Cluster formation in the W 40 and Serpens South complex triggered by the expanding H ii region

Abstract

We present the results of mapping observations covering a large area of 1 square degree around W 40 and Serpens South carried out in the ¹²CO (J = 1–0), ¹³CO (J = 1–0), C¹⁸O (J = 1–0), CCS (J_N = 8₇–7₆), and N₂H⁺ (J = 1–0) emission lines with the 45 m Nobeyama Radio Telescope. W 40 is a blistered H ii region, and Serpens South is an infrared dark cloud accompanied by a young cluster. The relation between these two regions, which are separated by ∼20′ on the sky, has not been recognizable so far. We found the C¹⁸O emission is distributed smoothly throughout the W 40 and Serpens South regions, and that the two regions seem to be physically connected. We divided the C¹⁸O emission into four groups in terms of the spatial distributions around the H ii region which we call 5, 6, 7, and 8 km s⁻¹ components according to their typical LSR velocity, and propose a three-dimensional model of the W 40 and Serpens South complex. We found two elliptical structures in the position–velocity diagrams, which can be explained as part of two expanding shells. One of the shells is small inner shell just around the H ii region, and the other is a large outer shell corresponding to the boundary of the H ii region. Dense gas associated with the young cluster of Serpens South is likely to be located at the surface of the outer shell, indicating that the natal clump of the young cluster is interacting with the outer shell being compressed by the expansion of the shell. We suggest that the expansion of the shell induced the formation of the young cluster.

1 Introduction

OB stars have significant influence on the physical and chemical environments of the natal clumps, and trigger formation of the new generation of stars (e.g., Elmegreen & Lada 1977). Investigation of star formation around the H ii regions is therefore essential to understand how star formation occurs and evolves by the influence of H ii regions.

W 40 and Serpens South, part of the Aquila cloud complex, are deeply embedded in dust and gas. Several infrared (IR) observations have revealed the global dust distributions of the two regions. In figure 1 we show an overview of the two regions in a three-color composite image of the 8.0 μm (red), 5.8 μm (green), and 4.5 μm (blue) emission lines from the Spitzer Space Telescope (hereafter, Spitzer).

Fig. 1.

The W 40 and Serpens South regions shown in a three-color composite image made with the 8.0 μm (red), 5.8 μm (green), and 4.5 μm (blue) data with Spitzer. The plus signs indicate the positions of the high-mass infrared sources IRS 1A South, IRS 2B, IRS 3A, and IRS 5 (Smith et al. 1985; Shuping et al. 2012). (Color online)

W 40 is a blister-type H ii region (e.g., Vallee 1987) excited by a massive OB star cluster and is characterized by the “hourglass-shaped structure” (Rodney & Reipurth 2008) which was revealed by infrared observations using the Midcourse Space Experiment and Spitzer. The hourglass-shaped structure consists of two interconnected cavities having an extent of 17′ × 30′ with one lobe oriented to the southeast and with the other to the northwest (Rodney & Reipurth 2008). Shuping et al. (2012) concluded that the O9.5 star named IRS 1A South is the dominant source of the Lyman continuum luminosity needed to power the H ii region, and is the likely source of the stellar wind that has made the hourglass-shaped structure. The dynamical age of the H ii region has been estimated to be 0.19–0.78 Myr (Mallick et al. 2013). There are dense molecular clumps in and around the H ii region (e.g., Dobashi et al. 2005; Dobashi 2011; Shimoikura et al. 2015; Rumble et al. 2016; Shimajiri et al. 2017). In our previous study on the ¹²CO (J = 3–2) and HCO⁺ (J = 4–3) emission lines (Shimoikura et al. 2015) we presented evidence that the velocity field of the region shows great complexity, consisting of at least four distinct velocity components, and that the two components at 5 km s⁻¹ and 10 km s⁻¹ are likely to be tracing dense gas interacting with the expanding shell around the H ii region. Hundreds of young stellar objects (YSOs) are associated with the region, indicating active ongoing star formation (e.g., Kuhn et al. 2010; Bontemps et al. 2010; Maury et al. 2011; Mallick et al. 2013). Kuhn et al. (2010) suggested that the YSOs are young, with an age of ≤1 Myr. These results suggest that the YSOs may be the second-generation stars born by the influence of the expanding shell of W 40. However, the previous studies were limited to a small area just around the H ii region, not covering the entire W 40 region. It is therefore not yet evident how far the interstellar medium around W 40 is affected by the ionization front.

About 20′ west of the W 40 region there is another star-forming region: Serpens South. This region can be recognized as a filamentary obscuring structure in the Spitzer image (see figure 1). A young cluster has been found around the center of the structure (Gutermuth et al. 2008). The cluster consists of a large fraction of protostars, some of which blow powerful collimated outflows out (Nakamura et al. 2011), indicating that they formed within the past 0.5 Myr (Gutermuth et al. 2008). The filamentary structure was found to consist of multiple smaller filaments identified by gas and dust observations (e.g., Kirk et al. 2013; Nakamura et al. 2014; Fernández-López et al. 2014; Könyves et al. 2015). These filaments show velocity gradients along their elongation (e.g., Kirk et al. 2013; Nakamura et al. 2014; Fernández-López et al. 2014). Kirk et al. (2013) suggested that the velocity gradient in one of the filaments is due to gas accretion toward the central young cluster. On the other hand, Fernández-López et al. (2014) pointed out that the velocity gradients along the filaments can be interpreted as a projection of large-scale turbulence, as they also found velocity gradients perpendicular to the main axes of the filaments. On a larger scale, results of low-resolution molecular observations (∼3′) covering both W 40 and Serpens South by Nakamura et al. (2017) suggest that the two regions might be physically connected. In that case, we wonder how much the Serpens South region can be influenced by the W 40 H ii region. Though each region has been widely studied individually, their relationship has remained unclear to date.

The IR observations as shown in figure 1 can reveal only part of the relation of W 40 and Serpens South, and they provide information neither on the internal structure nor the velocity field. To better illustrate the larger-scale view of the surroundings of W 40 and Serpens South, and to better understand the relation between the two regions, large-scale mapping observations in molecular emission lines with a high angular resolution are needed.

In this work we mapped the whole region shown in figure 1 in some molecular lines using the 45 m telescope at Nobeyama Radio Observatory (NRO). This study is based on “the Star Formation Legacy project” which is a large-scale survey project of molecular gas in star-forming regions. The outline of the project will be described in a forthcoming paper (Nakamura et al. 2019a). Results of wide-field polarimetric observations toward Serpens South are given in Kusune et al. (2019)—see also Sugitani et al. (2011). Results of molecular line observations toward other star-forming regions will be presented in separate articles (Dobashi et al. 2019; Nakamura et al. 2019b; Shimoikura et al. 2019; Tanabe et al. 2019).

The present study intends to understand the complex spatial and velocity structures of both the W 40 and Serpens South regions by analyzing the molecular data. Recent VLBA parallax measurements by Ortiz-León et al. (2017) indicate a distance of 436 ± 9 pc for W 40. Serpens South has the same LSR velocities as W 40 (e.g., Nakamura et al. 2017). Thus, we adopt 436 pc for both W 40 and Serpens South in this paper.

We present the observational procedures in section 2. In section 3, we show the global distributions of the molecular gas. We also describe results of analyses of the temperature distributions and velocity fields. In addition, we investigate the morphology and kinematics of the observed regions. Our conclusions are summarized in section 4.

2 Observations

2.1 Observations with the NRO 45 m telescope

Molecular observations were performed by using the NRO 45 m telescope. The observations were carried out for ∼150 hr in total during the period from 2015 April to 2017 March. The ¹²CO (115.271204 GHz), ¹³CO (110.201354 GHz), C¹⁸O (109.782176 GHz), CCS (93.870098 GHz), and N₂H⁺ (93.1737637 GHz) emission lines were observed simultaneously using the on-the-fly (OTF) mode. The half-power beam width (HPBW) of the telescope at 110 GHz is ∼15′, corresponding to ∼6500 au at a distance of 436 pc. We mapped an area of ∼1 square degree around W 40 and Serpens South in these lines.

We used the multi-beam receiver “FOREST” (FOur beam REceiver System on the 45 m Telescope; Minamidani et al. 2016) as the frontend to obtain the large maps. The typical system temperature was 170 K at 110 GHz. The backend used was the SAM45 (Spectral Analysis Machine for the 45 m telescope) digital spectrometers having a bandwidth of 31.25 MHz and a frequency resolution of 15.26 kHz. The frequency resolution corresponds to a velocity resolution of ∼0.04 km s⁻¹ at 110 GHz. For intensity calibration, the chopper-wheel method (Kutner & Ulich 1981) was used. The telescope pointing was checked by observing the SiO maser source V1111 Oph every 1.5 hr, which was found to be accurate within ∼5′. We also observed a small region (⁠|$\sim {1.^{\prime }5} \times {1.^{\prime }5}$|⁠) in Serpens South, every time we tuned the receiver to check the intensity calibration, and we found that the intensity fluctuations for all of the lines were less than 10%.

The spectra were reduced by fitting the baseline with linear functions, and the data were converted to three-dimensional data with a spheroidal convolution at an angular grid of |${7.^{\prime \prime }5}$| and a velocity resolution of 0.1 km s⁻¹. The velocity resolution was smoothed to this value to achieve a high signal-to-noise ratio. We used the NOSTAR software package (Sawada et al. 2008) for the reduction. We finally obtained the spectral data with an effective angular resolution of ∼22′ at 110 GHz.

Following the calibration formulae available at the NRO website,¹ we estimated the main beam efficiencies of the 45 m telescope for the ¹²CO, ¹³CO, C¹⁸O, CCS, and N₂H⁺ lines to be 0.416, 0.435, 0.437, 0.497, and 0.500, respectively, and used them to convert the obtained |$T_{\rm a}^*$| scale data to T_mb scale data. The one-sigma noise level of the reduced data is ΔT_mb = 0.3–0.9 K for the ∼0.1 km s⁻¹ velocity resolution. The observed molecular lines and the resulting noise levels are summarized in table 1.

Table 1.

Observed molecular lines.

Molecule	Transition	Frequency* (GHz)	Effective beam size (′)	ΔTmb† (K)
N2H+	J = 1–0	93.1737637	24	0.27
CCS	JN = 87–76	93.8700980	24	0.26
C18O	J = 1–0	109.7821760	22	0.33
13CO	J = 1–0	110.2013540	22	0.36
12CO	J = 1–0	115.2712040	22	0.87

*The rest frequency for the N₂H⁺ line is taken from Pagani, Daniel, and Dubernet (2009), and those of the other lines are taken from the website of Lovas (1992).

^† The one-sigma noise level at a velocity resolution of 0.1 km s⁻¹.

2.2 Archival data

We used Herschel archive data from the Herschel Gould Belt survey (HGBS), i.e., the 70, 250, 500 μm maps, dust temperature map, and column density map of the observed region (André et al. 2010; Könyves et al. 2015). We also used Spitzer data obtained from the public Spitzer archive.² These infrared data are compared with the molecular lines data.

3 Results and discussion

3.1 The molecular distributions

In figure 2 we show the ¹²CO, ¹³CO, and C¹⁸O spectra averaged over the observed region. The C¹⁸O spectrum has a single peak at V_LSR = 7.3 km s⁻¹. There is a dip in the ¹²CO and ¹³CO spectra at this velocity, which should be due to absorption by colder gas in the foreground.

Fig. 2.

¹²CO, ¹³CO, and C¹⁸O spectra averaged over the pixels detected at the >5 σ noise level. The vertical-dashed line indicates the peak radial velocity of the averaged C¹⁸O spectrum (V_LSR = 7.3 km s⁻¹). (Color online)

Figure 3 shows the integrated intensity maps of the C¹⁸O, CCC, and N₂H⁺ emission lines for the W 40 and Serpens South regions. As seen in figure 3a, the C¹⁸O distributions around the densest parts of the W 40 and Serpens South regions are similar to the structures found in our previous studies (e.g., Nakamura et al. 2014; Shimoikura et al. 2015). The figure clearly shows that the C¹⁸O emission is distributed over the entire observed area continuously. In addition, as we will show later, the radial velocities of the emission line change smoothly over the observed area (e.g., see figure 14). We therefore conclude that the W 40 and Serpens South regions are physically connected.

Fig. 3.

Velocity-integrated intensity maps of the (a) C¹⁸O, (c) CCS, and (d) N₂H⁺ emission lines. The velocity ranges in units of km s⁻¹ used for the integration are shown in the parentheses, and the lowest level/interval of the contour in units of K km s⁻¹ are shown in the brackets above each panel. In panel (b), the C¹⁸O map (white contours) is overlaid on a composite color image of Herschel 500 μm (red), 70 μm (green), and Spitzer 8.0 μm (blue). The white box denotes the area where the molecular data were not obtained due to the limited observing time.

In figure 3b we also show the C¹⁸O intensity map overlaid on the composite image of Herschel 500 μm (red), 70 μm (green), and Spitzer 8 μm (blue). The C¹⁸O emission is well correlated spatially with the dust emission revealed by Herschel showing several dust filaments. We also show a composite image around W 40 made from Spitzer 8 μm, 2MASS K_s, and H bands in figure 4. We can see some dark lanes absorbed by dust, which can be seen in the dust emission in the Herschel image. These dust filaments seem to be connected with Serpens South across the H ii region, and are associated with the C¹⁸O emission. One of the filaments is located at the waist of the hourglass structure, and it is known that many YSOs are associated with this dark lane (e.g., Kuhn et al. 2010; Maury et al. 2011; Mallick et al. 2013); this can be seen in figure 5, which shows the distributions of young cores around W 40 and Serpens South cataloged by André et al. (2010) and Könyves et al. (2015).

Fig. 4.

Composite image of W 40 made with Spitzer 8 μm (red), 2MASS K_s (green), and H (blue). Some dark lanes are shown with a white arrow.

Fig. 5.

Distributions of cores reported by André et al. (2010) and Könyves et al. (2015) shown on the C¹⁸O map. Starless, prestellar, and protostellar cores are indicated by black, pink, and red circles, respectively. The starless cores are the cores without YSO candidates; the prestellar cores are the cores with YSO candidates (see Könyves et al. 2015). (Color online)

In figure 3b, we also found that the H ii region is sharply outlined by a cavity of the C¹⁸O emission: There is less C¹⁸O emission on the southeast side of the W 40 region as well as in the region between W 40 and Serpens South, and such regions coincide well with the extent of the H ii region seen as the bright blue region in the Spitzer image.

In figures 3c and 3d, elongated filamentary structures are detected with the CCS and N₂H⁺ emission lines. The CCS and N₂H⁺ emission lines are dominant toward a part corresponding to the main body of Serpens South. The morphology of the N₂H⁺ emission around the main body matches well with that of the dust emission detected by Herschel (shown in red in figure 3b), which is consistent with the previous studies (Kirk et al. 2013; Nakamura et al. 2014). We also found that the region showing the N₂H⁺ emission around the OB cluster is compact, suggesting that the N₂H⁺ emission is enhanced around the OB cluster due to the high temperature. The CCS emission is not detected in the W 40 region above the 3 σ noise level, as we already found in our previous study of CCS (J_N = 4₃–3₂) lines (Shimoikura et al. 2015). In Serpens South, we found that the CCS intensity peak does not coincide with the N₂H⁺ peak corresponding to the position of the young cluster. The CCS emission is the strongest ∼8′ north of the N₂H⁺ intensity peak, and from there, it extends to the northwest in the map (figure 3c). Observations of HC₇N by Friesen et al. (2013) also show that the strong molecular emission extends to the north of the young cluster, showing distributions similar to those seen in our CCS map.

In earlier studies, the CCS and N₂H⁺ emission lines were surveyed only in a limited region around the OB cluster in W 40 (e.g., Pirogov et al. 2013; Shimoikura et al. 2015) or around the main body of Serpens South (e.g., Kirk et al. 2013; Nakamura et al. 2014). Our maps reveal the overall distributions of these molecular lines in W 40 and Serpens South, and we found that they are detected at some points along the dust filaments as well as in dust condensations in the H ii region, not only in the main body of Serpens South.

Based on the above results, we found that the ionization front clearly delineates the boundary between the ionized region and the molecular gas. The morphology of the C¹⁸O distribution shows that W 40 and Serpens South are molecular clouds in the same system, and that they are shaped by the expansion of the H ii region.

In addition, we found an extended C¹⁸O-emitting region, which may be a reflection nebula, west of Serpens South. The distribution is consistent with one of the Herschel dust filaments, and only a part of the nebula is detected in N₂H⁺ and CCS on the northern side of the filament.

3.2 The temperature distributions

In figure 3, we found that there is an anticorrelation between the C¹⁸O emission and the Spitzer image, and the molecular gas seems to be influenced by the H ii region. To investigate the influence of the H ii region, we first estimated the excitation temperature T_ex for each observed molecular line throughout the observed region. For the estimation, we assumed local thermodynamic equilibrium (LTE) and that the observed emission lines are optically thin.

Under the LTE condition, the main beam temperature T_mb is expressed by the following radiative transfer equation:

$$\begin{equation} T_ {\rm mb}=[J(T_ {\rm ex})-J(T_{\rm bg})]\lbrace 1-{\rm exp}[-\tau (v)]\rbrace , \end{equation}$$

(1)

where |$J(T)=T_{0} /(e^{T_{0} /T}-1)$|⁠, T₀ = hν/k, and the background temperature T_bg = 2.73 K. Here, ν is the rest frequency of the emission line, k is the Boltzmann constant, h is the Planck constant, and τ(v) is the optical depth as a function of velocity.

We estimated T_ex of C¹⁸O, denoted as |$T^{\rm {co}}_{\rm ex}$|⁠, from the T_mb data of the optically thick ¹²CO line (τ ≫ 1) by measuring the maximum value in the ¹²CO spectra. The line parameters of the molecules were taken from Splatalogue.³ The |$T^{\rm {co}}_{\rm ex}$| map obtained by using equation (1) is shown in figure 6a. In figure 6b, we also show the dust temperature map T_dust for comparison, which is obtained by the Herschel observations (André et al. 2010; Könyves et al. 2015).

Fig. 6.

(a) Excitation temperature map derived from the ¹²CO (J = 1–0) data. (b) Dust temperature map (André et al. 2010; Könyves et al. 2015). (c) Excitation temperature map of N₂H⁺ (J = 1–0) (middle), and close-up views of the W 40 region (left) and the Serpens South region (right). The C¹⁸O intensity map in figure 3a is overlaid in panels (a) and (b) by the white contours. The lowest contour and the contour interval are 4 K km s⁻¹ and 2 K km s⁻¹. (Color online)

For the N₂H⁺ molecule, we estimated the line parameters including the excitation temperature |$T^{\rm {N{_2}H^{+}}}_{\rm ex}$| as follows: The N₂H⁺ (J = 1–0) line has seven hyperfine components (e.g., Caselli et al. 1995). Because all of the components have the same |$T^{\rm {N{_2}H^{+}}}_{\rm ex}$| and line width ΔV, we simultaneously fitted seven Gaussian components to the observed N₂H⁺ (J = 1–0) spectrum at each position using equation (1). Here, τ(v) for the N₂H⁺ line can be expressed as

$$\begin{equation} {\tau }(v)={\tau _{\rm tot}}\sum _{i=1}^{7} C_{i} {\rm exp} \frac{{-(v-V_{0}+v_{i})^{2}}}{{2\, \sigma _{i}}^{2}} , \end{equation}$$

(2)

where C_i is the normalized relative intensity (i.e., |$\sum _{i=1}^{7} C_{i}=1$|⁠) and v_i is the artificial velocity difference due to the slight differences of the rest frequencies. The parameters for the estimations are taken from Tiné et al. (2000) and Splatalogue. σ_i is the velocity dispersion, which can be expressed as |${\Delta V}/\sqrt{8 \ln 2}$|⁠, where ΔV is the velocity dispersion measured at full width at half maximum (FWHM). We selected positions where the N₂H⁺ emission is detected with ≥10 σ, and performed the fitting to derive the total optical depth τ_tot of the seven hyperfine components, the peak LSR velocity V₀, ΔV, T_mb, and |$T^{\rm {N{_2}H^{+}}}_{\rm ex}$| for each position. We show the fitted values of |$T^{\rm {N{_2}H^{+}}}_{\rm ex}$| across the observed region in figure 6c. In figure 7, we also show an example of the observed and fitted N₂H⁺ spectra together with the C¹⁸O and CCS spectra measured at some positions. Here, the derived parameters of N₂H⁺ at the position shown in figure 7c are V₀ = 7.66 ± 0.01 km s⁻¹, τ_tot = 3.18 ± 0.06, ΔV = 1.72 ± 0.03 km s⁻¹, and |$T^{\rm {N{_2}H^{+}}}_{\rm ex}=10.16\pm 0.40\:$|K. Similarly, the parameters at the position shown in figure 7e are V₀ = 4.65 ± 0.01 km s⁻¹, τ_tot = 4.24 ± 0.11, ΔV = 0.70 ± 0.01 km s⁻¹, and |$T^{\rm {N{_2}H^{+}}}_{\rm ex}=8.47\pm 0.70\:$|K.

Fig. 7.

C¹⁸O, N₂H⁺, and CCS spectra observed at the positions indicated by the black plus signs. The results of the hyperfine fit of the N₂H⁺ spectra are also shown. The map in the middle represents the N(H₂) distribution derived from the Herschel data in units of cm⁻² (e.g., André et al. 2010). The lowest contour and the contour interval of the N(H₂) map are 3 × 10²² cm⁻². (Color online)

In figure 6a, |$T^{\rm {co}}_{\rm ex}$| reaches a value of 56 K in the vicinity of the OB cluster, whereas |$T^{\rm {co}}_{\rm ex}$| around the Serpens South region stays around ∼10 K and exceeds 15 K only around the young cluster. We note that the derived values of |$T^{\rm {co}}_{\rm ex}$| are the lower limits since the ¹²CO line often shows heavy self-absorption around ∼7 km s⁻¹, as seen in figure 2. The T_dust map clearly shows that the distribution is similar to the shape of the H ii region traced by Spitzer, showing that the temperature decreases with increasing distance from the ionized gas. Here, N₂H⁺ will be destroyed by CO evaporated from grain surfaces at 25 K (e.g., Lee et al. 2004). We found that T_dust around the OB cluster is higher than 25 K. As shown in figure 3d, the N₂H⁺ emission around the OB cluster is compact, and we thus suggest that the N₂H⁺ molecule is largely destroyed by CO evaporating from dust. In the Serpens South region, there is a tendency that the temperature becomes higher around the young cluster in figure 6c: |$T^{\rm {N{_2}H^{+}}}_{\rm ex}$| is >5 K, which increases to ∼11 K in the dense regions where the young cluster is located. The temperature around the young cluster is also high in the T_dust map.

The temperature distributions of |$T^{\rm {co}}_{\rm ex}$|⁠, T_dust, and |$T^{\rm {N{_2}H^{+}}}_{\rm ex}$| roughly show a similar tendency: All of the maps show a trend that the temperature decreases with increasing distance from the H ii region, suggesting that the molecular gas and the dust in W 40 are warmed by the radiation from the OB cluster.

3.3 The velocity structure

To investigate the global velocity structures, we made mean velocity V₀ maps of the C¹⁸O, CCS, and N₂H⁺ emission lines. For C¹⁸O and CCS, T_mb was measured through fitting the lines with a single Gaussian at each observed position. V₀ was then measured at each position as

$$\begin{equation} {V_0}=\frac{{\int T_{\rm mb}v\, dv}}{{\int T_{\rm mb}\, dv}} . \end{equation}$$

(3)

The results are shown in figures 8a and 8b. For the N₂H⁺ (J = 1–0) line in figure 8c, we show the values of V₀ fitted with equations (1) and (2), as described in the previous subsection.

Fig. 8.

Mean velocity maps of the (a) C¹⁸O, (b) CCS, and (c) N₂H⁺ emission lines. The map of N₂H⁺ was obtained by fitting all of the hyperfine lines (see text). (d) Close-up view of each map for the Serpens South region. (Color online)

The global spatial distributions of V₀ of the three emission lines show a similar velocity gradient between the inner regions of W 40 and the Serpens South region. We found that Serpens South is accompanied by the molecular gas with multiple velocity components, as reported in the previous studies for limited regions (e.g., Kirk et al. 2013; Fernández-López et al. 2014; Nakamura et al. 2014): In the southern part of Serpens South, the three emission lines are detected at ∼6–7 km s⁻¹, while these are detected at higher velocities in its northern part.

In figure 9 we also show the velocity dispersion ΔV maps for the three emission lines. The ΔV maps for the C¹⁸O and CCS emission lines are derived from a single Gaussian fitting for each line profile. In the case of C¹⁸O, ΔV around the H ii region is noticeably larger. The reason for the larger ΔV is that there are multiple velocity components, as can be seen in the spectra of the figure. We also found that ΔV's of CCS and N₂H⁺ tend to be large around the young cluster of Serpens South, which is also seen in the spectra shown in figure 7c. This is also thought to be due to the existence of multiple velocity components, as we provide a suggestion in subsection 3.5.

Fig. 9.

Velocity width (FWHM) maps of the (a) C¹⁸O, (b) CCS, and (c) N₂H⁺ emission lines. In panels (1)–(3), the C¹⁸O spectra observed at some positions indicated by the black arrows in panel (a) are shown, where short orange arrows denote distinct velocity components seen in the spectra, and two values in parentheses represent the coordinates of the position. (Color online)

3.4 The distribution of the fractional abundance

To understand the chemical characteristics of the observed region, we estimated the fractional abundances of C¹⁸O, CCS, and N₂H⁺. On the assumption of the LTE condition, the column densities of C¹⁸O and CCS—N(C¹⁸O) and N(CCS), can be approximated using the integrated intensity of the lines, ∫T_mbdv, according to the following formula (e.g., Hirahara et al. 1992; Mangum & Shirley 2015):

$$\begin{equation} N=\frac{3h}{8\pi ^{3}}\frac{Q}{\mu ^{2}S_{ij}}\frac{e^{E_{\rm u}/k{T}_{\mathrm{ex}}}}{e^{T_{0}/T_{\mathrm{ex}}}-1}\frac{\int \beta ^{-1}T{\rm _{mb}}dv}{J(T_{\mathrm{ex}})-J(T\mathrm{_{bg}})} , \end{equation}$$

(4)

where Q is the partition function approximated as Q = k T_ex/h B₀ + 1/3 (e.g., Mangum & Shirley 2015). B₀ is the rotational constant of the molecule. μ is the dipole moment, E_u is the energy of the upper level, and S_ij is the intrinsic line strength of the transition from i to j state. β is the escape probability related to the optical depth τ as β = (1 − e^−τ)/τ, and β = 1 for τ ≪ 1. For T_ex, we adopted |$T^{\rm {co}}_{\rm ex}$| for C¹⁸O, and a fixed value of 5 K for CCS measured in dark clouds (e.g., Hirota et al. 2009). The spectral-line parameters are taken from Splatalogue.

For the estimation of the column density of N₂H⁺, N(N₂H⁺), we used the following formula (e.g., Caselli et al. 2002; Mangum & Shirley 2015):

$$\begin{equation} N({\rm N}_{2}{\rm H}^{+})=\frac{8\pi ^{3/2}}{2\sqrt{\ln 2}}\frac{Q {\nu ^{3}}}{c^{3}A g_{\rm u}}\frac{\tau _{\rm tot}\Delta V}{1-e^{-{T_{0}/T_{\mathrm{ex}}}}} , \end{equation}$$

(5)

where g_u is the degeneracy of the upper level of a transition, and A is the Einstein coefficient for spontaneous emission. We calculated Q using the excitation temperature |$T_{\rm ex}^{\rm N_2H^+}$| derived in subsection 3.2. We summarize the line parameters used for this study in table 2.

Table 2.

Constants of the observed molecular lines.

Molecule	Transition	Sij	B  0 (GHz)	μ (Debye)	E  u (cm−1)	g  u	A (s−1)
N2H+	J = 1–0	—	46.586867	—	—	3	3.628 × 10−5*
CCS	JN = 87–76	7.97	6.47775036	2.88	13.82557	—	—
C18O	J = 1–0	1.00	54.8914	0.110	3.66194	—	—

* Pagani, Daniel, and Dubernet (2009).

We then estimated the fractional abundances of the three molecules, f (C¹⁸O), f (N₂H⁺), and f (CCS), by dividing their derived column densities by the H₂ column density derived from the dust data with Herschel. For this, we regridded and smoothed the H₂ column density data up to the same angular resolution as the 45 m data (∼22′). Figure 10 shows the resultant fractional abundance maps.

Fig. 10.

Fractional abundance maps of (a) C¹⁸O, (b) CCS, and (c) N₂H⁺. We also show the N(H₂) map (e.g., André et al. 2010) in panel (d). The lowest contour level and contour interval for the N(H₂) map are 1 × 10²² cm⁻². (Color online)

The value of f(C¹⁸O) reaches as high as ∼10⁻⁶ around the center of W 40, but it is apparently much lower in Serpens South, which may be due to the depletion of C¹⁸O onto dust in the cold and dense interior of the Serpens South filament. On the other hand, f (N₂H⁺) is high in Serpens South. f (CCS) is also high in Serpens South, which means that the Serpens South region is in an earlier evolutionary stage than the W 40 region (e.g., Shimoikura et al. 2012, 2018). In the Serpens South region, the low values of f (C¹⁸O), f (CCS), and f (N₂H⁺) are seen at the position of the young cluster. Because the difference in the fractional abundances suggests a difference in chemical reaction time (e.g., Suzuki et al. 1992), the northern part of the Serpens South filament is likely to be younger than the southern part. The next cluster formation may occur in the northern part of the filament, as suggested by Nakamura et al. (2014).

3.5 The four velocity components

To investigate the velocity distribution of the C¹⁸O emission in more detail, and to investigate the relation between the ionization front and the molecular gas, we drew channel maps of the C¹⁸O emission in steps of 0.3 km s⁻¹, and show them in figure 11 together with the Spitzer composite image (the same as figure 1).

Fig. 11.

Channel maps of the C¹⁸O emission drawn at every 0.3 km s⁻¹ in the velocity range of from 4.3 km s⁻¹ to 8.8 km s⁻¹. Each panel is overlaid on the image shown in figure 1. The velocity range (in units of km s⁻¹) used for the integration is indicated in brackets of each panel. The lowest contour level and the contour interval are 0.5 K km s⁻¹. The color scale for the C¹⁸O intensity is indicated in the top right corner. (Color online)

In the figure, the C¹⁸O distribution shows a complex velocity structure and varies from panel to panel. Most of the C¹⁸O emission is concentrated at ∼7 km s⁻¹, and the emission associated with Serpens South is also found around this velocity. At low velocities in panels (a)–(e), the C¹⁸O emission is found only around the OB cluster. At higher velocities in the other panels, the emission tends to be located on the periphery of the H ii region seen in the Spitzer image, for which we suggest that the C¹⁸O emission should trace the gas swept up by the expansion of the H ii region.

As shown in panels (1)–(3) of figure 9, a number of velocity components were detected in C¹⁸O in the observed region. Based on the inspection of the channel maps, we decided to categorize the emission into four groups in terms of the apparent spatial distribution around the H ii region. We shall call them 5, 6, 7, and 8 km s⁻¹ components according to their typical LSR velocity. These components show different features and are helpful in understanding the global velocity structure of the observed region. Figure 12 shows the distributions of the four components superposed on the Spitzer image (figure 1). In the following, we summarize the characteristics of the components seen in figures 8–12.

• 5 km s⁻¹ component: As seen in figure 12a, the molecular gas with LSR velocities in the range 4.3 ≲ V_LSR ≲ 5.9 km s⁻¹ is mainly located just around the H ii region. The excitation temperature of this component is very high in the H ii region, suggesting that the component is located near the exciting sources (the OB cluster) and is warmed by the radiation from them.

• 6 km s⁻¹ component: As seen in figure 12b, the molecular gas with LSR velocities of 5.9 ≲ V_LSR ≲ 6.8 km s⁻¹ is found around the H ii region and in the southern part of the main body of Serpens South. The component located around the H ii region shows a ring-like structure surrounding the 5 km s⁻¹ component. In addition, there is a filament at ∼6 km s⁻¹ with an elongation of ∼3 pc at the western most side in the observed area (as labeled “filament A”). This filament shows an anticorrelation with the 8 km s⁻¹ component shown in figure 12d.

• 7 km s⁻¹ component: As seen in figure 12c, the molecular gas with LSR velocities in the range 6.8 ≲ V_LSR ≲ 7.5 km s⁻¹ is found mainly around the Serpens South region as well as the periphery of the H ii region. The diffuse and extended emission is also detected in the H ii region, being coincident with the dark lane. The component associated with the dark lane extends from the Serpens South region to the W 40 region, as traced by the broken line.

• 8 km s⁻¹ component: As seen in figure 12d, the molecular gas with LSR velocities in the range 7.5 ≲ V_LSR ≲ 8.8 km s⁻¹ appears to lie in the boundary of the H ii region. The component is seen mostly on the northwestern side of the H ii region.

Fig. 12.

C¹⁸O intensity distributions of the four velocity components shown by the black/white contours (see text). The lowest contour and the contour interval are 1.4 K km s⁻¹. The background image is the same as that in figure 1. The velocity ranges in units of km s⁻¹ used for the integration of each component are shown in parentheses above each panel. (Color online)

In our previous study using the ¹²CO (J = 3–2) and HCO⁺ (J = 4–3) emission lines, performed only in the limited region around the W 40 H ii region, we reported that there are velocity components at V_LSR ≃ 3, 5, 7, and 10 km s⁻¹ (Shimoikura et al. 2015). These velocity components are mostly single-peaked components with well-defined radial velocities and are not categorized in the same way as the four components above, but the 5 and 7 km s⁻¹ components found in the previous study can be merged into those categorized in this work. The ∼3 and ∼10 km s⁻¹ components found in the previous work are not detected in C¹⁸O in this study with a good signal-to-noise ratio. However, they are clearly detected in the ¹²CO (J = 1–0) and ¹³CO (J = 1–0) lines. In figure 13, we show the distributions of the ∼3 and ∼10 km s⁻¹ components detected in the ¹³CO (J = 1–0) data; we shall call them the 3 and 10 km s⁻¹ component, respectively. These components are seen around the OB cluster, as we have already found in the earlier study. In this study, we further found that the 10 km s⁻¹ component is widely distributed over the observed area, especially on the periphery of the H ii region, like the 8 km s⁻¹ component in figure 12d.

Fig. 13.

Distributions of the 3 km s⁻¹ and 10 km s⁻¹ components identified based on the ¹³CO data (see text). The velocity ranges in units of km s⁻¹ used for the integration of each component are shown in parentheses above each panel. The lowest contour and the contour interval are 3.0 km s⁻¹. The background images are the same as in figure 1. (Color online)

In summary, the distributions of the identified components indicate interaction between the H ii region and the surrounding molecular gas. The components at higher velocities (i.e., 7, 8, and 10 km s⁻¹) tend to exhibit an arc-shaped structure and are located in the boundary of the H ii region, suggesting that they may be tracing the expanding shell of the H ii region.

3.6 The two expanding shells

To understand the spatial and velocity distributions of molecular gas well, and to investigate the effect of the expanding H ii region in the surroundings, we made position–velocity (PV) diagrams across the observed region using the C¹⁸O data. Figure 14 displays a series of PV diagrams measured along cuts centered at the position of IRS 1A South in various directions. In each PV diagram, the position of IRS 1A South as well as the positions separated by ±1000′ from the source, roughly corresponding to the boundary of the H ii region, are indicated by the white broken line.

Fig. 14.

PV diagrams of the C¹⁸O emission line taken along the cuts (a)–(r) shown in the C¹⁸O intensity map at the top-left panel. The lowest contour and the contour interval are both 0.6 K. The vertical red broken line indicates the systemic velocity (V_LSR = 7.3 km s⁻¹). The white broken lines indicate the position of IRS 1A South and those separated by ±1000′ from the source roughly corresponding to the boundary of the H ii region. Two elliptical structures are marked with a pink ellipse in panels (i) and (p). (Color online)

We found elliptical structures in the diagrams: we indicate two of them by an ellipse in figures 14i and 14p where the structures are evident. The elliptical structures should be due to expanding shell(s) of the H ii region centered at IRS 1A South or the OB cluster. We suggest that there are two shell-like structures: one is the small inner shell just around IRS 1A South shown by the ellipse in figure 14p, and the other is the large outer shell corresponding to the boundary of the H ii region delineated by the ellipse in figure 14i. The ellipses in the panels indicate that the inner and outer shells have radii of ∼0.5 pc and ∼2.5 pc, respectively, and an expanding velocity of ∼3 km s⁻¹. The expansion time scales of the two shells can be estimated, by dividing their radii by the expanding velocity, to be 1.6 × 10⁵ yr and 8.1 × 10⁵ yr for the inner and outer shells, respectively, which is consistent with the dynamical age of the H ii region, (1.9–7.8) × 10⁵ yr, estimated by Mallick et al. (2013) based on radio continuum observations. We also estimated the mass within the outer shell (the northwestern side of the W 40 region) to be ∼1 × 10³ M_⊙.⁴

We suggest a possibility that the two shell-like structures were created due to the inhomogeneous density distribution of the dense gas around the H ii region. As seen in figure 12, there are some patchy dense clumps (e.g., the 5–7 km s⁻¹ components) in the vicinity of IRS 1A South and the OB cluster. Expansion of the H ii region should be slowed or blocked by such clumps, and some parts of the expanding shell facing the clumps should appear to be an inner shell, while the rest of the shell should appear to be an outer shell, as illustrated in figure 15. We further suggest that this may be the very mechanism to create the hourglass-shaped structure of the H ii region.

Fig. 15.

Schematic representation of the W 40 H ii region, showing that the two shells were created by the expansion of the H ii region. The small inner shell is found in the vicinity of the H ii region, and the other shell is the larger outer one corresponding to the boundary of the H ii region.

3.7 The cluster formation in Serpens South

We investigated whether the expansion of the H ii region influenced the cluster formation in Serpens South. Figure 16a (same as figure 14b) shows the PV diagram measured along the cut b′–b crossing the position of the young cluster of Serpens South. The outer shell delineated by the ellipse in figure 16a is not seen very well in the C¹⁸O emission, so we show the PV diagrams of the ¹³CO and ¹²CO emission lines taken along the same cut in figures 16b and 16c, respectively, where the outer shell can be better recognized. As seen in the PV diagrams, it is very likely that the dense gas associated with the young cluster (evident in figure 16a) is located at the surface of the outer shell, indicating that the natal clump of the cluster can be affected by the shell. Actually, as seen in figure 16a, the velocity dispersion is enhanced abruptly at the position of the cluster (up to ∼2 km s⁻¹) compared to the adjacent positions, which can be also recognized in the ΔV maps presented in figures 9b and 9c. Though the observed enhancement of the velocity dispersion may be partially due to the feedback of star formation (e.g., outflows), this may support the idea that the natal clump of the young cluster is interacting with the outer shell being compressed by the expansion of the shell, which may have induced the formation of the young cluster.

Fig. 16.

(a) PV diagram of the C¹⁸O emission line (the same diagram as figure 14b). (b)–(c) PV diagrams of the ¹³CO emission line and the ¹²CO emission line taken along the same cut labeled b′–b. The lowest contour and contour interval are 1.0 K for panel (b) and 4.0 K for panel (c). (d) PV diagram of the C¹⁸O emission line (color scale + white contour) and CCS emission line (black contour) taken along the cut A–A′. The lowest contour and contour interval are 0.6 K for C¹⁸O and 0.4 K for CCS. The vertical broken line indicates the systemic velocity (V_LSR = 7.3 km s⁻¹) for panels (a)–(d). (Color online)

To further investigate the influence of the outer shell, we created a PV diagram for the C¹⁸O and CCS emission lines along the major axis of the Serpens South filament, which is shown in figure 16d. In the figure two major velocities components can be found, labeled as “X” and “Y,” which correspond to subfilaments extending to the southeast and northwest, respectively. The subfilament X has a velocity (∼6 km s⁻¹) slightly lower than the systemic velocity (∼7 km s⁻¹), and the subfilament Y has the same velocity as the systemic velocity. The young cluster is apparently located at the intersection of the subfilaments X and Y, and both are located around the surface of the outer shell. This positional coincidence suggests a scenario whereby the subfilament X has been accelerated by the expansion of the outer shell and recently collided against the subfilament Y to induce cluster formation at the intersection.

The previous studies (e.g., Kirk et al. 2013; Nakamura et al. 2014) also reported that Serpens South consists of molecular filaments with a different velocity. Nakamura et al. (2014) suggested that the collision of filaments with a different radial velocity may have triggered the cluster formation, as we discuss in the present study. The PV diagram of figure 16d shows a velocity gradient of ∼0.7 km s⁻¹ pc⁻¹ in the northwesterly direction, which is consistent with that found by the previous studies (Kirk et al. 2013; Fernández-López et al. 2014). Kirk et al. (2013) deduced that the velocity gradient is evidence of material flowing inside the filaments, leading to infall toward the intersection of the filaments where the cluster formation takes place. However, Fernández-López et al. (2014) suggested that the velocity gradients can also occur in scenarios where filaments are created by large-scale turbulence.

We basically agree with the scenario suggested by Nakamura et al. (2014) that the cluster formation was induced by collisions of filaments. Our subfilaments X and Y correspond to their filaments F8 and F1–3, respectively (see their figure 3). We further suggest that the outer shell of the W 40 H ii region has played a crucial role in accelerating the southeastern part of the filament (i.e., subfilament X) leading to the collision of the filaments. Also, the outer shell may have directly influenced the natal clump of the young cluster by compressing it to trigger the cluster formation.

3.8 The global 3D structure

In our previous work (Shimoikura et al. 2015), we investigated the three-dimensional structure of W 40 just around the H ii region. Because we are now reasonably confident that W 40 and Serpens South belong to the same system, we try to investigate the three-dimensional structure of the entire system based on the data obtained in this study.

In figure 17 we present the spatial distributions of each component together with the IR images. Considering the velocity difference of each component with respect to the systemic velocity at 7 km s⁻¹, as well as the expanding motion of the H ii region, the W 40 + Serpens South system should be structured basically in such a way that the 5 km s⁻¹ and 6 km s⁻¹ components are located on the near side of the H ii region facing toward us, and the 8 km s⁻¹ component is located on the far side.

Fig. 17.

Composite color images of W 40 and Serpens South made from the integrated intensity maps, with the 5 km s⁻¹ component in blue, the 6 km s⁻¹ component in pink, the 7 km s⁻¹ component in green, and the 8 km s⁻¹ component in red, respectively. The background image is the same as in figure 1 for panel (a), and the Herschel 250 μm image for panel (b).

Based on the above picture, we attempt to reveal the locations of the individual components in the system. The 5 km s⁻¹ component is distributed only around the center of the H ii region. It must be the dense gas close to and being blown away by the H ii region, and moving toward us. This component is probably located at the inner shell. The 6 km s⁻¹ component and part of the 7 km s⁻¹ component seen around the main body of Serpens South appear as absorption (i.e., dark) in the Spitzer image, delineating the shape of the Serpens South filaments. The absorption indicates that they must be the dense gas located in the foreground of the H ii region. The dark lane located in the waist of the H ii region at 7 km s⁻¹ also obscures the central part of the H ii region in the Spitzer image, and therefore it must be located in the foreground. The 8 km s⁻¹ component is found mainly in the northern part of the observed region, and it does not appear as absorption in the Spitzer image, suggesting that it is rather located behind the H ii region.

We further attempt to model the locations of the 3 and 10 km s⁻¹ components identified with the ¹³CO data in the same way. We found a high temperature of ∼30 K for the components, as seen in figure 6a, suggesting that they are located close to the central OB stars. As seen in the PV diagram of figure 16c, the 10 km s⁻¹ component is apparently located at the inner shell of the H ii region on the far side from the observer, which is consistent with the suggestion in our previous study (Shimoikura et al. 2015). The 3 km s⁻¹ component is also likely to be located at the inner shell but on the near side to the observer, the same as in the case of the 5 km s⁻¹ component.

In figure 18 we summarize the geometry of the entire W 40 and Serpens South system inferred from the above discussion. The morphology and kinematics around the W 40 H ii region are consistent with what we have proposed before (Shimoikura et al. 2015). Finally, using the Herschel-based N(H₂) data, we estimated masses of the four components to be ∼200, ∼1200, ∼5000, and ∼2000 M_⊙ for the 5, 6, 7, and 8 km s⁻¹ components, respectively.⁵

Fig. 18.

Schematic illustration of the proposed model for the three-dimensional geometry of the W 40 and Serpens South complex. As indicated in the box drawn with the broken line, panel (a) shows the view as observed on the sky (e.g., see figure 17), panel (b) shows the view observed from the declination axis perpendicularly to the line of sight, and panel (c) shows the view observed from the right ascension axis. There are two expanding shells created by the H ii region (see figure 15). In the model, the 3 and 5 km s⁻¹ components are located on the near side of the inner shell, and the 10 km s⁻¹ component is located on the far side of the inner shell. The 6 and 7 km s⁻¹ components are located around the surface of the outer shell. Both components associated with the young cluster of Serpens South are likely to be interacting with the outer shell. The 8 km s⁻¹ component is located on the far side of the outer shell. (Color online)

As seen in the PV diagrams (figure 14), the velocity distribution of each component is narrow with a line width of ∼1 km s⁻¹, except for the 5 km s⁻¹ component. Beaumont and Williams (2010), who observed 43 bubbles associated with H ii regions, indicated that there is often a ring of cold gas driven by massive stars, and that H ii regions form in flat molecular clouds with a thickness of a few parsec. We point out the possibility that some filamentary structures seen around the equator of the hourglass-shaped W 40 H ii region can be an edge-on view of a sheet-like or ring-like cloud, as suggested by Beaumont and Williams (2010) for other H ii regions.

In general, as an advancing ionization front wraps pre-existing clumps, its pressure triggers a gravitational collapse and star formation (e.g., Elmegreen & Lada 1977). The two shells we found in this study have been apparently created by the H ii region, and they may be such a place of sequential star formation, as it is actually evident in figure 5 showing the distributions of YSOs (André et al. 2010; Könyves et al. 2015). In the W 40 and Serpens South system, star formation probably occurred in W 40 first, and then the next-generation stars are forming in clouds along the expanding shells. We suggest that the young cluster in Serpens South was also induced by the expansion of the H ii region. This idea is consistent with the fact that the CCS emission is strongly detected in the Serpens South region while it is not detected in W 40, indicating that the Serpens South region is much younger than the W 40 region.

4 Conclusions

Large-scale mapping observations of the star-forming regions of W 40 and Serpens South were performed in ¹²CO, ¹³CO, C¹⁸O, CCS, and N₂H⁺ by using the NRO 45 m telescope, revealing the complex distributions of the molecular emission lines and their radial velocities in the regions. We found that the C¹⁸O emission is smoothly distributed over both the W 40 and Serpens South regions, suggesting that the two regions belong to the same system. The CCS and N₂H⁺ emission lines are enhanced in Serpens South. Serpens South is characterized by its filamentary structure in infrared images, such as by Spitzer and Herschel, and it is accompanied by a young cluster near the center of the structure. On the other hand, in W 40, the CCS emission is not detected significantly, and the N₂H⁺ emission is detected almost only around an OB cluster at the center.

Based on the C¹⁸O observations, we divided the molecular gas into four velocity components which we call the 5, 6, 7, and 8 km s⁻¹ components according to their typical LSR velocities. We found that two elliptical structures are seen in the PV diagrams, which can be explained as parts of two expanding shells. One of these shells is the small inner shell found in the vicinity of the H ii region, and the other is the larger outer shell corresponding to the boundary of the H ii region. The dense gas associated with the young cluster in Serpens South is likely to be interacting with the outer shell, for which we suggest that the cluster was induced to form in interaction with the expanding outer shell.

Strong CCS emission is detected in Serpens South while it is not detected in W 40, which means that the Serpens South region is younger than the W 40 region in terms of chemical reaction. This result is consistent with the scenario that the star formation initially occurred in W 40 and then the young cluster in Serpens South was induced to form in interaction with the expanding shell of the H ii region.

Based on the morphologies and velocity distributions of the four velocity components, as well as their appearance in the infrared images, we have made a three-dimensional geometrical model of the W 40 and Serpens South regions.

Acknowledgments

We are grateful to the referee for providing useful comments to improve this paper. We thank Yoshiko Hatano, Akifumi Yamabi, Sho Katakura, and Asha Hirose for their support of the observations. We are also grateful to the other members of the Star Formation Legacy project. This work was supported by JSPS KAKENHI Grant Numbers JP17K00963, JP17H02863, JP17H01118, JP16K12749. YS received support from the ANR (project NIKA2SKY, grant agreement ANR-15-CE31-0017). The 45 m radio telescope is operated by NRO, a branch of NAOJ. This research has made use of data from the Herschel Gould Belt survey (HGBS) project 〈http://gouldbelt-herschel.cea.fr〉. The HGBS is a Herschel Key Programme jointly carried out by SPIRE Specialist Astronomy Group 3 (SAG 3), scientists of several institutes in the PACS Consortium (CEA Saclay, INAF-IFSI Rome and INAF-Arcetri, KU Leuven, MPIA Heidelberg), and scientists of the Herschel Science Center (HSC).

Footnotes

1

〈https://www.nro.nao.ac.jp/∼nro45mrt/html/prop/eff/eff-intp.html〉.

2

〈http://archive.spitzer.caltech.edu/〉.

3

〈www.splatalogue.net〉.

4

The mass is derived from the Herschel-based N (H₂) data as M = αm_H∫_SN (H₂)dS, where S is the area of the H ii region determined within the radius of the outer shell (R = 2 pc), α is the mean molecular weight (taken to be 2.8), and m_H is the mass of a hydrogen atom.

5

We defined the extents of the four components by the 1.0 K km s⁻¹ contours of the C¹⁸O intensity (e.g., figure 12), and derived the masses of the components by integrating N (H₂) within their extents. For a pixel belonging to two velocity components or more, the mass in the pixel is shared by the components proportionally to the C¹⁸O intensities of the components.