EDP Sciences
Free Access
Issue
A&A
Volume 538, February 2012
Article Number A41
Number of page(s) 14
Section Interstellar and circumstellar matter
DOI https://doi.org/10.1051/0004-6361/201116774
Published online 30 January 2012

© ESO, 2012

1. Introduction

NGC 6357 is a complex of Hii regions and molecular clouds that form a very active star-forming region in the Sagittarius spiral arm. Optical, radio, and infrared (IR) images of NGC 6357 confirm that it contains a number of distinct Hii regions in different stages of evolution (e.g., Felli et al. 1990; Massi et al. 1997). Figure 1 shows a large cavity or a collection of smaller, connected cavities in the region, delineated by ionized gas. Weak and diffuse Hα emission permeates this feature. G353.2+0.9 is the bright emission region north of the cavity, seen in Fig. 1. Just 55″ south of G353.2+0.9, lies the massive open cluster Pismis 24 (hereafter Pis-24; Pişmiş 1959). This cluster is thought to be the main source of ionization of G353.2+0.9 (Massi et al. 1997; Bohigas et al. 2004). It contains at least ~20 early-type (OB) stars, plus 24 O-type candidates (Wang et al. 2007) and includes three stars that are amongst the brightest and bluest known in the Galaxy, of spectral types O3.5 III(f*), O3.5 If*, and O4 III(f+) (Maíz Apellániz et al. 2007). Massey et al. (2001) derived a distance of 2.56 kpc and an age of ~1 Myr for this cluster, and assuming that the molecular material is associated with Pis-24, we consider NGC 6357 to be at the same distance. The large cavity is unlikely to have been formed by Pis-24, because of its clearly off-centre position. The morphology and the size of the cavity seem to suggest that it was shaped by the winds and/or supernova events of one or more clusters (Wang et al. 2007) situated in the proximity of the centres of the smaller bubble-like structures.

Massi et al. (1997) performed a detailed study of the molecular emission associated with two of the Hii regions in NGC 6357, to wit: G353.2+0.6, and G353.2+0.9. The latter region is the younger one and exhibits signs of the presence of recently formed massive stars (e.g., ultra-compact Hii regions (UCHii), embedded sources with infrared (IR)-excess).

Our study is focused on G353.2+0.9. Felli et al. (1990) observed it with the VLA at λ = 6 cm with an HPBW of  (Fig. 7a of Felli et al. 1990). As for the Hα emission, the high-resolution interferometric radio continuum observations reveal a very complex structure of the ionized gas, with a well-defined sharp boundary running east-west (the “Bar”, in Fig. 2). The emission is characterized by a strong intensity gradient to the south, while showing a more gentle decrease to the north (Fig. 8 in Felli et al. 1990). Felli et al. (1990) found three UCHii (A, B, and C in Fig. 2).

The Ks band image in Fig. 2 shows that in the central part of the nebula there is an elephant trunk-like region of obscuration (clearly visible also in HST images; Hester & Desch 2005), with an UCHii region and IR source at its apex. This source shows a near-IR excess and X-ray emission: it was identified from HST observations to be in the evaporating gaseous globule (EGG) evolutionary phase (Hester et al. 1996), making it the first X-ray emitting EGG. This embedded object was classified as having a spectral type B0-B2 (Wang et al. 2007). The elephant trunk points toward Pis-24 and is thought to be formed by the radiation and stellar winds from the OB stars in this cluster. The IR emission is brightest along the sides of the trunk and on the south-western side of G353.2+0.9, facing Pis-24.

Massi et al. (1997) mapped G353.2+0.9 in CO(1–0) and 13CO(1–0). These data were complemented with observations of other molecules and transitions along strips in the north-south direction, to determine variations in physical parameters across the photon-dominated region (PDR). Massi et al. (1997) found that G353.2+0.9 is a face-on, blister-type Hii region, with most of the molecular material behind the Hii region and to the north of it.

thumbnail Fig. 1

The 8 μm emission image of NGC 6357, taken from the GLIMPSE survey (http://www.astro.wisc.edu/sirtf/, Benjamin et al. 2003). The cavity is clearly visible at the centre of the image. G353.2+0.9 is the bright region at its northern border. The location of Pis-24 is also shown in the figure. The coordinates are referred to the epoch J2000.

Open with DEXTER

Felli et al. (1990) suggested that G353.2+0.9 is not associated with Pis-24, arguing that the southern sharp boundary (the “Bar”) is produced by local ionization caused by embedded sources, that are also responsible for the radio emission of G353.2+0.9. They also concluded that the nebula is ionization-bounded to the south, implying that there are considerable quantities of molecular gas in the region south of the ionization front. Molecular-line observations do not however support this claim: Massi et al. (1997) found very little molecular emission at the location of the “Bar”.

Bohigas et al. (2004) found that this elongated structure has to have a considerable extent along the line-of-sight (1 − 5 pc, Bohigas et al. 2004). While this dimension is comparable to the extent in RA, it is much larger than the extent in DEC. This suggests that the “Bar” is a layer of ionized matter seen edge-on. It could be caused by the interaction of the photoionized photoevaporative flow with the free wind of the Pis-24 stars (Healy et al. 2004). This implies that the molecular gas in the region has already been swept by the stellar winds, i.e. the region south of the “Bar” should be nearly devoid of molecular material.

The present study follows up on the work described in Massi et al. (1997), which constitutes a first step in the study of the interface between the Hii region and the molecular cloud. Our aim is to clarify the morphology of the region, by observing optically thin molecular lines (e.g. C18O), and to confirm the absence of molecular material south of the main ionization front, which is identified as IF in Fig. 2 (see Sect. 3.11). This would strengthen the hypothesis of an association between G353.2+0.9 and Pis-24. The physical conditions of the gas are derived by means of a non-LTE analysis for those molecules with several observed transitions or with hyperfine structure, while for the remaining molecules we assumed LTE (local thermodynamic equilibrium). The observation of the continuum at 870 μm allows us to infer the dust column density and mass, and thus to determine those of the gas, by assuming a gas-to-dust ratio. With the H2 column densities determined in this way, we were able to calculate the abundance for the observed molecules.

thumbnail Fig. 2

Ks-band image of G353.2+0.9 obtained with SofI (Massi et al., in prep.). The actual ionization front is indicated by “IF” (see text). The locations of the “Bar”, Pis-24, and the elephant trunk are indicated. The red and green squares mark the location of some early-type stars in Pis-24 and the three UCHii regions identified by (Felli et al. 1990), respectively. The coordinates are referred to the epoch J2000.

Open with DEXTER

2. Observations and data reduction

The molecular-line observations were carried out between Sept. 1 and 9, 1999, with the 15-m Swedish-ESO Submillimeter Telescope (SEST; program 63.I-0189). This radio telescope was operational on La Silla in the period 1987–2004. The telescope was equipped with SIS receivers and a high-resolution acousto-optical spectrometer with a total bandwidth of 86 MHz, and a frequency resolution of about 42 KHz. The spectrometer was split into two parts, and we observed with two receivers simultaneously, one at a lower and the other at a higher frequency [e.g., C18O(1–0) and C18O(2–1)]. Detailed information about the molecular transitions observed are reported in Table 1. The columns indicate the molecules and transitions observed, their rest frequency, resolution in frequency and velocity, beam FWHM, main beam efficiency, spacing between points in the maps, and typical TMB rms noise per channel of the spectra, respectively.

Table 1

Molecular transitions observed.

Most observations were made in frequency-switching mode, with a switch-interval in frequency sufficiently small to make the emission appear in both the signal and reference cycles, but large enough to avoid overlap between them.

The SiO(5–4), CN(1–0), CN(2–1), and CH3CCH(6–5) lines were observed with position switching. The emission of the last two molecules exhibits hyperfine structure, making it necessary to observe in position-switch mode. The telescope pointing was checked every three hours on the nearby SiO maser source AH Sco, and was found to be accurate to within 5″. The same source was also used as the off-position. The calibration was made using the standard chopper-wheel method described in Kutner & Ulich (1981). Our molecular-line maps are centred on , (J2000), coinciding with the “Bar”. The angular extent of the observed region is about 5′ × 5′ for CS(2–1) and (3–2), while it is ~3′ × 3′ for the other molecules and transitions, with spacing between the raster point listed in Table 1 (Col. 7). The line intensities in this paper are expressed in terms of the main beam temperature, defined as . Data reduction and analysis for molecular-line data were performed with CLASS, part of the GILDAS (Grenoble Image and Line Data Analysis Software1) package.

The morphology and the distribution of the molecular gas can be investigated in greater detail by decomposing the emission profile into single Gaussian components at different VLSR. Given the limited number of velocity components, we decided to decompose the emission profiles at every position by fitting different Gaussian curves, starting from optically thin transitions (e.g. C18O). The characteristics of the components identified in this way were then used as a template for the decomposition of the emission profiles of the other molecules and transitions.

CLASS offers the possibility to fit lines with hyperfine structure, such as those of CN, by specifying the relative intensity of the hyperfine components in the case of optically thin emission and assuming that their ratios have their LTE values. CLASS uses the optical depth as a free parameter of the fit, and gives it as output of the procedure.

We retrieved a map of G353.2+0.9 at 870 μm (345 GHz), taken with APEX from the ATLASGAL survey (Schuller et al. 2009). The rms noise in the map is ~100 mJy/beam, determined in three regions free of emission around the Hii region. These data were analyzed with MOPSIC, the evolution of MOPSI (Map On-off Pointing Skydip Image), which was developed by Zylka (Obs. de Grenoble).

3. Results and discussion

3.1. Morphology

Figure 3 shows the maps of integrated line-emission towards G353.2+0.9. The molecular emission never extends significantly below Δδ = 0″, confirming the lack of molecular material south of the “Bar” and ruling out the possibility that this feature is an ionization front proceeding southward.

The molecular emission is concentrated between −10 and +1 km s-1 with, in some cases, strongly varying emission profiles between adjacent positions, as clearly visible in Fig. 4. In Fig. 3, it is possible to identify many different clumps. In higher-frequency transitions, such as CS(5–4) and CN(2–1), several clumps are resolved into two or more smaller clumps, or show an elongated appearance.

thumbnail Fig. 3

Maps of the integrated emission of the different molecular species and transitions, within the whole velocity range of emission. The first contour is the 3σ level. Molecule, transition, and integration limits are indicated above each map. The beam size is indicated by the filled circle. The observed positions are marked with a cross. The contours levels (in units of K km s-1) are, respectively (lowest (step) highest): a) 0.41 (1.0) 12.41; b) 0.51 (1.0) 11.51; c) 0.89 (2.0) 12.89; d) 0.4 (0.4) 2.4; e) 0.47 (0.7) 7.47; f) 0.81 (2.0) 16.81; g) 0.98 (3.0) 27.98; h) 1.25 (2.0) 19.25; i) 0.92 (1.5) 15.92. Coordinates are offsets (arcsec) with respect to (J2000).

Open with DEXTER

We found that the emission can be separated into 14 clumps in six velocity ranges (see Fig. 5; all in units of km s-1), i.e. −7.3 ≲ VLSR ≲  −6.1 (A and P), −6.1 ≲ VLSR ≲  −4.7 (B, C and L), −4.7 ≲ VLSR ≲  −3.3 (D and O), −3.3 ≲ VLSR ≲  −1.9 (E and F), −1.9 ≲ VLSR ≲  −0.4 (G and H), and −0.4 ≲ VLSR ≲  +0.8 (I, M and N). The names of the clumps do not correspond to those used in Massi et al. (1997), because of the higher spatial resolution in our present work and a different method of analysis (Gaussian decomposition versus integrated emission in velocity bins).

Figure 5 shows that the various clumps have slightly different position for different molecules and transitions, which could be the result of different excitation conditions, optical depth, chemical, or resolution effects. Furthermore, some clumps (such as P and O) are clearly visible only in high-density, high-frequency tracers. This is particularly clear in CS (5–4), which has the highest resolution. The emission of these tracers shows the presence of multiple high-density, small cores within a larger clump.

The clumps along the ionization front tend to have redder velocities than the others. This is especially true for low-density tracers (cf. Fig. 5). Clumps N, H, and F, and D, B, and A exemplify this behaviour. This can be understood by taking into account the radiative and mechanical action of the stars of Pis-24: when neutral gas is exposed to the intense energetic radiation of an early-type star, it becomes rapidly ionized near the surface of the cloud. This gas is heated to T ~ 10 000 K (a factor of ~100 with respect to cold, neutral gas in the cloud), consequently causing a comparable increase in pressure, thus leading to a rapid expansion. However, the expansion towards the neutral gas is stopped by the presence of the dense material of the cloud. In the opposite direction, the low-density ionized gas cannot halt the expansion of this overpressurized gas. The ionized material moves predominantly away from the cloud, having an equal and opposite effect on the cloud.

Therefore, the molecular emission indeed suggests that the ionized, overpressurized gas in G353.2+0.9, observed by Bohigas et al. (2004), is expanding, thus pushing the molecular material away from the observer.

thumbnail Fig. 4

TMB(C18O(2–1)) spectral map. Each frame corresponds to a single pointing with an offset in α and δ indicated in the figure. The x and y axes of each spectrum range from − 20 to 15 km s-1 and from − 0.2 to + 9 K, respectively.

Open with DEXTER

thumbnail Fig. 5

a) Integrated emission of C18O(1–0) superimposed on the Ks image. Each plot corresponds to a different VLSR. The VLSR of the clumps is indicated in the respective panel. The dashed (blue) lines indicate the observed area. The (red) contours show the integral under Gaussian components fitted to the line profiles. The beam of the transition is shown in the first panel. The first contour in each panel is the 3σ level. The integrated emission levels (in units of K km s-1) are (lowest (step) highest): ( − 6.7 km s-1) 0.17 (0.3) 2.27; (–5.5 km s-1, C) 0.19 (0.3) 1.99; ( − 5.5 km s-1, B) 0.20 (0.5) 4.70; ( − 4.0 km s-1) 0.18 (0.7) 6.48; ( − 2.5 km s-1) 0.19 (0.3) 2.89; ( − 1.0 km s-1) 0.18 (0.2) 1.78; ( + 0.2 km s-1) 0.15 (0.1) 0.65. b) As a), but for CS(5–4). The integrated emission levels (in units of K km s-1) are (lowest (step) highest): ( − 6.7 km s-1) 0.19 (0.3) 2.59; ( − 5.5 km s-1, C) 0.20 (1.0) 10.20; ( − 5.5 km s-1, B) 0.19 (0.5) 3.69; ( − 4.0 km s-1) 0.19 (0.7) 4.39; ( − 2.5 km s-1) 0.18 (1.0) 8.18; ( − 1.0 km s-1) 0.17 (0.2) 1.37; ( + 0.2 km s-1) 0.14 (0.1) 0.74. Note that names of the clumps do not correspond to those used in Massi et al. (1997).

Open with DEXTER

3.2. Temperatures

The first method used to derive the excitation temperatures uses the ratio of main beam temperatures of the (2–1) to (1–0) transitions of C18O. We first resampled the spectra to the same velocity resolution, and then convolved the C18O(2–1) data with a Gaussian beam to match the spatial resolution of the (1–0) transition.

We calculated the ratio R1 of TMB(2–1) to TMB(1–0) and determined Tex, assuming optically thin emission, from (Levreault 1988) (1)where Tex,21 is the excitation temperature of the C18O(2–1) transition. Assuming LTE, Tex is the same for all levels. Since both transitions were observed simultaneously, the systematic calibration uncertainties should not affect the temperature ratios, which are instead influenced by uncertainties in the Gaussian decomposition of the emission profile. Taking into account the uncertainties in the Gaussian fit performed with CLASS, we estimated that the uncertainties in TMB should not exceed 5−15%, where both lines are detected above 3σ. Assuming a typical uncertainty in TMB of 10%, the relative uncertainty in Tex is ~ 14%. We found that R1 is always between ~2 and ~4, indicative of optically thin emission throughout the region. The Tex(C18O) (derived from Eq. (1)) is fairly uniform, typically between ~15 K and ~25 K for all clumps. The higher values are found along the ionization front, while for clump C, which is associated with the elephant trunk, we have Tex ~ 20 K.

Another temperature probe is methyl acetylene (CH3CCH) (Bergin et al. 1994). This molecule has a high critical density, and its emission comes from high-density regions. Our observations are limited to four positions: (0″, 50″) (clump C), (75″, 75″) (clump E), (−75″, 125″) (clump B), and (0″, −25″). We detected CH3CCH only at (0″, 50″) and (75″, 75″), while the detection at ( − 75″, 125″) is uncertain. Only at the first position were four components of the K-ladder visible, giving a reliable temperature estimate. The Boltzmann plot analysis (Fig. 6) gave a Tex ~ 45 K for clump C, which is higher than that derived from C18O. At the other positions, we detected just two components, leading to very uncertain temperature estimates (22 K, clump E; 45 K, clump B) that, however, are roughly consistent with those determined from CS (see Sect. 3.8). This higher temperature could be due to the presence of internal heating sources in clumps B and C.

Although the excitation temperature is formally only a measure of the relative population of the energy levels in a transition, and therefore differs for different transitions, under the assumption of LTE, it provides a fair estimate of the kinetic temperature.

3.3. Column densities

Assuming LTE conditions, the total column density of the molecular gas can be obtained from observations of the J → J − 1 transition by means of the expression (Zielinsky 1999; Kramer & Winnewisser 1991) (2)where J(T) = (hνJ,J − 1/k)(ehνJ,J−1/kT − 1)-1, ηc is the efficiency with which the antenna couples to the source, μ is the dipole moment of the molecule, Z is the partition function, TBG is the background temperature, and TMB is the main beam temperature.

thumbnail Fig. 6

Boltzmann plot obtained from the CH3CCH data at (0″, 50″).

Open with DEXTER

Equation (2) implies that there is a linear relation between the column density and the integrated line intensity at fixed Tex. This equation holds only in case of negligible optical depth (τ). However, from the detection equation (3)one can derive a correction factor τ/(1 − e − τ), which makes it possible to calculate the column density while τ ≲ 2, with uncertainties less than 15% (Kramer 1988). In this expression, τ is the optical depth at the line centre (see Sects. 3.4 and 3.8).

In Table 2, we list the average value of the column density inside the 3σ contour of integrated intensity, derived from different molecules and transitions, for each distinct identified component. The derived excitation temperature is also listed, but we note that the column densities are computed assuming Tex = 20 K for all molecules and a τ derived according to Sect. 3.4. The assumption of a constant excitation temperature does not affect the results, because its range is small (15 − 25 K) and the variation in column density is only of the order of ≲ 15 − 20%.

3.4. Opacities and visual extinctions

We can derive the opacities of the transitions of C18O, from the detection Eq. (3). The C18O lines are usually optically thin (τ ~ 10-2), although they reach τ ~ 0.1 and τ ~ 0.2, respectively for the transition (1–0) and (2–1) in clump E, in the elephant trunk and in some of the clumps aligned with the bright emission to the west of the trunk, i.e. the ionization front. Assuming the standard value of ~8 for the abundance ratio X(13CO)/X(C18O), one derives values of τ for the of the order of 0.8 and 1.6, respectively, for the transition (1–0) and (2–1) at the same positions. This result confirms that the emission of is marginally thick, as argued by Massi et al. (1997). The higher optical depth indicates that the molecular material has accumulated in these regions. Furthermore, non-negligible opacities could explain the slightly lower Tex south of the clump C, with respect to those observed for the ionization front.

An estimate of the visual extinction can be obtained from the H2 column densities from the expression (Bohlin et al. 1978) (4)Typical values are 5−10 mag and 15−20 mag for C18O and H2CO, respectively. The maximum values of AV that we found, estimated from H2CO, were those of clumps C and E (~50 mag).

thumbnail Fig. 7

APEX image at 870 μm. The scale is expressed in Jy/beam. The APEX beam is shown as a filled circle. The locations of the main components identified are shown in the map. Component 5 was not fitted with a Gaussian, but was considered as a region of diffuse emission.

Open with DEXTER

3.5. Dust

From the 870 μm image, kindly provided by the ATLASGAL project2, we derived dust masses and densities, following Deharveng et al. (2009)(5)where S870 μm is the total flux density at 870 μm, D is the distance, B870 μm(Td) is the emission of a blackbody with temperature equal to Td at 870 μm, and κν ≡ κ0(ν/ν0)β is the dust opacity per unit mass at the indicated frequency. Ossenkopf & Henning (1994) recommend κ0 = 0.8 cm2 g-1 at 230.6 GHz, from which we derived κ870  μm ~ 1.8 cm2 g-1, assuming β = 2 (Hildebrand 1983). We assumed that the dust temperature is 30 K throughout the region. However, the resulting masses and column densities do not depend strongly on temperature, for Td = 20 K and Td = 50 K the difference is a factor of three. The dust mass resulting from the total 870 μm flux is ~21 M. κ870 μm has an uncertainty of a factor of two, which implies an uncertainty in the dust mass of the same factor. Furthermore, the uncertainty due to the dust temperature is of the same order of magnitude.

The surface brightness Fν at 870 μm also allows one to derive the gas column density. Following Deharveng et al. (2009)(6)F870 μm is the peak value of the surface brightness, Ωbeam is the solid angle covered by the beam, mH is the mass of a hydrogen atom, and we assumed a gas-to-dust ratio γ = 100. The contribution of the free-free emission, extrapolated from the radio data of Felli et al. (1990) assuming optically thin emission, is less than a few percent. The rms noise of the map (~100 mJy/beam) corresponds to a column density of 3.1 × 1021 cm-2 for T = 20 K, 1.9 × 1021 cm-2 for T = 30 K, and 1.0 × 1021 cm-2 for T = 50 K.

Table 2

VLSR, position, mean LTE column density, excitation temperature (C18O), FWHM of the lines, and diameter of the clumps.

We decomposed the emission into four bi-dimensional Gaussian components with MOPSIC, plus three regions of diffuse emission: Fig. 7 shows the condensations identified.

The morphology of the emission at this wavelength is very similar to that of the integrated molecular emission and to that of the regions of obscuration visible at near-IR wavelengths. The maxima in the molecular-line emission do not always coincide with those of the dust. In particular, clump E shows a large displacement. We suggest that this could be an effect of a non-negligible τ.

To associate the dust components with molecular cores, we superimposed the map of the 870 μm emission on that of the molecular emission, as shown in Fig. 8. The masses of the components identified in the APEX image and their associated molecular clumps are listed in Table 3. The total mass indicated in this table is computed by integrating the total 870 μm flux. The H2 column densities derived from the dust emission are on average in the range 3−7 × 1022 cm-2, while the maximum values are ~1023 cm-2 (components 1, 2, and 4).

thumbnail Fig. 8

CS(5–4) emission (black and gray contours) superimposed on the 870 μm emission. The beam size of CS(5–4) is indicated by the filled circle, while the open circle indicates the APEX beam size. Each plot corresponds to a different VLSR. The VLSR of the clumps is indicated in the respective panel. The contours (in units of K km s-1) for each clump are (lowest (step) highest): (−6.7 km s-1) 0.19 (0.3) 2.59; (−5.5 km s-1, C) 0.20 (1.0) 10.20; (−5.5 km s-1, B) 0.19 (0.5) 3.69; (−4.0 km s-1) 0.19 (0.7) 4.39; (−2.5 km s-1) 0.18 (1.0) 8.18; (−1.0 km s-1) 0.17 (0.2) 1.37; and (+0.2 km s-1) 0.14 (0.1) 0.74. The lowest contour corresponds to the 3σ level in .

Open with DEXTER

The visual extinctions corresponding to the maximum column densities are typically 100 mag. The average visual extinction lies between 15 and 40 mag, depending on the component. The corresponding mean and maximum volume densities, calculated assuming spherical symmetry for the clumps, are ~104 cm-3 and ~105 cm-3. These values are of the same order as those found from H2CO, under the assumption of LTE. The low-density layers of the clumps are not visible in the 870 μm map owing to its high rms noise.

thumbnail Fig. 9

Maps of the abundance relative to H2 of 13CO, C18O, and H2CO, derived from 13CO(1–0) (top left), C18O(1–0) (top right), C18O(2–1) (bottom left), and H2CO(21,2-11,1) (bottom right). The green contours shows the column density distribution of the molecules.

Open with DEXTER

3.6. Molecular abundances

In Fig. 9, we show the abundances of 13CO (not in the present dataset, taken from Massi et al. 1997), C18O and H2CO as a function of position. To derive these maps, we smoothed the APEX image to match the resolution of the molecular transition considered and we divided pixel per pixel the molecular column density map by the H2 column density map, obtained from the APEX image, using Eq. (6). We set a threshold of 3σ in both molecular and H2 column densities. We can clearly see that the abundances vary in the region, in a similar way for the three molecules. The abundances are lower near the IF and the elephant trunk, while they appear to increase further away from IF, away from Pis-24. The C18O abundance varies from ~9.0 × 10-8 in the region of the IF, to ~1.9 × 10-7 at the location of E, to ~2.3 × 10-7 more to the north, at roughly the location of D. In the same three regions, the abundances of 13CO and H2CO are ~1.0 × 10-6, ~1.5 × 10-6, and ~2.2 × 10-6; and ~6.0 × 10-10, ~1.5 × 10-9, and ~1.1 × 10-9, respectively. We expect that this variation is at least partially due to an increase in TK at the location of the ionization front. A different TK can modify the abundance through the H2 column density, which was derived from the 870 μm emission.

3.7. LTE masses and volume densities

The masses of the clumps are listed in Table 4. They were derived averaging the emission inside the 3σ contour to determine the average column density and then integrating over the beam-corrected area of the emission. We also took into account the optical depth τ (determined from the detection equation) and a correction for helium, which contributes with a factor 1.36. The total mass of the complex is ~2000 M, in agreement with that found from the 870 μm continuum emission. This mass can be compared to the total mass of ionized hydrogen (290 M, Bohigas et al. 2004) and to the total dust mass (21 M, see Sect. 3.5). The masses of single clumps range from a few tens to several hundreds of M.

The volume density n is determined assuming that the extent of the clump along the line-of-sight is equal to its size in the plane of the sky. The ratio of volume densities determined from low- and high-density tracers may be up to a factor of ~100, e.g., the mean volume densities derived from C18O are typically ~103 cm-3, while those of H2CO are ≈104 − 105 cm-3. The volume densities derived in this way are always of the same order of magnitude of the critical density of the molecular transition used to determine it.

3.8. Non-LTE analysis

The three transitions of CS, plus C34S(2–1), allowed a non-LTE analysis at offsets where data points were available for all transitions. For this purpose, we used the statistical equilibrium, radiative transfer code RADEX3, for the approximation of a uniform sphere. The model requires as input the kinetic temperature, the H2 number density, the molecular column density, and the FWHM of the line. We assumed the Solar value of 22.5 for the 32S/34S isotopic ratio. The model returns the brightness temperature of the lines, the opacity, and the excitation temperature. The FWHM of the CS line was determined by averaging those of the various observed lines, at one position. The brightness temperature of the line was estimated using (7)where ϑB is the beam size and ϑS is the source size. We assumed that all the transitions come from the same region, estimating the source size from the mean FWHM dimension of clumps in CS(5–4), that has the highest angular resolution. We varied the kinetic temperature between 10 K and 200 K, and the molecular hydrogen number density between 103 cm-3 and 107 cm-3, both in 50 equally spaced logarithmic steps. The column density was varied between 5.2 × 1012 cm-2 and 5.5 × 1015 cm-2 for CS and between 2.3 × 1011 and 2.4 × 1014 cm-2 for C34S, in 44 equal logarithmic steps.

To analyze the model results, we used a Bayesian approach. The Bayes theorem states that (8)

Table 3

Gas masses derived from dust emission.

where P(TK,N,n|D,m) is the probability of the parameters TK,N, and n, given the data and the model, called posterior, P(D|m,TK,N,n) is the probability of the data given the model and its parameters, or likelihood, and P(TK,N,n) is the probability of the model parameters, which is called prior. The parameter ϕ is a normalization constant given by the sum of the individual probability of each model, in order to have the posterior normalized to 1. We used a constant prior for the model parameters, thus giving equal weights to every value of the model parameters. The probability of measuring a certain value for the intensity of a line is assumed to be represented by a Gaussian curve centred on the value obtained from RADEX for specific physical conditions of the gas, and with a σ given by the uncertainty in the measured value. Therefore, we multiplied four Gaussian curves, one for each transition, and the probability density function (PDF) for τC34S, computed with JAGS4 assuming that CS(2–1) is optically thick, to obtain P(D|m,TK,N,n). The assumption of optical thickness for CS(2–1) is only used to derive the PDF for τC34S, but is not used further in the analysis of the model results. Explicitly, the expression for P(D|m,TK,N,n) is (9)where the index i runs over the three CS lines and C34S(2–1), li are the observed line intensities, μi are the modeled intensities, σl,i takes into account the rms of the spectrum and a 15% calibration uncertainty, and ϕ is a normalization constant.

Table 4

Single-clump and total gas masses.

To determine the physical conditions of the gas, i.e. the single parameters of the model, we had to integrate over the other parameters (marginalize) (10)From the PDF of the parameters, we derived the expectation values and the 1σ range. The results for CS are summarized in Table 5. The columns shows the offset of the spectrum used, the clump name, the kinetic temperature and its 1σ range, the number density of molecular hydrogen and its 1σ range, the column density and its 1σ range, the molecular abundance, the optical depth of CS(2–1), (3–2), (5–4) and C34S(2–1), and the excitation temperature of the same transitions, respectively.

From these analyses, we found that the volume density of molecular hydrogen ranges between several × 104 and few × 106 cm-3, while TK lies between ~11 K and ~45 K, for different clumps. The temperature derived from CH3CCH for clump C is higher than the kinetic temperature obtained from this analysis for (0′′, 50′′), while it is consistent with that at (0′′, 0′′), most probably because the CH3CCH beam takes in the emission from the region along the IF. However, CS(5–4) has its maxima at (0′′, 25′′) and at (25′′, 50′′) where we do not have data for CS(2–1) and (3–2), while at (0′′, 50′′) the emission of CS(5–4) is weak.

There is a clear trend showing an increase in the densities towards the south, in the direction of Pis-24 and in the clumps aligned with the IF (Fig. 2) west of the elephant trunk. This region shows typical densities of few × 105 cm-3 and TK ~ 30−40 K. The temperature for the point at (0′′, 0′′) appears to be higher than the surrounding points, with TK ~ 40 K, while (0′′, 50′′) has TK ~ 18 K. The column density ranges between ~2 × 1013 cm-2 and ~2 × 1014 cm-2, implying CS abundances of between 7.0 × 10-10 and 4.0 × 10-8, derived from the ratio with the H2 column densities determined from the APEX data. The opacity of CS(2–1), CS(3–2) and CS(5–4) lies between 0.5−4, 0.8−6, and 0.1−2.7, respectively.

The derived Tex-values are around 7−16 K for CS(2–1), 6−11 K for CS(3–2), 5−9 K for CS(5–4) and 6−13 K for C34S(2–1). The ratios of Tex for different transitions are always within a factor of two of each other.

A similar analysis was carried out for CN. In this case, P(D|m,TK,N,n) was calculated by comparing total fluxes rather than line temperatures, due to hyperfine splitting, which is not included in the present model. Therefore we have (11)where Fi is the measured flux, Fm,i is the output of the model, τtot,(1 − 0) is the optical depth of the (1–0) transition as measured from the hyperfine satellite ratios, and τm,(1 − 0) is that predicted by the model. The uncertainty σF,i takes into account the rms of the integral and a 15% calibration uncertainty, and στ is the uncertainty in τtot, as given by CLASS. We used a Gaussian prior on TK, centred on 35 K, with a σTK = 30 K, given the results of the analysis carried out for CS and the CO results in Massi et al. (1997). The parameter ϕ is the normalization constant. Also in this case, the column and the number densities are constrained quite well, while the temperature is much more uncertain.

The results obtained from RADEX for CN are summarized in Table 6. The columns shows the offset of the spectrum used, the clump name, the kinetic temperature and its 1σ range, the number density of molecular hydrogen and its 1σ range, the column density and its 1σ range, the molecular abundance, the optical depth of CN(1–0) and (2–1), and the excitation temperature of the same transitions, respectively. The values listed in Table 6 for τ10 are in very good agreement with those derived from the observations.

Owing to the poor constraints on TK, temperature maps are the most difficult to interpret. However, we obtain typical temperatures of between 25 K and 32 K for the whole region. The difference in TK found with CS at (0′′, 50′′) (18 K vs. 33 K) might be understood if one considers that CN is a good tracer of PDRs (Simon 1997), thus the emission may come from the more external material, directly heated and shocked by the interaction with the early-type stars of Pis-24. In all cases, these temperatures are usually within the 1σ interval also found for E and at the edges of D and B, nearly facing Pis-24.

In contrast, we derived a slightly lower temperature (~25 K) for C2 than for C1, which is nevertheless consistent with those cited above.

The clumps in the region have a typical number density in the range ~1−6 × 105 cm-3. The number density is highest along the IF (C1 and C2), in D and in E, with values up to a few × 106 cm-3. Clump E is particularly interesting, since it appears to have a resolved compressed layer facing Pis-24. The H2 number densities that we find are on average higher than those derived from CS, consistent with the idea that most of the emission comes from the high-density surface layers of the PDR.

thumbnail Fig. 10

Map of molecular abundance relative to H2, derived from CN data, through a non-LTE analysis. The green contours shows representative values of the column density, to clarify the distribution of the molecule.

Open with DEXTER

The column density of single clumps lies between ~5.0 × 1013 cm-2 and ~3.0 × 1014 cm-2.

After determining the column density for the various components, we were able to construct an integrated column density map, hence derive an abundance map similar to that of 13CO, C18O, and H2CO, as given in Fig. 10. The emerging abundance pattern closely resembles that of the other three molecules, indicating once more that the region of the elephant trunk and the brightest part of the IF is where the influence of Pis-24 is the strongest. The abundance of CN is ~5.0 × 10-9 in the region of C, ~8.0 × 10-9 around D, and ~7.9 × 10-9 around E. We found that CN is enhanced around clump C and the IF, where the NIR emission is at its strongest and the ionizing radiation more intense: this is suggested by a less pronounced variation in abundance than that found from other molecules. For CN, the column density peak of the gas and that of the dust for clump E coincide, showing that its displacement for the other molecules can be caused by optical depth effects (cf. Sect. 3.6) or that C18O and H2CO are frozen onto grains at the centre of the clump (given also the very low temperature derived from CS). However, a displacement is observed for the peak in the region of clumps A and B, again toward Pis-24. This and the temperatures comparable to those of C1 might indicate that the IF/PDR extends here, even though the brightness at NIR wavelengths is much lower than nearer Pis-24.

3.9. Virial masses

The total mass of a spherical system in virial equilibrium is given by (MacLaren et al. 1988) (12)where R is the radius and Δυ2 = 8ln2σ2, assuming a Gaussian velocity profile and a density profile described by a power law ρ(r) ∝ R − q with q < 3. The values of k2 are 210, 190, and 126, respectively for q = 0,1,2 (MacLaren et al. 1988). The uncertainty caused by the unknown density profile is approximately a factor of two. This expression neglects the influence of magnetic fields, rotation, and internal energy sources, which are usually non-negligible in molecular clouds.

To estimate R for Eq. (12), we used the “effective radius”, i.e. the radius of a circle with the same area as the clump above the FWHM level, corrected for the beam size. When the FWHM is smaller than the beam size, we assumed as an upper limit to the angular size, half of the actually observed FWHM. As a consequence of the large beam, some individual clumps may be blended and appear as one. For example, C is resolved into two different clumps in CS(5–4), and less clearly also in C18O(2–1), while it appears to be unresolved in the other transitions.

Figure 11 shows the virial parameter α = Mvir/MLTE as a function of MLTE, both determined from C18O(2–1), where we assumed that q = 2. The mass MLTE was derived within the FWHM contour in . All clumps with masses above 50 M, and also two with lower masses (M ~ 20 M G1 and O) have α ≈ 1, thus indicating that these clumps might be gravitationally bound. We note, however, that this is an oversimplification of the problem of stability and must be taken with caution.

thumbnail Fig. 11

Virial parameter α as a function of MLTE. Mvir and MLTE are both determined from C18O. MLTE is calculated within the FWHM contour in . The dashed line indicates α ~ 1, i.e. Mvir = MLTE.

Open with DEXTER

3.10. Selective photodissociation

G353.2+0.9 is illuminated by several early-type stars. These stars deliver a huge quantity of energetic photons that dominate the chemistry and the heating of the gas in the PDR. The incident far-ultraviolet (FUV) flux is usually measured in terms of the ratio of the FUV flux in the region and the mean interstellar flux (the Habing Field, 1.6 × 10-3 erg cm-2 s-1).

Considering the luminosities of the O3.5 If* [log (L/L) ~ 6], of the O3.5 III(f*) [log (L/L) ~ 5.9], and of the O4 III(f+) [log (L/L) ~ 5.8] stars given by Weidner & Vink (2010), we can assume that these stars dominate the emission of energetic photons in the region. The fraction of luminosity emitted in the FUV band was estimated using a simple black-body law, between 912 Å and 2067 Å. We obtained a total FUV luminosity of (13)and making use of (14)we found that G0 ~ 5.6 × 104, for a representative projected distance of the stars from the elephant trunk of 0.66 pc. On the other hand, for the ionization front, we obtained G0 ~ 2.0 × 104, using a projected distance of 1.12 pc between the IF and the cluster.

The high FUV flux may influence the ratio of C18O and 13CO, by means of selective photodissociation. This process tends to destroy the less abundant isotopologue, owing to the lower optical depth, thus less effective self-shielding capability.

We obtained the ratio of the transitions of the two isotopologues (13CO/C18O) for both observed transitions, combining 13CO and C18O values measured in positions less than half a beam apart. We corrected for the effects of opacity of the 13CO, evaluated through the detection equation (given in Eq. (3)). The ratio of the TMB of the lines should therefore be representative of the relative abundance of the two molecules. The 13CO/C18O ratio is everywhere consistent with the standard value of the relative abundance for the two isotopologues (~8), with three exceptions. The measured values of the 13CO/C18O ratio south of the ionization front, are a factor of two higher than the standard value. South of the elephant trunk, we did not detect C18O. However, assuming emission from C18O at rms intensity we obtain a lower limit to the ratio of 7 − 10. The model of Visser et al. (2009) shows that the relative increase in the 13CO/C18O ratio is about a factor 2–3, with the typical parameters of the region (AV ~ 5−10 mag, G0 ~ 104 − 105 and nH2 ~ 104 − 105 cm-3). The results reveal that selective photodissociation does indeed occur south of the ionization front, and there is an indication that selective photodissociation is also taking place south of the elephant trunk.

At the position of the elephant trunk, at the IF and for clump E the ratio is smaller than the standard value. In this case, the ratio may still be influenced by optical depth effects (in particular an underestimated opacity for 13CO, which was determined from Eq. (3), assuming optically thin emission).

3.11. Ionization front and geometry of the region

Neither the present study nor that carried out by Massi et al. (1997) found significant quantities of molecular material around the “Bar” (see Figs. 2 and 5). This rules out the possibility that the “Bar” is an ionization front eroding a molecular cloud. This also implies that we cannot exclude an association of G353.2+0.9 with Pis-24 based on the position of this feature (as argued by Felli et al. 1990 prior to the availability of molecular line maps of the region). Energetic, spectral, and excitation analyses (e.g. Massi et al. 1997; Bohigas et al. 2004) indicate that Pis-24 is indeed associated with G353.2+0.9 and that their proximity is not just a projection effect. The actual ionization front in G353.2+0.9 lies along the IR-bright feature labeled IF in Fig. 2 and is possibly associated with the UCHii region C (Felli et al. 1990).

We found a significant number of molecular cores along this bright ridge of emission and its continuation to the north-west, where the NIR brightness strongly diminishes. Furthermore, densities and temperatures are on average higher in this region. Our application of RADEX also revealed slight increases in TK, volume density of H2, and opacity here. There is intense radio continuum emission at 5 GHz (Felli et al. 1990) associated with this feature. This, together with the molecular gas distribution and physical conditions, confirm that IF is the main ionization front in G353.2+0.9. Here, the ionizing flux generated by the stars embedded in the region and by those of Pis-24, erodes the molecular cloud and pushes its material towards the north. When the resolution is high enough, one can see that the clumps near the ionization front, such as H, D, and C, are roughly parallel to it. High-density tracers are observed in rather small features at the edge of the ionization front, indicating the regions where the gas was compressed by the shock front.

Molecular emission strongly decreases south of this ionization front and no massive clumps are observed here, although some emission is visible, especially in low-density tracers (see Fig. 5). Bohigas et al. (2004) found that the region immediately surrounding the ionization front is characterized by a thin layer (10-3 pc) of very dense ionized material, where a photoevaporative flow is generated.

We detect only very faint emission from CN(2–1) along the “Bar”, which implies that there are very small clumps immersed in the PDR.

3.12. Pismis-24 13 (N36)

This star is located in the northern part of G353.2+0.9 (cf. Fig. 2) and is classified as a spectral type O6.5 V((f)) (Massey et al. 2001). Pismis-24 13 is worth noting because it seems to have produced its own Hii region in the molecular gas (see e.g., Fig. 3 in Hester & Desch 2005). This is confirmed by the radio continuum and ion-line observations (Bohigas et al. 2004), which reveal free-free emission following very well the outer edge of the cavity and an increase in electron density in coincidence with this feature. This Hii region appears to be in the foreground with respect to the elephant trunk and ionization front.

Dense star clusters often produce runaway OB stars with high radial velocities through two different mechanisms (Zinnecker & Yorke 2007): asymmetric supernova explosions and dynamical three-body encounters. The radial velocities of these objects exceed 40 km s-1. Gvaramadze et al. (2011) confirm that NGC 6357 is rich in OB runaway stars ejected from the clusters within the cavity. Pismis-24 13 could be one of these OB runaways, ejected from Pis-24 in the direction of the molecular clouds, where its radiation and wind then created the observed Hii region. To confirm this hypothesis we need a spectral measurement of the radial velocity of the star.

4. Summary

Table 5

Summary of RADEX results from CS data, for selected offsets.

Table 6

Summary of RADEX results from CN data, for selected offsets.

We have observed the Galactic Hii region G353.2+0.9 in several molecular lines, which has allowed us to distinguish at least 14 clumps in its associated molecular cloud. We have determined temperatures, densities, and masses of each clump. We also identified the location of the real ionization front in G353.2+0.9. There is a tendency for the clumps near to the ionization front to have redder velocities than those further to the north. This is especially noticeable in low-density tracers (cf. Fig. 5), and is caused by the expansion of the ionized, overpressurized gas pushing the molecular material away from the observer.

Excitation temperatures derived from the ratios of the line temperatures of the C18O(1–0) to the (2–1) lines indicate that Tex is in the range 15−25 K, with the higher values being found along the IF. The temperatures derived from CH3CCH, which is an effective tracer of kinetic temperature, are found to lie in the range of ~22−45 K.

Assuming LTE and a constant Tex = 20 K, we derived the molecular column densities of 13CO (from Massi et al. 1997), C18O, and H2CO, thus obtaining maps of molecular abundances, from their ratios with the H2 column density, which we derived from the APEX 870 μm image, assuming TK = 30 K and a gas-to-dust ratio γ = 100 (Fig. 9). The maps show similar features, with a decrease in the molecular abundance in the region of the elephant trunk and the IF with respect to other region of intense molecular emission [~ (75′′, 75′′), E; ~ (–100′′, 125′′), D, B]. The molecular abundances derived in this way are uncertain by at least a factor of two, owing to the variations in TK across the region. Nevertheless, the region of lower molecular abundance outlines the IF, clearly showing the area where the influence of early-type stars is the strongest.

Column densities of molecular hydrogen derived from C18O and H2CO, under the assumption of LTE and with the abundances calculated as described above, range between 1020−1023 cm-2. The visual extinctions are proportional to the H2 column density: typical values are in the range 5−30 mag depending on the transition used, while the maximum values, estimated from H2CO assuming LTE, are found in clumps C and E (~50 mag). The volume densities, determined assuming spherical geometry for the clumps, lie between ~103 cm-3 and ~105 cm-3. The volume density derived in this way is of the same order of magnitude as the critical density of the transition used to determine it.

The total mass of gas in the region is ~ 2000 M. Single clumps have masses in the range 10 − several    × 102 M. The uncertainty in the mass for a given clump is dominated by the different physical conditions probed by the different transitions.

A simple virial analysis shows that all the clumps with masses above 50 M, in addition to two with lower masses (M ~ 20 M G1 and O) have α ≈ 1, thus indicating that these clumps might be gravitationally bound.

We performed a non-LTE analysis with RADEX, considering the four transitions of the two CS isotopologues in one case and the two transitions of CN in the other.

To analyze the model results, we used a Bayesian approach, evaluating the likelihood P(D|m,TK,N,n) based on Eqs. (9) and (11). We used a constant prior for all the parameters for CS and a Gaussian prior for TK (μ = 35 K, σ = 30 K) for CN, according to our knowledge from CO, CH3CCH, and the results from CS, to reduce the degeneration on this parameter.

For CS, we found TK ~ 11−45 K, depending on the clump considered. The H2 number density typically ranges from several    × 104 cm-3 to few    × 105 cm-3, but exceeds 106 cm-3 for clump E, where TK ~ 11 K. The CS column density lies between ~2 × 1013 cm-2 and ~2 × 1014 cm-2. Making use of this result, we determined the abundance of CS, taking the ratio of the molecular column density to that of H2 derived from the 870 μm emission. The abundances were found to lie between 7.0 × 10-10 and 4.0 × 10-8 (Table 5). These are lower limits because we do not know the column density of each velocity component, but just that of the strongest one. However, the results should not change by much, given that at a certain offset there is usually a single velocity component that dominates the emission.

For CN, we were able to construct maps for TK, N, and n. We found that TK lies typically in the range of 25–32 K, with a maximum of ~45 K, but TK is the most poorly constrained parameter, thus making the maps quite difficult to interpret. The H2 number densities that we found are on average higher than those suggested by CS, in the range of ~1−6    × 105 cm-3, which is consistent with the idea that most of the emission comes from high-density surface layers of the PDR (Simon 1997). The temperature TK shows a similar behaviour: the temperatures found from CN are usually slightly higher than those derived from CS or C18O. In the region of the elephant trunk and the IF, the volume density is even higher, ~ few × 106 cm-3, which is similar to the values also reached for E and D. The column density lies between ~ 5.0 × 1013 cm-2 and ~ 3.0 × 1014 cm-2. Having a map of CN column density for each clump, we summed them at every position, obtaining an integrated NCN map, similar to those of C18O, 13CO, and H2CO. This allowed us to obtain an abundance map (Fig. 10) that resembles those of the other molecules. The abundances found are in the range 5.0 − 7.9 × 10-9 (Table 6). This map showed that CN is enhanced in the region of the trunk and along the IF, where the decrease in molecular abundance is less pronounced than for other molecules.

We did not find significant quantities of molecular material in the region near the “Bar”, which had been previously thought to be an ionization front. This is consistent with the idea that the “Bar” appears to be a layer of ionized matter seen edge-on (Bohigas et al. 2004), the result of the free wind from the massive stars of Pis-24 interacting with the photoevaporative flow generated at the true ionization front (see Fig. 2). The presence of a very faint feature in CN(2–1) along the “Bar” possibly suggests that the small quantity of molecular material in this region could be distributed in very small condensations inside the PDR.

The high incident FUV flux strongly influences the shape and the properties of G353.2+0.9. We investigated the presence of selective photodissociation of C18O making use of 13CO data of Massi et al. (1997). The 13CO/C18O ratio is nearly everywhere consistent with the standard value of the relative abundance of the two isotopologues. We found that south of the ionization front and the elephant trunk, C18O is underabundant with respect to 13CO, being photodissociated by the strong, energetic radiation from Pis-24 stars.

There seems to be a separate semispherical Hii region in the northern part of G353.2+0.9, associated with the star Pis-24 13 (N36), which is an O6.5 V((f)) (Massey et al. 2001). We propose that this star is a runaway O star, dissociating the molecular gas while making its way through it. Radial velocity measurements are needed to confirm this hypothesis.


Acknowledgments

We thank the ATLASGAL project for kindly providing the APEX data. This research made use of the NASA ADS, SIMBAD and CDS

(Strasbourg) databases. At the time the observations discussed here were performed, Achim Tieftrunk was employed by the I. Physikalische Institut in Köln. We also thank Malcolm Walmsley and the anonymous referee for their useful comments, that improved the initial paper.

References

All Tables

Table 1

Molecular transitions observed.

Table 2

VLSR, position, mean LTE column density, excitation temperature (C18O), FWHM of the lines, and diameter of the clumps.

Table 3

Gas masses derived from dust emission.

Table 4

Single-clump and total gas masses.

Table 5

Summary of RADEX results from CS data, for selected offsets.

Table 6

Summary of RADEX results from CN data, for selected offsets.

All Figures

thumbnail Fig. 1

The 8 μm emission image of NGC 6357, taken from the GLIMPSE survey (http://www.astro.wisc.edu/sirtf/, Benjamin et al. 2003). The cavity is clearly visible at the centre of the image. G353.2+0.9 is the bright region at its northern border. The location of Pis-24 is also shown in the figure. The coordinates are referred to the epoch J2000.

Open with DEXTER
In the text
thumbnail Fig. 2

Ks-band image of G353.2+0.9 obtained with SofI (Massi et al., in prep.). The actual ionization front is indicated by “IF” (see text). The locations of the “Bar”, Pis-24, and the elephant trunk are indicated. The red and green squares mark the location of some early-type stars in Pis-24 and the three UCHii regions identified by (Felli et al. 1990), respectively. The coordinates are referred to the epoch J2000.

Open with DEXTER
In the text
thumbnail Fig. 3

Maps of the integrated emission of the different molecular species and transitions, within the whole velocity range of emission. The first contour is the 3σ level. Molecule, transition, and integration limits are indicated above each map. The beam size is indicated by the filled circle. The observed positions are marked with a cross. The contours levels (in units of K km s-1) are, respectively (lowest (step) highest): a) 0.41 (1.0) 12.41; b) 0.51 (1.0) 11.51; c) 0.89 (2.0) 12.89; d) 0.4 (0.4) 2.4; e) 0.47 (0.7) 7.47; f) 0.81 (2.0) 16.81; g) 0.98 (3.0) 27.98; h) 1.25 (2.0) 19.25; i) 0.92 (1.5) 15.92. Coordinates are offsets (arcsec) with respect to (J2000).

Open with DEXTER
In the text
thumbnail Fig. 4

TMB(C18O(2–1)) spectral map. Each frame corresponds to a single pointing with an offset in α and δ indicated in the figure. The x and y axes of each spectrum range from − 20 to 15 km s-1 and from − 0.2 to + 9 K, respectively.

Open with DEXTER
In the text
thumbnail Fig. 5

a) Integrated emission of C18O(1–0) superimposed on the Ks image. Each plot corresponds to a different VLSR. The VLSR of the clumps is indicated in the respective panel. The dashed (blue) lines indicate the observed area. The (red) contours show the integral under Gaussian components fitted to the line profiles. The beam of the transition is shown in the first panel. The first contour in each panel is the 3σ level. The integrated emission levels (in units of K km s-1) are (lowest (step) highest): ( − 6.7 km s-1) 0.17 (0.3) 2.27; (–5.5 km s-1, C) 0.19 (0.3) 1.99; ( − 5.5 km s-1, B) 0.20 (0.5) 4.70; ( − 4.0 km s-1) 0.18 (0.7) 6.48; ( − 2.5 km s-1) 0.19 (0.3) 2.89; ( − 1.0 km s-1) 0.18 (0.2) 1.78; ( + 0.2 km s-1) 0.15 (0.1) 0.65. b) As a), but for CS(5–4). The integrated emission levels (in units of K km s-1) are (lowest (step) highest): ( − 6.7 km s-1) 0.19 (0.3) 2.59; ( − 5.5 km s-1, C) 0.20 (1.0) 10.20; ( − 5.5 km s-1, B) 0.19 (0.5) 3.69; ( − 4.0 km s-1) 0.19 (0.7) 4.39; ( − 2.5 km s-1) 0.18 (1.0) 8.18; ( − 1.0 km s-1) 0.17 (0.2) 1.37; ( + 0.2 km s-1) 0.14 (0.1) 0.74. Note that names of the clumps do not correspond to those used in Massi et al. (1997).

Open with DEXTER
In the text
thumbnail Fig. 6

Boltzmann plot obtained from the CH3CCH data at (0″, 50″).

Open with DEXTER
In the text
thumbnail Fig. 7

APEX image at 870 μm. The scale is expressed in Jy/beam. The APEX beam is shown as a filled circle. The locations of the main components identified are shown in the map. Component 5 was not fitted with a Gaussian, but was considered as a region of diffuse emission.

Open with DEXTER
In the text
thumbnail Fig. 8

CS(5–4) emission (black and gray contours) superimposed on the 870 μm emission. The beam size of CS(5–4) is indicated by the filled circle, while the open circle indicates the APEX beam size. Each plot corresponds to a different VLSR. The VLSR of the clumps is indicated in the respective panel. The contours (in units of K km s-1) for each clump are (lowest (step) highest): (−6.7 km s-1) 0.19 (0.3) 2.59; (−5.5 km s-1, C) 0.20 (1.0) 10.20; (−5.5 km s-1, B) 0.19 (0.5) 3.69; (−4.0 km s-1) 0.19 (0.7) 4.39; (−2.5 km s-1) 0.18 (1.0) 8.18; (−1.0 km s-1) 0.17 (0.2) 1.37; and (+0.2 km s-1) 0.14 (0.1) 0.74. The lowest contour corresponds to the 3σ level in .

Open with DEXTER
In the text
thumbnail Fig. 9

Maps of the abundance relative to H2 of 13CO, C18O, and H2CO, derived from 13CO(1–0) (top left), C18O(1–0) (top right), C18O(2–1) (bottom left), and H2CO(21,2-11,1) (bottom right). The green contours shows the column density distribution of the molecules.

Open with DEXTER
In the text
thumbnail Fig. 10

Map of molecular abundance relative to H2, derived from CN data, through a non-LTE analysis. The green contours shows representative values of the column density, to clarify the distribution of the molecule.

Open with DEXTER
In the text
thumbnail Fig. 11

Virial parameter α as a function of MLTE. Mvir and MLTE are both determined from C18O. MLTE is calculated within the FWHM contour in . The dashed line indicates α ~ 1, i.e. Mvir = MLTE.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.