3.1. Molecular Line Observations

The ¹²CO(2-1), ¹³CO(2-1), and C¹⁸O(2-1) data were acquired with the APEX-1 receiver of the Swedish Heterodyne Facility Instrument (SHeFI; ^{Vassilev et al. 2008}) in the Atacama Pathfinder EXperiment (APEX) telescope, located in the Puna de Atacama (Chile). The backend for the observations was the eXtended bandwidth Fast Fourier Transform Spectrometer2 (XFFTS2) with a 2.5 GHz bandwidth divided into 32768 channels. The main parameters of the molecular transitions (rest frequency ν ₀, halfpower beam-width θ, velocity resolution ∆v_res, and rms noise of the individual spectra obtained using the OTF mode) are listed in Table 1. The selected off-source position free of molecular emission was RA, Dec.(J2000)= (16:36:40.56, −42:03:40.6).

Calibration was done using the chopper-wheel technique. The antenna temperature scale was converted to the main-beam brightness-temperature scale by T _mb = T _A/η _mb, where η _mb is the main beam efficiency. For the SHeFI/APEX-1 receiver we adopted η _mb = 0.75. Ambient conditions were good, with a precipitable water vapor (PWV) between 1.5 - 2.0 mm.

The molecular spectra were reduced using the CLASS90 software of the IRAM’s GILDAS software package.

In addition, we used molecular data from the Millimetre Astronomy Legacy Team 90 GHz Survey (MALT90) taken with the Mopra spectrometer (MOPS). We used beam efficiencies between 0.49 at 86 GHz and 0.42 at 230GHz (^{Ladd et al. 2005}). The data analysis was conducted with CLASS90 software. Emission was detected in the HCO⁺(1-0), HNC(1-0), and N₂H⁺(1-0) lines, which were used to detect high density regions within SDC341.232-0.268. Their main parameters are included in Table 1.

3.2. Images in the Infrared

We used near- and mid-infrared (NIR, MIR) images from the Spitzer-IRAC archive at 4.5 and 8.0µm of the Galactic Legacy Infrared Mid-Plane Survey Extraordinaire (GLIMPSE)⁸ (^{Benjamin et al. 2003}), and the Multiband Imaging Photometer for Spitzer (MIPS) image at 24µm from the MIPS Inner Galactic Plane Survey (MIPSGAL)⁹ (^{Carey et al. 2005}) to delineate the IRDC and investigate their correlation with the EGO candidates.

To trace the cold dust emission we utilized farinfrared (FIR) images from the Herschel Space Observatory belonging to the Infrared GALactic (HiGAL) plane survey key program (^{Molinari et al. 2010}). The data were obtained out in parallel mode with the instruments PACS (^{Poglitsch et al. 2010}) at 70 and 160µm, and SPIRE (^{Griffin et al. 2010}) at 250, 350, and 500µm. The angular resolutions for the five photometric bands spans from 8^′′ to 35^′′ .2 for 70µm to 500µm. Herschel Interactive Processing Environment (HIPE v12¹⁰, ^{Ott 2010}) was used to reduce the data, with reduction scripts from standard processing. The data reduction and calibration (including zero-level and color correction) is described in detail in § 2.2 of ^{Cappa et al. (2016)}.

We also used images at 870µm from the APEX Telescope Large Area Survey of the Galaxy (ATLASGAL) (^{Schuller et al. 2009}) with a beam size of 19^′′ .2. This survey covers the inner Galactic plane, with an rms noise in the range 0.05-0.07 Jy beam⁻¹. The calibration uncertainty in the final maps is about 15%.

4.1. CO data: Morphological and Kinematical Description

To visualize the spatial distribution of the molecular emission the upper panel of Figure 2 exhibits the ¹³CO(2-1) brightness temperature distribution from ≈ −47.0 to −42.5 km s⁻¹ in steps of 0.3 km s⁻¹. The most intense ¹³CO(2-1) emission appears in the range −45.2 to −44 km s⁻¹. The bottom central panel of Figure 2 shows the average ¹³CO(2-1) emission in the range −47.0 to −42.8 km s⁻¹, revealing two molecular clumps centered at RA, Dec.(J2000)= (16:52:37.06, −44:28:13.9) (Clump A) and RA, Dec.(J2000)= (16:52:29.72, −44:28:40.1) (Clump B). The coincidence of the clumps with EGO1, 2, and 4 (indicated with crosses in this figure) is clear. The effective radii of the clumps are 37^′′ .6 and 45^′′ .6 for ClumpA and ClumpB, respectively, as obtained taking into account the average ¹³CO(2-1) emission higher that 6.2 K. Finally, the bottom left and right panels of the same figure display the C¹⁸O(2-1) and ¹²CO(2-1) average emission in velocity intervals similar to that in the image of ¹³CO(2-1). Both clumps A and B are detected at C¹⁸O(2-1) revealing the existence of dense molecular gas in the clumps. The ¹²CO(2-1) emission delineates the IRDC but the clumps are not well defined.

Fig. 2 Upper panel: T _mb-maps of the ¹³CO(2-1) emission within the velocity interval from -47.0 to -42.5km s⁻¹ in steps of 0.3km s⁻¹. Contours range from 8 to 17K in steps of 1K in T _mb. Lower left panel: Average C¹⁸O(2-1) emission in the range -47.2 to -43.8km s⁻¹. Contours are 0.8, 1.0, 1.3, 1.6, and 2.0K. The red crosses mark the position of the EGOs. Lower central panel: Average ¹³CO(2-1) emission in the range −47.0 to −42.8km s⁻¹. Contours range from 6.2 to 9.4K in steps of 0.8K. Lower right panel: Average ¹²CO(2-1) emission in the range −48.2 to −42.5km s⁻¹. Contours range from 10 to 13K in steps of 1K. The color figure can be viewed online. 

Figure 3 shows the averaged ¹²CO(2-1), ¹³CO(2-1), and C¹⁸O(2-1) spectra within the area corresponding to ClumpA and ClumpB. The rms noise of the averaged spectra is 0.019 K for ClumpA, and 0.017 for ClumpB. For ClumpA, the ¹²CO(2-1) line exhibits a multi-peak structure with components in the interval from −70 to −20 km s⁻¹, with the most intense features between −48 and −35 km s⁻¹. Three velocity componentes are detected within the latter velocity interval, peaking at −46.6 km s⁻¹, −42.2 km s⁻¹, and −36.8 km s⁻¹. The ¹³CO(2-1) and C¹⁸O(2-1) spectra peak at around −44.0 km s⁻¹, coincident with a minimum in the ¹²CO(2-1) spectrum. Components outside this range will be analyzed in § 6.2. For ClumpB, the ¹²CO(2-1) line also shows a multi-peak structure between -55 and -25 km s⁻¹, with the maximum at -45.4 km s⁻¹. The ¹³CO(2-1) and C¹⁸O(2-1) profiles peak at -44.0 km s⁻¹.

Fig. 3 Averaged ¹²CO(2-1), ¹³CO(2-1) and C¹⁸O(2-1) spectra towards ClumpA (left) and ClumpB(right). 

Taking into account that the C¹⁸O(2-1) emission is generally optically thin, we adopted systemic velocities (v _sys) of -44.0 km s⁻¹ for ClumpA and ClumpB. The adopted systemic velocity coincides with the velocity for EGO2 and 4 reported by ^{Yang et al. (2017)} (−44.6 and −44.9, respectively) from methanol maser emision, and from masers by ^{Chen et al. (2011)}. It is also compatible with the velocity of NH₃ clouds identified by ^{Purcell et al. (2012)} at −43.7 km s⁻¹ at a position distant 15^′′ from EGO3, and from the CS(2-1) line emission obtained by ^{Bronfman et al. (1996)}, who observed toward [RA, Dec.(J2000)= (16:52:34.2, -44:28:36.0)], revealing the presence of high density regions (n _crit = 3.0 × 10⁵ cm⁻³). Circular galactic rotation models predict that gas with velocities of −44 km s⁻¹ lies at the near kinematical distance of 3.6 kpc (see, for example, ^{Brand & Blitz 1993}). Adopting a velocity dispersion of 6km s⁻¹, the uncertainty in distance is 0.5kpc (15%). The distance coincides also with that of S24 (^{Cappa et al. 2016}) indicating that the IRDC belongs to the same complex.

4.2. Parameters Derived From CO Data

Bearing in mind that the spatial distribution of the ¹³CO(2-1) and C¹⁸O(2-1) emission coincides with the IRDC, we used these lines to estimate the main parameters of the molecular gas assuming local thermodynamic equilibrium (LTE). All the calculations were carried out from the Gaussian fits to the lines averaged in the area of ClumpA and ClumpB shown in Table 2. The columns list the central velocity of each component v, the full-width at halfmaximum, the peak temperature, and the integrated emission. We applied the following expression to estimate optical depths (e.g. ^{Scoville et al. 1986}):

where τ ¹³ is the optical depth of the ¹³CO(2-1) gas and 7.6 = [¹³CO(2-1)]/[C¹⁸O(2-1)] (^{Sanhueza et al. 2010}) is the isotope abundance ratio. The ¹³CO(2-1) optical depths are indicated in Column 2 of Table 3. To estimate the C¹⁸O(2-1) optical depth we used

where ∆v is the full-width at half-maximum of the C¹⁸O(2-1) and ¹³CO(2-1) profiles. The results are listed in Column 3 of Table 3. Bearing in mind that ¹³CO(2-1) is moderately optically thick we calculated the excitation temperature from the ¹³CO(2-1) line using

where T* = hνk, ν is the frequency of the ¹³CO(2-1) line and T _bg = 2.7 K. Values estimated for T _exc are listed in Table 3.

The column density for ¹³CO(2-1) was derived using ^{Rohlfs & Wilson (2004)}

In this case we used the approximation for τ ₁₃ > 1,

Fig. 4 HCO⁺(1-0), HNC(1-0) and N₂H⁺(1-0) profiles towards ClumpA (top) and ClumpB (bottom). The color figure can be viewed online. 

This approximation helps to eliminate to some extent optical depth effects. The integral was evaluated as T _mean − _mb ×∆v, where T _mean − _mb is equal to the average T _mb within the area of each clump.

To estimate N(H₂) (Column 6 in Table 3) we adopted an abundance [H₂][¹³CO]= 5×10⁵ (^{Dickman 1978}). Then, the molecular mass (Column 9 in Table 3) was calculated from the equation

where m _sun = 2×10³³ g is the solar mass, the mean molecular weight µ = 2.76 (which includes a relative helium abundance of 25% by mass, ^{Allen 1973}), and m _H is the hydrogen atomic mass. In this expression d is the distance, N(H ₂) is the H₂ column density, and A is the area of the source in cm⁻². The effective radius of each clump, as seen in the ¹³CO(2-1) line, and the volume densities are listed in Colums 7 and 9 in Table 3.

The ambient volume densities, n _H2, was calculated assuming a spherical geometry for the clumps, using the formula

Ambient densities are listed in Column 9 of Table 3. The parameters calculated in this work agree with ^{Bergin & Tafalla (2007)}.

The virial mass can be determined following ^{MacLaren et al. (1988)}:

where r and ∆v are the radius of the region and the velocity width measured from the Gaussian fit of the C¹⁸O(2-1) emission, and k ₂ depends on the geometry of the ambient gas in the region, being 190 or 126 according to ρ ∝ r ⁻¹ or ρ ∝ r ⁻², respectively. M _vir values are listed in Column 10 in Table 3. The ratios M _vir /M(H ₂) suggest that both clumps may collapse to form new stars; since the virial mass value is smaller than the LTE mass value this clump does not have enough kinetic energy to stop the gravitational collapse.

Uncertainties in both the molecular mass derived using LTE conditions, M(H₂), and the virial mass M _vir are affected by the distance indetermination (15%) yielding a 30% error in M(H₂) and 15% in M _vir. Inaccuracies in the borders of the clumps originate errors in their sizes and thus additional uncertainties in the masses, suggesting errors of 40% in M(H₂). Virial masses are not free of uncertainties, because of the existence of magnetic field support that may lead to an overestimate of the derived values (see ^{MacLaren et al. 1988}), and because of the unknown density profile of the clump.

4.2.1. Analysis of the MALT90 Data

The MALT90 data cubes of this region show emission of HCO⁺(1-0), HNC(1-0), and N₂H⁺(1-0) molecules, which are the most often detected molecules towards IRDCs (^{Sanhueza et al. 2012}; ^{Rathborne et al. 2016}). These molecules have critical densities of 2×10⁵, 3×10⁵ and 3×10⁵ cm⁻³, respectively (Sanhueza et al. 2012), allowing us to infer a lower limit for the density of the clumps. These values are higher than the H₂ ambient densities derived from CO lines (see Table 3).

Spectra of these lines, averaged over the area of the clumps, are shown in Figure 5. By applying a Gaussian fitting to the HCO⁺ and HNC lines, a mean velocity of −46 km s⁻¹ was derived. For the case of the N₂H⁺ line, a multi Gaussian function was applied to take into account hyperfine structure. This molecule has seven hyperfine components in two blended groups of three lines (groups 2 and 3) and one isolated component (group 1). The rest frequencies of groups 1, 2 and 3 are 93.176, 93.173, and 93.171 GHz, respectively. The mean velocity is −44 km s⁻¹. The velocities of these lines coincide with the systemic velocities derived from CO data.

Fig. 5 HCO⁺(1-0) (levs: 0.65, 0.7, 0.8 and 0.9 K), HNC(1-0) (levs: 0.7, 0.85, 1.0 and 1.15 K) and N₂H⁺(1-0) (levs: 0.8, 1.0, 1.2 and 1.4 K) maps. Red circles have radii equal to the effective radii of the clumps as seen in ¹³CO(2-1), and are located at their position. Crosses mark the location of the EGO candidates. The color figure can be viewed online. 

These tracers of dense gas provide slightly different information: HCO⁺ often shows infall signatures and outflow wings (e.g. ^{Rawlings et al. 2004}; ^{Fuller et al. 2005}). However, in our case no signs of infall can be identified. HNC is specially preponderant in cold gas and is a commonly used tracer of dense gas in molecular clouds. Finally, N₂H⁺ is more resistant to freeze-out on grains than carbon-bearing species like CO.

Figure 5 shows the HCO⁺(1-0), HNC(1-0) and N₂H⁺(1-0) maps, integrated from −47.6 to −44.0 km s⁻¹, −46.7 to −43.1 km s⁻¹, and −47 to −41.4 km s⁻¹, respectively. Both clumps are detected in the three molecular lines indicating mean densities of 10⁵ cm⁻³. Carbon-bearing species, like CO, tend to disappear from the gas phase in the high density centers of the cores, while nitrogen-bearing species like N₂H⁺ survive almost unaffected up to much higher densities. This is evidenced in the fact that the spacial distribution of the N₂H⁺ emission is similar to that of dust continuum emission (see Figure 6).

Fig. 6 FIR view of SDC341.232-0.268 from Herschel (70-500µm) and ATLASGAL data. The blue circles have the same meaning as in Figure 5 and show the areas where the flux densities were integrated. The crosses mark the position of the EGOs. Contours correspond to 20rms. The color figure can be viewed online. 

Following ^{Purcell et al. (2009)} we estimate the N₂H⁺ optical depth and column density. Assuming the line widths of the individual hyperfine components are all equal, the integrated intensities of the three blended groups should be in the ratio of 1:5:3 under optically thin conditions. The optical depth can then be derived from the ratio of the integrated intensities of any group using the following equation:

∫Tmb,1dv∫Tmb,2dv=1-exp-τ11-exp-τ2=1-exp-τ11-exp-aτ1 , (9)

where a is the expected ratio of τ ₁ /τ ₂ under optically thin conditions.

To determine the optical depth we used only the intensity ratio of group 1/group 2, as ^{Caselli et al. (1995)} report anomalous excitation of the F₁,F = 1,0 → 1,1 and 1,2 → 1,2 components (in group 3). Thus we obtained τ ₁ = 0.11 and 0.19 for ClumpA and ClumpB, respectively.

Based on the expression for T _mb given by ^{Rohlfs & Wilson (2004)}, the excitation temperature for N₂H⁺ can be calculated with the following formula:

Texc=4.47/ln[1+(Tmb4.47(1-exp(-τ))+0.236)-1], (10)

and the column densities can be derived using (^{Chen et al. 2013}):

N(N2H+)=3kW8π3νSμ2(TexcTexc-Tbg)(τ1-exp(-τ))×Q(Texc)exp(Eu/kTexc) (11)

where k is the Boltzmann constant, W is the observed line integrated intensity (obtained from a Gaussian fit), ν is the frequency of the transition, and Sµ ² is the product of the total torsion-rotational line strength and the square of the electric dipole moment. T _exc and T _bg are the excitation temperature and background brightness temperature, respectively. E _u /k is the upper level energy in K, Q(T _exc) is the partition function at temperature T _exc and τ is the optical depth. For group 1 the values of ν, Sµ ² and E _u /k are 93176.2526 MHz, 12.42D² and 4.47K, respectively. These values were taken from the SPLATALOGUE catalogue¹¹.

We derive T _exc = 17.4 and 12.5K, and N(N₂H⁺) = 8.7×10¹³ and 8.1×10¹³ for ClumpA and ClumpB, respectively. Considering an abundance [N₂H⁺]/[H₂] = 5×10⁻¹⁰ for dark molecular clouds (^{Ohishi et al. 1992}) we obtain H₂ column densities of 1.7 and 1.6 ×10²³ for ClumpA and ClumpB, respectively. These column densities are higher than those obtained from CO calculations. Ambient densities are n _H2 ≈ 1.7×10⁵ cm⁻³ for each clump with an effective radii in N₂H⁺ of 0.53pc. This difference may be explained by considering that the nitrogen-bearing species survive almost unaffected up to much larger densities that carbon-bearing molecules.

6.1. Search for Additional YSOs Coincident with the IRDC

To investigate the coincidence of the IRDC with other candidate young stellar objects (YSOs), we analyze the characteristics of the point sources in the WISE and Spitzer catalogues (^{Wright et al. 2010}; ^{Benjamin et al. 2003}) projected in the region. In Figure 8 we mark the positions of Class I and II YSO candidates identified in the area of the molecular clumps. An inspection of the figure reveals that EGOs 1, 3, and 4 coincide with identified YSOs.

Fig. 8 Overlay of YSOs identified in the Spitzer and WISE catalogues and the molecular clumps in ¹³CO(2-1). Contours have the same meaning as in Figure 2. Yellow and blue symbols show the position of Class I and Class II YSOs, while red crosses mark the location of EGOs. Plus signs correspond to Spitzer sources and diamonds to WISE ones. The color figure can be viewed online. 

To study in some detail the nature of the EGOs we plotted their spectral energy distributions (SEDs) using the Robitaille’s SED fitting tool (^{Robitaille et al. 2007}). To perform the analysis we were able to use available data from 2MASS, IRAC-GLIMPSE, and MIPSGAL at 24µm only. The obtained SEDs are shown in Figure 9. Clearly, EGO 1, 3 and 4 have characteristics of YSOs, and as seen in Figure 8 coincide with additional evidence of star formation. As regards EGO2, the fit is not good; it does not coincide with other sources or with the molecular clumps. So, its status is doubtful. In Table 5 we display the main parameters obtained from the SEDs. Column 1 gives the name of the source; Columns 2 and 3, the age and mass, M _stellar, of the central source; Column 4, the mass of the envelope, M _env; Column 5, the infall rate, Ṁ _acr; and Column 6, the total luminosity. To perform the SEDs we adopted d = 3.6±0.5 kpc and a visual extinction of 3-4 mag.

Fig. 9 Spectral energy distribution for the four candidate EGOs. The color figure can be viewed online. 

Following ^{Robitaille et al. (2007)} an estimate of the evolutionary stage of the sources can be obtained based on the ratio of the infall rate and the mass of the central source. For the four candidate EGOs the ratio Ṁ _acr /M _stellar > 10⁻⁶ indicates Stage 0/I. The age of some of the sources suggests that they are still immersed in their envelopes. According to the fitting, EGO4 would be the most massive object in the sample, and EGO2 seems to be the most evolved one. These results should be taken with caution because of the limitations of the fitting tool that may arise from: (1) YSOs are complex 3d objects with slightly non-axisymmetric density structures, so the models are incorrect compared to actual density distributions; (2) there are often mixtures of sources and many objects appear as a single YSO, but they are often two or more objects; and (3) even in the case of an isolated YSO variability is an additional complication, since the data with which SEDs are usually constructed belong to different surveys that have been performed in different years and can produce discontinuities in the SEDs (^{Robitaille 2008}; ^{Deharveng et al. 2012}; ^{Offner et al. 2012}).

As pointed out before, the emission at 24µm, detected toward EGO1, 3, and 4, suggests the existence of warm dust and embedded protostars (^{Jackson et al. 2008}).

6.2. Evidence for Inflow/Outflow Motions

One way to visualize molecular outflows or infalls is to compare the optically thick ¹²CO(2-1) emission line with the optically thin C¹⁸O(2-1) molecular line toward the candidate EGOs, as shown in Figure 10 for EGO3 and EGO4. The ¹²CO(2-1) line presents a double peak structure with the blue shifted peak brighter than the red-shifted one, and a minimum at the systemic velocity, while the C¹⁸O(2-1) line presents a single peak centered on the absorption dip of the optically thick line. According to ^{Chen et al. (2011)} these spectra display characteristics typical of infall motions: a double-peak structure with the blue-shifted peak brighter than the red-shifted one, while an emission peak of an optically thin line appears centered on the absorption dip of the optically thick line. According to these authors, the infall motion is the only process that would consistently produce the blue profile asymmetry. The blue-shifted emission can be explained as due to high-excitation approaching warm gas located on the far side of the center of contraction. The emission of this gas undergoes less extinction than the emission from the red-shifted receding nearside material, given that the excitation temperature of the molecules increases toward the center of the region (^{Zhou 1992}). Thus, the spectra of EGO3 and EGO4 suggest the presence of infall motions.

Fig. 10 ¹²CO(2-1) and C¹⁸O(2-1) profiles in the direction to EGO3 (left) and EGO4 (right). In both pannels red and green lines correspond with Gaussian fits from the ¹²CO(2-1) double peak, and blue lines indicates a Gaussian component fitting the possible blue outflow wings. The color figure can be viewed online. 

Following ^{Mardones et al. (1997)}, we use the asymmetry parameter δv (the velocity difference between the peaks of the optically thick line and the optically thin lines, normalized by the FWHM of the thin line), to quantify the blue asymmetry. The parameter is defined as δv = (v_thick - v_thin) / ∆v_thin. A statistically significant excess of blue asymmetric line profiles with δv < -0.25 indicates that the molecular gas is falling into the clump. We consider the velocity and width of the C¹⁸O(2-1), from Table 2, as v_thin and ∆v_thin, and calculate v_thick from the Gaussian fits of ¹²CO(2-1) profiles (-46.9 and -45.3 from ClumpA and ClumpB, respectively). We find, for both ClumpA and ClumpB δv ≈ -0.06, which supports the infall hypothesis.

The interpretation presented above is supported by the M _vir /M(H₂) ratio previously derived for the objects. As the classical virial equilibrium análisis establishes, a ratio M _vir /M(H₂) ^< _∼ 1 indicates that a clump has too much kinetic energy and is unstable against gravitational collapse. Then, the derived ratios suggest that ClumpA and ClumpB are unstable and could be collapsing.

For the sake of completeness, we calculate the eventual infall rate estimated from ^{Klaassen & Wilson (2007)}

where n _H2 is the H₂ volume density, r _clump is the linear radius of the infall region, and v_inf is the infall velocity of the material. One way to estimate the infall velocity is considering the two layer radiative transfer model of ^{Myers et al. (1996)}. From the ¹²CO(2-1) and C¹⁸O(2-1) spectra corresponding to ClumpA in Figure 3, we calculated v_inf = 0.53 km s⁻¹ using equation (9) in ^{Myers et al. (1996)}. In order to get a better estimate, we also calculated the infall velocity using the Hill5 model (^{De Vries & Myers 2005}) in the PySpecKit package (^{Ginsburg & Mirocha 2011}) ¹². The Hill5 model employs an excitation temperature profile increasing linearly toward the center, rather than the two-slab model of ^{Myers et al. (1996)}, so the Hill5 model is thought to provide a better fit to infall motions (^{De Vries & Myers 2005}). From this model we obtained v_inf = 1.53 ± 0.26 km s⁻¹. Considering this value of v_inf = and the parameters listed in Table 3, we obtained Ṁ _inf ≈ 7.47×10⁻⁴ M _⊙ yr⁻¹. This result is consistent with those derived from surveys of massive YSOs such as ^{Klaassen & Wilson (2007)}. ClumpB does not show clear observational evidences of collapse.

A typical outflow appears as spatially confined wings beyond the emission from the cloud core. For EGO3 and EGO4, we can distinguish the presence of ¹²CO(2-1) blue wings from ≈ -60 to -50 km s⁻¹, shown in the spectra fit in Figure 10, while the velocity range of the optically thin C¹⁸O(2-1) emission is ≈ -47 to -40 km s⁻¹. The dashed line marks the systemic velocity of the clumps, which coincides with the central velocity of the C¹⁸O(2-1) line. Red wings cannot be identified, although multiple components in the ¹²CO(2-1) emission are present.

The typical signature of outflows can be seen in the position-velocity diagram. As shown in Figure 11 for EGO4, the extension of the emission toward velocities of about -55 km s⁻¹would correspond to the blue wings shown in Figure 10. Note that the declination of this extension and the right ascension of the map coincide with the position of EGO4.

Fig. 11 Position-velocity diagram toward EGO4 in ¹²CO(2-1) (Dec. vs. velocity for a fixed value of R.A. = 16:52:35). The color scale indicates Tmb and the declination is indicated in fractions of degrees. The color figure can be viewed online. 

These characteristics suggest that, although they are not spatially resolved, there may be outflows associated with EGO3 and EGO4.