Physical properties of high-mass clumps in different stages of evolution
A. Giannetti, J. Brand, A. Sanchez-Monge, F. Fontani, R. Cesaroni, M. T. Beltran, S. Molinari, R. Dodson, M. J. Rioja
AAstronomy & Astrophysics manuscript no. AGiannetti˙Properties˙Massive˙Clumps c (cid:13)
ESO 2018October 1, 2018
Physical properties of high-mass clumps in different stages ofevolution (cid:63)
A. Giannetti , , J. Brand , , ´A. S´anchez-Monge , F. Fontani , R. Cesaroni , M. T. Beltr´an , S. Molinari , R. Dodson , andM. J. Rioja , INAF-Istituto di Radioastronomia, Via Gobetti 101, 40129, Bologna, Italy Dipartimento di Astronomia, Universit`a di Bologna, Via Ranzani 1, 40127, Bologna, Italy Italian ALMA Regional Centre, Via Gobetti 101, 40129, Bologna, Italy INAF-Osservatorio Astrofisico di Arcetri, Largo E. Fermi 5, 50125, Firenze, Italy INAF-Istituto di Astrofisica e Planetologia Spaziali, Via Fosso del Cavaliere 100, 00133, Roma, Italy ICRAR, University of Western Australia, Perth, Australia Observatorio Astron´omico Nacional (OAN), Apartado 112, E-28803, Alcala de Henares, SpainPreprint online version: October 1, 2018
Abstract
Context.
The details of the process of massive star formation are still elusive. A complete characterization of the first stages of theprocess from an observational point of view is needed to constrain theories on the subject. In the past 20 years we have made a thoroughinvestigation of colour-selected IRAS sources over the whole sky. The sources in the northern hemisphere were studied in detail andused to derive an evolutionary sequence based on their spectral energy distribution.
Aims.
To investigate the first stages of the process of high-mass star formation, we selected a sample of massive clumps previouslyobserved with the Swedish-ESO Submillimetre Telescope at 1 . . Methods.
With ATCA we observed the selected sources in the NH (1,1) and (2,2) transitions and in the H O(6 − ) maser line.Ammonia lines are a very good temperature probe that allow us to accurately determine the mass and the column, volume, and surfacedensities of the clumps. We also collected all data available to construct the spectral energy distribution of the individual clumps and todetermine if star formation is already occurring through observations of its most common signposts, thus putting constraints on theevolutionary stage of the source. We fitted the spectral energy distribution between 1 . µ m with a modified black body toderive the dust temperature and independently determine the mass. Results.
We find that the clumps are cold ( T ∼ −
30 K), massive ( M ∼ − M (cid:12) ), and dense ( n (H ) (cid:38) cm − ) and that theyhave high column densities ( N (H ) ∼ cm − ). All clumps appear to be potentially able to form high-mass stars. The most massiveclumps appear to be gravitationally unstable, if the only sources of support against collapse are turbulence and thermal pressure, whichpossibly indicates that the magnetic field is important in stabilizing them. Conclusions.
After investigating how the average properties depend on the evolutionary phase of the source, we find that the temperatureand central density progressively increase with time. Sources likely hosting a ZAMS star show a steeper radial dependence of thevolume density and tend to be more compact than starless clumps.
Key words.
ISM: Molecules – Stars: Formation – Stars: Massive
1. Introduction
Massive stars spend a significant part ( (cid:38) ffi cult toinvestigate their early evolutionary stages. The discovery of IR-dark clouds (IRDCs; e.g., Perault et al. 1996; Egan et al. 1998)seen in absorption against the mid-IR Galactic background madeit possible to identify the most likely birthplaces of high-massstars. These clouds are usually filamentary, hosting complexesof cold ( T (cid:46)
25 K) and dense ( n (cid:38) cm − ) clumps, withhigh H column densities ( N (cid:38) cm − ) and masses that usu-ally exceed 100 M (cid:12) (though not all of them will form massivestars; e.g., Kau ff mann & Pillai 2010). Clouds with such low tem- Send o ff print requests to : A. Giannetti ([email protected]) (cid:63) Partly based on observations with the Herschel satellite.
Herschel is an ESA space observatory with science instruments provided byEuropean-led Principal Investigator consortia and with important partici-pation from NASA. peratures have a spectral energy distribution (SED) that peaksat far-IR (FIR) wavelengths and are optically thin in the mil-limetre / submillimetre regime. The emission at these wavelengthsusually matches the IR absorption very well (e.g., Rathborneet al. 2006; Pillai et al. 2006), and makes it easy to identify thecold and dense gas concentrations. Some clumps within IRDCsshow signs of active star formation, such as 24 µ m emission,presence of extended excess emission at 4 . µ m , masers andSiO emission from outflows (e.g., Beuther et al. 2005; Rathborneet al. 2005; Chambers et al. 2009).The pre- / proto-stellar phase for high-mass stars is very short( (cid:46) × yr), according to statistical studies (Motte et al. 2007),when compared to the low-mass regime ( ∼ × yr, Kirk et al. The excess at 4 . µ m, typically named Extended Green Object or“green fuzzy” is commonly interpreted as arising from H and CO lines,likely tracing shocks (e.g., Noriega-Crespo et al. 2004; Marston et al.2004). 1 a r X i v : . [ a s t r o - ph . GA ] J u l . Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution ff erent evolutionary stages has been supportedby a large number of observations at both low- and high-angularresolution (e.g., Molinari et al. 1996, 1998a,b; Brand et al. 2001;Fontani et al. 2005; Beltr´an et al. 2006). Those with δ < ◦ have been observed with the SEST in the continuum at 1.2-mm(SIMBA) and in CS (Fontani et al. 2005; Beltr´an et al. 2006). Themm-continuum maps often show the presence of several clumpsaround a single IRAS source. A comparison with MSX (and laterSpitzer) images revealed that some of these clumps are associatedwith mid-IR emission, while others appear IR-dark (Beltr´an et al.2006).A first attempt to exploit such large amount of data to definean evolutionary sequence for the clumps and their embeddedsources was carried out by Molinari et al. (2008), who distin-guished three di ff erent types of objects, on the basis of their mmand IR properties: (Type 1) objects with dominant mm emission,and not associated with a mid-IR source; (Type 2) objects withboth IR and mm emission; and (Type 3) objects with clearlydominant IR emission. Using a simple model, the authors couldexplain these di ff erent types in terms of an evolutionary scenario,in order of increasing age: (Type 1) starless cores and / or high-mass proto-stars embedded in dusty clumps; (Type 2) deeplyembedded Zero-Age Main Sequence (ZAMS) OB star(s) stillaccreting material from the parental clump, and (Type 3) OBstars surrounded only by the remnants of the molecular cloud.From our recent ATCA 1.3 cm continuum and line (H O maserat 22 GHz) observations (S´anchez-Monge et al. 2013a) of a largenumber ( ∼ Omaser emission was found associated with 13%, 26%, and 3%of sources of Type 1, 2 and 3, respectively. These findings cor-roborate the evolutionary sequence derived by Molinari et al.(2008).In this paper we will explore how the gas and dust propertiesin massive clumps depend on the evolutionary phase, as deter-mined from the source type and the presence of signposts ofhigh-mass star formation.The paper is organized as follows: in Sect. 2 we briefly de-scribe the sample selection, in Sect. 3 we describe the observa-tions performed with the Australia Telescope Compact Array,and describe the data reduction procedure; in Sect. 4 we show theresults for the quantities directly derived from the observations; inSect. 5 we discuss how the sample is divided into “star-forming”and “quiescent” clumps, and into clumps likely hosting a ZAMSstar (Types 2 and 3) and clumps that are starless or with a deeplyembedded protostar (Type 1). The mean clump properties areinvestigated to search for di ff erences as a function of evolution-ary phase. In Sect. 6 we give a sketch of the di ff erent classesof objects identified and finally in Sect. 7 we summarize ourfindings. Table 1.
Central coordinates of the observed fields, names of theclumps in the field and their coordinates.
Phase Centre (J2000) Clump Clump Coordinates (J2000)RA DEC RA DEC08:49:35.13 − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − −
2. Sample and tracer
The 39 fields considered in this paper were selected fromthe Beltr´an et al. (2006) survey at 1 . / SIMBA towards IRAS sources, and contain 46 massivemillimetre clumps. The coordinates of the field centres and ofthe clumps are listed in Table 1. Each field contains at least onemassive clump. The selection of fields was done according tosimple criteria: (i) source declination δ < − ◦ ; (ii) comparablenumbers of MSX-dark and -bright sources; (iii) clumps as iso-lated as possible, i.e. with a separation greater than 1 SEST beam(24 (cid:48)(cid:48) ) between MSX and non-MSX emitters to limit confusion;(iv) masses in excess of ∼
40 M (cid:12) in the Beltr´an et al. (2006) cata-logue. In this work we will use the gas temperature derived fromammonia observations and the dust temperature derived from amodified black-body fit to the SED to obtain a more accurateestimate of the mass and related quantities. For a spectroscopictool we selected ammonia, which is an ideal tracer for cold, densegas, not depleting up to high number densities ( (cid:38) cm − ) and
2. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure 1.
Typical map for a source in which we filter out theextended emission. The contours are ±
3% of the intensity peakof 628 . τ . This allows a direct estimate of the column density.The five lines are further split into several closely spaced com-ponents by magnetic interactions between the nuclei; however,these lines are typically not resolved observationally.
3. Observations and data reduction
The fields were observed in the NH (1,1) and (2,2) inversion tran-sitions (23694 .
50 MHz and 23722 .
63 MHz, respectively) and inthe H O(6 − ) maser line (22235 .
08 MHz), with the AustraliaTelescope Compact Array (ATCA). The observations were per-formed between the 4th and 8th of March 2011, for a total tele-scope time of 48 hours. We used the array in configuration 750D,providing baselines from 31 m to 4469 m. The primary beamof the telescope at these frequencies is ∼ (cid:48) .
5. The flux densityscale was determined by observing the standard primary calibra-tor PKS1934 −
638 (0 .
78 Jy at 23650MHz), with an uncertaintyexpected to be (cid:46) −
441 wasused as the bandpass calibrator. Pointing corrections were derivedfrom nearby quasars and applied online. Weather conditions weregenerally good, with a weather path noise ∼ µ m or better.The total time on source was divided into series of snapshotswith a variable duration of between 3 and 5 minutes observedover a range as large as possible in hour angle, to improve theuv-coverage for each target. As a consequence of the observingstrategy, the total on source integration time varies, and is typ-ically between ∼
30 min and ∼ ∼ . − at ∼ . O maser line.The data were edited and calibrated with the MIRIAD soft-ware package, following standard procedures. Deconvolution and imaging were performed in AIPS with the “imagr” task, applyingnatural weighting to the visibilities. Ammonia emission lineswere visible only on the shortest baselines, thus we discarded allbaselines (cid:38)
30 k λ . In order to obtain images with the same angu-lar resolution, we reconstructed all of them with a clean circularbeam of diameter 20 (cid:48)(cid:48) , except for 17195 − − − (cid:48)(cid:48) × (cid:48)(cid:48) . Moreover, 16254 − − ff erent ways: froma circular area of diameter 20 (cid:48)(cid:48) centred at the peak emission, oraveraged over the (larger) region enclosed in the 3 σ contour ofthe NH (1,1) integrated emission. The spectra extracted from thedata cube were imported in CLASS , and the lines were fittedusing METHOD NH3 for the NH (1,1) inversion transition, thattakes into account the hyperfine splitting of the line, thus givingas output also the optical depth of the main line. This method wasalso used for the (2,2) transition, in order to obtain a better esti-mate of the full width at half maximum (FWHM) linewidth ∆ V and τ for the 9 sources for which we detected the (2,2) hyperfinestructure. The spectral rms ranges from 3 to 55 mJy, with typicalvalues around 10 mJy. The value of the rms for each spectrum isgiven in Table 2.H O maser emission is detected on all baselines, allowingus to achieve the highest angular resolution allowed by the arrayconfiguration ( ∼ − (cid:48)(cid:48) ). For 16254 − −
4. Results and analysis
The NH (1,1) integrated emission (zeroth moment) is shown inpanels (a) and (b) of Fig. C.1 together with the SEST 1 . (1,1) and (2,2); 43 have been detected in NH (1,1).Three clumps were not detected in NH at all: 15454 − − − (1,1) may showsignificant displacement with respect to the millimetre peak. Forsome sources this may be caused by a low signal-to-noise ratioof the NH emission. Alternatively, the o ff set could be the resultof optical depth or chemical e ff ects. Twelve clumps have opticaldepths larger than 1 . ff set. We also made maps of the emissionof one of the satellite lines, (that are likely to be optically thin,as their optical depth is ∼ ff set remains unchanged. Thus,optical depth e ff ects cannot be the dominant cause for the o ff set.The peak flux of the two ammonia transitions, the rms of thespectra, ∆ V , the optical depth τ , the systemic velocity V LSR of Part of the GILDAS (Grenoble Image and Line Data AnalysisSoftware http://iram.fr/IRAMFR/GILDAS/ ) package 3. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure 2.
Ratio of the line FWHM ∆ V for the NH (2,2) and (1,1)lines as a function of kinetic temperature, derived from our dataassuming ∆ V (2,2) = ∆ V (1,1).the line, the rotation temperature T rot , the kinetic temperature T K and the ammonia column density N (NH ) are listed in Table 2.The optical depth of the (1,1) transition ranges from (cid:28) ∼
4, showing that ammonia emission is moderately thick in theseobjects.Figure 2 shows that the FWHM ∆ V of the (2,2) transition ison average slightly larger than that of the (1,1), indicating thatthe two transitions do not trace exactly the same volume of gas,the (2,2) transition being more sensitive to regions with a higherdegree of turbulence due to its higher energy. The same result isfound by Rygl et al. (2010). To estimate a rotation temperature(e.g., Mangum et al. 1992; Busquet et al. 2009) from the lineratio the two ammonia transitions must trace the same volumeof gas. Thus, if this assumption is correct, the ∆ V should be theequal. The di ff erence in ∆ V that we measure is su ffi ciently smallnot to invalidate our assumption that the emission from NH (1,1)and (2,2) comes from the same region. We thus consider ourtemperature estimates reliable.The kinematic distances were recomputed for the clumps withthe V LSR derived from NH and the rotation curve of Brand &Blitz (1993), and were found to be in agreement with those givenin Beltr´an et al. (2006), except for 08477 − − − − < −
15% forvirtually all of our sources.In the inner Galaxy, objects along the line-of-sight on eitherside of the tangent point have the same radial velocity, whichleads to an ambiguity in the kinematic distance (“near” and “far”)for several of our targets. Thus, we checked all our sources forassociated 8 µ m absorption features in the Spitzer / GLIMPSEimages, for H i self-absorption observations towards them in theliterature and for the height with respect to the Galactic midplane.The near distance is chosen if the complex is observed in ab-sorption against the Galactic mid-InfraRed (MIR) backgroundor if the source at the far distance is further than 150 pc fromthe Galactic plane ( ∼ − ffi ths (2011) report H i self-absorption mea-surements for 7 H ii regions near our observed fields (containing12 clumps in total). They locate 3 H ii region / clump complexes atthe far distance. However, we are confident that 2 (15557-5215and 17040-3959) of those 3 are instead at the near distance as the8 µ m images show a clear absorption patch, and this is unlikelyif the sources were on the far side of the Galactic centre. Hencethe far distance was assigned to only 1 of our fields (containing 1clump). The distances adopted are listed in Table 2. Where thenear-far ambiguity could not be resolved (8 sources), the neardistance was assumed. We derive the rotation temperature ( T rot ), and the molecular col-umn density, following the method described in Busquet et al.(2009). This assumes that the transitions between the inversiondoublets can be approximated as a two level system (see Ho& Townes 1983), and that the excitation temperature and linewidths are the same for both the (1,1) and (2,2) transition (seeSect. 4.1). The kinetic temperature T K was then extrapolated from T rot using the empirical method outlined in Tafalla et al. (2004).This relation gives results accurate to a 5% level for temperatures (cid:46)
20 K. This procedure was used to derive the gas temperatureboth from the spectra extracted from an area equal to that of thebeam around the peak of the ammonia emission, and from thoseaveraged over the whole area of NH (1,1) emission.In order to also have an estimate of the uncertainty, the methodto derive T rot , T K and ammonia column density N (NH ) was im-plemented in JAGS (Just Another Gibbs Sampler). JAGS is aprogram for the analysis of Bayesian models, based on MarkovChain Monte Carlo simulations. It computes the posterior proba-bility distribution, summarizing our knowledge of the quantitiesconsidered, given a user-defined model (i.e. the equations andthe assumptions of Gaussianity for the quantities directly derivedfrom the fit in our case), the data and our prior knowledge of thequantities involved (Andreon 2011). This program was used toderive T rot , T K and N (NH ) and their uncertainty, propagatingthe Gaussian uncertainty of the parameters of the fit, as given byCLASS. Constant priors were used on these parameters, i.e. T τ and τ . To check the dependency of the results on the choice of the prior , we used also a Gaussian prior with a large σ . The resultsshow that the derived parameters are virtually independent of the prior choice.The temperatures and N (NH ) derived from the peak spec-trum in this way are listed in Table 2, with their uncertainties. Onthe other hand, Table C.6 shows the observed spectral parametersfor the spectra averaged over the whole NH (1,1) emission, therotation and kinetic temperatures and the average ammonia col-umn density, with the respective uncertainties. T K obtained fromthe spectra extracted from both the peak of the NH and those ob-tained from the whole area of emission are in the range between ∼
10 and ∼
28 K. T rot and T K calculated from the two sets ofammonia spectra agree very well in most cases. Few exceptionsexist, where the 68% credibility intervals for the kinetic temper-ature do not overlap (3 cases), but with di ff erences of ∼ (2,2) emission, averagedover the same area as the (1,1). Thus in the following we use the T K derived at the peak of ammonia emission http://mcmc-jags.sourceforge.net/
4. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure 3.
Mass of the clumps as a function of distance. The blacksolid line indicates the typical mass sensitivity for the SESTimages (see text). The sources with signs of active star formationare shown as red filled circles, those without as black open squares(see Sect. 5). Associated MSX emission is indicated as a blackplus, and radio-continuum emission as a black cross. The whitecross indicates 17195-3811c1 (see text).The gas temperatures derived from ammonia imply that theaverage ∆ V of the ammonia lines (between ∼ . . − )is well in excess of the thermal broadening in such cold gas( ∼ .
15 km s − for T K =
20 K), indicating that turbulence mayplay a major role in supporting the clumps.
To determine characteristic ammonia abundances we used thespectra averaged over all the NH emission. We derived N (H )from the average 1 . column density by N (H ). Thetotal range of abundances for all the sources in the sample lies be-tween ∼ − and ∼ − . For most of the objects the abundancesare in the typical range of ∼ − − − (cf. Wienen et al. 2012,and references therein). We compared the NH abundance derivedin this way with the abundance derived at the peak of ammoniaemission: we find this latter quantity is typically slightly greaterthan the former, with ratios in the range ∼ . −
10 and meanand median values of 2 . .
2, respectively. A sub-sample ofthe clumps observed in ammonia was also observed in C O andN H + with APEX (Fontani et al. 2012). The carbon monoxidewas found to be heavily depleted in these sources, showing thatthey are in an early phase of evolution. On the contrary, the ob-served ammonia abundances indicate that NH is not depleted ona large scale in these clumps, in agreement with studies of clumpsin low-mass star-forming regions, whereas CO is also depleted(e.g., Tafalla et al. 2002). Determining accurate masses for the clumps is crucial to deter-mine the evolutionary phase of the clump from a mass-luminosityplot (Molinari et al. 2008), and to see if the clump is massiveenough to form high-mass stars. With our temperature determina-tion (assuming that the gas, dust and kinetic temperatures T g , T d and T K are equal) we are able to compute more accurate massesthan those listed in Beltr´an et al. (2006), that were derived as-suming T d =
30 K. The clump masses are calculated from theintegrated 1 . gas = γ S D κ B ( T d ) , (1)where S is the total flux density at 250 GHz, D is the dis-tance, γ is the gas-to-dust ratio, B ( T d ) is the emission ofa black body with temperature equal to T d at 250 GHz, and κ ≡ κ (250 GHz /ν [GHz]) β is the dust opacity per unit massat the indicated frequency. We used κ = . − at ν = . β was derived from the modified black-bodyfit to the spectral energy distribution of the clumps, using onlythe SEST and Hi-GAL fluxes (see Sect. 4.5), where the data wereavailable, otherwise we chose β =
2, as in Beltr´an et al. (2006).The masses and their uncertainties are again estimated withJAGS, taking into account the probability distribution of T K , asderived from the ammonia observations, the uncertainty of the in-tegrated 1 . . σ flux density of 100 mJy / beam and a T d =
15 K. In this work weadopted the mass computed within the FWHM contour, in orderto consider only the inner regions of the clump, excluding the ex-ternal envelope (see Sect. 5), and because the measured diameterdepends on the signal-to-noise ratio. Therefore when we generi-cally speak of the mass we refer to masses computed within theFWHM contour. In this way the mass could be underestimatedby a factor of 2, if the source is Gaussian and the envelope con-tribution is negligible. Our mass estimates are thus conservative,and the possible variation is indicated in figures 4 and 10 for com-parison. The masses are listed in Table C.1. For completeness,masses and densities computed within the 3 σ contour contourare shown in Table C.2.Angular diameters were derived from 1 . θ of the clumps at FWHMlevel are estimated assuming that the sources are Gaussian, usingthe relation θ = √ FWHP − HPBW , with FWHP = √ A /π ,where A is the area within the contour at half peak intensity, andHPBW is the SEST half-power beam width. If the angular sizederived in this way is less than half the beam size, the source isdeemed unresolved and we set an upper limit to its size equalto half the HPBW (Wilson et al. 2005). The linear diameters atFWHM level range from ∼ . ∼ . ff mann & Pillai (2010) derived an empirical relation be-tween mass and radius to separate the clumps that are able to formhigh-mass stars from those that are not. With the mass now muchbetter constrained, we can use this relation to test if our clumpshave the potential of forming massive stars. In Fig. 4 we showthe mass and size of our clumps, compared to the Kau ff mann& Pillai (2010) relation, scaled to the same dust opacity as usedin the present work. From the figure we observe that the vastmajority of our sources lie above the Kau ff mann & Pillai (2010)relation, indicated as a dashed line, corroborating the idea thatthe whole sample is constituted of similar objects and suggesting
5. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure 4.
Mass-radius plot; r FWHM is the beam-corrected radiusat FWHM level. The symbols are the same as in Fig. 3. Theboundary for massive-star formation derived by Kau ff mann &Pillai (2010) ( M [M (cid:12) ] = r / pc) . , rescaled to match ourdust opacity) is indicated as a dashed line. Sources above this lineare able to form massive stars. The uncertainty in mass is shownfor each point. A variation of a factor of 2 in mass (see text) isindicated in the bottom right corner.that virtually all of them could form massive stars. This makesour sample a good one to study the evolution of massive clumpspotentially able to form massive stars.The column- and volume-densities of molecular hydrogenwere calculated using the mass and the diameters, assumingspherical symmetry and correcting for helium ( ∼
8% in number;e.g., Allen 1973). These two quantities are found to lie between ∼ . − × cm − , and 0 . − × cm − , respectively:values like these are typical of IRDCs (e.g., Egan et al. 1998;Carey et al. 1998, 2000; Pillai et al. 2006). The surface density Σ was determined by averaging the mass over the deconvolvedFWHM area of emission at 1 . Σ for the clumps in thissample is found to lie between 0 .
03 and 1 . − .The mass, volume-, column-, and surface densities with their68% credibility intervals for all the clumps are listed in Table C.1. Important insights in the evolutionary state of a source can begained through its L / M ratio, as proposed by Saraceno et al.(1996) for the low-mass regime and by Molinari et al. (2008)for the high-mass regime. We constructed the Spectral EnergyDistribution (SED) for the sources in our sample complement-ing the SEST data with Herschel / Hi-GAL (500 µ m, 350 µ m,250 µ m, 160 µ m, 70 µ m; Molinari et al. 2010), MIPSGAL(24 µ m; Carey et al. 2009), MSX (band A 8 . µ m, C 12 . µ m,D 14 . µ m, E 21 . µ m; Price et al. 2001) and GLIMPSE(8 . µ m, 5 . µ m, 4 . µ m, 3 . µ m; Benjamin et al. 2003;Churchwell et al. 2009) data. We smoothed all the images toa common resolution of 25 (cid:48)(cid:48) (the approximate resolution of the350 µ m image) except that at 500 µ m, which has a resolution of ∼ (cid:48)(cid:48) . Two di ff erent polygons were defined for each wavelength:one to derive the flux density of the source, and the other for Pilbratt et al. 2010. Here we use the PACS (Poglitsch et al. 2010)and SPIRE (Gri ffi n et al. 2010) instruments. the background in the region. The bolometric luminosity wascalculated by integrating the fluxes over frequency, interpolatinglinearly between the measured fluxes at di ff erent frequencies inlogarithmic space. The uncertainty was estimated by simply inter-polating between the lower and upper limit of the 68% credibilityinterval of the fluxes used to derive the luminosity, respectively.The fluxes at the longest wavelengths in the SED can beused to infer the typical properties of the dusty envelope. Withthis in mind we fitted the SED fluxes, down to 70 µ m with amodified black body. The point at 70 µ m was included in thefit to better constrain the temperature in the case where the SEDhas its peak shortward of 160 µ m. The inclusion of the flux at70 µ m implies that we will measure a higher T d , because weare tracing the warm dust layers near the embedded (proto-)star.The fit procedure is described in Appendix B. The results of themodified black-body fit, the luminosity derived integrating theSED from 1 . . µ m, and the uncertainties in thesequantities for each clump are listed in Table 3. In the table welist only the clumps with data in all the five Herschel / Hi-GALbands. Figures C.5 and C.6 show the SED with the fit results.17195-3811c1 is on the edge of the Herschel / HiGAL 160 / µ mmaps, thus is not included here. However we use the lower limitson the fluxes at these wavelengths for the fit with the Robitaillemodels (see below) for this latter source.From Fig. 5 we can see that the characteristic T d obtainedfrom the fit of a modified black body to the SED down to 70 µ mis usually in good agreement with T K derived from the ammoniaobservations. Seven sources have | T K − T d | ≥ ff erence, implying astatistically significant discrepancy. Four of these objects have T d > T K : these are also the cases where T d is high, always above20 K. The discrepancy may arise from a combination of di ff erentcauses: the fact that the ratio of the two lowest transitions ofammonia is optimal to derive temperatures only up to 20 −
25 K,that ammonia and dust emission are probing di ff erent regionsof the clump, and that the strong emission in these clumps fromwarm dust, heated by the central star and visible at 70 µ m, isbiasing the modified black-body fit towards higher T d .The gas masses obtained from the modified black-body fitusually agree, within the uncertainties, with those derived simplyfrom the 1 . σ and that within the FWHM contour (Fig. 6). As not all sourceshave a complete SED, and considering the reasonable agreementbetween masses determined from the 1 . µ m in the smoothed im-ages. Otherwise, only the modified black-body fit is done. To fitthe SED with the Robitaille models, we make use of the online
6. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure 5.
Comparison between the kinetic temperature ( T K ) de-rived from ammonia and the dust temperature ( T d ) from theSED-fit. The dashed line indicates equal temperatures and theyellow-shaded region shows a di ff erence of ± Figure 6.
Comparison between the mass derived from the 1 . M SED areindicated. The bars for the 1 . σ contour. The dashed line indicates equal massesand the yellow-shaded region shows a di ff erence of a factor oftwo. Table 4.
Summary of the properties derived from the fit of theRobitaille models to the SED of the objects with significant mid-IR emission. Our estimated range in L is shown for comparison.The ranges in M ∗ , L Rob and M env are those spanned by the bestten models. The masses are derived with a di ff erent dust model,and thus deviate from our estimate. Clump M ∗ L Rob
L M env (M (cid:12) ) (10 × L (cid:12) ) (10 × L (cid:12) ) (10 × M (cid:12) )13560-6133c1 9 . − . −
43 26 −
34 15 . − . . − . −
49 24 −
30 5 . − . . − . −
28 10 −
13 1 . − . . − . −
665 241 −
291 16 . − . . − . −
234 68 −
84 5 . − . . − . −
15 6 − . − . . − . −
334 76 −
94 14 . − . . − . −
116 30 −
38 23 . − . . − . −
781 77 −
96 16 . − . . − . −
209 78 −
102 19 . − . . − . −
638 234 −
279 9 . − . . − . − − . − . . − . −
34 17 −
21 2 . − . SED fitting tool (Robitaille et al. 2007). In the tool, we allowedthe foreground interstellar extinction to range between 1 and2 mag kpc − (e.g., Allen 1973; Lynga 1982; Sche ffl er 1982).The stellar masses obtained from the fit range between ∼ ∼
30 M (cid:12) , corresponding to spectral types approximately B7-O8.The luminosities derived vary between ∼
600 and 65000 L (cid:12) ,agreeing with those derived by simply integrating the SED, inter-polating linearly in the log-log space. Our estimate of L tends tobe lower, as the linear interpolation in the log-log space gives alower limit for the luminosity and because of the model assump-tions. However, for consistency, in the following we will use ourestimate of the luminosity for all sources.The envelope mass from the fit of the Robitaille models isgreater than that derived either from the modified black-bodyfit or from the 1 . ff ner et al.(2012) compared synthetic SEDs of deeply embedded proto-stars,derived from simulated observation obtained with a 3D radiativetransfer code for dust emission, for a vast range of parameters,with the best fit obtained from the standard grid of Robitaillemodels. These authors showed that usually the fit recovers thetrue luminosity and stellar mass, although with large uncertainties,but systematically overestimates the mass of the envelope, mainlydue to the assumption of a di ff erent dust model. The di ff erenceis more pronounced in the mm-regime, strongly influencing themass determination. Our assumption of dust opacity is similar tothat used by O ff ner et al. (2012) in the mm-regime, explainingthe discrepancy between our mass estimates and the envelopemass from the fit with the online SED fitting tool.A summary of the results of the fits with the Robitaille modelsis shown in Table 4. To investigate the stability of the clumps, we performed the sim-plest virial analysis, without taking into account magnetic orrotational support.To derive the virial mass we assumed a constant density pro-file, using Eq. (3) in MacLaren et al. (1988), which implies that M vir [M (cid:12) ] = R [pc] ∆ V [km s − ] , where R is the radius and ∆ V is the FWHM of the line. In this way we obtain an upper http://caravan.astro.wisc.edu/protostars/
7. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure 7.
Virial parameter α = M vir / M as a function of M . Thesymbols are the same as in Fig. 3. The dashed lines indicates α = α = n (H ) ∝ ( r / r ) − reduces the virial mass by about a factor of2 (cf. MacLaren et al. 1988). The virial parameter α = M vir / M is used as an indicator for gravitational stability; α < α = α are upper limits. Figure 7 shows that for virtually allthe clumps we find α (cid:46)
1, implying that they are dominated bygravity. In Sect. 5.2.6 we discuss in detail the observed values of α . First order moment maps reveal the presence of velocity gradi-ents, typically around 1 − − pc − . A more detailed analysisfor the most interesting clumps is the subject of a forthcomingpaper. Thirteen clumps in our sample show maser emission. The spectrawere extracted from the data cubes at each position where emis-sion was detected (see Fig. C.4) within a polygon comparable tothe beam dimensions (between ∼ (cid:48)(cid:48) and ∼ (cid:48)(cid:48) ), and imported inCLASS. The lines were then fitted with Gaussians. Two sourceshave very strong lines, reaching nearly ∼
100 Jy. We typicallyfind multiple velocity components (up to 22) towards a singleclump. A summary of the maser emission properties is shown inTable 5. The water maser range of velocities usually straddles thesystemic velocity of the clump, as shown in Fig. 8 (cf. Brand et al.2003). The positions of the maser spots are indicated in Fig. C.2as white open squares. A comparison with S´anchez-Monge et al.(2013a) shows that 2 sources detected in their study are not de-tected in our observations, while 2 targets that we detect, werenot detected in S´anchez-Monge et al. (2013a) (cf. Table C.3), asexpected because of the well-known variability of water masers(e.g., Felli et al. 2007). All of these sources show other signs ofactive star formation.
Figure 8.
Range of velocity of the water maser vs. V
LSR of theclump. The dashed line indicates V H O = V LSR , i.e. a maservelocity equal to the systemic velocity of the clump.
IR RCGFM
All: 4IR + GF + RC:2GF + RC + M:0IR + RC + M: 1IR + GF + M: 5IR + GF: 3RC + M: 1GF + RC: 0IR + M: 1GF + M: 1IR + RC: 6IR: 6M: 0RC: 1GF: 1
Figure 9.
Summary of specific star formation indicators in theclumps. The labels correspond to green fuzzies (GF), mid-IRemission (IR), H O maser (M), and radio continuum emission(RC) (see Sect. 5 for details). For example, the small central circleshows the combination IR + GF + RC, while the “triangular” areawith a darker shade surrounding the small circle represents thepresence of all four signposts of star formation.
5. Discussion
In order to investigate how the clump properties depend on theirevolutionary state, the sources were first separated into two sub-samples, according to the presence or absence of signposts ofactive star formation. In particular, we considered the presenceof water maser(s), “green fuzzies”, 24 µ m and radio-continuumemission. – µ m emission: We overlaid the Spitzer MIPSGAL imagesat 24 µ m and the SEST 1 . µ m source is found at the location of the 1 . µ m as shown in Fig. C.2, and 3
8. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution ti m e time6.2 8.013.518.028.0 Figure 10.
Mass-Luminosity plot for the sources in our samplewith Hi-GAL observations. The mass is computed within theFWHM contour of 1 . ff erent masses; thelines are labeled with the final mass of the most massive star (inM (cid:12) ). Time increases from bottom to top and from right to left,as indicated. Radio-continuum emission and MSX emission arefound nearly exclusively in clumps near the ZAMS. A variationof a factor of 2 in mass (see text) is indicated in the bottom leftcorner.among those without Spitzer 24 µ m data show emission inthe 21 µ m MSX image. – “Green Fuzzies”: extended 4 . µ m emission produced byshock-excited molecular lines, commonly associated withClass II CH OH masers (Cyganowski et al. 2009). To identifythe “green fuzzies”, we followed the procedure described inChambers et al. (2009). Again, only 41 clumps out of the46 are covered by GLIMPSE data. Sixteen clumps show thepresence of extended, excess 4 . µ m emission. – Water masers: 13 clumps in our sample show maser emis-sion. The water maser is a known indicator of star formation,thought to appear in the early stages of the process (Breen &Ellingsen 2011), and is observed both in low- and high-massstar formation regions. – Radio-continuum: 40 of the clumps in our sample were ob-served with ATCA at 22 GHz and 18 GHz (S´anchez-Mongeet al. 2013a). Twelve sources were detected, and 3 more havea tentative detection at about 3 σ at one of the frequencies.The radio-continuum alone is not always considered su ffi cientto classify a source as star forming. This is because we findone source (16061 − / Hi-GALimages. This special case is discussed in the Appendix.If any of these signposts is observed (except radio continuumin the special case discussed in the Appendix), a clump is indi-cated as star-forming. Based on these criteria, of the 46 objectsobserved, 31 were classified as star-forming and 15 as quiescent.A summary of the clumps with specific indicators of ongoingstar formation is given in Fig. 9. All the star formation signposts
Figure 11.
Normalized histogram of the luminosities for the SFS(green) and for the QS (black). 500 L (cid:12) and 8000 L (cid:12) are indicatedby the dashed lines.in each clump are presented in Table C.3. Figure C.2 shows allthe observed fields of view and clumps (dashed and solid redcircles, respectively), with the SEST emission (contours) super-imposed on the MIPS 24 µ m image, and the position of maserspots (white open squares) and “green fuzzies” (green open cir-cles). Removing those clumps with no NH (2,2) detection, 26and 10 clumps remain in the star-forming and in the quiescentsub-samples, respectively.In the following, we compare the average properties of theclumps using the parameters we have derived, for the two sub-samples, to look for systematic di ff erences in the physical prop-erties characterizing the two classes. Note that hereafter the qui-escent and star-forming sub-samples will be denoted with theacronyms QS and SFS , respectively.
To refine the separation in di ff erent evolutionary phases we makeuse of the mass-luminosity ( M − L ) plot, which is an e ffi cientand well-established diagnostic tool to disentangle the di ff erentevolutionary phases of star formation in the low-mass regime(Saraceno et al. 1996). Molinari et al. (2008) proposed that itcould also be used for high-mass stars, under the hypothesis thatstar formation at high- and low-mass proceeds in a similar fashion,with accretion from the surrounding environment playing a majorrole (e.g., Krumholz et al. 2009). Molinari et al. (2008) built asimple model for the evolution of a clump, based on the turbulentcore prescriptions of McKee & Tan (2003), ranging from the earlycollapse phase to the complete disruption of the dusty envelopeby the central object.Figure 10 shows the M − L plot for the sources in our samplefor which we were able to derive the luminosity (see Sect. 4.5and Table 3) from the integration of the SED (with the linearinterpolation in the log-log space) and the mass derived fromthe 1 .
9. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution locus as determined by Urquhart et al. (2013). The latter authorsconsider clumps showing methanol maser emission and with aluminosity from the Red MSX Sources survey (Urquhart et al.2008). 90% of the sources studied by Urquhart et al. (2013) have L > L (cid:12) , the remaining 10%, with L < L (cid:12) , have low gasmasses (10 − M (cid:12) ), and could be forming intermediate-massstars. Objects classified as either YSO and UCH ii regions wereused to determine the ZAMS locus, as no statistically significantdi ff erence is found between the two classes. The di ff erence inslope between the two ZAMS-lines could be due to the fact thatin Molinari et al. (2008) the luminosities were determined fromthe Robitaille models, using IRAS fluxes in the FIR, and couldthus be overestimated. The grey curves show the evolution of thesource predicted by the simple model, for di ff erent final masses,from ∼ ∼
30 M (cid:12) . Time increases in the direction of thearrows shown in the upper right part of the plot. In the collapsephase, before the central object reaches the ZAMS, the mass ofthe core envelope does not change by much, but the luminosityincreases rapidly. After the central star reaches the ZAMS, theluminosity does not vary much, and the energetic radiation andwind begin to destroy the parental clump, visible as a steadydecrease in envelope mass.Comparing the distribution of sources in the plot with theevolutionary tracks of the Molinari et al. (2008) model, our ob-jects span the total range in envelope masses, for final masses ofthe star between about 6 and 30 M (cid:12) . The stellar masses are inagreement with those derived from the fit of the SED with theRobitaille models for the objects near the ZAMS, for which thefinal mass is similar to the current mass, as the main accretionphase is over. Massive stars appear to form virtually always inclusters (e.g., Lada & Lada 2003). In both the Molinari et al.(2008) and the Robitaille models the luminosity is considered asbeing dominated by the most massive object formed in the cluster.The derived bolometric luminosity and stellar mass could thus beoverestimated. A detailed study of the stellar population in theseclumps will be carried out in a subsequent paper, by means ofmid-IR and near-IR data.A first macroscopic di ff erence between the SFS and the QSsources is the luminosity. As can be seen in Fig. 11, the clumpsin the QS have low luminosities, distributed between ∼
100 and500 L (cid:12) . On the other hand, the sources in the SFS show a peakat ∼
500 L (cid:12) , but also a second peak at L ∼ × L (cid:12) (bothindicated as dashed lines in the figure). The distribution of thesources in the M − L plot indicates that part of the sources inthe SFS (10 out of 20, including 17195-3811c1, with the meanluminosity derived with the online SED fitting tool) are likelyhosting a ZAMS star and have stopped the accelerating accretionphase, according to the model of Molinari et al. (2008). Thesources with signs of active star formation, but well below theZAMS loci, are essentially indistinguishable from the quiescentones in terms of luminosities. Figure 10 shows that in our sampleall sources with L > L (cid:12) have strong IR (MSX) and / or radiocontinuum emission, while those below this threshold do not. Theradio emission from the sources near the ZAMS locus, with finalmasses greater than 8 M (cid:12) shows that the interpretation of the M − L plot is essentially correct, and that the prediction of theend of the accelerating accretion phase is reasonably good.The small range of luminosity and its low average value forthe QS shows that this is a homogeneous sample, with all theclumps in an early phase of evolution. On the other hand, theSFS appear to include clumps in widely di ff erent evolutionarystages: the points in the diagram go from clumps similar to thoseof the QS, to clumps containing a ZAMS star and beyond, wherethe star is dispersing the envelope. For our sample, a simple Figure 13.
Correlation between luminosity and kinetic tempera-ture. The uncertainties for L and T K are indicated in the figure.criterion in luminosity is su ffi cient to separate the sources thatlikely have an embedded ZAMS star from the rest. Thus, inthe following we will refer to objects containing a ZAMS staras those with L > L (cid:12) . In Fig. 10 these objects fall in theregion encompassed by the ZAMS loci, except 13560 − − M − L plot (see Molinari et al. 2008);the classification of sources in one of the three types is shownin Table C.3. The clumps between the two ZAMS loci, with L > L (cid:12) would be Type 2 clumps (including the two sourcesslightly below the Urquhart et al. ZAMS locus, 13560 − − ii region may be quenched by the high accretion rates.The leftmost point in the mass luminosity plot (Fig. 10) identifiesthe most evolved source in our sample, 17355 − − M − L plot, when the parent cloud is dispersed bythe destructive action of the central star. This source is discussedin more detail in the Appendix.Thus, the M − L plot and the signposts of active star formationgive complementary information about the evolutionary state ofthe clump, allowing us to refine the Molinari et al. (2008) clas-sification, separating objects likely hosting a ZAMS star fromthe other sources in the SFS. Our original sample is thus finallydivided into three di ff erent classes (without considering Type 3objects): Type 1 quiescent clumps, apparently starless, Type 1with signs of active star formation, but still a low luminosity andType 2 sources, hereafter QS, SFS-1 and SFS-2, respectively.Table C.5 shows explicitly that SFS-1 and SFS-2 have very dif-ferent luminosity-to-mass ratios. The clumps in the SFS-1 can bein a very early phase of the process of formation of a high-massobject; alternatively, the signposts of active star formation couldbe generated by more evolved lower-mass stars.
10. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution (a) (b) (c) Figure 12.
Normalized histogram (to the total number of sources in each class) of kinetic temperature for (a) the SFS (green), and theQS (black); (b) the same as (a) , but the SFS-2, with L > L (cid:12) (in red) are separated from the SFS-1 (in blue), the red triangle showsthe temperature for the Type 3 source; (c) the same as (a) , but the SFS was divided into clumps with (magenta) and without (cyan)radio-continuum emission. Mean and median values of the temperature are indicated in each panel. The total number of sources ineach class is shown above each panel. (a) (b) Figure 14.
Histogram of the diameters of the clumps. Panel (a) shows the diameters of the SFS (green) vs. the QS (black). Panel (b) shows the diameters for the SFS-2 is indicated in red, the SFS-1 is indicated in blue, and the QS in black. The total number of sourcesin each class is shown above each panel.
The temperature of gas and dust may be influenced by the pres-ence of a (proto-)star deeply embedded in a clump. Figure 12ashows the histogram of temperatures for the two samples. Thenormalized counts of the SFS are shown in green, and those ofthe QS in black. We find that it is possible to observe tempera-ture di ff erences on a large scale, comparing the average valuesof T K and T d of the QS and the SFS. The typical T K of the star-forming clumps is greater than that of the QS, with mean values of T K = . + . − . K and 14 . + . − . K, for the SFS and QS, respectively.Thus, the average temperature increases as evolution proceeds.In Fig. 12b we can see that the SFS-2 (in red) show aslightly higher T K ( T K = . + . − . K) than the SFS-1 (in blue)( T K = . + . − . K). Figure 12b shows also the QS, to underlinethat both the SFS-1 and SFS-2 sources on average have a higher T K . Figure 12c shows that the SFS with 1 . (1,1) and(2,2) at high angular resolution in cores in clustered high- andintermediate-mass star forming regions. They find that starlesscores have an average T ∼
15 K, lower than T ∼
21 K found forproto-stellar cores. These values are very close to those found inthis work. S´anchez-Monge et al. (2013b) show that the highertemperatures in starless cores in clustered environments with re-spect to more isolated cases can be explained considering externalheating from the nearby massive stars. Also Rygl et al. (2010)and Urquhart et al. (2011) find that actively star-forming clumpsare slightly hotter than the quiescent ones. A behaviour similarto that of the kinetic temperature is observed for T d as a functionof evolutionary phase, not unexpected given the good agreementof the two temperatures (see Sect. 4.5; Fig. 5). The mean valuesare T d = . + . − . ; 15 . + . − . ; 11 . + . − . K for SFS-2, SFS-1 and QS,respectively.
11. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure 15. . | T K − T d | ≥ T K and L bol was also found by e.g.,Churchwell et al. (1990), Wu et al. (2006), Urquhart et al. (2011),and by S´anchez-Monge et al. (2013b), for cores in clusteredenvironments, using the luminosity of the whole region. From the panel (a) of Fig. 14 we note that the SFS tend to havesmaller FWHM diameters than the QS. From the distribution ofsizes shown in panel (b) we note that the SFS-2 have a peak atthe smallest linear dimensions. Plotting the area within a specificintensity contour of 1 . . ff ect of the lower temperature on the clump size atthe typical noise levels of the SEST maps.This procedure shows that with our noise levels we miss ∼
30% of the emission area for a typical QS, while only ∼ −
15% is lost for a typical SFS-2. The SFS-1 usually showsan intermediate behaviour. The larger fraction of emission areabelow the noise level for the QS confirms that we are not able todetect the external envelope of the coldest clumps, and that theactual linear size of QS sources is very similar to that of SFS-2objects. On the other hand, the area within the FWHM contour issmaller for SFS-2 sources, possibly indicating that sources host-ing a ZAMS object are more compact and centrally concentrated.We investigate an alternative possibility, namely that the observedFWHM size may be T -dependent, performing a simple test us-ing a 1D simulation with RATRAN (Hogerheijde & van der Tak2000), constructing a clump with a typical radial dependence ofthe density ( ∝ r − . , cf. Beuther et al. 2002; Mueller et al. 2002),and comparing the continuum at 250 GHz with and without acentral luminous heating source. The radial temperature depen-dence is assumed to be ∝ ( r / r ) − . (e.g., Wolfire & Cassinelli Figure 16.
Normalized histogram of volume density of molecularhydrogen averaged within the FWHM contour. We show the SFSin green and the QS in black. The total number of sources in eachclass is shown above the panel.1986). We observe that the clump with the embedded source, hasa smaller (by 20 − The mean column-, volume- and surface-densities are comparedfor QS and SFS, and for SFS-1 and SFS-2 in Fig. 16 and 17,respectively.The histogram of the SFS is shifted towards higher densities;Chambers et al. (2009) obtain the same result, with star-formingsources on average denser than the quiescent ones. Taking into ac-count the separation into SFS-1 and SFS-2, the density histogramsshow that the SFS-2 usually have higher values of column-,volume- and surface-densities than either QS or SFS-1; the lattertwo classes have density distributions peaking at similar values,with that of the SFS-1 showing a tail with values similar to thoseof the SFS-2. Thus, the clumps hosting a ZAMS star have higherdensities, indicating that as star formation proceeds, the clumpsappear to become denser in the central parts. The same is foundby Butler & Tan (2012), comparing their sample of starless coresto a sample of more evolved objects (from Mueller et al. 2002).We caution that, as the source size could be underestimated due tothe presence of a central heating source, causing the clump to ap-pear more centrally peaked (see Sect. 5.2.2), the densities couldbe overestimated for the SFS-2, thus explaining the observeddi ff erences with both QS and SFS-1.Rygl et al. (2013) argue that star formation signposts arenot present in clumps with a column density below a value of4 × cm − . No such threshold e ff ect for the onset of starformation is observed in our sample, but only two of our clumpsof any type have column densities well below 4 × cm − (cf.Table C.1).
12. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution (a) (b) (c) Figure 17.
Normalized histograms of (a) column-, (b) volume- and (c) surface-densities of molecular hydrogen averaged within theFWHM contour. We show the SFS-2 in red, SFS-1 in blue and the QS in black. The total number of sources in each class is shownabove the panels.We note that the values of the mass surface density are typ-ically lower than the theoretical threshold of
Σ = − formassive star formation, on average by a factor 2 −
6. The the-oretical threshold for Σ given by Krumholz & McKee (2008)holds for a single core, stabilized against fragmentation only byradiative heating. L´opez-Sepulcre et al. (2010) show that massivestar formation, indicated by the presence of massive molecularoutflows, most probably driven by massive YSOs, is occurringalso in clumps with a much lower mass surface density, of theorder of Σ ∼ . − . Butler & Tan (2012) show that similarvalues of the mass surface density Σ are typical also of massivestarless cores, assuming a gas-to-dust ratio of 150, thus consis-tent with our average value of ∼ . − for the QS, for agas-to-dust ratio of 100.We investigate in more detail the possibility that the radial de-pendence of the density changes in di ff erent classes of objects, byfitting a power law to the average radial 1 . . x . Following the notation inWard-Thompson et al. (1994) (and references therein), the powerlaw indices of the mm emission m , of the density p and of thetemperature q are related by m = p + Q ( ν, T ) q −
1, where Q ( ν, T )is a coe ffi cient depending on the wavelength and the temperature,near to unity for h ν/ ( k B T ) (cid:28) m (cid:39) . − . ∼ . − . T ∝ ( r / r ) q , with q = − . p for volume density is ∼ . − . p ∼ . − .
4. The value of p is highly uncertain, because ofthe assumptions made and because of the very limited numberof points for each source, however we are mainly interested inthe di ff erence between the QS and SFS-2, and not the specificvalue of p . As the clumps in QS and SFS have similar massesand total sizes (see Sect. 4.4 and 5.2.2), this di ff erence in density profile suggests that SFS-2 clumps may indeed be denser in thecentral regions. Other authors investigated the di ff erences in thedensity power law index for sources in di ff erent stages of evolu-tion. Beuther et al. (2002) also derive an average radial power lawdependence for density with p ∼ . − .
0. These authors findthat the density distribution is flatter for sources in the very earlystages ( p ∼ . p ∼ . p ∼ . p ∼ . ff erent technique,with the same quantity derived by Mueller et al. (2002) for moreevolved objects ( p ∼ . The masses measured for our clumps are in the range 10 − (cid:12) , which indicates that we are probing the low-massend of the Kau ff mann & Pillai (2010) relation (cf. their Fig. 2band Fig. 4 in this paper). With respect to more massive clumps,those in this sample will probably form a very limited number ofstars with M > (cid:12) , making the assumption of a single massivestar in the clump plausible. On the other hand, we know thatmassive stars are being formed in some objects, as we observecompact radio continuum emission, likely arising in H ii regions(S´anchez-Monge et al. 2013a).Comparing the source size and densities (Fig. 14 and 17)SFS-2 sources are typically the most compact, suggesting thatan evolution of the clumps towards more compact entities withevolution might indeed occur. However, as time proceeds and themassive stars dissociate the molecular gas they move down and tothe right in the M − r plot (Fig. 4), as shown by 17355 − The velocity gradients found for the clumps (1 − − pc − ;Sect. 4.6) are comparable for the SFS and QS, and also the frac-
13. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure 18.
Average 1 . x . In the bottom left corner the slope m is indicated. Figure 19.
Example of NH (1,1) second moment map. Leftpanel shows Spitzer / MIPSGAL 24 µ m image in colourscaleand NH (1,1) integrated emission in contours, while the rightpanel shows the second moment map.tion of clumps showing gradients is similar: about 50% of theobjects.From the second moment map we find that the clumpsin the SFS have larger linewidths, with an average value of2 . − compared to 1 . − for the sources in theQS. For 08477 − − − − − µ m source and the rest ofthe clump of ∼ − µ m source. Thesesources always have a ∆ V of the (2,2) line larger than that ofthe (1,1) at this position. This indicates that we are probing theregions where the embedded source is injecting turbulence. Allthese sources have a luminosity in excess of 10 L (cid:12) , except15557 − L =
740 L (cid:12) ), and 08477 − ∼ . − and ∼ . − for starless and proto-stellar cores, respectively. Figure 7 presents the virial parameter α ≡ M vir / M as a function of M . As already noted, all clumps in our sample appear dominatedby gravity. Moreover, these are upper limits for α as the virialmass should be reduced by up to a factor of 2 (see MacLarenet al. 1988) as a consequence of the density gradients found in theclumps; another factor of 2 may arise because the mass was wascomputed within the FWHM contour. Two sources in the SFS has α (cid:38)
2: this could be due to the action of the embedded YSO(s)disrupting the parental cloud. The value of the virial parametertend to decrease as the clump mass increases.Let’s explore a few possibilities to explain α < / dust temperature, and are thus overestimating themass from the 1 . (1,1) and (2,2) essen-tially trace cold gas and they could be optically thick, this may bea possibility. An independent determination of the temperaturein the clumps is given by the dust temperature derived from themodified black-body fit of the mm-FIR part of the SED. Dustemission is optically thin in this regime, thus probing even theinner regions of the clump. Comparing T K and T d we find thatthe two temperatures are usually in good agreement, as discussedin Sect. 4 and shown in Fig. 5. Therefore we do not expect thebulk of the gas in the clump to have temperatures systematicallyhigher than those adopted here, and consequently lower masses.Moreover, also quiescent clumps have α (cid:28)
1, and the temper-ature in this case should not be underestimated, as no heatingsource is present in these sources. Even if it were the case, itis di ffi cult to account for a factor of 3 −
10. The independentestimate of the mass through the SED essentially confirms thatthe mass of the clumps is correct.Another possibility is that we are underestimating the radiusof the clump, leading us to underestimate the virial mass. Panagia& Walmsley (1978) show that if the source is not Gaussian, wemay underestimate the radius by even a factor of 2. However, it isnot clear why the most massive clumps should be di ff erent fromthe others. The same holds for the uncertainty connected to thegas-to-dust ratio, assumed to be 100.
14. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Fontani et al. (2002) and L´opez-Sepulcre et al. (2010) reachopposite conclusions regarding the stability of the clumps intheir samples. In the former work the authors show that the ratiobetween M vir / M reaches values as low as ours, using as a tracerCH CCH, while L´opez-Sepulcre et al. (2010) obtain α ∼ O line emission. Also Hofner et al. (2000) find α < O. Some of the clumps in our samplehave been observed in C O(3 −
2) by Fontani et al. (2012), andthese authors suggest that the clumps are in virial equilibriumor even have virial masses greater than the clump mass (i.e., α (cid:38) O and the NH linewidths we findthat the former are larger, typically by a factor of 1 . ∼ .
6. As a consequence, the virial masses calculatedusing the C O linewidth would be larger by a factor 2 . . α would increase accordingly. S´anchez-Monge et al.(2013b) find linewidths for cores similar to those found in thisstudy. The di ff erence in linewidth between NH and C O maybe explained by the fact that C O is tracing a more extendedand di ff use region, where ∆ V may be larger due to the e ff ectsof the environment in which the clump is embedded, or dueto the presence of several gas concentrations along the line ofsight (i.e., additional indistinguishable velocity components in thespectra), not dense / massive enough to be visible in NH , or by thepresence of velocity gradients across the clumps. Moreover, C Ois more prone to be entrained in outflows driven by objects alreadyformed in the clump. In conclusion, α may be underestimated,nevertheless, even taking into account the possibilities discussedabove, α would still be < t f f = (cid:112) π/ (32 G ρ ) ∼ × yr, this may suggest that magneticfield plays a significant role in stabilizing the clumps. Followingthe relation for virial equilibrium given in McKee et al. (1993),neglecting the surface pressure term, we get the expression forthe equilibrium magnetic field B = . × − (cid:32) M
100 M (cid:12) (cid:33) (cid:32) R (cid:33) − . × (cid:32) M
100 M (cid:12) (cid:33) − . (cid:32) R (cid:33) (cid:32) ∆ V − (cid:33) . [ G ] , (2)where M is the clump mass, R is the clump radius and ∆ V is theFWHM of the line. Using appropriate numbers for these param-eters we find that |−→ B | ≈ . − ffi cient to stabilize theclumps. Such values of the magnetic field have been observed to-wards regions undergoing massive star formation (e.g., Crutcher2005; Girart et al. 2009).
6. Observational classification of high-mass clumps
Figure 21 shows a sketch of the di ff erent evolutionary phasesidentified in this work, with representative values for the ob-served properties indicated in the yellow rectangles. The arrowsconnecting the di ff erent source types show how the evolutionproceeds. In addition to the properties listed in the figure, wefind that the SFS-2 have smaller FWHM diameters, that can bedue to an intrinsically smaller size and / or just an observationale ff ect caused by the presence of a temperature gradient gener-ated by the heating of the embedded massive (proto-)stars (see Figure 20.
Ammonia abundance as a function of Galactocentricdistance. The symbols are the same as in Fig. 3.Sect. 5.2.2). The density profile seems to be steeper in the SFS-2clumps than in the QS, even when allowing for a temperaturegradient in the former sub-sample. Thus the central density mayindeed be higher for clumps with a similar size and mass. Moredetailed studies are needed to confirm this result.Mean and median values for various parameters of the QSand SFS, and of the SFS-1 and SFS-2 are listed in Tables C.4 andC.5.
7. Summary and conclusions
From ATCA NH observations of 46 clumps previously observedwith the SEST in the 1 . (1,1) and (2,2). With a reliable and independent temperatureestimate through the NH (1,1) and (2,2) line ratio, we determinedthe mass of the gas.We performed the simplest virial analysis to investigate thestability of the clumps against gravity. All sources, but one, showa virial parameter α (cid:46) α < |−→ B | ≈ . − − − − ,but within the canonical values of ∼ − − − for the vastmajority of the sample, showing that this molecule is not depleted,as opposed to CO in these clumps (Fontani et al. 2012).These data were complemented with Herschel / Hi-GAL,Spitzer / MIPSGAL, MSX and Spitzer / IRAC data, to constructthe SEDs of the sources (Sect. 4.5). From the SEDs we derivedthe luminosity of 32 sources (i.e. those with Herschel / Hi-GALdata), out of which we have 29 sources with a reliable ammoniadetection in both lines, so that we could locate the clumps in a M − L plot (Fig. 10). The sample was divided into sub-samplesof clumps in di ff erent phases of evolution on the basis of thepresence or absence of signs of ongoing star formation and onthe location of the sources in the M − L plot. To classify a clumpas star-forming, we considered the presence of at least one of the
15. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution QS SFS-1SFS-2Type 3 T K ≈ T d ∼
13 K N (H ) ∼ × cm − n (H ) ∼ × cm − L / M ∼ (cid:12) M (cid:12)− Σ ∼ . − T K ≈ T d ∼
17 K N (H ) ∼ × cm − n (H ) ∼ × cm − L / M ∼ (cid:12) M (cid:12)− Σ ∼ . − T K ≈ T d ∼
23 K N (H ) ∼ × cm − n (H ) ∼ × cm − L / M ∼
24 L (cid:12) M (cid:12)− Σ ∼ . − Clump Cores High-mass Proto-starMassive Outflow Low-mass (Proto-)starHigh-mass ZAMS StarHC / UCH ii RegionH ii Region
Figure 21.
A simple sketch of the evolutionary phases considered in this paper for massive star formation. The representativeproperties of clumps in the di ff erent evolutionary stages as derived in this work are listed in the yellow rectangles.following tracers: 24 µ m emission, 1 . . µ m).The star-forming sub-sample (SFS) includes these sources, whilethe quiescent sub-sample (QS) contains those without detectablesigns of ongoing star formation.We used the M − r relation found by Kau ff mann & Pillai(2010) to assess if also the clumps without clear signs of ongoingmassive star formation in our sample are potentially able to formhigh-mass stars. Virtually all sources lie above of the empiri-cal relation, confirming that our sample is a good one to studyevolution in the first stages of the formation of high-mass stars.We explored if and how the average properties of a clumpdepend on the presence of active star formation and on its evolu- tionary phase (Sect. 5). The information from the M − L plot andthat on the presence of ongoing star formation are complemen-tary: we find that Type 1 sources include both the star formingclumps with a low luminosity, well below the ZAMS loci in the M − L plot, and the clumps in the QS, while Type 2 sources arethe clumps in the SFS hosting a ZAMS star. For our sample, aconvenient criterion based on L was enough to separate Type 1and Type 2 objects (cf. Fig. 10): sources with L > L (cid:12) arelikely to host a ZAMS star. This idea is corroborated by the largefraction of these sources that are encompassed by the two deter-minations of the ZAMS locus (Molinari et al. 2008; Urquhartet al. 2013), and show radio-continuum and strong mid-IR emis-sion, i.e. an UCH ii region. Therefore we define the following
16. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution classes of objects using both the information obtained from the M − L plot and from classical signposts of star formation: QS forquiescent Type 1 sources, SFS-1 for Type 1 sources with signs ofactive star formation, and SFS-2 for Type 2 sources. A sketch ofthese phases with the typical values for the physical parametersderived in this work is shown in Fig. 21. Analyzing the typicalproperties of the clumps in our sample we find that they dependon the evolutionary phase of the source. The di ff erences foundcan be summarized as follows: – SFS-2 sources always show radio continuum or strong mid-IR emission, suggesting the presence of an H ii region, whileQS and SFS-1 objects do not. – The average temperatures (both kinetic and dust) of the threeevolutionary classes slowly increases from the QS, to SFS-1,to SFS-2 sources, with typical values of ∼
13 K, 17 K and ∼
23 K, respectively. – The temperature of the clumps appears to be correlated withthe luminosity of the source (Fig. 13). – SFS-2 objects have smaller FWHM diameters (median valuesof 0 . vs. . – As a consequence, clumps classified as SFS-2 on averagehave higher volume-, column- and surface-densities inside theFWHM intensity contour of the 1 . – Assuming that density (for all clumps) and temperature (forSFS-2 clumps) both vary as power laws as a function ofclump radius, we derived the power law indices for density,and found them to be steeper in more evolved sources. Typicalpower law indices for the molecular hydrogen volume densityare p ∼ . − . ∼ . − . – The fact that SFS-2 sources are the most extreme in terms ofcompactness suggests that QS sources are still contracting.
Acknowledgements.
The Australia Telescope is funded by the Commonwealthof Australia for operation as a National Facility managed by CSIRO. This workis based in part on observations made with the Spitzer Space Telescope, whichis operated by the Jet Propulsion Laboratory, California Institute of Technologyunder a contract with NASA. This research made use of data products from theMidcourse Space Experiment. Processing of the data was funded by the BallisticMissile Defense Organization with additional support from NASA O ffi ce ofSpace Science. This research has also made use of the NASA / IPAC InfraredScience Archive, which is operated by the Jet Propulsion Laboratory, CaliforniaInstitute of Technology, under contract with the National Aeronautics and SpaceAdministration. This research made use of the NASA ADS database.
References
Adams, F. C. 1991, ApJ, 382, 544Allen, C. W. 1973, Astrophysical quantitiesAndreon, S. 2011, ArXiv e-prints: 1112.3652Beltr´an, M. T., Brand, J., Cesaroni, R., et al. 2006, A&A, 447, 221Benjamin, R. A., Churchwell, E., Babler, B. L., et al. 2003, PASP, 115, 953Beuther, H., Schilke, P., Menten, K. M., et al. 2002, ApJ, 566, 945Beuther, H., Sridharan, T. K., & Saito, M. 2005, ApJ, 634, L185Brand, J. & Blitz, L. 1993, A&A, 275, 67Brand, J., Cesaroni, R., Comoretto, G., et al. 2003, A&A, 407, 573Brand, J., Cesaroni, R., Palla, F., & Molinari, S. 2001, A&A, 370, 230Brand, J. & Wouterloot, J. G. A. 1994, A&AS, 103, 503Breen, S. L. & Ellingsen, S. P. 2011, MNRAS, 416, 178Busquet, G., Palau, A., Estalella, R., et al. 2009, A&A, 506, 1183Butler, M. J. & Tan, J. C. 2012, ApJ, 754, 5Carey, S. J., Clark, F. O., Egan, M. P., et al. 1998, ApJ, 508, 721Carey, S. J., Feldman, P. A., Redman, R. O., et al. 2000, ApJ, 543, L157Carey, S. J., Noriega-Crespo, A., Mizuno, D. R., et al. 2009, PASP, 121, 76 Chambers, E. T., Jackson, J. M., Rathborne, J. M., & Simon, R. 2009, ApJS, 181,360Churchwell, E., Babler, B. L., Meade, M. R., et al. 2009, PASP, 121, 213Churchwell, E., Walmsley, C. M., & Cesaroni, R. 1990, A&AS, 83, 119Crutcher, R. M. 2005, in IAU Symposium, Vol. 227, Massive Star Birth: ACrossroads of Astrophysics, ed. R. Cesaroni, M. Felli, E. Churchwell, &M. Walmsley, 98–107Cyganowski, C. J., Brogan, C. L., Hunter, T. R., & Churchwell, E. 2009, ApJ,702, 1615Dame, T. M. & Thaddeus, P. 1994, ApJ, 436, L173Dame, T. M., Ungerechts, H., Cohen, R. S., et al. 1987, ApJ, 322, 706Egan, M. P., Shipman, R. F., Price, S. D., et al. 1998, ApJ, 494, L199Felli, M., Brand, J., Cesaroni, R., et al. 2007, A&A, 476, 373Fontani, F., Beltr´an, M. T., Brand, J., et al. 2005, A&A, 432, 921Fontani, F., Cesaroni, R., Caselli, P., & Olmi, L. 2002, A&A, 389, 603Fontani, F., Giannetti, A., Beltr´an, M. T., et al. 2012, MNRAS, 423, 2342Girart, J. M., Beltr´an, M. T., Zhang, Q., Rao, R., & Estalella, R. 2009, Science,324, 1408Green, J. A. & McClure-Gri ffi ths, N. M. 2011, MNRAS, 417, 2500Gri ffi n, M. J., Abergel, A., Abreu, A., et al. 2010, A&A, 518, L3Hildebrand, R. H. 1983, QJRAS, 24, 267Ho, P. T. P. & Townes, C. H. 1983, ARA&A, 21, 239Hofner, P., Wyrowski, F., Walmsley, C. M., & Churchwell, E. 2000, ApJ, 536,393Hogerheijde, M. R. & van der Tak, F. F. S. 2000, A&A, 362, 697Kau ff mann, J. & Pillai, T. 2010, ApJ, 723, L7Kirk, J. M., Ward-Thompson, D., & Andr´e, P. 2005, MNRAS, 360, 1506Krumholz, M. R., Klein, R. I., McKee, C. F., O ff ner, S. S. R., & Cunningham,A. J. 2009, Science, 323, 754Krumholz, M. R. & McKee, C. F. 2008, Nature, 451, 1082Lada, C. J. & Lada, E. A. 2003, ARA&A, 41, 57L´opez-Sepulcre, A., Cesaroni, R., & Walmsley, C. M. 2010, A&A, 517, A66Lynga, G. 1982, A&A, 109, 213MacLaren, I., Richardson, K. M., & Wolfendale, A. W. 1988, ApJ, 333, 821Mangum, J. G., Wootten, A., & Mundy, L. G. 1992, ApJ, 388, 467Marston, A. P., Reach, W. T., Noriega-Crespo, A., et al. 2004, ApJS, 154, 333McKee, C. F. & Tan, J. C. 2003, ApJ, 585, 850McKee, C. F., Zweibel, E. G., Goodman, A. A., & Heiles, C. 1993, in Protostarsand Planets III, ed. E. H. Levy & J. I. Lunine, 327Molinari, S., Brand, J., Cesaroni, R., & Palla, F. 1996, A&A, 308, 573Molinari, S., Brand, J., Cesaroni, R., Palla, F., & Palumbo, G. G. C. 1998a, A&A,336, 339Molinari, S., Pezzuto, S., Cesaroni, R., et al. 2008, A&A, 481, 345Molinari, S., Swinyard, B., Bally, J., et al. 2010, PASP, 122, 314Molinari, S., Testi, L., Brand, J., Cesaroni, R., & Palla, F. 1998b, ApJ, 505, L39Motte, F., Bontemps, S., Schilke, P., et al. 2007, A&A, 476, 1243Mueller, K. E., Shirley, Y. L., Evans, II, N. J., & Jacobson, H. R. 2002, ApJS,143, 469Noriega-Crespo, A., Morris, P., Marleau, F. R., et al. 2004, ApJS, 154, 352O ff ner, S. S. R., Robitaille, T. P., Hansen, C. E., McKee, C. F., & Klein, R. I.2012, ApJ, 753, 98Ossenkopf, V. & Henning, T. 1994, A&A, 291, 943Palla, F., Brand, J., Comoretto, G., Felli, M., & Cesaroni, R. 1991, A&A, 246,249Panagia, N. & Walmsley, C. M. 1978, A&A, 70, 411Perault, M., Omont, A., Simon, G., et al. 1996, A&A, 315, L165Pilbratt, G. L., Riedinger, J. R., Passvogel, T., et al. 2010, A&A, 518, L1Pillai, T., Wyrowski, F., Carey, S. J., & Menten, K. M. 2006, A&A, 450, 569Poglitsch, A., Waelkens, C., Geis, N., et al. 2010, A&A, 518, L2Price, S. D., Egan, M. P., Carey, S. J., Mizuno, D. R., & Kuchar, T. A. 2001, AJ,121, 2819Rathborne, J. M., Jackson, J. M., Chambers, E. T., et al. 2005, ApJ, 630, L181Rathborne, J. M., Jackson, J. M., & Simon, R. 2006, ApJ, 641, 389Reid, M. J., Menten, K. M., Zheng, X. W., et al. 2009, ApJ, 700, 137Robitaille, T. P., Whitney, B. A., Indebetouw, R., & Wood, K. 2007, ApJS, 169,328Robitaille, T. P., Whitney, B. A., Indebetouw, R., Wood, K., & Denzmore, P.2006, ApJS, 167, 256Rygl, K. L. J., Wyrowski, F., Schuller, F., & Menten, K. M. 2010, A&A, 515,A42Rygl, K. L. J., Wyrowski, F., Schuller, F., & Menten, K. M. 2013, A&A, 549, A5S´anchez-Monge, ´A., Beltr´an, M. T., Cesaroni, R., et al. 2013a, A&A, 550, A21S´anchez-Monge, ´A. et al. 2013b, MNRAS submittedSaraceno, P., Andre, P., Ceccarelli, C., Gri ffi n, M., & Molinari, S. 1996, A&A,309, 827Sche ffl er, H. 1982, in Landolt-Bornstein: Group 6: Astronomy, ed. P. Biermann,H. H. Fink, K. J. Fricke, W. Gliese, M. Grewing, W. K. Huchtmeier, B. F.Madore, H. Netzer, J. Rahe, H. Sche ffl er, L. D. Schmadel, J. Schmid-Burgk,
17. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution G. A. Tammann, J. Tr˜amper, R. Wielen, A. Witzel, & G. Zech, 45Schuller, F., Menten, K. M., Contreras, Y., et al. 2009, A&A, 504, 415Swinyard, B. M., Ade, P., Baluteau, J.-P., et al. 2010, A&A, 518, L4Tafalla, M., Myers, P. C., Caselli, P., & Walmsley, C. M. 2004, A&A, 416, 191Tafalla, M., Myers, P. C., Caselli, P., Walmsley, C. M., & Comito, C. 2002, ApJ,569, 815Thompson, R. I. 1984, ApJ, 283, 165Urquhart, J. S., Hoare, M. G., Lumsden, S. L., Oudmaijer, R. D., & Moore,T. J. T. 2008, in Astronomical Society of the Pacific Conference Series, Vol.387, Massive Star Formation: Observations Confront Theory, ed. H. Beuther,H. Linz, & T. Henning, 381Urquhart, J. S., Moore, T. J. T., Schuller, F., et al. 2013, MNRAS, 431, 1752Urquhart, J. S., Morgan, L. K., Figura, C. C., et al. 2011, MNRAS, 418, 1689Ward-Thompson, D., Scott, P. F., Hills, R. E., & Andre, P. 1994, MNRAS, 268,276Wienen, M., Wyrowski, F., Schuller, F., et al. 2012, A&A, 544, A146Wilson, T. L., Rohlfs, K., & Huettermeister, S. 2005, Tools of Radio Astronomy,ed. SpringerWolfire, M. G. & Cassinelli, J. P. 1986, ApJ, 310, 207Wu, Y., Zhang, Q., Yu, W., et al. 2006, A&A, 450, 607
Appendix A: Comments on individual sources
A.1. 16061 − − − − ii region produced by the central YSO,but an ionization front generated by a nearby massive star (mostlikely the corresponding IRAS source), rather than by the YSO.A similar situation is observed in 16435 − (2,2) was notdetected. A.2. 17355 − As evolution proceeds, the massive ZAMS star delivers hugequantities of ionizing photons and energetic particles into theparent molecular cloud. This process disperses the clump, leavingonly remnants of the original cloud, around an H ii region.17355 − . / M ratio is the highest among all the observedclumps, above 120 L (cid:12) M (cid:12)− . The virial parameter α ( ∼ .
5) is the highest. This could be due to the destructive action ofthe central ZAMS star, dispersing the clump from which it wasformed. On the other hand, this object does not show compactradio-continuum emission. The M − L plot shows that the massof the central ZAMS star should be M ∼ (cid:12) , correspondingto a B5 star, comparing its position in the M − L plot with theevolutionary tracks. The output of Lyman continuum photonsof such a star is low, and below the detection limit of the radioobservations (cf. Thompson 1984; S´anchez-Monge et al. 2013a).However the uncertainty on the stellar mass is quite large. TheSED fit with Robitaille models suggests a stellar mass of ∼ (cid:12) .Another possibility to explain the lack of radio continuum emis-sion is that it is extended and filtered out by the interferometer.Finally, we remind that in both cases the stellar mass is an upperlimit, under the assumption that the luminosity is dominated bythe most massive object.The SED of this object is di ff erent from that typical of theother objects with appreciable emission in the mid-IR, beingalmost flat in terms of energy up to 3 . µ m. Also the Robitaillemodel indicates that this is an older object. A.3. Undetected sources
A number of sources do not have detections in ammonia, es-pecially in the (2,2) transition. Among the SFS, they are weakalso in (1,1), suggesting low beam-averaged NH column densi-ties. 16164 − − . ii region, di ff erent from the nearby clumpsdetected in ammonia. 13563 − − − (1,1) detections, in-dicating very low gas temperatures. These clumps might be justtoo cold to be detected in (2,2) or simply not massive enough toform stars with M > (cid:12) .Finally, also 15454 − and visual inspection of the SEST map suggest that thismight be just a noise spike in the SEST 1 . Appendix B: SED fit
The modified black-body fit was done with a simple Bayesianapproach, considering Gaussian uncertainties on the fluxes, takinginto account the rms of the image and the calibration uncertaintyat each wavelength (15% for SEST and SPIRE, 20% for PACSred and 10% for PACS blue fluxes: Beltr´an et al. 2006; Swinyardet al. 2010; Poglitsch et al. 2010). We adopted a modified blackbody as a model, comparing the observed and predicted fluxesat each wavelength, and multiplying the probability for eachpoint, to obtain the total probability of the model. We took aconstant prior on the mass, while we used a Gaussian prior on β , with mean value µ = σ =
2, andon the dust temperature T d , with µ =
20 K and σ =
15 K. Weconsidered temperatures between 5 and 50 K in 200 equal linearsteps, masses between 10 and 10000 M (cid:12) in 200 equal logarithmic
18. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution steps, and β between 0 . prior , we also tried touse a constant prior on all the parameters of the fit. The resultsshow that T d , M and β are not sensitive to this choice, within theuncertainties. To derive the probability distribution, and thus theuncertainty on the single parameters we integrated over the othertwo parameters. Appendix C: Tables and figures
19. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Table 2.
Parameters of the ammonia spectra extracted from an area equal to that of the beam, around the NH emission peak. Thecolumns indicate the clump name, the peak flux of the (1,1) transition and the rms of the spectrum, the V LSR of the emission, the ∆ V of NH (1,1), the opacity of the (1,1) line and its uncertainty, the peak flux of the (2,2) transition and the rms of the spectrum, and the ∆ V of NH (2,2), T rot , T K and ammonia column density, with their uncertainties, the near and far kinematic distance. The clumpsabove the horizontal line are those classified as star-forming, while the clump below it are those classified as quiescent (see Sect. 5). C l u m p F m a x ( , ) R M S V L S R ∆ V ( , ) τ ( , ) σ τ ( , ) F m a x ( , ) R M S ∆ V ( , ) T r o t % i n t . T K % i n t . N ( NH ) % i n t . D n ea r D f a r N o t e s ( m J y )( m J y )( k m s − )( k m s − )( m J y )( m J y )( k m s − )( K )( K )( K )( K )( × c m − )( × c m − )( kp c )( kp c ) - c . . . . . . . . . . . − . . . − . . . − . . − - c . . . . . . ... . ... −−−−−− . − ( ) - c . . − . . . . . . . . . − . . . − . . . − . . . - c . . . . ......... . ... −−−−−− . − ( ) - c . . − . . . . . . . . . − . . . − . . . − . . . - c . . − . . . . ... . ... −−−−−− . a . ( ) - c ... . ............... . ... −−−−−− . c . ( ) - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . . . − . . . . . . . −−−−−− . a . ( ) - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . . b - c . . − . . . . . . . . . − . . . − . . . − . . c . - c . . − . . . . . . . . . − . . . − . . . − . . c . - c ... . ............... . ... −−−−−− . a . ( ) - c . . − . . . . . . . . . − . . . − . . . − . . a . ( ) - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . ( ) - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . ......... . ... −−−−−− . c . ( ) - c . . − . . . . . . . . . − . . . − . . . − . . . - c . . − . . . . . . . . . − . . . − . . . − . . c . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c ... . ............... . ... −−−−−− ...... ( ) - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . . . . . . − . . . − . . . − . . c . - c . . − . . . . . . . . . − . . . − . . . − . . a . - c . . − . . . . ... . . −−−−−− . a . ( ) - c . . − . . . . ... . . −−−−−− . a . ( ) - c . . − . . . . ... . . −−−−−− . c . ( ) - c . . − . . . . . . . . . − . . . − . . . − . . c . N o t e s . ( ) S ou r cee x c l ud e d fr o m t h ea n a l y s i s du e t onon - d e t ec ti on i n NH ( , )( s i gn a l - t o - no i s e < ) . ( ) S ou r ce w it h a s i ng l e ob s e r v a ti on : s p ec t r ae x t r ac t e d fr o m non - c l ea n e d m a p s . ( a ) T h e d i s t a n cea m b i gu it y w a s r e s o l v e d f o r t h i ss ou r ce . W ec ho s e t h e n ea r d i s t a n ce . ( b ) T h e d i s t a n cea m b i gu it y w a s r e s o l v e d f o r t h i ss ou r ce . W ec ho s e t h e f a r d i s t a n ce . ( c ) D i s t a n cea m b i gu it yno t r e s o l v e d . T h e n ea r d i s t a n ce w a s a ss u m e d f o r t h i ss ou r ce .
20. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Table 3.
Parameters derived from the modified black-body fit of the SED down to 70 µ m, and luminosity of the clumps (integratedfrom 1 . . µ m). The clumps above the horizontal line are those classified as star-forming, while the clump below itare those classified as quiescent (see Sect. 5). The columns shows the dust temperature ( T d ), the mass ( M ) of the gas and the dustemissivity index β from the modified black-body fit, and the luminosity of the clumps derived integrating the SED, with theiruncertainties. Clump T d
68% int. M
68% int. β
68% int. L
68% int. (K) (K) (10 × M (cid:12) ) (10 × M (cid:12) ) (10 × L (cid:12) ) (10 × L (cid:12) ) . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . .
35 0 . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . . . − . ff erent stages of evolution Table 5.
Summary of the emission characteristics of H O masers. The columns show the clump name, the number of Gaussiancomponents in the spectrum, the range of V LSR over which we detect emission, the flux density peak, the integrated emission, thewater maser luminosity, the V LSR and ∆ V of the strongest component (S.C.), and the o ff set of the maser spot with respect to the phasecentre (Ph.C.). No correction was performed for the primary beam. Clump Maser V LSR (min , max) F peak (cid:82) Fd ν L H O V LSR
S.C. ∆ V S.C. O ff set (Ph.C.) (km s − ) (Jy) (Jy km s − ) 10 − × L (cid:12) (km s − ) (km s − ) ( (cid:48)(cid:48) ) .
5; 14 . .
67 2 . . . . , − − .
2; 26 . .
78 37 . . − . . − , − . − . . . . − . . , − . − . .
95 2 . . − . . , − − . − . .
93 9 . . − . . − , − . − . .
57 1 . . − . . − , − − . − . .
54 7 . . − . . , − − . − . .
03 11 . . − . . , − − . − . .
56 2 . . − . . , − . − . .
18 0 .
15 0 . − . . , − . − . . . . − . . − , − . − . .
59 0 .
67 3 . − . . − , − . − . . . . − . . − , − . − . .
47 0 .
51 2 . − . . − , − . − . .
50 2 . . − . . , − − . − . .
54 1 . . − . . , − − . − . .
46 0 .
75 3 . − . . , − − . − . .
70 2 . . − . . − , − − . − . .
30 2 . . − . . − a − . − . .
39 10 . . − . . − a Notes. ( a ) These sources have poor uv-coverage, thus no position for the maser spots is given.22. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Table C.1.
Properties of each clump, averaged within the FWHM contour. The columns show respectively: clump name, mass,volume- and column-density of H , mass surface density, with their 68% credibility interval, and the beam-corrected diameters of theclumps. The clumps above the horizontal line are those classified as star-forming, while the clumps below it are those classified asquiescent. Clump M
68% int. n (H ) 68% int. N (H ) 68% int. Σ
68% int. Diameter Notes (10 × M (cid:12) ) (10 × cm − ) (10 × cm − ) (10 − × g cm − ) ( pc) − . . − . . . − . . . − . .
29 1 . − .
69 0 . − . . − . . . − . . . − . .
39 1 . − .
64 0 . − . . − . . . − . . . − . .
09 2 . − .
43 0 . − . . − . . . − . . . − . .
07 0 . − .
35 0 . − . . − . . . − . . . − . .
10 12 . − .
80 0 . − . . − . . . − . . . − . .
90 0 . − .
08 0 . − . . − . . . − . . . − . .
77 1 . − .
38 0 . − . . − . . . − . . . − . .
11 0 . − .
28 1 . − . . − . . . − . . . − . .
78 1 . − .
62 0 . − . . − . . . − . . . − . .
19 1 . − .
61 0 . − . . − . . . − . . . − . .
06 3 . − .
68 0 . − . . − . . . − . . . − . .
63 2 . − .
43 0 . − . . − . . . − . . . − . .
89 2 . − .
37 0 . − . . − . . . − . . . − . .
25 1 . − .
47 0 . − . . − . . . − . . . − . .
46 1 . − .
69 2 . − . . − . . . − . . . − . .
58 2 . − .
03 0 . − . . − . . . − . . . − . .
81 0 . − .
95 1 . − . . − . . . − . . . − . .
47 2 . − .
49 0 . − . . − . . . − . . . − . .
83 1 . − .
07 0 . − . . − . . . − . . . − . .
21 1 . − .
38 0 . − . . − . . . − . . . − . .
25 6 . − .
84 0 . − . . − . . . − . . . − . .
47 1 . − .
91 0 . − . . − . . . − . . . − . .
65 0 . − .
00 0 . − . . − . . . − . . . − . .
13 1 . − .
38 0 . − . . − . . . − . . . − . .
15 0 . − .
33 0 . − .
15 0 . − . . . − . . . − . .
31 0 . − .
37 0 . − . . − . . . − . . . − . .
13 1 . − .
86 0 . − . . − . . . − . . . − . .
01 0 . − .
22 0 . − . . − . . . − . . . − . .
25 0 . − .
40 1 . − . . − . . . − . . . − . .
56 1 . − .
07 0 . − . . − . . . − . . . − . .
14 0 . − .
45 1 . − . . − . . . − . . . − . .
45 0 . − .
55 1 . − . . − . . . − . . . − . .
08 1 . − .
57 1 . − . . − . . . − . . . − . .
37 0 . − .
69 0 . − . . − . . . − . . . − . .
43 0 . − .
78 0 . − . . − . . . − . . . − . .
56 1 . − . < . Notes. (1) Source with very poor uv-coverage in NH : spectrum extracted from non-cleaned maps.(2) Lower limit for the mass and the derived quantities.(3) Volume-, column- and surface-densities are lower limits, as the source is unresolved. 23. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Table C.2.
Properties of each clump, averaged within the 3 σ contour. The columns show respectively: clump name, mass, volume-and column-density of H , mass surface density, with their 68% credibility interval, and the diameters of the clumps. The clumpsabove the horizontal line are those classified as star-forming, while the clumps below it are those classified as quiescent. Clump M
68% int. n (H ) 68% int. N (H ) 68% int. Σ
68% int. Diameter Notes (10 × M (cid:12) ) (10 × cm − ) (10 × cm − ) (10 − × g cm − ) ( pc) − . . − . . . − . . . − . .
87 0 . − .
98 0 . − . . − . . . − . . . − . .
53 0 . − .
63 1 . − . . − . . . − . . . − . .
85 0 . − .
10 2 . − . . − . . . − . . . − . .
46 0 . − .
58 0 . − . . − . . . − . . . − . .
20 1 . − .
40 1 . − . . − . . . − . . . − . .
51 0 . − .
57 1 . − . . − . . . − . . . − . .
91 0 . − .
10 1 . − . . − . . . − . . . − . .
57 0 . − .
61 2 . − . . − . . . − . . . − . .
30 0 . − .
60 2 . − . . − . . . − . . . − . .
65 0 . − .
73 1 . − . . − . . . − . . . − . .
60 0 . − .
63 1 . − . . − . . . − . . . − . .
64 0 . − .
76 1 . − . . − . . . − . . . − . .
92 0 . − .
10 1 . − . . − . . . − . . . − . .
71 0 . − .
77 1 . − . . − . . . − . . . − . .
51 0 . − .
56 5 . − . . − . . . − . . . − . .
97 0 . − .
10 1 . − . . − . . . − . . . − . .
65 0 . − .
76 2 . − . . − . . . − . . . − . .
60 0 . − .
77 1 . − . . − . . . − . . . − . .
53 0 . − .
62 2 . − . . − . . . − . . . − . .
39 0 . − .
45 1 . − . . − . . . − . . . − . .
30 1 . − .
50 0 . − . . − . . . − . . . − . .
75 0 . − .
83 0 . − . . − . . . − . . . − . .
66 0 . − .
80 1 . − . . − . . . − . . . − . .
85 0 . − .
86 1 . − . . − . . . − . . . − . .
71 0 . − .
77 1 . − . . − . . . − . . . − . .
24 0 . − .
27 0 . − . . − . . . − . . . − . .
20 0 . − .
30 1 . − . . − . . . − . . . − . .
85 0 . − .
99 0 . − . . − . . . − . . . − . .
99 0 . − .
10 1 . − . . − . . . − . . . − . .
73 0 . − .
83 1 . − . . − . . . − . . . − . .
89 0 . − .
10 1 . − . . − . . . − . . . − . .
44 0 . − .
51 1 . − . . − . . . − . . . − . .
71 0 . − .
86 2 . − . . − . . . − . . . − . .
86 0 . − .
99 1 . − . . − . . . − . . . − . .
84 0 . − .
05 1 . − . . − . . . − . . . − . .
94 0 . − .
10 0 . Notes. (1) Source with very poor uv-coverage in NH : spectrum extracted from non-cleaned maps.(2) Lower limit for the mass and the derived quantities.24. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Table C.3.
Summary of star formation signposts for all the observed clumps. The clumps above the horizontal line are those classifiedas star-forming, while the clump below it are those classified as quiescent. For the “green fuzzies”, the notation Y + indicates multiple“green fuzzies” associated with the clump. In the last column we show the Type of the source (see text). Clump 24 µ m Emission MSX Emission GF 1 . O Maser Type08477 − − − − − − − − − − + N N 115470 − − + N Y 115557 − c − − + Y Y 216061 − + Y Y 216061 − + Y Y c − − + N N d − − b N N 116164 − − d − − − a − − a − a . . .17195 − a . . .17355 − − − − − − − − − − − − − − − − Notes. ( a ) These sources have poor uv-coverage, thus a firm assignment to a clump in the region was not possible. ( b ) Visual inspection of theMid-IR images showed that this GF could be generated by the Spitzer PSF. However, we still consider the clump as star forming to be conservative,and because of the presence of a strong 24 µ m source within the 3 σ contour of the 1 . ( c ) Sourcedetected in our work, but not in that of S´anchez-Monge et al. (2013a). ( d ) Source detected in the work of S´anchez-Monge et al. (2013a), but not inour work. 25. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution T a b l e C . . M ea n a nd m e d i a nv a l u e s f o r t e m p e r a t u r e s , l u m i no s it y , du s t e m i ss i v it y i nd e x β , L / M r a ti o , g a s p r op e r ti e s a ndd i a m e t e r s o f t h ec l u m p s i n t h e Q S a nd t h e SFS a v e r a g e d w it h i n t h e F W H M c on t ou r o f t h e . mm c on ti nuu m e m i ss i on . T h e p a i r s o f c o l u m n ss ho w r e s p ec ti v e l y : k i n e ti c t e m p e r a t u r e , du s tt e m p e r a t u r e , m a ss , d i a m e t e r , c o l u m nd e n s it y , vo l u m e d e n s it y , s u rf ace d e n s it y , l u m i no s it y , du s t e m i ss i v it y i nd e x a nd L / M r a ti o . P a r a m e t e r T K T d M D N H n H Σ L β L / M ( K )( K )( M (cid:12) )( p c )( × c m − )( × c m − )( g c m − )( × L (cid:12) )( L (cid:12) / M (cid:12) ) SFS Q SSFS Q SSFS Q SSFS Q SSFS Q SSFS Q SSFS Q SSFS Q SSFS Q SSFS Q S M ea n19 . + . − . . + . − . . + . − . . + . − . + − + − . . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . M e d i a n18 . . . . . . . . . . . . . . . . . . T a b l e C . . M ea n a nd m e d i a nv a l u e s f o r t e m p e r a t u r e s , l u m i no s it y , L / M r a ti o , g a s p r op e r ti e s a ndd i a m e t e r s a v e r a g e d w it h i n t h e F W H M c on t ou r o f t h e . mm c on ti nuu m e m i ss i on o f t h ec l u m p s i n t h e SFS , d i v i d e d i n SFS - a nd SFS - . T h e p a i r s o f c o l u m n ss ho w r e s p ec ti v e l y : k i n e ti c t e m p e r a t u r e , du s tt e m p e r a t u r e , d i a m e t e r , c o l u m nd e n s it y , vo l u m e d e n s it y , s u rf ace d e n s it y , l u m i no s it y a nd L / M r a ti o . P a r a m e t e r T K T d D N H n H Σ LL / M ( K )( K )( p c )( × c m − )( × c m − )( g c m − )( × L (cid:12) )( L (cid:12) / M (cid:12) ) SFS - SFS - SFS - SFS - SFS - SFS - SFS - SFS - SFS - SFS - SFS - SFS - SFS - SFS - SFS - SFS - M ea n21 . + . − . . + . − . . + . − . . + . − . . . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . M e d i a n21 . . . . . . . . . . . . . . . .
26. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Table C.6.
Parameters of the ammonia spectra averaged over the whole NH (1,1) area of emission. The columns indicate the clumpname, the peak flux of the (1,1) transition and the rms of the spectrum, the V LSR of the emission, the ∆ V of NH (1,1), the peakflux of the (2,2) transition and the rms of the spectrum, and the ∆ V of NH (2,2), T rot , T K and ammonia column density, with theiruncertainties. The clumps above the horizontal line are those classified as star-forming, while the clump below it are those classifiedas quiescent (see Sect. 5). Note that, contrary to Table 2, we do not include clumps not detected in NH (1,1). S ou r ce F m a x ( , )r m s V L S R ∆ V ( , ) F m a x ( , )r m s ∆ V ( , ) T r o t % i n t . T K % i n t . N ( NH ) % i n t . ( m J y )( m J y )( k m s − )( k m s − )( m J y )( m J y )( k m s − )( K )( K )( K )( K )( × c m − )( × c m − ) - c . . . . . . . . . − . . . − . . . − . - c . . . . ... . ... −−−−−− - c . . − . . . . . . . − . . . − . . . − . - c . . . . ... . ... −−−−−− - c . . − . . . . . . . − . . . − . . . − . - c . . − . . ... . ... −−−−−− - c . . − . . ... . ... −−−−−− - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − .
19 10088 - c . . − . . ... . ... −−−−−− - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . ... . ... −−−−−− - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . . . . . . − . . . − . . . − . - c . . − . . ... . ... −−−−−− - c . . − . . . . . . . − . . . − . . . − . - c . . − . . ... . ... −−−−−− - c . . − . . ... . ... −−−−−− - c . . − . . ... . ... −−−−−− - c . . − . . . . . . . − . . . − . . . − .
27. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.1. (a) :Colorscale: SEST 1 . (1,1) zeroth moment. The contours start from 15% of the peak, in step of15% of the peak integrated emission. The peak flux and the ∆ V of the line are given in Table 2. The position of each clump in oursample is indicated by a black cross. The SEST beam is indicated as a white circle, and the ATCA beam is shown as a black circle.The IRAS name is indicated above each panel, and the clump names are shown in the figures. The fields with poor uv-coverage (seeSect. 3) are not used to produce overlays, as the maps produced are not as reliable as the others.
28. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.1. (b) : Colorscale: NH (1,1) zeroth moment; contours: SEST 1 .
29. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.2.
SEST 1 . / MIPS 24 µ m images. The IRAS name is indicatedabove each panel, and the clump names are shown in the figures. Dashed and solid red lines indicate the ATCA field and the observedclumps, respectively. H O maser spots are shown as white open squares, while “green fuzzies” are indicated as green open circles. Inthe last four panels the SEST emission is superimposed on MSX images. Purple lines indicate the 3 σ and FWHM polygons used toextract fluxes from the SEST images.
30. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.2.
Continued.
31. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.2.
Continued.
32. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.3. NH (1,1) and (2,2) spectra extracted at the peak of NH3(1,1) emission for all clumps. The clump name and transition areindicated above each panel.
33. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.3.
Continued.
34. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.3.
Continued.
35. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.3.
Continued.
36. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.3.
Continued.
37. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.4. H O maser spectra. The clump name and the o ff set of the maser w.r.t. the phase centre are indicated above each panel.
38. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.5.
Spectral energy distributions for the SFS. Uncertainties in the fluxes are indicated. The yellow-shaded area shows theregion encompassed by the extreme values of the fit-derived parameters for the modified black body.
39. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.5.
Continued.
40. Giannetti et al.: Physical properties of high-mass clumps in di ff erent stages of evolution Figure C.6.