THE PHYSICAL ENVIRONMENT AROUND IRAS 17599–2148: INFRARED DARK CLOUD AND BIPOLAR NEBULA

Radio continuum observations of IRAS 17599–2148 at 0.61 and 1.28 GHz bands were obtained using the GMRT on 2012 December 29 and 2012 December 23 (Proposal Code: 23_054; PI: L. K. Dewangan), respectively. The radio data reduction was performed using the AIPS software, following a similar method to that described in Mallick et al. (2012, 2013). The 0.61 and 1.28 GHz maps have rms noise of 0.34 and 0.38 mJy beam^–1, respectively. The synthesized beam sizes of the final 0.61 and 1.28 GHz maps are ∼56 × 52 and ∼28 × 24, respectively.

Note that a large amount of emission from the Galactic plane was expected during the observations, which influences the low-frequency observations and increases the effective antenna temperature. Since IRAS 17599–2148 is located in the Galactic plane, our GMRT 0.61 and 1.28 GHz maps are corrected for system temperature (see Omar et al. 2002; Mallick et al. 2012, 2013; Baug et al. 2015). The details about the adopted procedure of system temperature corrections are described in Mallick et al. (2012, 2013).

The imaging observations of IRAS 17599–2148 were taken with the 8.2 m VLT with the NAOS-CONICA (NACO) adaptive optics system (Lenzen et al. 2003; Rousset et al. 2003) in H band (λ_c = 1.66 μm, Δλ = 0.33 μm), K_s band (λ_c = 2.18 μm, Δλ = 0.35 μm), and L' band (λ_c = 3.80 μm, Δλ = 0.62 μm). We downloaded these data from the ESO-Science Archive Facility (ESO proposal ID: 089.C-0455(A); PI: João Alves). The single-frame exposure time was 21 s in L' band and 24 s in H and K_s bands. We used five H frames, five K_s frames, and 11 L' frames. In each band, the final NACO image was obtained through a standard analysis procedure such as sky subtraction, image registration, combining with the median method, and astrometric calibration, using IRAF and STAR-LINK software. The astrometric calibration of the NACO images was done using the UKIRT NIR Galactic Plane Survey (GPS; Lawrence et al. 2007) K-band point sources. The VLT/NACO H, K_s, and L' images have resolutions of 02 (∼840 au), 02 (∼840 au), and 01 (∼420 au), respectively.

Deep NIR photometric JHK magnitudes of point sources have been retrieved from the UKIDSS GPS sixth archival data release (UKIDSSDR6plus). The UKIDSS observations (resolution ∼08) were carried out using the UKIRT Wide Field Camera (Casali et al. 2007). The UKIDSS GPS photometric data were calibrated using the Two Micron All Sky Survey (2MASS; Skrutskie et al. 2006) data. We extracted only reliable NIR photometric data, following the recommendations given in Lucas et al. (2008) and Dewangan et al. (2015a). To avoid saturation, the sources were obtained fainter than J = 12.1 mag, H = 11.1 mag, and K = 10.0 mag in our selected GPS catalog. Furthermore, the 2MASS photometric magnitudes were retrieved for bright sources that were saturated in the GPS catalog.

On a larger scale, an IRDC and a BN, associated with the bright extended emission, are visually observed in the MIPSGAL 24 μm image (see Figure 1(a)). In the MIPSGAL 24 μm image, the embedded stellar sources are also seen toward the IRDC. Fourteen ATLASGAL dust continuum sources at 870 μm are also marked in Figure 1(a). The ATLASGAL sources are found toward the IRDC and the BN, tracing cold and dense regions. It appears as a chain of dense clumps/cores in our selected target region (e.g., Tafalla & Hacar 2015). In Figure 1(a), the 24 μm image is saturated near the IRAS position, where the radio continuum peak position was reported (Becker et al. 1994; Dewangan et al. 2012). Figure 1(b) shows a three-color composite image made using Herschel 160 μm in red, Herschel 70 μm in green, and GLIMPSE 8.0 μm in blue. The IRDC appears as bright emission regions at wavelengths longer than 24 μm. The BN is also seen as the brightest extended emission in Figure 1(b). Figure 2(a) shows the ATLASGAL and Herschel submillimeter images of the region, which trace the cold dust emission (see Section 3.2 for quantitative estimate). The BN is traced in all the Herschel images. Interestingly, we infer an elongated filamentary morphology (extension ∼21 pc) in the ATLASGAL and Herschel submillimeter images, where several condensations are found. Furthermore, the IRDC and the G8.14+0.23 H ii region (including the BN) are part of the filamentary structure (see a dashed curve in Figure 2(a)), which is embedded in the cloud associated with IRAS 17599–2148 (i.e., M008.2+0.2; see Figures 14(b) and 15 in Takeuchi et al. 2010). The integrated NANTEN ¹²CO intensity map (beam size ∼26) traces a continuous velocity structure in the direction of IRAS 17599–2148 (Takeuchi et al. 2010). It includes the eastern and western sides of the bubble, as well as the bubble itself (i.e., entire elongated filamentary structure [EFS]; see Figure 14(b) in Takeuchi et al. 2010). Hence, these previous molecular line data, even with coarse beam size, further confirm the existence of the EFS. In Figure 2(a), the Herschel images trace the faint dust emission (or infrared bridges) between the ATLASGAL dust continuum sources, which are located within the EFS. Taken together, the submillimeter images and the previously published CO line data reveal the existence of a single EFS in the region probed in this paper.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In Figure 2(b), we present the JCMT CO(J = 3–2) gas emission in the direction of the IRAS position. Note that we do not have JCMT molecular line observations for the entire field of view, which is probed in this paper. The JCMT CO data are available only toward the BN. Considering the better resolution of the JCMT HARP CO(3–2) data compared to the previously published NANTEN ¹²CO data, we are able to observe the bipolar morphology in the integrated JCMT CO intensity map (in a velocity range of 14–26 km s⁻¹). The kinematics of molecular gas are described in Section 3.3. In Figure 2(b), the 850 μm continuum emission, which traces the densest materials, is also overlaid on the integrated CO map and is mostly distributed toward the waist of the BN. The 850 μm continuum map (beam size ∼14'') reveals two peaks (i.e., dense clumps/cores), and one of them coincides with a Class II 6.7 GHz MME. The 6.7 GHz MME is a reliable tracer of massive young stellar objects (MYSOs; e.g., Walsh et al. 1998; Minier et al. 2001; Urquhart et al. 2013).

In order to trace the distribution of ionized emission, we present high-resolution GMRT radio continuum maps at 0.61 GHz (beam size ∼56 × 52) and 1.28 GHz (beam size ∼28 × 24) of the region around IRAS 17599–2148. Figure 3 shows the GMRT 0.61 GHz (top panel) and 1.28 GHz (bottom panel) contours overlaid on the Spitzer 4.5 μm image. Note that the radio emission is not detected toward the IRDC (not shown here) and is found only toward the previously known G8.14+0.23 H ii region. However, these new maps provide more insight into the radio peak, due to the high-resolution map, compared to the previously reported radio continuum map at 20 cm (beam ∼6''; Dewangan et al. 2012). Both GMRT maps show similar morphology of the ionized emission, and their peak positions are offset from the 6.7 GHz MME. The extended morphology and a single peak at 0.61 GHz map are similar to what was previously seen in the 20 cm map (Dewangan et al. 2012). In addition to the extended morphology, due to high-resolution observations, the 1.28 GHz map has resolved a single peak (as seen in the 0.61 GHz map) into two compact peaks (see Figure 3(b)) (i.e., cp1 [α₂₀₀₀ = 18^h03^m0149, δ₂₀₀₀ = −21°48'1223; peak flux density = 22.52 mJy/beam; angular size = 2878 × 1395; deconvolved size = 2866 × 1371] and cp2 [α₂₀₀₀ = 18^h03^m0138, δ₂₀₀₀ = −21°48'1382; peak flux density = 22.63 mJy/beam; angular size = 1425 × 993; deconvolved size = 1402 × 957]). Using the 1.28 GHz map, we computed the integrated flux densities equal to 1355.50 and 480.12 mJy for the cp1 and cp2 peaks, respectively. These integrated flux densities are utilized to compute the number of Lyman continuum photons (N_uv). The equation of N_uv is given by (Matsakis et al. 1976),

$$\begin{eqnarray}{N}_{{uv}}({s}^{-1}) & = & 7.5\times {10}^{46}\left(\displaystyle \frac{{S}_{\nu }}{\mathrm{Jy}}\right){\left(\displaystyle \frac{D}{\mathrm{kpc}}\right)}^{2}{\left(\displaystyle \frac{{T}_{e}}{{10}^{4}{\rm{K}}}\right)}^{-0.45}\\ & & \times {\left(\displaystyle \frac{\nu }{\mathrm{GHz}}\right)}^{0.1},\end{eqnarray} \tag{ 1 }$$

where S_ν is the measured total flux density in Jy, D is the distance in kpc, T_e is the electron temperature, and ν is the frequency in GHz. The analysis is done for a distance of 4.2 kpc and for an electron temperature of 10,000 K. We find N_uv (or log N_uv) to be ∼1.86 × 10⁴⁸ s⁻¹ (48.27) and ∼6.6 × 10⁴⁷ s⁻¹ (47.82) for cp1 and cp2, respectively. These estimates correspond to a single ionizing star of spectral type O8V–O8.5V and O9V–O9.5V (see Table 1 in Martins et al. 2005 for theoretical values) for cp1 and cp2, respectively. Based on these results, within 18'' of the IRAS position, two radio O spectral type sources and a 6.7 GHz MME (without radio peak) are present, indicating the presence of different early evolutionary stages of massive star formation. The search of the driving sources of the radio peaks and the 6.7 GHz MME is presented in Section 3.4.

Zoom In Zoom Out Reset image size

We also computed the integrated flux densities of the extended radio morphology seen (at 3σ) in both radio maps. Using Equation (1), the estimation of N_uv is also performed for both radio frequencies (0.61 and 1.28 GHz) separately. Taking into account D = 4.2 kpc, T_e = 10,000 K, and S_1.28 = 3.39 Jy, we find N_uv (or log N_uv) = 4.7 × 10⁴⁸ s⁻¹ (48.67). Similarly, we obtain N_uv (or log N_uv) = 6.9 × 10⁴⁸ s⁻¹ (48.84) for D, T_e, and S_0.61 = 5.39 Jy. The estimates of N_uv values at different frequencies correspond to a single ionizing star of O6V–O6.5V spectral type (Martins et al. 2005), which is consistent with the previously reported spectral type of the ionizing source of the G8.14+0.23 H ii region (e.g., Kim & Koo 2001).

Within the EFS, the radio emission is detected toward the IRAS 17599–2148 position and is absent toward the IRDC, revealing different evolutionary stages of star formation.

In the column density map, the clumpfind algorithm (Williams et al. 1994) has been applied to identify the clumps and their total column densities. Twelve clumps are identified, which are labeled in Figure 4(b), and their boundaries are also shown in Figure 4(c). We estimated the masses of these clumps using their total column densities. The mass of a single clump can be computed using the formula,

$$\begin{eqnarray}&&{M}_{\mathrm{clump}}={\mu }_{{{\rm{H}}}_{2}}{m}_{H}{\mathrm{Area}}_{\mathrm{pix}}{\rm{\Sigma }}N({H}_{2}),\end{eqnarray} \tag{ 2 }$$

where ${\mu }_{{{\rm{H}}}_{2}}$ is assumed to be 2.8, Area_pix is the area subtended by 1 pixel, and ΣN(H₂) is the total column density. The mass of each Herschel clump is listed in Table 1. The table also contains an effective radius of each clump, which is provided by the clumpfind algorithm. The clump masses vary between 410 and 7024 M_⊙. The EFS is associated with at least six massive clumps (i.e., IDs 1–5 and 10), indicating the fragmentation of the cloud. The most massive condensation (i.e ID 1) hosts two compact radio continuum sources and the 6.7 GHz MME, which is located at the waist of the BN. At least three massive and cold condensations (i.e., IDs 3–5) are seen toward the IRDC. At least four condensations (i.e., IDs 6–8 and 12) are also found at the edges of the BN.

Table 1.  Summary of the Properties of the Identified Herschel Clumps within the Region Probed in This Paper Using the Herschel Data (see Figures 4(b) and (c))

ID	R.A.	Decl.	Rca	Mclump	Pclump	${t}_{\mathrm{ff},\mathrm{sph}}$	Vpeak	ΔV	Mvir
	[J2000]	[J2000]	(pc)	(M⊙)	(10−10 dynes cm−2)	(Myr)	(km s−1)	(km s−1)	(M⊙)
1	18:03:04.7	−21:48:13.0	2.0	7023.7	29.2	0.39	18.6b	3.36b	2830
2	18:02:57.7	−21:47:58.6	1.8	4147.6	14.8	0.47	18.58b	3.39b	2625
3	18:03:18.8	−21:45:53.6	2.5	5454.5	6.9	0.57	18.7c	3.1c	3047
4	18:03:33.9	−21:42:52.2	2.5	6418.6	9.6	0.52	18.9c	2.7c	2306
5	18:03:27.8	−21:46:36.0	2.9	6422	5.5	0.60	⋯	⋯	⋯
6	18:03:02.7	−21:45:38.9	1.4	1192.6	3.6	0.66	⋯	⋯	⋯
7	18:02:56.6	−21:50:46.5	1.2	763.8	3.0	0.70	⋯	⋯	⋯
8	18:03:00.6	−21:50:18.7	1.4	1046.2	2.8	0.70	⋯	⋯	⋯
9	18:03:22.7	−21:51:29.7	2.4	3377.9	3.3	0.70	⋯	⋯	⋯
10	18:02:40.6	−21:48:25.7	1.2	776.8	2.3	0.74	⋯	⋯	⋯
11	18:03:15.7	−21:50:47.4	1.8	1678.5	2.5	0.73	⋯	⋯	⋯
12	18:02:56.7	−21:46:06.6	0.9	410	2.1	0.76	⋯	⋯	⋯

Notes. Column (1) lists the IDs given to the clump. The table also contains positions, deconvolved effective radius (R_c), clump mass (M_clump), self-gravitating pressure (${P}_{\mathrm{clump}}\approx \pi G({M}_{\mathrm{clump}}/\pi {R}_{c}^{2}$)²), spherical free-fall time (${t}_{\mathrm{ff},\mathrm{sph}}\approx 16.6{({R}_{c}/\mathrm{pc})}^{3/2}{({M}_{\mathrm{clump}}/{M}_{\odot })}^{-1/2}$), peak velocity (V_peak), line width (ΔV), and virial mass (M_vir = 126R_cΔV²).

^aEstimated using the clumpfind algorithm. ^bNH₃(1, 1) derived line parameters from Wienen et al. (2012). ^cHCO⁺(3–2) derived line parameters from Shirley et al. (2013).

Download table as:  ASCIITypeset image

In order to examine the internal dynamical properties of these clumps in more detail, we computed their physical parameters (such as spherical free-fall time, virial mass, virial parameter, self-gravitating pressure, and Mach number). Using the mass (M_clump) and radius (R_c) of each clump, we can estimate the mean mass density ρ₀ and the spherical free-fall time ${t}_{\mathrm{ff},\mathrm{sph}}$ = (3π/32Gρ₀)^1/2 (=16.6 (R_c/pc)^3/2 (M_clump/M_⊙)^−1/2 Myr). We estimated ${t}_{\mathrm{ff},\mathrm{sph}}$ values for our identified clumps ranging from 0.4 to 0.77 Myr, which are tabulated in Table 1.

In each clump, the pressure exerted by the self-gravity of the molecular gas can be estimated using ${P}_{\mathrm{clump}}\approx \pi G({M}_{\mathrm{clump}}/\pi {R}_{c}^{2}$)². The value of P_clump is tabulated in Table 1. In general, the pressure associated with a typical cool molecular cloud (P_MC) is found to be ∼10⁻¹¹–10⁻¹² dynes cm⁻² (for a temperature of ∼20 K and a particle density of ∼10³–10⁴ cm⁻³) (see Table 7.3 of Dyson & Williams 1980, p. 204), which is much lower than P_clump. This comparison indicates an acting of self-gravity and a signature of star formation activity in these clumps.

To infer whether the identified clumps are stable or unstable against gravitational collapse, the virial mass analysis is employed. As mentioned above, we find six clumps associated with the filamentary structure, and two of them are located at the waist of the BN. The molecular line observations were reported toward four out of these six clumps. We estimated virial mass (M_vir) and virial parameter (M_vir/M_clump) for these four clumps where line width values were measured. The virial mass of a clump of radius R_c (in pc) and line width ΔV (in km s⁻¹) is defined as M_vir (M_⊙) = kR_cΔV² (MacLaren et al. 1988), where the geometrical parameter k = 126 for a density profile ρ ∝ 1/r². A virial parameter less than 1 suggests a clump prone to collapse, and greater than 1 suggests one resistant to collapse. For these clumps, the virial parameter is less than 1, indicating unstable clumps against gravitational collapse.

Following Dewangan et al. (2016), we estimated thermal sound speed (a_s) and nonthermal velocity dispersion (σ_NT) of the two clumps where gas kinetic temperature (T_kin) values were reported (Wienen et al. 2012). Wienen et al. (2012) estimated T_kin for clumps 1 and 2 to be 22.86 and 17.71 K, respectively. The sound speed a_s $(={({{kT}}_{\mathrm{kin}}/\mu {m}_{H})}^{1/2}$) is computed using T_kin and μ = 2.37 (approximately 70% H and 28% He by mass). The nonthermal velocity dispersion is defined by

$$\begin{eqnarray}&&{\sigma }_{\mathrm{NT}}=\sqrt{\displaystyle \frac{{\rm{\Delta }}{V}^{2}}{8\mathrm{ln}2}-\displaystyle \frac{{{kT}}_{\mathrm{kin}}}{17{m}_{H}}}=\sqrt{\displaystyle \frac{{\rm{\Delta }}{V}^{2}}{8\mathrm{ln}2}-{\sigma }_{{\rm{T}}}^{2}},\end{eqnarray} \tag{ 3 }$$

where ΔV is the measured line width of the observed NH₃ spectra, σ_T (=(kT_kin/17m_H)^1/2) is the thermal broadening for NH₃ at T_kin (e.g., Dunham et al. 2011), and m_H is the mass of the hydrogen atom. We obtain a_s and σ_NT to be 0.28 and 1.42 for clump 1, respectively, while a_s = 0.25 and σ_NT = 1.44 for clump 2, respectively. These values give us Mach numbers (σ_NT/a_s) of 5 and 5.8 for clumps 1 and 2, respectively, indicating that these clumps are supersonic where nonthermal motions are dominated. The nonthermal motions could be explained by turbulence and outflows from YSOs.

Figure 6(b) shows a three-color composite image made using the GLIMPSE images (5.8 μm in red; 4.5 μm in green; 3.6 μm in blue), which is also overlaid with the 1.28 GHz and 850 μm emission. A point-like source appears the reddest and brightest in the Spitzer images and is located very close to the submillimeter emission peak (a projected linear separation of ∼8'') and the 6.7 GHz MME (a projected linear separation of ∼4''). The source was previously designated as IRS 1 and was characterized as a candidate MYSO (stellar mass ∼10 ± 2 M_⊙) based on its SED modeling from the NIR to submillimeter data (see Figure 11 in Dewangan et al. 2012). IRS 1 is barely observed in the GPS H-band image (not shown here); however, it is seen as a single object in the GPS K-band image (resolution ∼08; not shown here) and the GLIMPSE 3.6–8.0 μm images (resolution ∼2'') (see Figure 6(b)). IRS 1 is an embedded source located very close to the peak of the 6.7 GHz MME and is associated with a clump having the highest column density. Additionally, considering IRS 1 as a single bright source in the GLIMSE images, we consider the IRS 1 source as an infrared counterpart (IRc) of the 6.7 GHz MME. At the sensitivity of the observed GMRT radio map (3σ ∼ 1.14 mJy beam⁻¹), IRS 1 is situated at a projected linear separation of ∼10'' from the radio peaks seen in the 1.28 GHz map, which suggests that IRS 1 is not associated with any radio continuum peak (or an ultracompact (UC) H ii region; see Figure 6(b)). Hence, IRS 1 cannot be the powering source of the H ii region associated with the IRAS 17599–2148 region. As mentioned before, the 6.7 GHz MME is a reliable tracer of MYSOs (e.g., Walsh et al. 1998; Minier et al. 2001; Urquhart et al. 2013). Altogether, IRS 1 could be a genuine massive protostar candidate (mass ∼ 10 ± 2 M_⊙) in a very early evolutionary stage, before the onset of a UCH ii region. Based on the presence of two compact radio continuum sources (with the radio spectral class of O-type stars) and the MME, the IRAS 17599–2148 region appears to be a very similar system to W42 (see Figure 2 in Dewangan et al. 2015a).

In order to study the small-scale environment of IRS 1, we present the ESO-VLT archival NIR images. In Figure 7, we show the VLT/NACO adaptive optics images of IRS 1 in H, K_s, and L' bands. The VLT/NACO H-band image detects a faint point-like source, while K_s and L' images resolve the IRS 1 into at least three point-like sources (i.e., designated as IRS 1a, IRS 1b, and IRS 1c; see Figure 7(b)) within a scale of ∼4200 au, and one of the sources (i.e., IRS 1a) is associated with the diffuse emission (see Figures 7(b) and (c)). We find these sources in the separation range from ∼033 (1390 au) to ∼057 (2400 au). De Villiers et al. (2015) summarized that the 6.7 GHz masers switch on after the onset of the outflow. The diffuse emission seen below the ∼4200 au scale with IRS 1a might be linked to the outflow. In general, the NIR emission is attributed to scattered light escaping through the outflow cavity (e.g., Zhang & Tan 2011). Hence, the diffuse emission feature could be a jet, which appears as a narrow emission feature and may be located inside an outflow cavity as traced in the L' image. Additionally, Urquhart et al. (2013) suggested the exclusive association of Class II 6.7 GHz methanol masers with MYSOs. Hence, IRS 1a could be the main massive protostar among other point-like sources in the IRS 1 system.

Zoom In Zoom Out Reset image size

Additionally, the multiplicity of point-like sources has been observed around IRS 1 within a scale of 4200 au. Massive stars are generally found in binary and multiple systems (Duchêne & Kraus 2013). For example, in M8 and W42 star-forming regions, the well-characterized "O" stars are resolved into multiple sources using the high-resolution NIR images (Goto et al. 2006; Dewangan et al. 2015a). Around IRS 1, the multiple sources are found in the separation range from ∼033 (1390 au) to ∼057 (2400 au). Due to the coarse beam of the JCMT-SCUBA-2, the 850 μm map is unable to resolve the dense core associated with the IRS 1 source. Based on the separation of the sources in the IRS 1 multiple system, the possibility of core fragmentation may explain the origin of multiple systems (see Tobin et al. 2016, and references therein).

Based on these indicative results, further detailed studies of IRS 1 are encouraged using high-resolution interferometric observations at longer wavelength.

To trace the IRcs of two compact radio peaks (cp1 and cp2), Figure 8(a) shows a three-color composite image using 4.5 μm (red), 3.6 μm (green), and 2.2 μm (blue), which is also overlaid with the 1.28 GHz emission. The source seen close to the peak cp1 is barely observed in the GPS K-band image. The source seen close to the peak cp2 is also observed in the GPS K-band image (position and JHK photometry: α₂₀₀₀ = 18:03:01.38, δ₂₀₀₀ = −21:48:13.96; m_J = 18.22 ± 0.07 mag; m_H = 17.39 ± 0.08 mag; m_K = 17.79 ± 0.3 mag). In Figure 8(b), we present the VLT/NACO adaptive optics NIR images (H and K_s; below the 8000 au scale) toward the radio peaks (i.e., cp1 and cp2). A point-like source is seen toward each compact radio peak in the VLT/NACO images. There are no nebular features seen toward these sources in the NACO images.

Zoom In Zoom Out Reset image size

A high-resolution spectroscopic study will be very helpful to confirm the ionizing sources of the radio peaks.

The study of YSO populations is a powerful tool to depict the picture of ongoing star formation activity in a given star-forming region. In this section, we identify and classify the YSO populations using their infrared excess for a wider field of view around IRAS 17599–2148. Previously, Dewangan et al. (2012) also presented the photometric analysis of point-like sources in the IRAS 17599–2148 region, which was restricted only near the BN. In the following, we give a brief description of YSO identification and classification schemes adopted in this work.

1. In this scheme, the GLIMPSE-MIPSGAL color–magnitude diagram ([3.6]–[24]/[3.6]) is used to identify the different stages of YSOs (Guieu et al. 2010; Rebull et al. 2011; Dewangan et al. 2015a), where the sources have detections in the MIPSGAL 24 μm and GLIMPSE 3.6 μm bands. The 24 μm image is saturated near the IRAS position; therefore, this scheme is not used to trace YSOs near the IRAS position (see Figure 1(a)). Following the conditions given in Guieu et al. (2010), the boundary of different stages of YSOs is also highlighted in Figure 9(a). In our selected region, we find 170 sources that have detections in both the 3.6 and 24 μm bands. We identify 39 YSOs (4 Class I; 7 Flat-spectrum; 28 Class II) and 131 Class III sources. Furthermore, Figure 9(a) exhibits the boundary of possible contaminants (i.e., galaxies and diskless stars; also see Figure 10 in Rebull et al. 2011). We also find that our identified YSO populations are free from the contaminants. We also identify an embedded source in the IRDC, which is detected at wavelengths longer than 3.6 μm and is designated as MG008.2845+00.1664 in the MIPSGAL photometry catalog (Gutermuth & Heyer 2015). The source has 4.5–24 μm band photometry and is a YSO candidate with a color (m_8.0–m₂₄) of 3.54 mag.

Zoom In Zoom Out Reset image size

2. In this scheme, the GLIMPSE color–color diagram ([3.6]–[4.5] vs. [5.8]–[8.0]) of sources is employed that has detections in all four GLIMPSE bands. Following the Gutermuth et al. (2009) schemes, YSOs and various possible contaminants (e.g., broad-line active galactic nuclei, PAH-emitting galaxies, shocked emission blobs/knots, and PAH-emission-contaminated apertures) are identified in our selected region. The possible contaminants are also removed from the selected YSOs. Furthermore, these YSOs are classified into different evolutionary stages based on their slopes of the Spitzer-GLIMPSE SED (α_3.6–8.0) computed from 3.6 to 8.0 μm (i.e., Class I [α_3.6–8.0 > −0.3], Class II [−0.3 > α_3.6–8.0 > −1.6], and Class III [−1.6 > α_3.6–8.0 > −2.56]; e.g., Lada et al. 2006). The GLIMPSE color–color diagram ([3.6]–[4.5] vs. [5.8]–[8.0]) is shown in Figure 9(b). More details on the YSO classifications can be found in the work of Dewangan & Anandarao (2011, and references therein). This scheme leads to 15 YSOs (5 Class I; 10 Class II), 2282 photospheres, and 288 contaminants.

3. In this scheme, the color–color space ([4.5]–[5.8] vs. [3.6]–[4.5]) of sources is used that has detections in the first three GLIMPSE bands (except the 8.0 μm band). The color conditions, [4.5]–[5.8] ≥ 0.7 and [3.6]–[4.5] ≥ 0.7, are adopted to select protostars, as given in Hartmann et al. (2005) and Getman et al. (2007). A total of 11 protostars are found in our selected region (see Figure 9(c)).

4. In this scheme, the color–magnitude space (H–K/K) is adopted for selecting additional YSOs. To find a color condition, we explored the color–magnitude space of the nearby control field (size ∼5' × 5'; central coordinates: α_J2000 = 18^h02^m207, δ_J2000 = −21°50'311) and selected a color H–K value (i.e., ∼2.2) that separates large H–K excess sources from the rest of the population. Considering this color H–K cutoff criterion, 409 embedded YSOs are identified in our selected region (see Figure 9(d)).

Finally, all these schemes give us a total of 474 YSOs in our selected region (as shown in Figure 1). The positions of all these YSOs are shown in Figure 10(a). Several embedded sources are detected toward the EFS.

Zoom In Zoom Out Reset image size

The spatial distribution of young stellar populations is often examined using their surface density analysis (e.g., Gutermuth et al. 2009; Bressert et al. 2010), which allows us to infer the young stellar clusters. Previously, using the nearest-neighbor (NN) technique, Dewangan et al. (2012) also obtained the surface density map of YSOs in the IRAS 17599–2148 region, which was studied only near the BN. The surface density map can be generated by dividing the selected field using a regular grid and computing the surface density of YSOs at each grid point. The surface number density at the jth grid point is defined by ρ_j = (n − 1)/A_j (e.g., Casertano & Hut 1985), where A_j represents the surface area defined by the radial distance to the n = 6 NN. Following this procedure, we generated the surface density map of all 474 selected YSOs, using a 5'' grid and six NNs at a distance of 4.2 kpc. In Figure 10(b), the resultant surface density contours of YSOs are presented. The contour levels are drawn at 4, 6, 9, 15, and 25 YSOs pc⁻², increasing from the outer to the inner regions. The YSO clusters are mainly seen toward the IRDC and the BN (see Figure 10(b)). The star formation activity is found toward all the condensations as traced in the Herschel column density map.

We also obtained the clustered populations from distributed sources using a statistical analysis of YSOs. We estimated an empirical cumulative distribution (ECD) of YSOs as a function of NN distance (see Chavarría et al. 2008; Gutermuth et al. 2009; Dewangan & Anandarao 2011, for more details). In the ECD analysis, a cutoff length (also referred to as the distance of inflection d_c) is selected for delineating the low-density/distributed populations. For the present case, we selected a cutoff distance of d_c ∼ 44'' (0.9 pc at a distance of 4.2 kpc). The ECD analysis provided a clustered fraction of about 72% YSOs (i.e., 344 from a total of 474 YSOs).