Assessing the Stellar Population and the Environment of an H ii Region on the Far Side of the Galaxy^*

Assuming that VVV CL177 is a real cluster, we determine its angular size from the radial density profile (RDP) based on our VVV stellar catalog, using the coordinates of the water maser as the central coordinates. To increase the contrast between the density of the cluster candidate and the background level, we exclude the stars with VVV (H−${K}_{S}\,\lt 2.4$) mag from the catalog, as they are more likely to be field stars, judging by the color and magnitude of stars more than 2' away from the center of the cluster candidate. The result is presented in Figure 3. The darker horizontal band marks the background level with its uncertainty. To better guide the determination of the VVV CL177's angular size, we fit a two-parameter King profile adapted to star counts (as in King 1966, but using star count instead of surface brightness). The fit obtained and its uncertainty are plotted in lighter gray. The radius derived by the King profile fitting is 0.45'. Of course, such a cluster would not be isolated, self-gravitating, and in equilibrium, as required for use of the King profile, yet this value can be used as a first estimate. As a comparison, the RDP seems to reach the background level between 0.7' and 1.05'.

Zoom In Zoom Out Reset image size

Interestingly, the RDP can be compared with the profile of the nebular emission. In Figure 3, we present the average flux measured in the four quadrants of the 8 μm-emission GLIMPSE II image centered on the maser coordinates as a function of radius. The morphology can vary significantly between quadrants, as they show a strong central cavity and local clumps, most likely sculpted by strong stellar winds. Yet all the profiles reach the background level at the same radius, which is close to 08, and the nebula remains mostly spherical, centered on the water maser.

Both the RDP and the radial 8 μm-emission profile seem to correspond, as if the density of stars (with the selected colors) is corresponding to the density of nebular gas. Without additional information, this observation could be explained in two possible ways. Either the nebula is in the foreground, reddening the stars behind it and giving the impression of a cluster of stars with comparable colors, or both the nebula and the stars with a (H−K_S ) color greater than 2.4 mag are related to the water maser, and are part of a giant star-forming region.

In the following sections, we will verify which of these scenarios is the most likely. Assuming it is the latter, and using the distance of ∼20 kpc from the water maser parallax, the cluster diameter would be at least a minimum of 5 pc, but potentially up to 9–10 pc if one includes all the 8 μm emission, and considers the farthest radius where the RDP reaches the background. In that case, the nebula is only visible from 8 μm and redder due to extreme reddening along the line of sight.

In order to verify if the stars that share the same highly reddened color within the cluster candidate radius are bound and associated to the water maser, or not, we inspect their proper motion (PM) distribution. We use the VVV InfraRed Astrometric Catalogue version 2 data (VIRAC2; Smith et al. 2018, Section 4.1). We apply the Gaussian Mixture Model (GMM; Everitt et al. 2011) technique, which is based on the assumption that the star distribution within an overdensity can be described by a superposition of multivariate Gaussian distributions (de Souza et al. 2017). The GMM was implemented in R ²⁵ using the mclust library (Scrucca et al. 2016). The mixture model includes a noise component for the field stars, which is initialized using the nearest neighbor cleaning method from Byers & Raftery (1998), via the NNclean function in the prabcus library of R. We apply the GMM technique in the proper motion space, using only the stars within the cluster candidate radius. We use 10 multivariate Gaussian components. The results are insensitive to the exact number of Gaussians, as long as it is large enough to disentangle the background from the cluster candidate members. The two most populated Gaussians (44 and 24 objects each) have considerable mixing probabilities (<28%). Nevertheless, their proper motion values (${\mu }_{{\rm{R}}.{\rm{A}}{.}_{1}}\,=-1.28\pm 0.29$ mas yr⁻¹, ${\mu }_{\mathrm{decl}{.}_{1}}=-2.05\pm 0.43$ mas yr⁻¹, ${\mu }_{{\rm{R}}.{\rm{A}}{.}_{2}}=-3.01\pm 0.44$ mas yr⁻¹, and ${\mu }_{\mathrm{decl}{.}_{2}}=-6.87\pm 0.33$ mas yr⁻¹) are between 3σ and 6σ from the water maser PMs (${\mu }_{{\rm{R}}.{\rm{A}}.}=-2.43\pm 0.1$ mas yr⁻¹, ${\mu }_{\mathrm{decl}.}=-4.43\pm 0.16$ mas yr⁻¹; Sanna et al. 2017).

This result remains the same when we select the stars within an annulus around the water maser coordinates. As can be seen in Figure 4 (left), the distribution of PMs, plotted in green contours, shows two groups: one with PMs more negative than the maser, and the other with PMs more positive. The stars within the cluster candidate radius (045) are plotted with gray or black dots, depending on their color (H−K_S ) being lower or higher than 2.4 mag, respectively. The distribution of the redder star PMs is surrounding the water maser PM, but also mimics the reference field. Though, note that there are potential systematic issues in the transformation of the PMs into the Gaia system, and errors may be underestimated.

Zoom In Zoom Out Reset image size

Our K-band spectra in Figure 2 cover four stars with a (H−K_S ) color greater than 2.4 mag. These spectra present three major features: H recombination emission lines (Brδ and Brγ at 1.95 $\mu {\rm{m}}$ and 2.16 $\mu {\rm{m}}$, respectively), H₂ lines in emission and the ${}^{12}$CO ${\rm{\Delta }}\nu =2$ bandheads (${}^{12}$CO (2,0) at 2.29, ${}^{12}$CO (3,1) at 2.32, and ${}^{12}$CO (4,2) at 2.35 $\mu {\rm{m}}$).

Star 2 shows only the H recombination and H₂ lines. Stars 3 and 4 show CO-bands in absorption and with a depth similar to that observed in giant stars. By contrast, star 1 shows the CO-bandheads clearly in emission, features related to accretion by young stellar objects (YSOs; Scoville et al. 1979). This spectrum confirms that at least one of the stars is an accreting YSO and could be associated with the star-forming region at the source of the maser emission. The spectra of stars 1 and 2 are typical of fairly massive embedded YSOs in the literature (Cooper et al. 2013).

Note that for star 1, the line ratios of 1-0 S(2), 2-1 S(3), 2-1 S(2), 3-2 S(3), and 1-0 S(0) to 1-0 S(1) give 0.45 ± 0.06, 0.19 ± 0.01, 0.17 ± 0.01, 0.09 ± 0.01, and 0.49 ± 0.07, respectively, once corrected for extinction (determined in the next section), which has an effect on the same scale as the uncertainties. Such ratios correspond to a UV fluorescence excitation mechanism (Wolfire & Königl 1991).

The spectral resolution and the signal-to-noise ratio of the spectra do not allow for radial velocity (RV) measurements more precise than 40 km s⁻¹. Using the ${}^{12}$CO bandheads observed in stars 1, 3, and 4, we find ${{\rm{V}}}_{{LSR}}=-13\pm$ 45 km s⁻¹, $-249\,\pm$ 40 km s⁻¹, and $-155\pm$ 40 km s⁻¹, respectively. While the RV for star 1 is in agreement with ${V}_{\mathrm{LSR}}=-16\pm$ 4 km s⁻¹ for the water maser associated with GAL 007.47+0.06 (Sanna et al. 2017), stars 3 and 4 are off by a significant factor, ruling out their membership of the cluster completely. Note that no RV could be obtained for star 2, as it only displays narrow emission lines. Those cannot be trusted, because the H recombination lines (Brδ and Brγ) typically arise either in a fast wind or gas falling onto the star at freefall velocity, and the H₂ lines typically arise in an outflow or wind that is often offset from the systemic velocity (Guo et al. 2020).

Figure 4 (right) shows star 1 ${{\rm{V}}}_{{LSR}}$ (blue square) compared to that of the maser (white square). The distance estimate to star 1 is ${21.06}_{-3.31}^{+6.77}$ kpc, and determined following the method described in Reid et al. (2014, Section 4), using the Galactic parameters from model A5 in Reid et al. (2009).

In summary, while at least some stars are YSOs with an RV that does not exclude a distance of 20 kpc, there are no indications based on PMs that the stars that share the same highly reddened color and are grouped a few arcseconds from the position of the water maser are members of a unique cluster.

Even if the PMs do not support the hypothesis that all the reddest stars within VVV CL177 estimated radius are members of a unique cluster, we can still assume that some of the stars may be members of a cluster associated with the water maser.

To determine the value for the extinction, we exploit the ratio of 1-0 S(1) and Q(3) emission lines of H₂. Since Q(3)/S(1) = 0.70 from the spontaneous decay rates (Turner et al. 1977; Wolniewicz et al. 1998), any deviation from that value is due to reddening, as it is affecting the S(1) more than the Q(3) line. For star 1, the star most likely associated with the water maser, Q(3)/S(1) = 1.7 ± 0.1, which, using the equations from O'Connell et al. (2005), corresponds to ${A}_{{\rm{V}}}$ = 38.0 ± 2.5 mag.

Figure 5 (left) shows the color–magnitude diagram (CMD) H−K_S as a function of K_S of the cluster within the estimated radius. Stars 1–4 from Figure 2 are marked. With no trustworthy membership information available, we simply apply the same strict cut at a color of H−K_S  = 2.4 mag as in Section 3.1, giving a rough first look at what VVV CL177's CMD could look like. There is no detected flux in the J band for most of the stars, as the brightest star barely reaches J ∼ 18 mag (corresponding to a photometric error greater than 0.1 mag in VVV).

Zoom In Zoom Out Reset image size

Using the water maser parallax distance and our value for extinction, we plot isochrones of young clusters from Ekström et al. (2012) over the CMD on Figure 5 (left). Interestingly, the position of the isochrones corresponds to the color cut used to select the stars representing the putative cluster, even if we now know that most of them may not be members of VVV CL177 according to their PMs. Also, there are no stars near those isochrones in the CMD obtained using an annulus at a 2' radius from the center of VVV CL177 as a comparison field (Figure 5, right). Based on this first estimate, would VVV CL177 be a real cluster, it would be younger than 20 Myr old, and only the most massive stars would be visible in our data.

We mapped the ¹³CO emission with the Argus instrument on the 100 meter Green Bank Telescope. Argus is a 16-pixel focal plane array operating between frequencies of 74–116 GHz, designed for fast mapping of molecular gas. While there were several ¹³CO components along the line of sight, there was an isolated line between −20 km s⁻¹ and −10 km s⁻¹ associated with the LSR maser velocity of −16 ± 4 km s⁻¹ from Sanna et al. (2017). We integrated only along this isolated ¹³CO component to obtain the map displayed in Figure 1 (right), and the flux distribution is corresponding to the position of the 8 μm nebula. There is some offset, of course, since the ¹³CO emission traces cold gas that has not yet formed into stars. Such a match indicated that both the cold molecular gas and the nebula are situated at the same distance as the water maser. For the rest of this section, we will adopt a distance of ∼20 kpc.

We assume a ¹³CO-to-H₂ molecular gas conversion factor of ${{\rm{X}}}_{{}^{13}\mathrm{CO}}$ = 10 × 10²⁰ cm⁻² (K km s⁻¹)⁻¹ (Cormier et al. 2018), which corresponds to a ¹³CO mass conversion factor, ${\alpha }_{{}^{13}\mathrm{CO}}$, of 21.62. Using this conversion factor, our integrated intensity (W${}_{{}^{13}\mathrm{CO}}$) of 3.26 K km s⁻¹, an assumed distance (D) of 20 kpc, and the solid angle of the source (${{\rm{\Omega }}}_{{\rm{S}}}$) of 5.44 × 10⁻⁸ Sr, we can estimate a total molecular gas mass from Bolatto et al. (2013): ${M}_{\mathrm{mol}}[{{\rm{H}}}_{2}]={\alpha }_{{}^{13}\mathrm{CO}}\,{L}_{{}^{13}\mathrm{CO}}={\alpha }_{{}^{13}\mathrm{CO}}\,{W}_{{}^{13}\mathrm{CO}}\,{D}^{2}\,{{\rm{\Omega }}}_{{\rm{S}}}$. This results in a total molecular gas mass of 1.60 × 10³ ${M}_{\odot }$. Separately, the Herschel/HiGal survey detected the far-infrared emission of the dust in the associated molecular clump (Elia et al. 2017). Unsure of the distance at that time, the survey team calculated a nominal mass of 36.72 M ${}_{\odot }$ at a standardized distance of 1 kpc. Scaling their nominal mass to the maser parallax distance, we find ${M}_{\mathrm{mol}}[{{\rm{H}}}_{2}]$ = 1.50 × 10⁴ ${M}_{\odot }$.

Using MAGPIS 1.4 GHz radio continuum data we find a flux density of 1.8 Jy, which at the assumed distance is equivalent to a Lyman continuum photon production rate of 6.5 × 10⁴⁹ s⁻¹ (Rubin 1968), assuming an electron temperature of 8000 K. This Lyman continuum luminosity is more than the output of a single main-sequence O3 star (Martins et al. 2005) and makes the region one of the most luminous in the Milky Way. The spatial scale is approximately 3 pc, i.e., a classical H ii region rather than a UC H ii region. It is more luminous than any of the ∼1000 first-quadrant H ii regions in Makai et al. (2017) or any of the 213 UC H ii regions in Urquhart et al. (2013) though the G055.114+02.422 region (Armentrout & Anderson 2017) is almost as luminous and located at an even larger Galactocentric radius, also in the Outer Scutum-Centaurus arm. Furthermore, the radio flux density is likely a lower limit because given the molecular gas morphology, many photons may be leaking out of the region and therefore not contributing to the radio flux.

In that scenario, all of our observations are the result of a coincidental alignment of many components, as suggested by Yamauchi et al. (2016). This would explain the lack of coherence in the star PMs. The group of stars that this study focused on share the same (H−K_S ) color because of the extinction of a foreground object. Finally, stars 1 and 2 are YSOs associated with a different star-forming region than the one at the source of the water maser emission.

Were that scenario the right one, at least in part, we would still be confident that our data show that the 8 μm nebular emission and the ¹³CO emission are associated with the water maser.

Our data indicate that the group of stars that share the same (H−K_S ) color and that are near the water maser coordinates are not likely clustered. But it remains possible that some of the stars are associated with the star-forming region at the source of the maser emission. When we put all the pieces together, we have:

• 1.  
  A water maser source ${20.4}_{-2.2}^{+2.8}$ kpc away from the Sun and associated with GAL 007.47+00.06 (Sanna et al. 2017).
• 2.  
  The RDP corresponding to the 8 μm emission profile, giving a radius between 045 and ∼1'.
• 3.  
  The confirmation of the YSO nature of at least one of the redder stars (star 1), a few arcsecs from the maser, plus a reliable YSO candidate (star 2). The spectra of both objects resemble the Cooper et al. (2013) massive YSO spectra.
• 4.  
  The RV of star 1 corresponds to that of the maser, and the ¹³CO emission at that velocity corresponds to the 8 μm emission.
• 5.  
  An extreme extinction near ${A}_{{\rm{V}}}=38.0\pm$ 2.5 mag determined from the 1-0 S(1) and Q(3) emission lines of H₂ flux ratio for star 1.

In that scenario, we could detect only a few of the brightest stars of VVV CL177. The other stars are foreground stars, which is quite likely in that line of sight. With ${K}_{S}\sim 13$–14 mag, these stars would have absolute magnitudes $-3.5{M}_{{K}_{S}}+1.5$. This is comparable to many members of the Cooper et al. (2013) sample of embedded massive YSOs, which have $-8.5{M}_{{K}_{S}}\,+3$, comparable extinction in many cases, and luminosities $L={10}^{3}$ to 10⁵ ${L}_{\odot }$.

On can note that the extreme extinction of star 1 is compatible with a large distance. Unfortunately, the VVV high-resolution 3D extinction maps (Chen et al. 2013) do not reach a distance of 20 kpc, but at 10 kpc, they give ${A}_{{\rm{V}}}$ ∼ 14.6 ± 1.5 mag. And at 6 kpc, one of the distances to GAL 007.47+0.06 by Wink et al. (1982), ${A}_{{\rm{V}}}$ ∼ 4.0 ± 0.7 mag, which is an order of magnitude off, discarding a closer distance for both GAL 007.47+0.06 and star 1.