New insights in the HII region G18.88$-$0.49: hub-filament system and accreting filaments

We examine the infrared, sub-millimeter (mm), and radio continuum images of G18.88$-$0.49 to investigate its physical environment (see Figure 2). In Figures 2a–2i, we display images at 3.6 $\mu$m, 8.0 $\mu$m, 70 $\mu$m, 160 $\mu$m, 250 $\mu$m, 350 $\mu$m, 500 $\mu$m, 870 $\mu$m, and 1950 MHz, respectively. The positions of the 6.7 GHz methanol maser and the RRL observations are also marked in the multi-wavelength images. The IRDCs seen in the images at 3.6 and 8.0 $\mu$m appear as bright emission regions in the images at 70–870 $\mu$m. As highlighted earlier, SDC18 is associated with the 6.7 GHz methanol maser and the EGO. In the sub-mm images, we find a bright condensation/clump hosting the 6.7 GHz methanol maser/EGO. The Herschel and ATLASGAL images show the existence of potential filaments in our target site (see dashed arrows in Figures 2g and 2h). A shell-like feature is also seen in these images (see a contour and a solid arrow in Figure 2g), and appears to be connected with potential filaments (see Sections 4.1.4 and 4.2 for quantitative information). In the direction of the shell-like feature, the NH₃ (1,1) emission has been observed (see Figure 6 in Li et al., 2019), suggesting the presence of dense gas (see plus symbols in Figure 2g).

The positions of 17 ATLASGAL dust clumps are also shown in Figure 2h (see empty and filled hexagons). In Table 2, we provide the positions and physical parameters of the clumps from Urquhart et al. (2018). All these clumps are traced in a V_lsr range of $\sim$[61, 67.5] km s^-1. We also compute the average volume density of each clump as $n_{\mathrm{H}_{\mathrm{2}}}$ = 3M_clump/(4$\pi$$R_{\mathrm{clump}}^{3}$$\mu_{\mathrm{H}_{\mathrm{2}}}m_{\mathrm{H}}$) (i.e., assuming that they are approximately spherical). Here, $R_{\mathrm{clump}}$ is the clump effective radius, $M_{\mathrm{clump}}$ is the clump mass, $m_{\mathrm{H}}$ is the mass of a hydrogen atom and the mean molecular weight $\mu_{\mathrm{H}_{\mathrm{2}}}$ is assumed to be 2.8. The values of $n_{\mathrm{H}_{\mathrm{2}}}$ for all the clumps are also tabulated in Table 2, and vary between $\sim$2 $\times$ 10² and $\sim$1.27 $\times$ 10⁴ cm^-3. We trace a dust temperature range of $\sim$13.7–25 K. We find 8 clumps with $M_{\mathrm{clump}}$ $>$ 10³ M_⊙ (see clump IDs with daggers in Table 2).

In Figure 2h, we also highlight six clumps (see filled hexagons) for which their physical parameters were derived from the NH₃ line data from Wienen et al. (2012); such parameters are listed in Table 3. Using these parameters, we also derive the sound speed ($a_{\mathrm{s}}$), thermal velocity dispersion ($\sigma_{T}$), non-thermal velocity dispersion ($\sigma_{\mathrm{NT}}$), Mach number ($M$ = $\sigma_{\mathrm{NT}}$/$a_{\mathrm{s}}$), and ratio of thermal to non-thermal gas pressure ($R_{\mathrm{p}}$ = ${a^{2}_{\mathrm{s}}}/{\sigma^{2}_{\mathrm{NT}}}$; see Lada et al., 2003, for more details). These parameters are also included in Table 3. The value of the sound speed $a_{\mathrm{s}}$ (= $(kT_{\mathrm{kin}}/\mu m_{\mathrm{H}})^{1/2}$) is computed considering $\mu$ = 2.37 (approximately 70% H and 28% He by mass) and the kinematical temperature ($T_{\mathrm{kin}}$) as given in Table 3. The non-thermal velocity dispersion is given by the following equation:

where $\Delta V$ is the measured full width at half maximum (FWHM) linewidth of the observed NH₃ spectra, $\sigma_{\mathrm{T}}$ (= $(kT_{\mathrm{kin}}/17m_{\mathrm{H}})^{1/2}$) refers to the thermal broadening for NH₃ (e.g. Dunham et al., 2011; Dewangan et al., 2016). The values of $R_{\mathrm{p}}$ range from 0.04 to 0.12 with an average of 0.09 in six clumps, while the range of Mach numbers of these clumps is [2.9, 4.8] with an average of 3.5. These physical parameters suggest that non-thermal pressure and supersonic non-thermal motions (e.g., turbulence, outflows, shocks, and/or magnetic fields) are dominant in these clumps (e.g., Lada et al., 2003).

Figure 2f displays the overlay of the NVSS 1.4 GHz continuum emission contours (beam size $\sim$45^′′; noise level $\sim$0.45 mJy beam^-1; Condon et al., 1998) on the Herschel 350 $\mu$m image. In the NVSS map, radio clumps are clearly seen toward the H ii regions G18.88$-$0.49 and G18.937$-$0.434. However, no radio continuum peak is seen toward the position of the 6.7 GHz methanol maser. Using the clumpfind IDL program (Williams et al., 1994), we estimated the integrated fluxes at NVSS 1.4 GHz of the radio sources containing these H ii regions. Using the integrated flux at NVSS 1.4 GHz, we computed the number of Lyman continuum photons ($N_{\mathrm{UV}}$) and the dynamical age ($t_{\mathrm{dyn}}$) of each H ii region. Following the equations and analysis adopted in Dewangan et al. (2017a), we compute the value of $\log{N_{\mathrm{UV}}}$ to be 48.22 and 47.90 for the H ii regions G18.88$-$0.49 and G18.937$-$0.434, respectively (see Matsakis et al., 1976, for equations). Based on these values, the H ii regions G18.88$-$0.49 and G18.937$-$0.434 are excited by O9V–O9.5V and O9.5V–B0V stars, respectively (e.g., Panagia, 1973). Considering the typical value of the initial particle number density of the ambient neutral gas ($n_{\mathrm{0}}$ = 10³ (10⁴) cm^-3), the dynamical age of the H ii regions G18.88$-$0.49 and G18.937$-$0.434 is estimated to be $\sim$0.4 (1.4) and $\sim$0.4 (1.3) Myr, respectively (e.g., Dyson & Williams, 1980). Kerton et al. (2013) also reported similar ages and spectral types of the powering stars of both the H ii regions. The electron temperature of 10⁴ K, the isothermal sound velocity in the ionized gas (= 11 km s^-1; Bisbas et al. (2009)), and the distance of 5 kpc are adopted in this calculation. The analysis assumes that the H ii regions are uniform and spherically symmetric.

The THOR survey provides six radio continuum maps at 1–2 GHz (beam size $\sim$25^′′; noise level $\sim$0.3–1 mJy beam^-1; Bihr et al., 2016) and spectral index of radio sources. The THOR maps at 1–2 GHz have a better resolution and sensitivity than the NVSS map at 1.4 GHz. In Figure 2i, we show the THOR 1950 MHz continuum map overlaid with the THOR 1950 MHz continuum contours. Figures 2f and 2i allow us to compare the observed radio emission of the H ii regions G18.88$-$0.49 and G18.937$-$0.434. In the THOR 1950 MHz continuum map, we see several peaks that are not detected in the NVSS radio continuum map. We also do not find any radio continuum peak toward the position of the 6.7 GHz methanol maser in the continuum map at 1950 MHz. The positions of the ATLASGAL clumps are also highlighted in the THOR continuum map, allowing us to compare the positions of the dust clumps with the ionized emission.

In Figure 2i, the positions of the radio continuum sources from the THOR survey are also marked. The spectral index is measured using the radio peak fluxes at 1.06, 1.31, 1.44, 1.69, 1.82, and 1.95 GHz (e.g., Bihr et al., 2016). The determination of the value of $\alpha$ helps us to infer whether the radio continuum emission is thermal or non-thermal (e.g., Rybicki & Lightman, 1979; Longair, 1992). Thermally emitting sources show a positive or near zero spectral index. SNRs exhibit non-thermal emission with $\alpha$ $\approx$ $-$0.5, while extragalactic objects generally have a steeper $\alpha$ $\approx$ $-$1 (e.g., Bihr et al., 2016). Based on the values of $\alpha$, two THOR sources G18.875$-$0.360 ($\alpha$ $\sim-1.05$) and G18.755$-$0.497 ($\alpha$ $\sim-0.77$) are more likely to be extragalactic objects. The H ii region G18.88$-$0.49 ($\alpha$ $\approx$ 0.09) shows thermal radio continuum emission, while the H ii region G18.937$-$0.434 ($\alpha$ $\approx$ $-$0.4), powered by an O-type star, displays non-thermal radio continuum emission (see Section 5.2 for more details). Despite the value of $\alpha$ $\approx$ $-$0.4, the source G18.937$-$0.434 was reported as an H ii region and not a SNR candidate (see source G18.939$-$0.443 in Wang et al., 2018). As mentioned earlier, this source was also classified as a Galactic H ii region (Anderson et al., 2011; Wenger et al., 2013). The observed spectral index of $\approx$ $-$0.4 is consistent with a combination of thermal and non-thermal emission. The estimation of the spectral type of the powering source of this H ii region appears to be overestimated because the total radio continuum flux of the H ii region has been treated as thermal emission (see also Section 5.2).

Figure 3a shows the Herschel temperature ($T_{\mathrm{d}}$) map of our selected target area around G18.88$-$0.49. We find the presence of warm dust emission ($T_{\mathrm{d}}$ = 23.5–25.5 K) toward both H ii regions, G18.88$-$0.49 and G18.937$-$0.434. In Figure 3b, we display the Herschel column density ($N(\mathrm{H}_{2})$) map overlaid with the $N(\mathrm{H}_{2})$ contours at [14.1, 21] $\times$ 10²¹ cm^-2. Here, to trace different features in the Herschel column density map, we employ different $N(\mathrm{H}_{2})$ contour levels through visual examination. The $N(\mathrm{H}_{2})$ contour at 14.1 $\times$ 10²¹ cm^-2 enables us to trace potential filaments in our selected target area. In the column density map, we identify the shell-like feature (extension $\sim$12 pc $\times$ 7 pc) at $N(\mathrm{H}_{2})$ = 21 $\times$ 10²¹ cm^-2 (see the black contour in Figure 3b; and also Section 4.1.1). The existence of the shell-like feature is also confirmed by the previously published NH₃ (1,1) line data (see plus symbols in Figure 2g), which is associated with the dense gas. With the help of the Herschel $N(\mathrm{H}_{2})$ map, the total mass of the shell-like feature is determined to be $\sim$2.9 $\times$10⁴ M_⊙ using the equation $M_{area}=\mu_{H_{2}}m_{H}A_{pix}\Sigma N(H_{2})$, where $\mu_{H_{2}}$ is defined earlier (i.e., 2.8), $A_{pix}$ is the area subtended by one pixel (i.e., 6^′′/pixel), and $\Sigma N(\mathrm{H}_{2})$ is the total column density (see also Dewangan et al., 2017a, for more details of the analysis). The clumpfind IDL program is employed to estimate the value of $\Sigma N(\mathrm{H}_{2})$ (e.g., Dewangan et al., 2017c, 2018b).

In Figure 3c, we display a color-composite map with the Spitzer 8.0 $\mu$m and the THOR 1950 MHz images. One of the MIR bubbles (MWP-1G018879-004949 (radius = 1^′.53)) surrounds the H ii region G18.88$-$0.49. In other words, the MIR bubble is filled with the ionized gas. The color-composite map is also overlaid with the $N(\mathrm{H}_{2})$ contours at [14.1, 26.4] $\times$ 10²¹ cm^-2. At $N(\mathrm{H}_{2})$ = 26.4 $\times$ 10²¹ cm^-2, we identify at least four subregions around the H ii region G18.88$-$0.49, which are labeled as A (mass $\sim$10.5 $\times$10³ M_⊙), B (mass $\sim$8 $\times$10³ M_⊙), C (mass $\sim$1.1 $\times$10³ M_⊙), and D (mass $\sim$0.8 $\times$10³ M_⊙). We followed the same steps as mentioned above to estimate the total mass of each sub-region. In Figure 3d, we highlight five potential filaments (fk, fl, fm, fn, and fo) having typical lengths of five to eight parsecs, the shell-like feature, and four subregions (see also Section 4.2.1). These proposed filaments, which have low column density, are associated with the shell-like feature having high column density. The association of the proposed filaments and the shell-like feature reveals an interesting configuration similar to a “hub-filament” system (e.g., Myers, 2009; Schneider et al., 2012; Baug et al., 2015; Dewangan et al., 2015, 2016, 2018a). We find that the ends of the proposed filaments, which are approaching the hub, are associated with the dense parts of the shell-like feature. This finding is further explored using the analysis of the molecular line data in Section 4.2. The H ii region G18.88$-$0.49 excited by an O-type star is seen almost at the center of the shell-like feature, where no dust emission at $N(\mathrm{H}_{2})$ = 21 $\times$ 10²¹ cm^-2 is detected (see Figures 3c and 3d). These findings suggest that the energetics of a young massive O-type star (e.g., ionizing radiation, stellar winds, and radiation pressure) might have interacted with its immediate surroundings. Additionally, some radio peaks are seen toward this shell-like feature (see subregions A, B and C). Furthermore, the H ii region G18.937$-$0.434 does not appear to be part of the hub-filament system.

We extracted ¹³CO spectra toward nine small fields (i.e., t1–t9) that are highlighted in Figure 3d. These are selected toward the regions around the shell-like feature. Figures 4a–4i present the averaged ¹³CO spectra toward nine small fields t1–t9. The spectra show that the molecular gas is well traced in a velocity range of [60, 70] km s^-1, confirming the existence of a single velocity component in the direction of our selected target site.

To infer the morphology of the molecular cloud associated with the target site, the integrated ¹²CO, ¹³CO, and C¹⁸O intensity maps (or moment-0 maps) are presented in Figures 5a, 5b, and 5c, respectively. The molecular gas is integrated in a velocity range of [60.2, 70.6] km s^-1. In Figure 5b, the column density contour at 14.1 $\times$ 10²¹ cm^-2 is shown to highlight the proposed Herschel filaments. The black contour (at $N(\mathrm{H}_{2})$ = 21 $\times$ 10²¹ cm^-2) represents the shell-like feature in Figure 5b. The ¹³CO emission follows quite well the morphology of the shell-like feature as investigated in the Herschel column density map. The moment-0 maps of ¹²CO and C¹⁸O also trace the shell-like feature, which is also well depicted by the dense gas tracer NH₃ (see plus symbols in Figure 5c). The C¹⁸O and NH₃ emissions are not very strong toward the proposed filaments, indicating that these filaments are not very dense compared to the shell-like feature. However, low intensity ¹³CO emission is clearly seen toward all these filaments at $N(\mathrm{H}_{2})$ = 14.1 $\times$ 10²¹ cm^-2 (see a thin contour in Figure 5b). No molecular emission is found toward the other H ii region G18.937$-$0.434.

The moment-1 maps (i.e., the intensity-weighted central velocity) of the ¹²CO, ¹³CO, and C¹⁸O emission are shown in Figures 5d, 5e, and 5f, respectively. Based on the comparison of the moment-1 maps to the integrated intensity maps, one can depict the noticeable velocity variations toward different parts of the shell-like feature. In Figures 5e and 5f, we see a noticeable sharp border between the two velocity fields toward the shell-like feature, revealing a velocity gradient.

In Figure 6a, we highlight different axes (i.e., “i1–i2”, “j1–j2”, “k1–k2”, “l1–l2”, “m1–m2”, “n1–n2”, and “o1–o2”) in the integrated intensity map of ¹³CO. Along each of these axes, a position-velocity diagram is presented in Figure 9. The axes “i1–i2” and ”j1–j2” are selected in the direction of the major and minor axes of the shell-like feature, respectively. Figure 6b presents a zoomed-in area around the shell-like feature using a color-composite map with the moment-0 map of ¹³CO and the THOR 1950 MHz images. Here, the moment-0 map of ¹³CO is processed through an edge detection algorithm (i.e. Difference of Gaussian (DoG); see Gonzalez & Woods, 2011; Assirati et al., 2014; Dewangan et al., 2017b). It enables us to visually probe five filaments (i.e., fk, fl, fm, fn, and fo) as seen in the Herschel column density map (see Figure 3d). The ionized gas is distributed almost at the center of the shell-like feature (extension $\sim$12 pc $\times$ 7 pc; aspect ratio $\sim$1.7), which is well traced with the NH₃ emission (see plus symbols). The aspect ratios of the selected five molecular filaments vary between 2.3 to 4. Hence, this molecular map supports the existence of the hub-filament system around the H ii region G18.88$-$0.49. Figure 6c displays the hub-filament system overlaid with the positions of the ATLASGAL clumps, the shell-like feature, and the filaments. The positions of Class I YSOs and flat-spectrum sources are also marked in Figure 6c. The identification of these YSOs is discussed in Section 4.3.

Figure 7 displays the integrated ¹³CO velocity channel maps (starting from 56.3 km s^-1 at intervals of 1.3 km s^-1). In Figure 8, we present the velocity channel maps of the C¹⁸O emission (starting from 59.6 km s^-1). In Figures 7 and 8, the shell-like feature and four Herschel subregions (“A–D”) are also highlighted. Both channel maps trace the shell-like feature, and it appears that some parts of the shell-like feature appear clumpy (i.e., subregions). The ¹³CO and C¹⁸O gas motion is also evident in the shell-like feature (including the Herschel subregions). The higher intensity of C¹⁸O is found toward the sub-region “B”, while the sub-region “A” is traced with the higher intensity of ¹³CO. In Figure 7, the ¹³CO emission is also seen toward all the proposed filaments (see five solid lines and also the panel at [66.7, 68.0] km s^-1).

In this section, we produce the position-velocity diagrams of ¹³CO to explore the velocity structure of the filaments. Figures 9a, 9b, 9c, 9d, 9e, 9f, and 9g show the position-velocity diagrams of ¹³CO along the axes “i1–i2”, “j1–j2”, “k1–k2”, “l1–l2”, “m1–m2”, “n1–n2”, and “o1–o2”, respectively. In these axes, the positions k2, l2, m2, n2, and o2 are selected toward the central hub. In direction of the shell-like feature (see Figures 9a and 9b), a velocity spread is observed. Figures 9c, 9d, 9e, 9f, and 9g show the existence of velocity gradients (about 0.8–1.8 km s^-1 pc^-1) along the filaments fk, fl, fm, fn, and fo, respectively. The end of each filament that reaches the hub is associated with the dense parts of the shell-like feature (see the positions k2, l2, m2, n2, and o2 in Figure 6). The implication of the hub-filament configuration concerning the star formation process in G18.88$-$0.49 is presented in Section 5.