Giant molecular cloud G18.1-0.3+51 associated with HII regions and supernova remnant in the 3-kpc expanding ring

In order to obtain the general view of the region around G18, we show in figure 1 the radio continuum image a wide region around G18 at 90-cm taken with the VLA (Very Large Array) (Brogan et al., 2006) and a composite map of far-infrared, 20-cm and 90-cm radio continuum survey maps extracted from the Multi-Array Galactic Plane Imaging Survey (MAGPIS) (Helfand et al., 2006), and VLA Galactic Plane Survey (VGPS) (Stil et al., 2006).

For molecular cloud analysis, we made use of the FUGIN survey data (Umemoto et al., 2017). The full beam width at half maximum of the telescope was 15″at the ¹²CO ($J=1-0$)-line frequency (115 GHz), and the velocity resolution was 1.3 km s^-1. The effective beam size of the final data cube is 20″, and the rms noise levels are $\sim 1$ K. The 3D FITS cubes have a pixel size of $(\Delta l,\Delta b,\Delta V_{\rm lsr})=$ (8”.5, 8”.5, 0.65 km s^-1). Fig. 2 shows examples of the channel maps of ¹²CO -line brightness temperature ($T_{\rm B})$ and longitude-velocity ($(l,V_{\rm lsr})$) diagrams at several latitudes and velocity channels.

Fig. 1 illustrates the location of G18 in a 90-cm radio continuum survey map reproduced from the SNR survey with the VLA (Brogan et al., 2006). G18 is marked by a circle, which is a compact association of several HII regions and an SNR in the northern top. The right panel enlarges a $30^{\prime}\times 30^{\prime}$ area around $l=18^{\circ}.15$, $b=-0^{\circ}.5$ in 20 cm radio continuum by red, 90 cm by blue, and 8 $\mu$m infrared emission by green color, as extracted from MAGPIS.

Figure 2 shows channel maps of the ¹²CO line brightness and longitude-velocity diagrams (LVDs) at various latitudes. Figs. 3 shows a wide area map of the ¹²CO  intensity, showing an extremely lopsided distribution of the CO line emission in the observed region with respect to the Galactic plane, indicating that the molecular gas near this velocity, $\sim 50$ km s^-1, is almost empty in the northern side of the Galactic plane.

Figure 4 shows intensity maps (moment 0 maps) around G18 integrated from $V_{\rm lsr}=40$ to 60 km s^-1in the ¹²CO , ¹³CO , and C¹⁸O line emissions along with an overlay of these figures in RGB (red, green, blue) color. The bottom panel shows an intensity profile as a function of the distance from the hub center averaged among the major spokes.

High concentration of molecular gas toward G18 at $(l,b)=(18^{\circ}.15,-0^{\circ}.30)$ is seen as an isolated peak in the field, which we name it here the "hub" according to the terminology of the molecular cloud structure (Myers, 2009). Several molecular branches extend from the hub extending in the directions at $PA\sim 20^{\circ},70^{\circ},\ 150^{\circ},\ 220^{\circ}$ and $\sim 300^{\circ}$, which we call "spokes" A, B, C, D, and E, respectively, as marked by the white lines in Figs. 3 and 4. The system composes a structure often called the hub-filament system (HFS) of interstellar dust lanes (Myers, 2009; Kumar et al., 2020). We here introduce a term ’spoke’ to represent the constructive role of the giant and straight structures in conjunction to the ’hub’, while it has the same meaning as ’giant filament’ in the filament paradigm (Hacar et al., 2022; Pineda et al., 2022).

The spokes compose a large X-shaped structure extending over $\pm 0^{\circ}.6$. Spoke A is densest but shortest, composing multiple cavity and molecular bubble (shell) structures. Spoke B extends almost straightly for more than $\sim 0^{\circ}.6$ toward NE, making a counter part to the also straight spoke D to the SW. Spokes B and E approximately trace the southern rim of the big molecular shell noticed in the earlier study marked as Shell P (Paron et al., 2013).

Spoke A toward the NW is connected to several open cavities, as marked by the white circles, which we name the molecular bubbles 1, 2, 3 and 4. The molecular bubbles are more clearly seen in the moment 0 map shown in figure 5(a) for the core $30^{\prime}\times 30^{\prime}$ area, where the cut-off brightness was set to $T_{\rm B}\leq 5$ K.

Molecular bubble 1 (hereafter, "bubble") is almost perfect, and the bubble’s center coincides in position with the strong radio continuum source (spot) of the HII region G18.146-0.280 as shown in figure 6 (b). The inner edge of the bubble exactly traces the southern rim of the HII shell surrounding the radio spot. Their coherent and concentric orientation strongly support the picture that they are formed by caving compression of the molecular hub by a high pressure HII sphere produced around the hot spot, possibly nesting O stars (Paron et al., 2013).

Bubbles 2 and 3 coincide with fainter HII regions, apparently overlapping with the western rim of SNR G18.1-0.1 on the same sight line. Bubble 4 encircles the entire HII region and SNR area, which might not be an independent object from the others.

The multiple bubbles with their intensity getting fainter according to their increasing sizes and distances from the GMC center (hub) suggests a sporadic injection of kinetic energy in the core region of the GMC along the scenario of sequential star formation (Elmegreen & Lada, 1977; Paron et al., 2013).

The moment 1 map (velocity field) in figure 5 shows a velocity gradient from the south (blue shifted) to north (red shift). This indicates either large-scale rotation on the SN plane, or an outflow from the center toward the north and away from the Sun.

The moment 2 map shows increasing velocity dispersion toward the hub center, attaining the highest value near the center at two positions. The NW peak of the velocity dispersion coincides with the CO intensity peak and the brightest spot of the HII region (Paron et al., 2013).

Note that the moment 2 map here is biased to indicate smaller velocity dispersion due to the cut-off brightness at $T_{\rm B}\leq 5$ K, which suppressed extended low brightness emission with broader velocity width.

Fig. 8 shows LVD along the Galactic plane from the Columbia ¹²CO line survey (Dame et al., 2001) and an LVD at $b=-0^{\circ}.3$ made from the FUGIN ¹²CO survey. Two major Galactic structures are indicated by dashed lines: the 4-kpc molecular ring (4-kpc MR) and 3-kpc ER drawing an inclined ellipse with the expanding velocity of 50 km s^-1at $l=0^{\circ}$. The molecular complex G18 is located at $(l,v)\sim(18^{\circ},+51)$ km s^-1) near the intersection of the 4-kpc MR and the 3-kpc ER. The FUGIN LVD demonstrates how G18 is compact and particular in the plotted area. There are three possible solutions about the kinematic location of G18 on the LVDs. One and second are the near- and far-side arms on the 4-kpc MR in ordinal circular rotation, and the other is the approaching side of the 3-kpc ER.

Using the CO-line profiles shown in figure 10 we measured the radial velocity of the molecular cloud to be $V_{\rm lsr}=51\pm 2$ km s^-1. This velocity agrees with that measured for the hydrogen recombination lines H$110\alpha$ of 52 to 53 km s^-1(Downes et al., 1980; Quireza et al., 2006).

We here use the most accurate rotation curve (CO) of the first quadrant of the inner Milky Way as shown in figure 7 (Sofue, 2021). The radial velocity $v_{\rm r}$ and distance $d$ to the object in circular rotaiton are related by

and

where $V(R)$ is the rotation velocity at Galacto-centric distance $R$, $V_{0}=238$ km s^-1, and $R_{0}=8.2$ kpc is the Galactic constants. By iteration we obtain $R=4.7\pm 0.2$ kpc, and the near side solution of $d$ is

and a far side solution

Because of the absence of the HI line absorption against continuum of the HII region, as described in the previous subsection, we may choose here the near solution of $d=3.9\pm 0.2$ kpc for the circular rotation.

As shown by the LVDs, G18 is located at the intersection of LVD ridges. So, another possibility is that G18 is located on the 3-kpc ER. If we assume that the source is moving with the expanding ring at $V_{\rm expa}$ and rotating at $V_{\rm rot}$ obeying the RC, we have

where $\theta_{\pm}$ is the GC angle satisfying

For the approaching motion of G18, we can attribute it in the near side of the 3-kpc ER and uniquely solve the equation to yield

Inserting $R_{\rm GC}=R_{0}\sin\ l_{\rm tan}=3.07\pm 0.14$ pc with $l_{\rm tan}=22^{\circ}\pm 1^{\circ}$ as read by an ellipse fit to the 3-kpc ER’s LV ridge in figure 8 , we obtain

Fig. 8 shows a longitude-velocity diagram of the ¹²CO -line emission at $b=-0^{\circ}.3$ in the central disc at $30\geq l\geq-30^{\circ}$. The two arm structures are evident: One is the 4-kpc MR composed of dense molecular complexes, running straightly from $(l,v)\sim(-30^{\circ},-50{\rm km\ s^{-1}})$ to $(+30^{\circ},70{\rm km\ s^{-1}})$, as marked by the red dashed line. The straight LV behavior indicates a circular rotating ring. The other is the 3-kpc ER, composing a tilted ellipse on the LV plane, as traced by the red dashed line, which represents an expanding ring of radius 3 kpc and expansion velocity 50 km s^-1. It is stressed that G18 is located near the intersection of the two rings on the LV plane.

The $\Sigma-D$ (surface brightness-diameter) relation at 1 GHz has been obtained for Galactic SNRs with known distances (Case & Bhattacharya, 1998; Pavlović et al., 2013). We here adopt the most recent relation by (Pavlović et al., 2013)

where $\Sigma_{\rm 1GHz}$ and $D$ are the surface radio brightness at 1 GHz and linear diameter of the SNR, respectively. In order to make the contamination by the thermal emission from the HII regions overlapping the SNR eastern shell as small as possible, we use here the lower frequency (90-cm continuum) map from VGPS survey (Stil et al., 2006). We measure the mean brightness on the map and converted it to that at 1 GHz assuming a spectral index of $\alpha=-0.6$. Using the thus obtained $\Sigma_{\rm 1\ GHz}$, we determined the linear diameter by applying the $\Sigma-D$ relation to obtain $D=20.1^{+22.9}_{-9.3}$ pc and the distance

This distance locates the SNR farther than G18 molecular complex. However, the large error allows for a possible minimum distance of $d\sim 10.1-4.7=5.4$ kpc.

On the other hand the HI -line absorption measurement toward the SNR has shown that the continuum radio emission from the SNR G18.15-0.17 exhibits absorption at $V_{\rm lsr}\simeq 55$ and 100 km s^-1, but no absorption is found at the terminal velocity 130 km s^-1(Leahy et al., 2014). The absorption at 55 km s^-1is attributed to the 3-kpc ER at 6.1 kpc. The 100 km s^-1absorption is due to the HI disc arm in circular rotation in front of the 3-kpc ER, and the kinematic distance is determined to be $5.6\pm 0.4$ kpc as the near solution, in good agreement with the current determination using HI -line absorption (Leahy et al., 2014). Another constraint is given by the absence of absorption at the terminal velocity, indicating that the SNR is located nearer than the tangent point at $\leq 7.6$ kpc.

From these estimations, we can give a firm constraint on the distance of the SNR as $4.0\leq d_{\rm SNR}\leq 7.6$ kpc. This means that it is reasonable to locate the SNR in the 3-kpc ER at a distance of 6.07 kpc. Besides the distance, the apparently compact location on the sky of the SNR in touch with the HII regions and encircled by the molecular shells (Figs. 6) as well as the isolation of the entire G18 system on the sky (Fig. 1) supports the idea that the SNR is physically associated with G18. In more detail, the sharp HI absorption at 55 km s^-1(Leahy et al., 2014) indicates that the SNR is located in the far side of G18 as illustrated in figure 8.

In table 1 we list the possible solutions for the distance to G18, and plot them in figure 9. We have obtained four possible distances:
(a) If the GMC and HII regions are rotating on a circular orbit, we have $d_{\rm MC,\ near}=3.9\pm 0.2$ kpc as the near solution.
(b) Or, $d_{\rm MC,\ far}=11.7\pm 0.2$, as the far side solution. However, this is not allowed, because no HI absorption is observed at the terminal velocity (Dey et al., 2022).
(c) HI -line absorption of radio continuum emission at $V_{\rm lsr}=55$ km s^-1by the 3-kpc ER (Leahy et al., 2014) indicates a distance of $d\geq 6.1$ kpc.
(d) If MC and HII regions are located on the 3-kpc ER, we uniquely obtain $d_{\rm MC}=6.07\pm 0.13$ for the near side solution with approaching (negative) radial velocity.
(e) The SNR’s distance is estimated by the $\Sigma-D$ relation as $d_{\rm SNR}\sim 10.1^{+11.5}_{-4.7}$ kpc, which allows a minimum distance of 6.1 kpc.

We thus obtain a unique combination of GMC, HII and SNR distances consistent with each other, (b) + (c) + (d) + (e), but (a) is not favoured. Figure 9 demonstrates the consistency of the solutions, except for those assuming the circular rotation of G18. The compact and isolated appearance of G18 in the LVD is also particular, showing an exceptionally high brightness clump among the others in Fig. 8. Another point is the LV location, where G18 is located near the intersection of the 3-kpc ER and 4-kpc MR, but displaced from the latter. From these facts, we conclude that the distance to G18 is $d=6.07$ kpc at a Galacto-centric distance of $R_{\rm GC}=3.1$ kpc.

We thus obtain a unique combination of GMC, HII and SNR distances consistent with each other is (b)+(c)+(d)+(e), and (a) is not favoured. Figure 9 demonstrates the consistency of the solutions, except for those assuming the circular rotation of G18. The compact and isolated appearance of G18 in the LVD is also particular, showing an exceptionally high brightness clump among the others in figure 8. Another point is the LV location, where G18 is located near the intersection of the 3-kpc ER and 4-kpc MR, but displaced from the latter. From these facts, we conclude that the distance to G18 is $d=6.07$ kpc at a Galacto-centric distance of $R_{\rm GC}=3.1$ kpc.

Given the distance of $d=6.07$ kpc, we then estimate the molecular mass of the region enclosed by a circle centered on the peak intensity position at G18.15-0.30. We measure the full velocity width of half maximum of the ¹²CO line emission $\sigma_{v}$, and the peak brightness temperature $T_{\rm p}$ using the line profiles averaged in each circle. Then the averaged integrated intensity inside the circle is calculated by

which yields the mean column density of $H_{2}$ by

The molecular mass inside radius $r$ is estimated by

where $\mu=1.38$ is the mean molecular weight per hydrogen including metals, $A=\pi r^{2}$ and $r=d\ {\rm tan}\theta_{\rm dia}/2$. We adopted the conversion factor with radial gradient given by

with $r_{e}=6.7$ kpc (Arimoto et al., 1996), which yields the conversion factor at $R_{\rm GC}=3$ kpc as $X_{\rm CO}=0.96\times 10^{20}$ for the local value of ${X_{\rm CO}}^{0}=2\times 10^{20}$ and $1.4\times 10^{20}$ H₂ cm ^-2 [K km s${}^{-1}]^{-1}$for $d=3.9$ kpc.

Table 2 lists the measured parameters and calculated masses for various radii, and figure 11 plots them. In order to compare the results with those by the current studies, we also present the values calculated for the near distance of 3.9 kpc. Note that the velocity width $\sigma_{\rm v}$ in this table varies rather slowly with the radius compared to the steeper increase of velocity dispersion toward the center in the moment 2 map in figure 5. The reason for the difference is because the values here are means within the averaging circles, while moment 2 map indicates individual local values and is biased to pick up peaky profile component due to the cut-off brightness at $T_{\rm B}\leq 5$ K, leading to smaller velocity widths.

Regardless the object’s distance, the calculated mass increases with the radius, so that the total mass is not uniquely determined from the observation. We here define the cloud’s size as the region inside a circle of diameter $0^{\circ}.4$ corresponding to $r_{\rm cloud}=22$ pc, and total mass to be $M_{\rm G18}\sim 2.8\times 10^{5}M_{\odot}$ for $d=6.07$ kpc.

The mass estimated here may be compared with that obtained in a wider area from ¹³CO intensity (Paron et al., 2013) of $5\times 10^{5}M_{\odot}$ for a distance of 4 kpc and radius 24 pc, which may be converted to $1.2\times 10^{6}M_{\odot}$ and 38 pc at our distance of 6.1 kpc. The difference will be partially due to the adopted conversion factor, which is about a half in the present calculation, and to the defined cloud size on the sky, which is also about a half here.

The half-width-of-half-maximum velocity related to the observed full velocity width $\sigma_{v}$

and radius $r$ are then used to calculated the Virial mass of the cloud inside by

We define the specific volume densities in a sphere of radius $r$ by

and

We plotted the calculated values against $r$ in figure 11 .

We introduce a ratio of the Virial to molecular masses $\mathcal{R}=M_{\rm vir}/M_{\rm mol}$. The ratio is about $\sim 2$ in the centre, and decreases with the radius at $\mathcal{R}=4(r/1{\rm pc})^{-0.6}$ as indicated by the dashed line in the figure. The cloud is more unstable near the center, and is stable, or gravitationally bound, at the edge of the cloud. Such a radial variation may be attributed to a high rate of injection of kinetic energy in the central region. The energy source may be the expanding HII shell produced by the massive star formation in the centre observed as the high-brightness HII region at G18.15-0.30, coinciding in position with the cloud density maximum (Fig. 1).

We have calculated the kinetic energy of the cloud with the derived molecular mass

and compared it with the gravitational energy

which are also listed in table 2 along with their ratio, $E_{\rm k}/E_{\rm g}$. We also calculated their density

and

as well as their ratio. In figure 11 we plotted their values as a function of the radius from the cloud center. The energies of the entire cloud to the edge is on the order of $\sim 10^{50-51}$ erg, We point out that the gravitational energy increases with the radius more steeply than the kinetic energy. The estimated values depend on the molecular mass, and hence on the conversion factor $X_{\rm CO}$, while the general trend of mutual radial variations do not change by the $X_{\rm CO}$ variation within a factor of $\sim 2$.

In the entire cloud, $E_{\rm g}$ significantly exceeds $E_{\rm k}$, or the total energy $U=E_{\rm k}-E_{\rm g}$ is negative. On the contrary, the core region inside $\sim 2$ pc has positive $U$, so that the gas is unstable or expanding. From these energetics, we may consider that the GMC is gravitationally bound and contracting, whereas the central region is expanding due to some injection of kinetic energy from inside. This property does not depend on the cloud’s distance.

The increase of kinetic energy toward the center is related to the increase of velocity dispersion. Although the averaged values over the cloud at different radii are not so variable, the moment 2 map in figure 5 shows an increase in the specific velocity dispersion toward the center.

The kinematic behaviour and gaseous motion can be examined using longitude-velocity diagrams. The right panel of figure 2 shows ¹²CO -line LVD from FUGIN at various latitudes, and figure 12 enlarges the LVD across the peak intensity position of the cloud along longitude, latitude, and the major axis of the molecular hub at $PA=45^{\circ}$. Impressive in these figures is the oval feature symmetrically elongated in the direction of the velocity, composing an ellipse in the LVD.

The elongated feature in the LVD may be explained by various mechanisms such as expanding shells due to the pressure of HII regions (Paron et al., 2013), cloud collision (Dey et al., 2022). interaction of two filaments (Kumar et al., 2020), or by the gravitational contraction.

We here notice that the LVD oval in Fig. 12 is extending for $\sim\pm 5{\rm km\ s^{-1}}$ symmetric with respect to the systemic velocity $V_{\rm lsr}=51$ km s^-1. Such symmetry of velocity may be naturally attributed to an expansion of a shell rather than to a cloud collision, which is characterized by a lopsided bridge feature extending from the targeted GMC to the colliding body (Haworth et al., 2015).

So, we next examine if the observed position-velocity diagrams and morphology of the molecular bubbles can be explained by expanding shells. The top panels show expanding-shell model by star-forming activity at the edge of a molecular cloud represented by the white ellipse. The expansion of a gaseous shell was traced by the adiabatic shock wave approximation in order to examine the qualitative behavior of the expanding bubble (Sakashita, 1971; Sofue, 2022). The left panel shows the shock fronts at two epochs at $t\sim 0.5$ and 1 My as projected on the $(x,z)$ plane, and the right shows corresponding position-velocity diagram seen from the $z$ direction of the shell. Here, the kinetic energy is injected continuously at the coordinate origin by an injection rate of $dE/dt\sim 2\times 10^{49}$ erg My^-1. The parent cloud is assumed to have an off-center Gaussian density profile as indicated by the white ellipse with the center density of $10^{4}$ H₂cm^-3. The parameter combination corresponds to a linear scale of the figure frame of $\sim 20$ pc and velocity frame $\sim 10$ km s^-1.

The middle panels show the integrated ¹²CO intensity (K m s^-1) of G18 molecular bubble1 1 to 3 and the averaged LV diagram near bubble 1 from Fig. 12. The bottom panels show an enlarged (three times than the middle panel) channel map $T_{\rm B}$ (K) at $V_{\rm lsr}=50$ km s^-1around bubble 1 and LVD across the strong continuum source of the HII region shown by the contours in the left panel. The LVD shows an expanding oval feature indicative of a spherical expansion of a shell (bubble) at $\sim\pm 8$ km s^-1. Such a symmetric and high velocity expansion cannot be explained by a cloud collision, which postulates a lopsided oval with a bridge feature Haworth et al. (2015).

The cavity structure opening from the cloud’s edge toward the north and the position-velocity oval elongated in the velocity direction are well simulated by this simple model. The bridge feature, often raised as the evidence for the collision model (Fukui et al., 2021), is also produced in the expanding bubble model. This model is also consistent with the location of the brightest HII region with a hot spot G18.144-0.282 as the energy injection source inside the cavity as observed in figure 6, where a couple of O stars are reported to be located (Paron et al., 2013). It is also pointed out that molecular bubbles 2 and 3 are approximately centred by HII region G18.19-0.18 overlapping with the SNR. These multi-bubble structure is recognized as corresponding multi-LVD ovals in figure 12. We also stress that the outer shell-shaped HII regions G18.18-0.40 (N21) and G18.26-0.31 (N22) are located surrounded by CO shell or arcs.