H ii REGION G46.5-0.2: THE INTERPLAY BETWEEN IONIZING RADIATION, MOLECULAR GAS, AND STAR FORMATION

As mentioned above, about 10' toward the southwest of the H ii region (∼12 pc at the assumed distance of 4 kpc), some pillar-like features can be seen emitting at 8 μm, pointing to the ionizing source of G46 and apparently embedded in the molecular cloud GRSMC G046.34-00.21. Even though it seems to be a large distance, several pillars were found as far from the ionizing source as those analyzed in this work (e.g., in the Vulpecula rift; Billot et al. 2010). What is really remarkable in this case is that these pillars-like features, resembling heaps and corrugations, are well outside the H ii region, contrary to what is usually found: bubble-like structures with pillars inside or over their boundaries. Taking into account that the pillars point toward the G46 open border, it is suggested that they were produced by UV photons escaping from the H ii region.

By inspecting ¹³CO J = 1–0 toward the pillar-like features (Figure 4 (right)), we find that there are two associated molecular structures. The eastern-most IR pillar-like feature has associated molecular gas between 54.0 and 56.5 km s⁻¹ (white contours in Figure 4 (right)), while the western one is related to a molecular structure ranging between 56.5 and 59.0 km s⁻¹ (yellow contours in Figure 4 (right)). The different velocity ranges for these molecular structures may either indicate that both features are located at slightly different positions along the line of sight or that they have somewhat different kinematics. There is remarkably good agreement between the morphology of the molecular gas and the pillar-like features seen in IR. Moreover, the molecular structures present a morphology consisting of a dense head with a less dense tail, as is usually found by observations toward and predicted by models of this kind of structures in the area surrounding H ii regions (Pound et al. 2005, 2007; Schuller et al. 2006). However, following the work of Mackey & Lim (2010), the appearance of these structures resembles a previous evolutionary stage in the formation of pillars because a well-formed tail cannot be identified behind their heads. In what follows, we characterize the interaction between the ionization radiation escaping from G46 and the pillar-like features.

In order to study the radiation influence over the tips of the pillar-like features and the possibilities of triggered star formation via RDI (e.g., Bertoldi 1989; Lefloch et al. 1997; Kessel-Deynet & Burkert 2003), we evaluate the pressure balance between the ionized gas stalling at the head of the pillars and the neutral gas of their interiors. Assuming that the ionizing photons came from an O7V-type star located at the center of G46 (see Section 3.1), we use the predicted ionizing photon flux for an O7V star from Schaerer & de Koter (1997) and the projected distance between the star and the pillars to estimate roughly the amount of UV photons arriving at the surface of the pillars in ${{{\Phi }}_{{\rm pre}}}\sim 2\times {{10}^{8}}$ cm⁻² s⁻¹. This value represents an upper limit due to the fact that the projected distance between the star and the pillars is a lower limit of the actual distance between them. Our predicted ionizing photon flux is similar to those measured toward several bright-rimmed clouds (Thompson et al. 2004a, 2004b) which, similar to the pillars, are molecular clouds sculpted by the radiation leaking from H ii regions, and in many cases they have the same large-scale morphology as the pillars (Thompson et al. 2004a).

Using the ${{{\Phi }}_{{\rm pre}}}$ obtained above and following Thompson et al. (2004b), we estimate an upper limit for the electron density of about 64 cm⁻³ at the expected ionized boundary layer (IBL) at the tip of both pillar-like structures. This is almost three times greater than the critical value of ∼25 cm⁻³ above which an IBL is able to develop around a cloud (Lefloch & Lazareff 1994). To obtain this value, we consider that the pillar heads have radii of 05 and assume an effective thickness of the IBL of $\eta =0.1$. Then, using a typical sound speed for the ionized gas of 11.4 km s⁻¹, we determine that the pillars tips are supporting an external pressure of ${{P}_{{\rm ext}}}/k\sim 2.1\times {{10}^{6}}$ cm⁻³ K. It is important to note that the obtained P_ext represents strictly an upper limit because the predicted Φ is an upper limit and due to the used η (see Thompson et al. 2004b).

On the other side, integrating the ¹³CO J = 1–0 emission over the area containing each pillar head and using the typical local thermodynamic equilibrium (LTE) formulae to derive the ¹³CO and H₂ column densities by assuming ${{T}_{{\rm ex}}}=10$ K and [H₂/¹³CO] $=\;5\times {{10}^{5}}$ (see, e.g., Yamaguchi et al. 1999), we estimate the mass of each pillar head to be about 450 and 300 ${{M}_{\odot }}$ for the eastern and western pillars, respectively. The area of each pillar head was assumed taking into account the radius of their tip curvatures, which is a circular area with a radius of 05 for both pillar heads. Thus, assuming spherical shapes, we estimate molecular densities of about $7.0\times {{10}^{3}}$ and $4.6\times {{10}^{3}}$ cm⁻³, respectively. Then, considering the velocity interval in which each pillar structure extends, we obtain the velocity dispersion ${{\sigma }_{v}}\sim 1.05$ km s⁻¹, which is in agreement with those predicted by models of pillar formation (Gritschneder et al. 2010; Dale et al. 2012). Finally, using the obtained densities and ${{\sigma }_{v}}$, we derive the internal pressure for each pillar head as ${{P}_{\operatorname{int}}}/k\sim 1.8\times {{10}^{6}}$ and $\sim 1.2\times {{10}^{6}}$ cm⁻³ K for the eastern and western pillars, respectively. As Thompson et al. (2004b) state, these pressure values are very likely underestimated because the ¹³CO J = 1–0 line underestimates the true density and this molecule may be depleted by selective photodissociation at the boundary of the clouds. Following these authors, the internal pressure is likely underestimated by no more than a factor of 15; thus we conclude that ${{P}_{\operatorname{int}}}/k$ should be between $2\times {{10}^{6}}$ and $3\times {{10}^{7}}$ cm⁻³ K.

In conclusion, we obtain P_int > P_ext, which suggests that the diluted ionization front stalls at the pillar heads, probably until the effects of mass evaporation and increasing recombination within the IBL raises the ionized gas pressure to equilibrium with the interior pressure (Lefloch & Lazareff 1994). This result shows that it is unlikely that a shock is propagating farther into the molecular gas, discarding that the RDI mechanism is ongoing in the pillars' interiors.

With the purpose of investigating the molecular material toward the open border of G46, we used higher angular resolution data of several molecular species extracted from the JCMT database acquired toward this region (see the rectangle in Figure 4 (left)). Figure 5 shows the integrated emission of ¹²CO, ¹³CO, and C¹⁸O J = 3–2, while Figure 6 displays the HCO⁺ and HCN J = 4–3 integrated emission. The ¹²CO and ¹³CO J = 3–2 emission show clumpy, elongated molecular features with a high density filament extending to the southwest of the mapped region. No considerable molecular emission appears toward the open border of G46, suggesting the presence of a pre-existing region with scarce molecular gas, or suggesting that the UV photons have carved the molecular cloud. In any case, this can be the path followed by the UV photons escaping from G46 to reach the farther pillar-like structures.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The C¹⁸O J = 3–2, HCO⁺, and HCN J = 4–3 emissions are concentrated in a small compact clump related to the infrared dark cloud IRDC 046.424-0.237 (Peretto & Fuller 2009) and to the millimeter continuum source BGPS G046.427-00.237 (see the Bolocam Galactic Plane Survey v2; Ginsburg et al. 2013), suggesting that the emission lines emanate from a deeply embedded clump. Moreover, this structure is associated with the cold high-mass clump G46.43-0.24 (Wienen et al. 2012) in which the (1, 1) and (2, 2) ammonia lines were detected at v$_{{\rm LSR}}\sim 52.3$ and 52.9 km s⁻¹, respectively. The detection of these molecular species, mainly the HCN J = 4–3 line with its critical density that can be between 10⁶ and 10⁸ cm⁻³ (Takakuwa et al. 2007; Greve et al. 2009), indicates the presence of very high density gas. Considering that this clump is a potential site of star formation, we analyze it in the following.

The line parameters of the observed molecular transitions toward the center of the clump are given in Table 2 as derived from the Gaussian fits from the spectra shown in Figure 7. ¹²CO J = 3–2 was fitted using three Gaussians, while the parameters of the other lines were obtained from single-component Gaussian fits which coincide in velocity with the main ¹²CO component. The different velocity components observed in the ¹²CO spectrum are present over almost the whole region. Thus, we conclude that they correspond to different molecular components along the line of sight, reflecting the clumpiness in the region.

Zoom In Zoom Out Reset image size

In order to obtain a rough estimate of the molecular clump mass, we assume LTE. We calculate the excitation temperature from

$$\begin{eqnarray}&&{{T}_{{\rm ex}}}(3\to 2)=\frac{16.59\;{\rm K}}{{\rm ln} [1+16.59\;{\rm K}/({{T}_{{\rm max} }}{{(}^{12}}{\rm CO})+0.036\;{\rm K})]},\end{eqnarray} \tag{ 1 }$$

where ${{T}_{{\rm max} }}{{(}^{12}}{\rm CO})$ is the ¹²CO peak temperature toward the clump center at ∼52 km s⁻¹, obtaining ${{T}_{{\rm ex}}}\sim 18$ K. We derive the ¹³CO and C¹⁸O optical depths ${{\tau }_{13}}$ and ${{\tau }_{18}}$ from (e.g., Buckle et al. 2010)

$$\begin{eqnarray}&&\frac{^{13}{{T}_{{\rm mb}}}}{^{18}{{T}_{{\rm mb}}}}=\frac{1-{\rm exp} (-{{\tau }_{13}})}{1-{\rm exp} (-{{\tau }_{13}}/X)},\end{eqnarray} \tag{ 2 }$$

where ¹³T_mb and ¹⁸T_mb are the peak temperatures of the ¹³CO and C¹⁸O J = 3–2 line at the center of the region, and X = 8.4 is the assumed isotope abundance ratio [¹³CO/C¹⁸O] (Frerking et al. 1982; Wilson 1999), providing ${{\tau }_{13}}\sim 3.5$ and ${{\tau }_{18}}\sim 0.4$, which indicate that the C¹⁸O J = 3–2 line appears to be moderately optically thin. Thus, we estimate its column density from

$$\begin{eqnarray}\begin{array}{ccccccccccccccc} N({{{\rm C}}^{18}}{\rm O}) & = & 8.26\times {{10}^{13}}{{e}^{\frac{15.81}{{{T}_{{\rm ex}}}}}}\frac{{{T}_{{\rm ex}}}+0.88}{1-{{e}^{\frac{-15.81}{{{T}_{{\rm ex}}}}}}} \\ {} & {} & \times \frac{1}{J({{T}_{{\rm ex}}})-J({{T}_{{\rm BG}}})}\int {{T}_{{\rm mb}}}\;dv \\ \end{array}\end{eqnarray} \tag{ 3 }$$

with

$$\begin{eqnarray}&&J(T)=\frac{h\nu /k}{{\rm exp} (\frac{h\nu }{kT})-1}.\end{eqnarray} \tag{ 4 }$$

To obtain the molecular hydrogen column density N(H₂), we assume an abundance ratio of [H₂/C¹⁸O] $=\;5.88\times {{10}^{6}}$ (Frerking et al. 1982; Wilson 1999). Finally, the mass was derived from

$$\begin{eqnarray}&&M=\mu {{m}_{{\rm H}}}\mathop{\sum }\limits_{i}\left[ {{D}^{2}}\;{{{\Omega }}_{i}}\;{{N}_{\;i}}({{{\rm H}}_{2}}) \right],\end{eqnarray} \tag{ 5 }$$

where Ω is the solid angle subtended by the beam size, ${{m}_{{\rm H}}}$ is the hydrogen mass, the mean molecular weight, μ, is assumed to be 2.8 taking into account a relative helium abundance of 25%, and D is the distance. Summation was performed over all of the beam positions on the molecular structure observed in C¹⁸O displayed in the contours in Figure 5. The obtained mass is about 375 ${{M}_{\odot }}$. Assuming an ellipsoidal volume with a semi-axis of 45'', 20'', and 20'', we derive a density of n $\sim 1.5\times {{10}^{4}}$ cm⁻³. The density should be higher in the innermost region of the clump where the HCN J = 4–3 line emanates, showing a density gradient. Using the deconvolved radius of about 0.5 pc calculated from

$$\begin{eqnarray}&&{{R}_{{\rm clump}}}=\sqrt{\frac{S-{\rm beam}\;{\rm area}}{\pi }},\end{eqnarray} \tag{ 6 }$$

where S is the area inside the clump, we note that the mass obtained above is slightly over the mass–size threshold for massive star formation in IRDCs presented in Kauffmann & Pillai (2010), suggesting that massive YSOs can be formed within this clump.

Using the derived HCO⁺ and HCN J = 4–3 parameters listed in Table 2, we performed a non-LTE study of these molecular species with the code RADEX, which uses the mean escape probability approximation for the radiative transfer equation (van der Tak et al. 2007). Using the measured Δv, we ran the code to fit ${{T}_{{\rm mb}}}$ and estimate the column densities. Taking into account that Wienen et al. (2012) measured a kinetic temperature of ${{T}_{k}}\sim 16$ K toward this clump from the ammonia lines, we fix this parameter in 20 K and assume densities between 10⁵ and 10⁷ cm⁻³ to obtain the values presented in Table 3. From the obtained column densities, we observe HCN/HCO⁺ abundance ratios of about 4.4, 3.8, and 1.2 for ${{n}_{{{{\rm H}}_{2}}}}={{10}^{5}}$, 10⁶, and 10⁷ cm⁻³, respectively. In all cases, we obtain a N(HCN)/N(HCO⁺) ratio larger than unity, as found toward several clumps in IRDC G48.66-0.22 (Pitann et al. 2013) and the active star-forming region W49A (Roberts et al. 2011). As these authors point out, the steady-state chemical models for molecular species in gas-phase predict HCN > HCO⁺ only for ${{T}_{k}}\lt 25$ K, with a density of ${{n}_{{{{\rm H}}_{2}}}}={{10}^{6}}$ cm⁻³, which is consistent with our results.

Table 3.  Radex Results from the HCO⁺ and HCN J = 4–3 Lines using ${{T}_{k}}=20$ K and the Indicated ${{n}_{{{{\rm H}}_{2}}}}$

Input ${{n}_{{{{\rm H}}_{2}}}}$	N(HCO+)	${{\tau }_{{\rm HC}{{{\rm O}}^{+}}}}$	N(HCN)	${{\tau }_{{\rm HCN}}}$	${{X}_{{\rm HCN}/{\rm HC}{{{\rm O}}^{+}}}}$
(cm−3)	(cm−2)		(cm−2)
105	$1.9\times {{10}^{14}}$	7.50	$8.4\times {{10}^{14}}$	4.50	4.42
106	$1.7\times {{10}^{13}}$	0.70	$6.5\times {{10}^{13}}$	0.37	3.82
107	$5.4\times {{10}^{12}}$	0.14	$6.8\times {{10}^{12}}$	0.09	1.25

Download table as:  ASCIITypeset image

YSOs always show an excess in their infrared emission. The level of excess in the infrared can be effectively used to discriminate YSOs from field stars and distinguishing different evolutionary stages. At a considerably early evolutionary stage, protostars are mostly embedded in dust envelopes and exhibit large excesses of infrared emission and an infrared spectral index ${{\alpha }_{{\rm IR}}}\gt -0.3$ indicative of flat or ascending SED at wavelengths longward of 2 μm (Lada 1987; Greene et al. 1994). For pre-main-sequence stars which possess optically thick disks, the SEDs tend to descend and the infrared spectral indices are in the range $-1.6\lt {{\alpha }_{{\rm IR}}}\lt -0.3$. Finally "transition disk" sources are more evolved YSOs where the inner parts of the disks have been cleared by photoevaporation of the central stars or by planet forming processes. For such YSOs, we expect to detect IR excesses at wavelengths longer than 16 μm (Strom et al. 1989). These properties of YSOs make photometric observations in the near- to mid-infrared plausible mechanisms for discriminating them from field stars. Different color-based source identification and classification schemes have been developed and verified in practice. In this section, we identify potential YSOs following the scheme proposed by Gutermuth et al. (2009). The resulting YSOs are classified as Class I (protostars, including Class 0, Class I, and "flat spectrum" sources), Class II, and "transition disk" sources.

There are several kinds of contaminants that could be misidentified as YSOs in our original sample. Extragalactic contamination could stem from star-forming galaxies and broad-line active galactic nuclei, which show PAH emission features yielding very red 5.8 and 8.0 μm colors (Stern et al. 2005; Gutermuth et al. 2008). In our own Galaxy, unresolved knots of shock emission and resolved PAH emission are often detected in the IRAC bands, yielding additional contaminants. Following the techniques presented in Gutermuth et al. (2009), we exclude contaminants, and then identify and categorize potential YSOs as described in the following.

In the first phase, only sources with valid detections in all four IRAC bands were considered. Any source fulfilling the color criteria of $[3.6]-[4.5]\gt 0.7$ and $[4.5]-[5.8]\gt 0.7$ is regarded as a Class I YSO. In the remaining pool, Class II sources were picked out based on the constraints of (i) $[3.6]-[4.5]-{{\sigma }_{1}}\gt 0.15$, (ii) $[3.6]-[5.8]-{{\sigma }_{2}}\gt 0.35$, (iii) $[4.5]-[8.0]-{{\sigma }_{3}}\gt 0.5$, and (iv) $[3.6]-[5.8]$ + ${{\sigma }_{2}}\leqslant \frac{0.14}{0.04}\times (([4.5]-[8.0]-{{\sigma }_{3}})-0.5)+0.5$. Here, ${{\sigma }_{1}}=\sigma ([3.6]-[4.5])$, ${{\sigma }_{2}}=\sigma ([3.6]-[5.8])$, and ${{\sigma }_{3}}=\sigma ([4.5]-[8.0])$ are combined errors added in quadrature.

In the second phase, sources with 24 μm data in the remaining pool are re-examined. Sources with colors of $[5.8]-[24]\gt 2.5$ or $[4.5]-[24]\gt 2.5$ are classified as "transition disks." With the potential contaminants and YSOs identified above excluded, there still remain some sources with bright 24 μm emission. For sources lacking valid photometric data in one or more of the IRAC bands, these are picked out as additional Class I YSOs once fulfilling $[24]\gt 7$ and $[X]-[24]\gt 4.5$ MAG, where $[X]$ is the longest wavelength IRAC detection that we have.

The classification scheme of Gutermuth et al. (2009) includes three phases, using 2MASS data to identify more YSOs in addition to the aforementioned methods. However, in our case, as G46 is much farther than the sources in Gutermuth et al. (2009), 2MASS photometry would be severely affected by foreground interstellar extinction, and the use of these near-IR data to identify YSOs would induce heavy contamination from foreground field stars. Thus, we do not use the NIR bands in our YSOs search.

The above identification procedures have resulted in 106 YSOs. They are classified into 22 Class I, 60 Class II, and 24 "transition disks" objects. A summary of the results is given in Table 4. We note that the color-based classification scheme would lead to misidentification. A Class II YSO viewed at high inclination would show features resembling those of a Class I source. An edge-on Class I YSO can have infrared color similar to a Class 0 source. Thus, all Class I, II, and "transition disk" sources identified in this paper are YSO candidates. The distribution of these YSO candidates on distinct color–color spaces are presented in Figures 8 and 9. It is important to note that we also use the point sources extracted from Gutermuth & Heyer (2015) to identify YSO candidates and reach identical results.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Figure 10(a) shows the large-scale spatial distribution of the 106 YSOs identified in the region. Among the 82 Class I and Class II YSOs, more than 40 (about 50%) are located in projection onto the molecular cloud GRSMC G046.34-00.21 (see Figure 4), which fills only the 20% of the whole survey field. Such a concentration of YSOs is suggestive of active star formation taking place in the cloud. On the other hand, from the 24 "transition disks" found in the whole field, none are detected associated with the molecular gas, indicating the relatively young nature of the YSOs embedded in this cloud. Interestingly, all of the YSOs associated with the molecular cloud are placed in projection between the open border of G46 and the head of the pillars. Moreover, there are no young sources in the farthest part of the cloud behind the heads of these pillars (see Figure 10(b)).

Zoom In Zoom Out Reset image size

From Figure 10(a), only two concentrations of young objects can be noted over the whole field, one centered at 19^h16^m30^s, +11°51'00'' (J2000), consisting of 15 Class II type sources, and the other centered at 19^h17^m00^s, +11°47'00'' (J2000) mostly composed by Class I type YSOs (nine sources), which are located just ahead of the pillar-like features. A closeup view of the pillar-like feature regions is shown in Figure 10(b). A point source is revealed at 24 μm which is too faint in the IRAC bands and not listed in the GLIMPSE catalog. We have marked this 24 μm point source using a red circle in Figure 10(b). Weak at short wavelengths, this point source would be younger than the others and could be a Class 0 candidate.

The fact that the Class II concentration is located closer to the open border of G46 than the Class I group strongly suggests an age gradient in the YSO distribution. On the other hand, the absence of young sources inside and behind the pillar-like structures could indicate that the H ii region's influence has not reached the molecular material behind the pillars, which is in agreement with the result of Section 3.2.1. We suggest a scenario in which the propagation of ionization radiation escaping from the H ii region triggered the star formation observed in the molecular cloud through an RDI mechanism and then stalled at the surface of the pillars heads. We cannot discount that a growing density of the IBL reaches equilibrium with that of the cloud, and thus the shock front will continue its propagation into the cloud (Thompson et al. 2004a).

Additionally, in the region surveyed with JCMT (see yellow rectangle in Figure 4), there are five YSO candidates. One of these candidates, a Class I source, lies exactly at the center of the dense molecular clump mapped with HCO⁺ and HCN J = 4–3 (see Figure 6), which is in agreement with the molecular gas conditions studied in Section 3.2.2.