources associated with IRS 1 as well as the 12 CO outflow lob es are not detected in ammonia. The NH3 (2,2) transition is clearly detected towards the eastern side of MME, where 2 p eaks are resolved. The NH3 (2,2) line emission around IRS 3 and towards the west is clumpy, p eaking towards MMA and near MMB and IRS3-SW.

From a comparison of the N-b earing to the C-b earing sp ecies, b oth C2 H and c-C3 H2 highlight clearly the cavity associated with IRS 3 (see top panels in Fig. 3), while NH2 D and NH3 trace emission extending east-west passing through IRS 3 (see Fig. 5 and b ottom panels of Fig. 3) more similar to that seen in the 3 mm continuum map. Thus, there seems to b e a chemical dichotomy in I20343 among Carb on- and Nitrogen-b earing molecules. The exception is represented by N2 H+ , detected away from IRS 3. This molecule app ears to trace the part of the cloud less disrupted by the expanding cavity. The sp ectra of the NH3 (1,1) and NH3 (2,2) emission, integrated within the 5 contour p olygon of the 3 mm continuum image (except for MMB and MMD, where we used the 3 contour) are shown in Fig. 6. For IRS3-SW, we used the 5 rms contour of the NH2 D (1­1) line integrated emission (Fig. 3, b ottom-right panel). To compare the different molecular sp ecies, in Fig. 6 we also show the integrated sp ectra of
c 2011 RAS, MNRAS 000, 1­17


Dense gas in IRAS 20343+4129

7

Figure 4. Integrated emission of molecular tracers (white contours) superimposed on the image of I20343 in MAMBO at 1.2 mm (Beuther et al. 2002). The solid white contour depicts the 15% level of the zero-order moment map of C2 H (1­0). The dashed white contour corresponds to the 15% level of the zeroorder moment map for the NH2 D (1­1) line. The ellipse in the bottom-left corner shows the CARMA synthesised beam for the C2 H map and it is almost coincident to that of the NH2 D map. The grey dotted-dashed circles represent the primary beams of the CARMA 10 and 6 m dishes (73 and 121 ), respectively.

NH2 D and N2 H+ extracted using the same p olygons. We do not show the sp ectra towards MMC b ecause this core is undetected in NH2 D and NH3 (2,2), and marginally detected in the other lines. Among all the NH3 (1,1) sp ectra, MMA shows the strongest emission, and MME shows the broadest lines, of up to 2.2 km s-1 . Such a large line broadening in MME could b e due to a double velocity comp onent, as suggested by the N2 H+ sp ectrum which has 10 times b etter sp ectral resolution. Concerning MMB, MMD, and IRS3-SW NH3 (1,1) sp ectra, it is striking the anomaly seen in the inner satellite hyp erfine lines, with one inner satellite clearly detected ab ove 5 and the other inner satellite remaining undetected. The anomaly for the non-LTE case due to hyp erfine selective photon trapping affects only the outer satellites (red stronger than blue, Stutzki & Winnewisser 1985), allowing us to rule out this p ossibility in I20343. Rather, anomalies of one inner satellite b eing stronger than the other have b een observed in several works (Lee et al. 2002, Longmore et al. 2007, Purcell et al. 2009) and explained as b eing due to systematic motions, following the theoretical work of Park (2001). Park (2001) shows that if the core is contracting the inner blue satellite should b e stronger than the inner red satellite for a systematic motion in the range of 0.4­1 km s-1 , and for a range of H2 numb er densities and NH3 column densities which are consistent with those derived by us (as we will show in Sect. 3.4). This is the case of MMD and IRS3-SW. On the other hand, if the core is undergoing expanding motions, the prediction is that the inner red satellite should b e stronger than the inner blue satellite, as seen for the case of MMB. Thus, it seems that for these three clumps the NH3 (1,1) anomalous intensity of the hyp erfine
c 2011 RAS, MNRAS 000, 1­17

comp onents is consistent with contracting/expanding motions. The MMA NH3 (1,1) sp ectrum is a very sp ecial case, as the inner satellites are detected at an intensity smaller than that exp ected for LTE conditions (maximum main-tosatellite ratio in LTE is 3.6, while the ratio for MMA is 4.3). This could b e explained if the opacity is high and the excitation temp erature for the main line and the satellites is different, with the satellites having a lower excitation temp erature (and hence a smaller main b eam temp erature). A detailed discussion on the temp erature ratio b etween main line and satellites leads to a p ossible non isothermal core made of two layers, with the external one b eing hotter than the inner one. If this interpretation is correct, core MMA could b e heated externally, p erhaps by IRS 3 and/or the infrared sources west of I20343 (see right panel in Fig. 1 and Fig. 3). We give details of this explanation in App endix A. We fitted the NH3 , NH2 D, and N2 H+ sp ectra in order to derive the physical parameters of the gas traced by these molecules. To take into account the line hyp erfine structure, we followed the method describ ed in the CLASS user manual4 . The NH3 (1,1), NH2 D (1­1) and N2 H+ (1­0) lines were fitted this way, while we fitted the NH3 (2,2) lines with Gaussians. The derived fit parameters are rep orted in Table 2, except the line velocities that will b e extensively discussed in Sect. 3.3. These parameters have b een used to derive the molecular column densities, the derivation of which will b e describ ed in Sect. 4.2. To fit the hyp erfine structure of MMB, MMD, and IRS3-SW, we used only the detected satellite and the main line, in order to obtain reliable opacities. For NH3 , in Table 2 we list also the rotation temp erature, Trot , obtained from the (2,2)-to-(1,1) intensity ratio following the method outlined in Busquet et al. (2009), which is based on the discussion presented in Ho & Townes (1983). These range from 13 K (in MMD) to 23 K (in IRS3-SW). In order to obtain a first approach to temp erature variations of the dense gas across I20343, we computed the ratio of the integrated NH3 (2,2)/(1,1), which has b een shown to b e a reasonable approach to the kinetic temp erature (e.g., Torrelles et al. 1993; Zhang et al. 2002). The result is shown in Fig. 7. As can b e seen from the figure, the ratio is largest towards three main regions: near IRS 3 (MMB, and IRS 3SW), to the north of MMA, and to the eastern side of MME. Thus, the sp ectra toward MMA, which can b e explained with external heating, is consistent with the 22/11 ratio map which reveals that the heating comes most likely from the north of MMA. Interestingly, the north of MMA is spatially coinciding with the redshifted CO outflow lob e driven by IRS 1, suggesting that the outflow and the dense gas are interacting (Fig 7). Most intriguing is the high ratio seen at the eastern side of MME, which extends all along the northsouth direction and is seen westwards of the H2 extended emission rep orted by Kumar et al. (2002). Such a spatial coincidence is suggestive of a p ossible relation b etween the high 22/11 ratio and the H2 emission. Finally, the high 22/11 ratio near IRS 3 could b e indicative of direct heating by the UC Hi i region associated with IRS 3.

4

The CLASS program is part of the GILDAS software, developed at the IRAM and the Observatoire de Grenoble, and is available at http://www.iram.fr/IRAMFR/GILDAS


8

F. Fontani et al.
Table 2. NH3 , N2 H+ , and NH2 D line parameters for the 3 mm continuum cores (except MMC) and the molecular core IRS3-SW. The parameters have been derived from the spectra shown in Fig. 6 using the fitting procedure described in Sect. 3.2.2. Core Tex (K) ­b 9.7 21.5 19.8 7.9 NH v (km s-1 ) 1.52 1.03 1.21 2.2 1.03
3

> 3. 1. 1. 3.

a

MMA MMB MMD MME IRS3-SW
a b

1 5 4 5 2

Trot (K) 18(1) 20(7) 13(2) 18(3) 23(7)

Tex (K) 7 ­ 14 6 13

N2 H+ v (km s-1 ) 1.63 ­ 0.9 0.7 0.7



a

4.6 ­ 5.2 10 3.3

Tex (K) 17 20 41 26 23

NH2 D v km s-1 1.05 1.03 0.89 1.01 0.72



a

2.8 2.5 0.38 3.2 1.8

= Total opacity of the line; = in this core the excitation temperature for the main and the satellite lines of the NH3 (1,1) transition is probably different. See Appendix A for the detailed discussion on the derivation of the physical parameters from NH3 .

Figure 6. Spectra MMC, undetected the 5 rms contour MMD; the 5 rms

of the detected transitions of NH3 in all the lines) and the molecular of the 3 mm continuum image for contour of the integrated emission

, NH2 D and N2 H+ at the position of the 3 mm continuum condensations (except condensation IRS3-SW. The polygon used to extract the spectra corresponds to: MMA and MME; the 3 rms contour of the 3 mm continuum image for MMB and of NH2 D (1­1) for IRS3-SW.

c 2011 RAS, MNRAS 000, 1­17


Dense gas in IRAS 20343+4129

9

Figure 7. Colorscale: ratio of the integrated NH3 emission (2,2)/(1,1). Red triangles indicate the 3 mm continuum peaks, which correspond to (from east to west) MME, MMD, MMC, MMB and MMA. The black-dashed contours indicate the VLA 1.3 cm emission (2.5, 5, 7.5, 12, 15 times 50 µJy beam-1 ). The blue/red solid contours depict the high-velocity CO (2­1) blueshifted/redshifted emission (Palau et al. 2007b). Blue contours range from 10 to 99% in steps of 12% of the peak intensity (21.335 Jy beam-1 km s-1 ); red contours are 6, 12 to 99% in steps of 12% of the peak intensity (41.781 Jy beam-1 km s-1 ).

following the method describ ed in Sect. 3.2.2 to take into account their hyp erfine structure. 3.3.1 Line peak velocities

Figure 5. Top panel: integrated emission map of the ammonia (1,1) inversion transition (contours) observed with the VLA towards I20343, superimposed on the 3 mm continuum (CARMA, grey-scale). First contour level and step are the 10% of the peak (224 Jy beam-1 m s-1 ). The ellipse at bottom-left represents the VLA synthesised beam (4. 20 â 3. 19, at P.A. 11 ). Purple triangles pinpoint the 3 mm continuum peaks (see Fig. 2). All other symbols (crosses, arrows and semiellipse) are the same as in Fig. 2. Bottom panel: same as top panel for the (2,2) inversion transition. First contour level and step are the 15% of the peak (57 Jy beam-1 m s-1 ). The synthesised beam in this case is 4. 25 â 3. 19, at P.A. 13 .

Fig. 8 shows maps obtained from the line p eak velocities. In all tracers the radial velocities are predominantly blueshifted to the west and red-shifted to the north-east. The east-west velocity gradient is not uniform and suggests a p ossible torsion of the gas. Interestingly, the inclination of this gradient with resp ect to the line of sight is opp osite to that of the outflow associated with IRS 1, since the blue lob e of the outflow is located on the side where the dense gas is red-shifted, and vice-versa. In general, all tracers with emission near IRS 3 (in MMB) show that the gas is blueshifted at this p osition. 3.3.2 Line widths

3.3

Kinematics

To insp ect the velocity field, we have extracted from the interferometric channel maps sp ectra of the molecular transitions detected on grids with regular spacings (1.5 â 1.5 for C2 H, 2.5 â 2.5 for c-C3 H2 , 1 â 1 for N2 H+ , 1.6 â 1.5 for NH2 D and NH3 ). The sp ectra extracted have b een fitted following different methods: for C2 H (1­0), c-C3 H2 (2­1) and NH3 (2,2) we assumed Gaussian lines, while for N2 H+ (1­0), NH2 D (1­1) and NH3 (1,1) we fitted the lines
c 2011 RAS, MNRAS 000, 1­17

Maps of the line widths are presented in Fig. 9. The measured line broadenings are generally a factor > 3 higher than the thermal broadening, exp ected to b e of the order of 0.1 - 0.3 km s-1 , indicating that the gas kinematics is largely dominated by non-thermal motions. This finding confirms previous studies in similar intermediate- to high-mass protoclusters (e.g. Palau et al. 2007a, Fontani et al. 2009, Busquet et al. 2010), and represents one of the most imp ortant differences b etween these dense cores and those observed in low-mass star forming regions, where the line widths are dominated by thermal broadening, even in clustered environments (e.g. Kirk et al. 2007, Walsh et al. 2007, Bourke et al. 2011). In summary we highlight three regions where the gas is more turbulent (Fig. 9): (i) around IRS 3, esp ecially in b etween IRS 1 and IRS 3 (see the C2 H and c-C3 H2 line widths in Fig. 9); (ii) north of MMA, near the red lob e; (iii) towards MMA. The turbulence enhancement around IRS 3 is easily explained by the expading cavity, while in


10

F. Fontani et al.

MMA and north of it could rather b e due to the red-lob e of the outflow. On the other hand, we stress that towards MME and in b etween IRS 1 and IRS 3 some sp ectra show a double-velocity comp onent (see e.g. the isolated hyp erfine comp onent in the N2 H+ (1­0) line towards MME in Fig. 6). These are esp ecially evident in the N2 H+ , NH3 and NH2 D sp ectra, but the second comp onent can b e seen also in some sp ectra of C2 H and c-C3 H2 close to the cavity. Therefore the large broadening ab ove IRS 3 and in MME could just b e due to the sup erp osition of two unresolved velocity comp onents. Interestingly, for C2 H and N2 H+ the line broadening is relatively small close to the blue lob e of the outflow associated with IRS 1 and larger b esides the red lob e, while the opp osite seems to occur for the c-C3 H2 (2­1) line (top-middle panel in Fig. 9). This could indicate a selective influence of the flow on the different molecules in the surrounding material. However, we stress that the c-C3 H2 emission is very faint at the b orders of the region plotted in Fig. 9, where the fit results are affected by large relative errors. 3.3.3 Position-velocity plots

In order to further study the velocity field of the NH3 emission near IRS 3, we computed p osition-velocity plots for the NH3 (1,1) line in the east-west direction, and centred at offset 0 ,-2 with resp ect to the CARMA phase centre. The final plot is shown in panel "a" of Fig. 10. The NH3 (1,1) emission in the p osition-velocity plot shows two main p eaks, one corresp onding to the eastern cloud and the other corresp onding to the western cloud, and b oth p eaks are linked through fainter emission which overall shows a U-like structure. Such a feature resembles the shap e predicted by the model of Arce et al. (2011) for an expanding bubble. In this model, an expanding shell would app ear as a ring in the (pv) plots (see their Fig. 5), while we only see the blue-shifted half of it. However, a U-like feature can b e explained if the source driving the bubble is slightly displaced (b ehind the bubble) with resp ect to the surrounding molecular gas, so that we mainly see the gas which is moving towards us. In fact, the tails of the U-like feature are found, as exp ected, at approximately the systemic velocity. This suggests that IRS 3 may b e pushing the surrounding dense material away (either through its winds/radiation or through the associated Hi i region), with an expansion velocity (difference b etween the 'tip' of the U-like feature and the 'tails') of ab out 2 km s-1 (see panel "a" in Fig. 10). A similar expanding shell was recently found around an infrared-source at the centre of a region devoid of gas emission in the intermediate- to highmass protocluster IRAS 05345+3157 (Fontani et al. 2012). 3.4 3.4.1 Physical parameters from 3 mm and 1.3 cm continuum emission 3 mm

Figure 10. a) NH3 (1,1) position-velocity plot centred at offset (0,-2 ) with respect to the CARMA phase centre and along a cut in the east-west direction. The NH3 (1,1) data cube was convolved with a beam of 5 â2 and PA=0 to increase signal-to-noise. Contours are: -3, 3, 6, 9, 15, 21, and 27 times 0.001 Jy beam-1 . b) NH3 (1,1) integrated emission map (same contours as in Fig. 5) with the cut of the position-velocity plot of panel "a" indicated by a dashed line. Contours start at 12% of the peak intensity, 224 Jy beam-1 m s-1 , and increase in steps of 10%. The VLA synthesised beam is shown in the bottom-right corner.

and ma jor axes of the CARMA synthesised b eam (see Table 1). Because the contours at half maximum were blended, at the edge b etween two cores we decided to separate the emission arising from the different condensations identifying the p eaks and considering the first unblended contour. The same criterion was applied to derive the integrated flux density, F , given in Col. 5 of Table 3. From F , we have computed the mass of the condensations under the assumptions that the dust millimetercontinuum emission is optically thin, and that the dust temp erature equals the gas kinetic temp erature. This latter hyp othesis implies coupling b etween gas and dust, which is a reasonable assumption for H2 volume densities ab ove 105 cm-3 . Under these assumptions, the gas mass can b e derived from the formula: M= F d2 , B (T )k (1)

We have identified 5 main condensations in the 3 mm continuum emission map (see Sect. 3.1). Their p eak p ositions are given in Cols. 2 and 3 of Table 3. We also list the angular () and linear (D) diameters (Cols. 6 and 7 in Table 3) computed assuming the sources are Gaussians, and deconvolving the contour at half maximum with a Gaussian b eam with HPBW corresp onding to the geometric mean of the minor

where d is the source distance, B (T ) is the Planck function at dust temp erature T , and k is the dust opacity p er unit dust mass. For this latter, we extrap olated the value at 230 GHz given by Kramer et al. (1998), k230 = 0.005 cm2 g-1 (which assumes a gas-to-dust ratio of 100), through the p ower-law k = k230 [ (GHz)/230] . We have assumed = 2, which is a typical value derived for dusty envelop es of
c 2011 RAS, MNRAS 000, 1­17


Dense gas in IRAS 20343+4129

11

Figure 8. Maps of the line and N2 H+ (1­0). Contours for N2 H+ . Bottom panels range from: 9.5 to 13.5 km in the bottom-right corner (Sect. 3.1). The position of as in Fig. 2.

peak velocity for all detected molecular lines. Top panels show (from left to right): C2 H (1­0), c-C3 H2 (2­1) are in steps of 1 km s-1 , and range from: 8 to 11 km s-1 for C2 H; 10.5 to 13.5 for c-C3 H2 ; 9 to 12 km s-1 show (from left to right): NH3 (1,1) NH3 (2,2) and NH2 D (1­1). Contours are in steps of 1 km s-1 , and s-1 for NH3 (1,1); 9 to 13 km s-1 for NH3 (2,2); 8.5 to 12.5 km s-1 for NH2 D. In each panel, the ellipse represents the CARMA or VLA synthesised beam. The black triangles pinpoint the 3 mm continuum peaks IRS 1 and IRS 3, as well as the associated outflow and fan-shaped H2 emission, are depicted in each panel

Figure 9. Same as Fig. 8 for the line widths. Contours are in steps of 0.5 km s-1 and range from: 0.5 to 3 km s-1 for C2 H; 0.6 to 3.1 km s-1 for c-C3 H2 ; 0.4 to 2.4 km s-1 for N2 H+ ; 0.6 to 2.6 km s-1 for NH3 (1,1); 0.8 to 3.2 km s-1 for NH3 (2,2) ; 0.3 to 1.3 km s-1 for NH2 D (1­1).

c 2011 RAS, MNRAS 000, 1­17


12

F. Fontani et al.
index is consistent with optically thin free-free emission favoring the interpretation that the centimetre emission comes from an Hi i region rather than a thermal radio jet (which typically have steep er sp ectral indices). We calculated the physical parameters of the ionised region at 1.3 cm assuming the emission is optically thin, and obtained an emission measure of 3.4 â 104 cm-6 p c, characteristic of UC Hi i regions, and a flux of ionising photons of 2 â 1044 s-1 , consistent with the Hi i region b eing ionised by an early-typ e B2 star (Panagia 1973). Interestingly, the extension of such an UC Hi i region towards the east is similar (although larger in size) to the extension seen in [NeI I] emission by Campb ell et al. (2008), and which is interpreted as due to the expansion of the ionised gas and disruption of the natal envelop e. In this context, the secondary centimetre p eak near MMD could b e related as well to the expanding ionised gas. We estimated the p ossible contribution of free-free emission to the flux measured at 3 mm for MMB, and is of 0.3 mJy (4 of 1.3 cm observations, and using the sp ectral index of 0.1), out of 4.6 mJy, or 6%. Thus, thermal dust emission is the main contribution to the 3 mm continuum emission in MMB. Finally, we estimated an upp er limit for the 3 mm emission associated with IRS 3 of 2.3 mJy, measured as the 3 mm flux density inside the 4 contour of the centimeter emission (Fig. 11), and we cannot rule out the p ossibility of the UC Hi i region b eing still associated with remnant natal dust, although a pro jection effect could b e also p ossible.

10

IRAS 20343+4129 IRS3
emission measure: 3.4 10 cm pc 3 -3 electron density: 1.0 10 cm ionizing photons: 2.0 10 s (B2 ZAMS) spectral index: 0.1+/-0.2
44 -1 4 -6

Flux density (mJy) 1

1.3 cm

7 mm

3 mm

10 Frequency (GHz)

100

Figure 11. Spectral energy distribution of IRS 3 in the cm/mm range. The physical parameters have been derived assuming the emission at 1.3 cm comes from an optically thin Hii region.

massive (proto)stellar ob jects (e.g. Molinari et al. 2000, Hill et al. 2006). As gas temp erature, we have taken the kinetic temp erature obtained by extrap olating the rotation temp erature derived from the ammonia (2, 2)/(1, 1) line ratio for each core (see Sect. 3.2.2 and Table 2) following the empirical approximation method describ ed in Tafalla et al. (2004). The kinetic temp eratures derived this way are listed in Col. 4 of Table 3, and are in b etween 14 and 25 K, higher (a factor 1.5 ­ 2) than the values measured typically in low-mass clustered starless cores (e.g. Andr´ et al. 2007, Foster et al. 2009) e which are around 10­13 K. For MMC, undetected in NH3 (2,2), we decided to give a range of masses computed in the temp erature interval 15­30 K. The resulting masses are listed in Col. 8 of Table 3. All fragments have masses consistent with intermediate-to high-mass emb edded ob jects. The most massive one is MME (23 M ). By assuming spherical and homogeneous cores, we have derived the average H2 volume and column densities. The average volume and column densities (given in Cols. 9 and 10 of Table 3) are of the order of 106-7 cm-3 and 1023-24 cm-2 . Such high column densities are consistent with b eing the birthplaces of intermediate- and/or high-mass objects (e.g. Krumholz & McKee 2008). 3.4.2 1.3 cm

4 4.1

DISCUSSION Column densities of the PDR tracers C2 H and c-C3 H2

In Sect. 3.1 we showed that the centimeter emission is dominated by one single source associated with IRS 3 and extending towards the east. In addition, a secondary source near MMD was also identified. A 3.6 and 6 cm source associated with IRS 3 is already rep orted by Miralles et al. (1994) and Carral et al. (1999). However, from these two measurements only, and taking into account the uncertainties, the sp ectral index of the centimeter source associated with IRS 3 could not b e well determined (e.g., Palau et al. 2007b). Our new measurement at 1.3 cm allows to b etter constrain the sp ectral index of the source to 0.1 ± 0.2 (Fig. 11). Such a sp ectral

The two carb on-b earing sp ecies C2 H and c-C3 H2 are among the most abundant simple carb on-chain molecules detected in the interstellar medium, and are b elieved to b e good tracers of PDRs (Lucas & Liszt 2000, Pety et al. 2005, Gerin et al. 2011). C2 H is formed either from photodissociation of acetylene (C2 H2 ) followed by dissociative recombination of C2 H+ (Mul & McGowan 1980) or through neutral-neutral 2 reaction b etween C and CH2 in hot gas (Sakai et al. 2010). cC3 H2 is b elieved to b e formed by dissociative recombination of c-C3 H+ . Both sp ecies b enefit from the presence of atomic 3 carb on not locked in CO, and a good correlation b etween the two tracers has b een found at the illuminated surface of the Horsehead nebula (Pety et al. 2005, Gerin et al. 2009), as well as in b oth diffuse and translucent clouds (Lucas & Liszt 2000, Gerin et al. 2011). We have investigated the relation among the two sp ecies in I20343. For this purp ose, we have extracted the sp ectra of C2 H (1­0) and c-C3 H2 (2­1) on a grid of spacing 2.5 â 2.5 (roughly half of the CARMA synthesised b eam at the frequency of the c-C3 H2 (2­1) transition), and fitted the sp ectra with Gaussian lines. Then, from the integrated intensity obtained from the fits, we have computed the column densities assuming that b oth lines are optically thin. This assumption is mandatory b ecause the opacity of the lines cannot b e directly measured (for the C2 H (1­0) line we observed only the main hyp erfine comp onent, and we do not have isotop ologues for c-C3 H2 ). However, given that the line profiles generally do not show effects due to high optical
c 2011 RAS, MNRAS 000, 1­17


Dense gas in IRAS 20343+4129
Table 3. Peak position, angular and linear diameter, integrated flux density, mass, 3 millimeter condensations detected with CARMA. The masses are computed for derived from ammonia. The H2 volume and column densities are calculated assuming to the deconvolved level at half maximum. F has been obtained by integrating the contour level. Core Peak position R.A.(J2000) Dec.(J2000) 20h 36m 06s .37 +41 39 59. 1 07s .54 +41 39 55. 6 s .72 40 07. 2 07 +41 08s .23 +41 40 02. 0 s .26 39 54. 7 08 +41 Tk
a

13

H2 volume and column density of the = 2, and assuming the temperatures a spherical source with diameter equal continuum flux density inside the 3

F



MMA MMB MMC MMD MME
a b c d

(K) 22b 25 ­ 14 22

(mJy) 7.4 4.6 1.9 3.7 8.3

Diameter s D ( ) (pc) 7.3 0.049 6.8 0.046 5.5 0.037 6.3 0.043 8.0 0.054

M

cont

n

H2

N (H2 ) (â1023 cm-2 ) 10.3 6.4 7.1c ,3.3d 11.4 9.6

(M ) 21 11 8c ,4d 17 23

(â106 cm-3 ) 6.7 4.5 6.2c ,2.9d 8.7 5.7

= = = =

Derived from Trot as explained in Sect. 3.4; see Appendix A for the derivation of Tk from ammonia for this core; assuming a temperature of 15 K; assuming a temperature of 30 K.

depths, we are confident that the assumption is reasonable. We used the general formula for optically thin transitions (compare to, e.g., Eq. A.3 in Pillai et al. 2007): Nt 3h Q(Tex ) W = 8 3 S µ2 J (Tex ) - J (T e
BG
EJ kTex

ot

) eh

/kTex

-1

(2)

where: EJ and S are energy of the upp er level and line strength, resp ectively, W is the integrated intensity of the line, Q(Tex ) is the partition function at the temp erature Tex , the line rest frequency, J (Tex ) and J (TBG ) are the equivalent Rayleigh-Jeans temp erature at frequency computed for the excitation and background temp erature (TBG 2.7 K ), resp ectively; µ the molecule's dip ole moment (0.77 Debye for C2 H and 3.27 Debye for c-C3 H2 ). For C2 H, in Eq. (2) W has b een obtained by multiplying the integrated emission of the hyerfine comp onent observed for its relative intensity (0.416). As excitation temp erature, we have assumed a reasonable value of 20 K based on the excitation temp eratures computed for the other lines (see Table 2). The values of EJ , S µ2 and Q have b een taken from the Cologne Database for Molecular Sp ectroscopy (CDMS5 , Muller et al. 2001). For this latter, we have extrap olated the ¨ values tabulated to an excitation temp erature of 20 K. For c-C3 H2 , the ortho-/para- ratio is included in the partition function. The results are shown in Fig. 12. The column density of C2 H is of the order of 1014 cm-2 across the cloud (Fig. 12, top panel), while that of c-C3 H2 is of the order of 1012 cm-2 (Fig. 12, middle panel). The C2 H column densities are generally larger than those found by Gerin et al. (2011), who measured column densities of 1013 cm-2 , while those of c-C3 H2 are more consistent. Their ratio is on average of the order of 200 - 400 (Fig. 12, b ottom panel), i.e. one order of magnitude larger than the value 20 - 30 measured in translucent clouds (Gerin et al. 2011), as well as in diffuse high latitude clouds (Lucas & Liszt 2000) and in the Horsehead nebula (Pety et al. 2005). On the other hand, the chemical models of PDRs in Gerin et al. (2011) seem to b e more consistent with our observational results rather than with theirs, b e-

cause the models predict total column densities consistent with our values for b oth sp ecies, and ratios of the order of 100 or even more (see their Table 5). Interestingly, we find the largest ratios (around 400­800) close to the outflow lob es and to the east and west of the cavity walls, where the gas is probably less dense b ecause most disrupted. Significant enhancement can b e noticed also in the eastern clump, in b etween MMD and MME, where 1.3 cm emission is detected (see Sect. 3.1 and Fig. 2). This would confirm strongly that b oth molecules are produced in PDR regions, and that they are mayb e tracing a low density envelop e in which the dense cores detected in N2 H+ and NH3 are emb edded.

4.2

Chemical differentiation and nature of the 3 mm continuum cores

5

http://www.astro.uni-koeln.de/cdms/

Studies of intermediate- and high-mass star-forming regions suggest that the relative abundance ratio NH3 -to-N2 H+ is an evolutionary indicator for dense cores (e.g. Palau et al. 2007a, Fontani et al. 2012): cores with no signs of star formation typically have larger NH3 -to-N2 H+ column density ratio than cores associated with active star formation. Also, Fontani et al. (2008) and Busquet et al. (2010) have measured that the deuterated fraction (i.e. the column density ratio of a deuterated sp ecied to that of the hydrogenated counterpart) of NH3 and N2 H+ is of the order of 0.1 in pre­protostellar core candidates, as high as in low-mass pre­ stellar cores, while it is lower in more evolved ob jects. With this in mind, we have investigated the evolutionary stage of the millimeter cores in I20343 based on the column density ratios NH3 -to-N2 H+ and NH2 D-to-NH3 . From the line parameters derived in Sect. 3.2.2 (see Table 2) we have computed the total column densities of NH3 , NH2 D and N2 H+ from Eq. (A1) of Caselli et al. (2002) b ecause all lines are optically thick (except NH2 D (1­1) in MMD, but the opacity is well-constrained). As for the other parameters, for a detailed discussion on the NH3 data of MMA see the App endix A. In Table 4 we rep ort the column densities of NH3 , NH2 D and N2 H+ , and the column density ratios NH2 D-to-NH3 and NH3 -to-N2 H+ . The NH2 D-to-NH3 ratio is of order 0.1 (from 0.07 to 0.15), and does not change greatly from

c 2011 RAS, MNRAS 000, 1­17


14

F. Fontani et al.
+

Table 4. Total column densities of NH3 , NH2 D and N2 H IRS3-SW. core MMA MMB MMD MME IRS3-SW N (NH3 ) (â1015 cm-2 ) > 2.416 2.3(0.6) 3.1(0.7) 4.4(0.7) 1.8(0.4) N (N2 H+ ) (â1013 cm-2 ) 7(2) ­ 13(4) 5.5(0.7) 6(3)

for the 3 mm continuum cores, except MMC, and the additional core

N (NH2 D) (â1014 cm-2 ) 2. 2. 2. 6. 2. 2(0. 6(0. 1(0. 8(0. 1(0. 2) 3) 2) 7) 2)

N (NH2 D)-to-N (NH3 ) < 0.09 0.11(0.04) 0.068(0.002) 0.15(0.04) 0.12(0.04)

N (NH3 )-to-N (N2 H+ ) > 34 ­ 24(13) 80(23) 30(22)

associated with intermediate- to high-mass star-forming regions (e.g. Busquet et al. 2010, Pillai et al. 2011). Interestingly, relatively high values are found also close to the UC Hi i region associated with IRS 3, in MMB and IRS3-SW. Because NH2 D is efficiently formed on dust grains, a strong UV radiation can heat the dust and cause NH2 D evap oration thus increasing its abundance. On the other hand, a strong UV radiation could (at least partly) decrease the abundance of NH3 , due to its interaction with H+ to form NH+ (e.g. 3 Fuente et al. 1993). Concerning the NH3 -to-N2 H+ ratio, we find the largest value in MME ( 80). The enhancement of the NH3 -toN2 H+ ratio can b e understood when freeze-out of sp ecies heavier than He b ecomes imp ortant (see e.g. Flower et al. 2006), so that it is exp ected to increase when the starless core gets closer to the onset of star formation. In this scenario, the fact that MME has the largest NH3 -to-N2 H+ ratio suggests that this core could b e close to the onset of gravitational collapse, i.e. MME could b e a candidate massive pre­stellar core. However, Palau et al. (2007b) measured with the SMA a mass of only 0.7 M from the 1.3 mm continuum, while we find 23 M . This discrepancy likely comes from extended flux filtered out by the SMA, which means that the core is quite flat and not centrally-p eaked as exp ected for a pre­stellar core. Based on the results of this work, we prop ose our final interpretation for the nature of each one of the 3 mm condensations: · MMA is probably a protostar candidate. Although it does not show any emb edded infrared source, its relatively high Trot , large line broadening, and NH2 D-to-NH3 lower than in other cores suggest that this condensation is evolved. · MME is likely a pre­stellar core, b ecause it shows high NH2 D-to-NH3 and NH3 -to-N2 H+ ratios, is more quiescent than MMA and it does not app ear fragmented into smaller condensations when observed at higher angular resolution (Palau et al. 2007b). Assuming a typical star formation efficiency of 30%, the core, the mass of which is 23 M , has the p otential to form an intermediate- to high-mass ob ject. · the nature of MMB, MMC and MMD is less clear. Due to the low NH2 D-to-NH3 and NH3 -to-N2 H+ ratios, MMD could b e a protostellar ob ject, consistent with clear hints of contraction motions seen in the NH3 (1,1) sp ectrum, while for MMB we found hints of expansion due to asymmetric emission in the two inner satellites (see Sect. 3.2.2). Certainly, all condensations are p erturb ed (MMB by the ionisation front from IRS 3, MMC by IRS 1 and the outflow associated with it, MMD p erhaps by a combination of b oth).
c 2011 RAS, MNRAS 000, 1­17

Figure 12. From top to bottom: Column density map of C2 H and c-C3 H2 , and their ratio, derived from the CARMA observations. For the column density derivation of both species, an excitation temperature of 20 K has been adopted. The other symbols are the same as in Figs. 8 and Fig. 9.

core to core. Such values are much larger than the cosmic D/H ratio, estimated to b e 10-5 (Linsky et al. 2006), and comparable to those measured towards low-mass pre­stellar cores (Roueff et al. 2005) and infrared dark clouds (Pillai et al. 2007). This implies that the deuteration in the cores of I20343 is as high as in colder and more quiescent environments, despite the relatively higher gas temp erature and turbulence, and confirms previous findings in other dense cores


Dense gas in IRAS 20343+4129
Only higher sensitivity and angular resolution observations will allow to b etter understand the nature of these cores.

15

4.3

Interaction of IRS 1 and IRS 3 with the dense gas: an expanding cavity

The most striking result of this work is the clear confirmation of a cavity op ened by IRS 3 in the molecular surrounding gas, and a tight interaction b etween this cavity and the surrounding dense gas. We have found several evidences of this: (i) the morphology of all the molecular tracers, esp ecially in C2 H and c-C3 H2 , delineates a cavity around IRS 3 and the 1.3 cm continuum map resolving the ionised gas p erfectly matches the profile of the cavity; (ii) the NH3 integrated intensity (2,2)/(1,1) ratios are large near IRS 3; (iii) the line widths are also large near IRS 3, sp ecially in C2 H and c-C3 H2 ; (iv) the p osition-velocity plot of NH3 shows a U-structure typical of an expanding shell; (v) in the MMB core we found hints of expansion in the NH3 (1,1) sp ectrum due to different intensity of the two inner satellites. These evidences of such an interaction are shown for the first time in this work. If we put together all the results obtained, we sp eculate ab out a p ossible scenario that describ es the star formation history in I20343: IRS 1 and IRS 3, b oth having b olometric luminosities of ab out 1000 L , seem to come from the same natal cloud while b eing clearly in different evolutionary stages, which p oints towards different generations of (intermediate- to high-mass) star formation in I20343. In this context, IRS 3 could have induced the formation of IRS 1, as could b e inducing star formation on the west (in MMA). On the other hand, in this bright-rimmed cloud the star formation probably has not b een triggered by the UV radiation from the Cygnus OB2 association stars, b ecause IRS 3, the massive star that formed first, is relatively distant from the bright rim, and the dense gas where we find the bulk of the current star formation activity is around IRS 3 and away from the bright rim. Therefore, the star formation seems to b e dominated by IRS 3, which has b een caught in the act of pushing away and disrupting its natal cloud.

· We confirm the presence of an expanding cavity driven by IRS 3 demonstrated mainly by the shap e of the emission in the two PDR tracers C2 H and c-C3 H2 , as well as by hints of expanding motions from b oth the p osition-velocity diagrams and the asymmetric intensity of the two inner satellites of the NH3 (1,1) line of the millimetre core closest to IRS 3 (MMB). · The non-thermal line widths across the filament indicate that the gas kinematics is dominated by turbulence, similarly to other intermediate- and high-mass star-forming regions and different from low-mass dense starless cores. · The rotation and kinetic temp eratures derived from ammonia are on average larger than those typically found in cores associated with low-mass star-forming regions, esp ecially around the cavity walls. The most massive and extended millimeter core, MME, shows physical and chemical signatures of an intermediate- to high-mass pre­stellar core candidate. · We have b etter constrained the sp ectral index of the radio-continuum emission associated with IRS 3, which turns out to b e flat, and thus the ionised gas comes from a region photoionised by the B2 ZAMS star at the centre of IRS 3. · the column density ratio C2 H/c-C3 H2 is of the order of 200-400 across the source and is higher where the dense gas is getting disrupted. · the deuterated fraction NH2 D-to-NH3 is of the order of 0.1 in all cores, as large as in low-mass pre­stellar cores and infrared dark clouds. We find high levels of deuteration also close to the cavity driven by IRS 3. We sp eculate that evap oration of NH2 D and NH3 destruction caused by the UV radiation field can influence this high deuteration. These findings undoubtedly confirm a tight interaction in I20343 b etween the most massive and evolved ob jects and the dense surrounding starless cores in several resp ects (kinematics, temp erature, chemical gradients), and suggest that IRS 3 could have induced the formation of IRS 1 and of the other gaseous condensations accumulated on the cavity walls. However, the large-scale morphology of the molecular tracers suggests that we are likely seeing only a limited p ortion of a very extended gaseous filament. Only a large p c-scale mosaic will allow us to fully trace the distribution of the dense gas in the region and delineate a complete view of the core p opulation.

5

SUMMARY AND CONCLUSIONS

The protocluster associated with the centre of I20343 is an excellent location where the interaction b etween evolved intermediate- and high-mass protostellar ob jects and dense (starless) cores can b e studied. We have derived the physical and chemical prop erties of the dense gas by means of CARMA and VLA observations of the millimeter and centimeter continuum, and of several molecular tracers (C2 H, cC3 H2 , NH3 , NH2 D, N2 H+ ). Below, we summarise the main results. · Morphologically, the dense gas is distributed in a filament oriented east-west that passes in b etween IRS 1 and IRS 3, the two most massive and evolved ob jects. We resolve the dense gas into five millimeter continuum condensations. All of them show column densities consistent with p otentially b eing the birthplace of intermediate- to high-mass objects, although the masses derived from continuum suggest that they can form intermediate-mass ob jects more likely.
c 2011 RAS, MNRAS 000, 1­17

ACKNOWLEDGMENTS Supp ort for CARMA construction was derived from the Gordon and Betty Moore Foundation, the Kenneth T. and Eileen L. Norris Foundation, the James S. McDonnell Foundation, the Associates of the California Institute of Technology, the University of Chicago, the states of California, Illinois, and Maryland, and the National Science Foundation. Ongoing CARMA development and op erations are supp orted by the National Science Foundation under a coop erative agreement, and by the CARMA partner universities. We acknowledge supp ort from the Owens Valley Radio Observatory, which is supp orted by the National Science Foundation through grant AST 05-40399. AP is grateful to Inma Sepulveda for insightful discussions. AP is supp orted by ´ the Spanish MICINN grant AYA2008-06189-C03 (co-funded


16

F. Fontani et al.
Lee, H.-G., Koo, B.-C., Park, Y.-S., Ho, P. T. P. 2002, JKAS, 35, 105 Linsky, J. L., et al. 2006, ApJ, 647, 1106 Longmore, S. N., Burton, M. G., Barnes, P. J. et al. 2007, MNRAS, 379, 535 Lucas, R. & Liszt, H.S. 2000, A&A, 358, 1069 MacLaren, I., Richardson, K.M. & Wolfendale, A.W. 1988, ApJ, 333, 821 Miralles, M.P., Rodriguez, L.F., Tapia, M., et al. 1994, A&A, 282, 547 Molinari, S., Brand, J., Cesaroni, R., & Palla, F. 2000, A&A, 355, 617 Mul, P.M. & McGowan, J.W. 1980, ApJ, 237, 749 Muller, H. S. P., Thorwirth, S., Roth, D. A., & Win¨ newisser, G. 2001, A&A, 370, L49 Padovani, M., Walmsley, C.M., Tafalla, M., Galli, D., Muller, H.S.P. 2009, A&A, 505, 1199 ¨ Palau, A., Estalella, R., Girart, J.M., et al. 2007a, A&A, 465, 219 Palau, Aina, Estalella, R., Ho, P.T.P., Beuther, H., Beltr´n, a M.T. 2007b, A&A, 474, 911 ´ Palau, Aina, S´nchez-Monge, A., Busquet, G., et al. 2010, a A&A, 510, A5 Panagia, N. 1973, ApJ, 78, 929 Park, Y.-S. 2001, A&A, 376, 348 Peeters, E., Hony, S., Van Kerckhoven, C. et al. 2002, A&A, 390, 1089 Pety, J., Teyssier, D., Foss´, D. et al. 2005, A&A, 435, 885 e Pillai, T., Kauffmann, J., Wyrowski, F. et al. 2011, A&A, 530, 118 Pillai, T., Wyrowski, F., Hatchell, Gibb, J.A.G., Thompson, M.A. 2007, A&A, 467, 207 Purcell, C. R., Minier, V., Longmore, S. N. et al. 2009, A&A, 504, 139 Roueff, E., Lis, D.C., van der Tak, F.F.S., Gerin, M., Goldsmith, P.F. 2005, A&A, 438, 585 Rygl, K.L.J., Brunthaler, A., Sanna, A. et al. 2011, arXiv1111.7023 Sakai, N., Saruwatari, O., Sakai, T., Takano, S., Yamamoto, S. 2010, A&A, 512, 31 Sridharan, T. K., Beuther, H., Schilke, P., Menten, K. M., Wyrowski, F. 2002, ApJ, 566, 931 Stutzki, J. & Winnewisser, G. 1985, A&A, 144, 13 Tafalla, M., Myers, P.C., Caselli, P., & Walmsley, C.M. 2004, A&A, 416, 191 Torrelles, J.M., Verdes-Montenegro, L., Ho, P.T.P., Rodriguez, L.F., Canto, J. 1993, ApJ, 410, 202 Walsh, A. J., Myers, P. C., Di Francesco, J., et al., 2007, ApJ, 655, 958 Zhang, Q., Hunter, T.R., Brand, J. et al. 2005, ApJ, 625, 864 Zhang, Q., Hunter, T.R., Sridharan, T.K., Ho, P.T.P. 2002, ApJ, 566, 982

with FEDER funds) and by a JAE-Doc CSIC fellowship cofunded with the Europ ean Social Fund under the program `Junta para la Ampliaci´n de Estudios'. GB is funded by an o Italian Space Agency (ASI) fellowship under contract numb er I/005/07/01. We are grateful to the anonymous referee for his/her valuable comments and suggestions.

REFERENCES Andr´, Ph., Belloche, A., Motte, F., Peretto, N. 2007, e A&A, 472, 519 Arce, H.G., Borkin, M.A., Goodman, A.A., Pineda, J.E., Beaumont, C.N. 2011, ApJ, 742, 105 Beuther, H., Semenov, D., Henning, Th., Linz, H. 2008, ApJ, 675, L33 Beuther, H., Schilke, P., Menten, K.M. et al. 2002, ApJ, 566, 945 Bourke, T.L., Myers, P.C., Caselli, P. et al. 2011, arXiv:1111.4424 Briggs, D.S. 1995, 187th American Astronomical Society Meeting, Vol. 27, p.1444 Busquet, G., Palau, Aina, Estalella, R., et al. 2009, A&A, 506, 1183 Busquet, G., Palau, A., Estalella, R. et al. 2010, A&A, 517, L6 Campb ell, M. F., Sridharan, T.K., Beuther, H. et al. 2008, ApJ, 673, 954 Carral, P., Kurtz, S., Rodriguez, L.F. et al. 1999, RMxAA, 35, 97 Caselli, P., Walmsley C.M., Zucconi, A. et al. 2002, ApJ, 565, 344 Flower, D. R., Pineau Des For^ts, G., & Walmsley, C. M. e 2006, 456, 215 Fontani, F., Caselli, P., Zhang, Q. et al. 2012, arXiv:1203.3412 Fontani, F., Zhang, Q., Caselli, P., Bourke, T.L. 2009, A&A, 499, 233 Fontani, F., Caselli, P., Bourke, T.L., Cesaroni, R., Brand, J. 2008, A&A, 477, L45 Foster, J.B., Rosolowsky, E.W., Kauffmann, J., et al. 2009, ApJ, 696, 298 Friesen, R.K., Di Francesco, J., Shirley, Y.L., Myers, P.C. 2009, ApJ, 697, 1457 Fuente, A., Martin-Pintado, J., Cernicharo, J., Bachiller, R. 1993, A&A, 276, 473 Gerin, M., Ka´mierczak, M., Jastrzebska, M. et al. 2011, z A&A, 525, A116 Gerin, M., Goicoechea, J.R., Pety, J., Hily-Blant, P. 2009, A&A, 494, 977 Henneb elle, P.; Commer¸on, B.; Joos, M. et al. 2011, A&A, c 528, 72 Hill, T., Thompson, M.A., Burton, M.G. 2006, MNRAS, 368, 1223 Ho, P.T.P. & Townes, C.H. 1983, ARA&A, 21, 239 Kirk, H., Johnstone, D., & Tafalla, M. 2007, ApJ, 668, 1042 Kramer, C., Alves, J., Lada, C., et al. 1998, A&A, 329, L33 Krumholz, M.R., Klein, R.I., McKee, C.F. 2011, ApJ, 740, 74 Krumholz, M. & McKee, C.F. 2008, Nature, 451, 1082 Kumar, M.S.N., Bachiller, R., & Davis, C. J. 2002, ApJ, 576, 313

APPENDIX A: ANALYSIS FOR THE CASE OF DIFFERENT Tex FOR THE MAIN AND THE INNER SATELLITE LINES OF NH3 (1, 1) Let us assume that the linewidth of the magnetic hyp erfine comp onents of the NH3 (1, 1) inversion transition is large
c 2011 RAS, MNRAS 000, 1­17


Dense gas in IRAS 20343+4129
T
Tex (1, 1; is) < Tex (1, 1; m) m 1 is > 1 External heating
MB

17
(A-4)

(1, 1; m)

T k (1 - e

- m

),

giving that the optical depth of the main line must b e m 3.1. Thus, taking into account that is = 0.28 m , we obtain 0.61 1 - e - i s 1 - e - m Tex (1, 1; is) Tex (1, 1; m) 1, (A-5)

Observer

resulting in 0.23 0.38. (A-6)

Figure A-1. Geometry of MMA

enough to make them unresolved, so that only the five electric hyp erfine lines, i.e. one main line ("m"), two inner satellites ("is"), and two outer satellites ("os"), are resolved. Assuming that the excitation temp erature is Tex Tbg , the ratio of intensities of the main line to the inner satellite line is TMB (1, 1; is) Tex (1, 1; is)(1 - e-is ) = . TMB (1, 1; m) Tex (1, 1; m)(1 - e-m ) (A-1)

The result is that the deep er layer traced by the satellite line is colder than the outer layer traced by the main line. If we assume that the kinetic temp erature of MMA, obtained from the ratio TMB (2, 2)/TMB (1, 1), is tracing the outer layer, the outer layer temp erature is 22 K, while the inner layer temp erature is b etween 5 K and 8 K. The higher temp erature of the outer layer is indicative of external heating, as discussed in the text.

The usual assumption is that we are observing a homogeneous isothermal region, so that b oth excitation temp eratures, Tex (1, 1; is) and Tex (1, 1; m), are equal, and that is = 0.28 m . In this case, 0.28 TMB (1, 1; is) TMB (1, 1; m) 1. (A-2)

The lower limit corresp onds to the optically thin case, while the upp er limit is the optically thick case. In Sect. 3.2.2 we have shown NH3 (1, 1) sp ectra for each of the 3 mm continuum clumps. For the case of MMA, the intensity ratio of the inner satellites and the main line is 0.23 ± 0.05, lower than the optically thin limit, 0.28. The assumption that the observed ratio is close to 0.28, and that the emission in MMA is optically thin, leads to inconsistent results. For an optical depth of the main line of m < 0.1, we obtain that Tex > 230 K, which is much higher than the kinetic temp erature estimated from the intensities ratio TMB (2, 2)/TMB (1, 1) (see text), Tk = 22 K. This result is improbable, since we exp ect the excitation temp erature to b e, in general, lower than the kinetic temp erature. The intensity of the (1, 1; m) is close to the kinetic temp erature, indicating that the optical depth of the main line is probably m 1. The optical depth of the satellite, however, can b e lower, so that b oth lines are tracing the emission of different layers of the region observed: the main line, the outer layer facing the observer; and the satellite line, a deep er layer of material (see Fig. A-1). The easiest explanation of the anomalous ratio TMB (1, 1; is)/TMB (1, 1; m) is to assume that the region is not isothermal, and that the two layers at different physical depths, have different temp eratures. So, the two excitation temp eratures, Tex (1, 1; is) and Tex (1, 1; m), are not equal. Thus, Tex (1, 1; is) 1 - e-is TMB (1, 1; is) = 0.23. = Tex (1, 1; m) 1 - e-m TMB (1, 1; m) Assuming that Tex (1, 1; m) is lower than Tk ,
c 2011 RAS, MNRAS 000, 1­17

(A-3)