Discovery of a Giant Molecular Cloud at the Midpoint of the Galactic Bar Dust Lanes: M4.7–0.8

The midpoint of the near-side Galactic Bar Dust Lane is observed to be one of the brightest regions observed in the ¹²CO data set (Figure 1(a), highlighted by the red box). Based on the geometrical model presented in M. C. Sormani & A. T. Barnes (2019), the Midpoint is located at R/R_bar ∼ 0.5, at a distance of 7.0 kpc (we discuss this 3D geometry in more detail in Section 5.1). Gas located at the midpoint is observed to have a line-of-sight velocity of ∼200 km s⁻¹ (Figure 1(b); R. Fux 1999; D. J. Marshall et al. 2008; M. C. Sormani & A. T. Barnes 2019). Based on previous work from M. C. Sormani & A. T. Barnes (2019), the ¹²CO at this location is estimated to have a mass of  ∼10⁶ M_⊙ and is suggested to accrete into the CMZ in 5–6 Myr, potentially forming a Sgr B2-like complex after CMZ accretion. This accretion is in the Westward direction for the near-side Dust Lane (see Figure 1(c)), indicating the plane-of-sky velocity is also Westward. Figure 1(c) shows the ¹²CO contours overlaid on the Herschel 250 μm dust continuum. The solid white box denotes the region we will be referring to as the “Nexus” in this paper. The Nexus coincides with the brightest ¹²CO emission and appears correlated with the nexus of the dust continuum and ¹²CO emission. The dashed white box highlights the feature we will be referring to as the “Filament.” This narrow dust continuum feature appears to be an extension off of the Nexus of the cloud and shows a filamentary morphology.

In this paper, we target these two fields (white boxes in Figure 1(c)) of the midpoint region of the Galactic Bar Dust Lanes using the Robert C. Byrd Green Bank Telescope (GBT). We target multiple transitions of ammonia (NH₃) and cyanobutadiyne (HC₅N) to investigate the higher-resolution molecular gas morphology and kinematics of this region. We discuss the observations and data reduction in Section 2. We present the detection of dense gas in these observations in Section 3 and analyze the morphology and kinematics of the cloud. We compare our dense gas observations to previous dust continuum studies in Section 4. Lastly, we investigate the 3D location, star formation, and other cloud properties in Section 5. We summarize these conclusions in Section 6.

The seven beams of the KFPA were used for simultaneous, dual circular-polarization observations over the  ∼3 footprint of the receiver. Flux calibration was achieved via off-source “blank-sky” and switched noise-diode measurements, resulting in an estimated antenna temperature accuracy of  ∼20%. Atmospheric opacity values were determined from NWS grid values using a combination of three local sites, depending upon the established weather conditions (i.e., wind direction and agreement between the sites) at the time of observation. The VEGAS spectrometer backend was used in mode 20 (23.44 MHz, 200 km s⁻¹, bandwidth per spectral window with 5.7 kHz, 0.07 km s⁻¹, spectral resolution). Table 1 shows the line and imaging information (e.g., rest frequency, angular resolution) for the spectral lines presented in this study. More information on the imaging stage of the data reduction is discussed in Section 2.3.

Maps were performed via an on-the-fly process with Nyquist-sampled rows in Galactic longitude. The “Off” positions were observed at intervals corresponding to the observation of four complete map rows. This allowed for sky emission compensation on a timescale of  ∼10 minutes, adequate for observations at the K band. Due to the high level of extended emission in directions toward the GC, large position offsets were necessary in order to ensure emission-free “Off” measurements. These offsets were typically 5° and 3° in longitude and latitude, respectively. Pointing and focus calibrations were performed at the beginning of each observing session and approximately hourly thereafter.

Data were reduced primarily using the GBTIDL package (P. Marganian et al. 2006).¹⁴ The “Off” position measurements were found to be stable over the duration of each individual session and so were averaged together in order to provide increased signal-to-noise ratio (SNR) in the final reduction. Each individual Nyquist-sampled integration was smoothed via a Gaussian convolution function. The resulting spectral resolution is 0.7 km s⁻¹. The “Off” spectra were further smoothed to a resolution of 0.84 km s⁻¹ before being subtracted from the reference spectra and scaled via the system temperature determined from the “flickered” noise-diode measurements. Atmospheric corrections (found as described above) were applied in order to place our data on the corrected antenna temperature (T_A*) scale.

An initial spectral baseline was fit to emission-free regions of each integration using orthogonal polynomials of the order of 7. The subsequent output was then gridded via the gbtgridder¹⁵ routine using a Gaussian imaging kernel and a pixel width of 9″, approximately one-third of the observed beam size. The FITS cubes resulting from each session were examined individually on the basis of rest frequency as well as on the individual KFPA beam/feed used. In some cases, it was found that individual beams were unstable and/or contributing significantly higher levels of system noise. Where this was found to be detrimental to image quality, these beams were excised from the overall reduction.

It was found that, over three observing sessions performed on the “filament” section of our overall region, there were significant disagreements in the positional metadata of our maps. This resulted in considerable differences in the antenna temperature measurements when compared pixel-to-pixel. It is believed that this problem arose as a result of poor elevation pointing calibration of the GBT, below elevations of  ∼20°.¹⁶ This was addressed in our data via careful examination of pointing offset data from peak scans and corrective offsets being applied to those data before gridding. The maps performed during the pilot phase of the study showed no evidence of this issue. Therefore, we are confident in the astrometry of those maps as a reference and determine our pointing accuracy to be the same as our pixel width, 9″. Following the final combination of data from all individual sessions, a final baseline removal process was applied to each cube as a whole. This was done using a polynomial function fit to emission-free regions. The order of the polynomial fit was chosen to be as low as necessary to remove any obvious “ripple” structure on the scale of the spectral window (23.44 MHz). For the NH₃ (1,1), (3,3), (4,4), and (5,5) lines, this was determined to be a third-order polynomial. For the NH₃ (2,2) and HC₅N transitions, this was found to be second and fifth order, respectively. No direct correlation was found between the level of sinusoidal variation in a transition’s bandpass and any known hardware properties. System temperatures and measured rms noise values were largely consistent across all observed frequencies.

Figures 2 and 3 show the maximum emission in the NH₃ (1,1)–(4,4) and the HC₅N (9−8) transitions in the M4.7–0.8 molecular cloud.¹⁸ The NH₃ (3,3) transition is the brightest out of all of our detected spectral lines, with a peak emission of 2.04 K (>200σ; Table 1). The HC₅N (9–8) transition is the faintest of our detected spectral lines, with a peak of 0.17 K (17σ; Table 1).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The dense molecular gas at the midpoint of the Galactic Bar Dust Lanes (M4.7–0.8) is observed to be extended in Galactic Longitude (∼045, ∼55 pc) and contains several bright clumps that are unresolved in these GBT observations (∼30″, ∼1 pc; a scale bar is shown in Figure 2, bottom). The large-scale structure is fairly consistent across all observed transitions, showing a similar morphology in all five spectral lines (Figures 2 and 3). However, when looking at the localized morphology in the cloud, we can observe interesting substructures. These features are annotated in the top NH₃ (1,1) panel of Figure 2, and they will be discussed throughout this paper.

As mentioned previously in this paper, we highlight two main regions in the cloud: a “Nexus” (the region associated with the ¹²CO peak) and a “Filament” (the dust continuum filament extending out from the Nexus). The Nexus contains several faint, diffuse, extended structures (005–015), with a main bright clump located at l = 474, b = −076, labeled as “Main Clump” in Figure 2.

The Filament is generally long (∼03,  ∼40 pc) and narrow (∼003,  ∼4 pc), resulting in a length-to-width ratio of roughly 10. While generally observed to be one continuous structure, we note a small discontinuity, labeled as the “Gap” (Figure 2; l = 461, b = −085), where the emission is relatively faint when compared with adjacent sections of the Filament. This decrease in emission at this location in the structure is observed across all five transitions. On the eastern side of the Gap, there is a shell-like structure, labeled as the “Shell,” which shows a brighter curved rim and a fainter cavity (dotted black line in Figure 2). The Shell appears to be brightest on the eastern rim with very low level emission on the western rim.

On the western-most region of the Filament, we observe five bright knots of emission in the NH₃ transitions, labeled as Knots A–E in Figure 2. These sources are roughly evenly spaced, with a separation of  ∼2 (4.4 pc) between adjacent knots. While bright in the NH₃ transitions, we do not observe evidence of these knots in the HC₅N line above 125 mK (∼12σ)—only in the large-scale structure of the Filament. Knot C is observed to be the brightest in every NH₃ transition, including the NH₃ (5,5) transition (Figure A1). Knot E is observed to be the faintest in our observed transitions, and is only faintly detected in the J ≤ 4 transitions. Knot E is also observed to be smaller (r = 20″; 0.7 pc) than Knots A–D and is unresolved in these observations.

Knots B, C, and D exhibit some similar characteristics when comparing the NH₃ (1,1) to NH₃ (3,3) emission. Here, the NH₃ (1,1) emission describes condensed circular emission peaks, which are toward the northern edge of the Filament. In comparison, the NH₃ (3,3) emission has a more extended morphology, with the NH₃ (1,1) peaks occurring toward the eastern end of each respective prolate knot. This elongation in the higher J transitions is annotated in the bottom panel of Figure 2. It should be noted that this is true for emission measured in the NH₃ (4,4) and (5,5) transitions as well (Figures 3 and A1, respectively). We also note that this is true for emission measured both as the peak temperature as well as the integrated intensity of the respective NH₃ transitions, as we will illustrate in the next section in our kinematic analysis.

In our study of the near-side Dust Lane, we not only seek to determine the gas structure of the observed region, as discussed in Section 3.1, but also aim to establish the bulk motions within that gas. Whether it is possible to confirm the transport of material along the Dust Lane with the presented data or to investigate any potential structural developments within the cloud as a single element, it is useful to examine the dynamic state of the observed gas. Here we present a kinematical analysis of the dense gas in the surveyed region using a thorough inspection of the observed spectra as a function of position.

The simplest way to investigate gas kinematics is by extracting spectra. We selected five regions to investigate: the Main Clump, the Shell, and Knots B, C, and D. The apertures used to extract the spectra for each region are shown as black circles in Figure 2.¹⁹ The spectra were then extracted and plotted in Figure 4 using the python programs SpectralCube²⁰ and pyspeckit²¹ (A. Ginsburg et al. 2022).

Zoom In Zoom Out Reset image size

The NH₃ molecule has a known hyperfine structure containing four primary satellite lines that are offset from the main component (e.g., P. T. P. Ho & C. H. Townes 1983, gray arrows in the leftmost panel of Figure 4). These satellite lines are clearly detected in the (1,1)–(3,3) transitions. However, drawing upon the (1,1) transitions as an example, we can see that the inner satellite lines are blended with the main component in all of the extracted spectra. It can be seen that the outer satellite lines are also not clearly separated, particularly in the cases of the Shell and Knot D. We attribute this blending to the particularly large linewidths seen in this region and discuss this further in Section 3.2.1. The NH₃ (3,3) line in the Knot B region shows a narrow-line component around 195 km s⁻¹ that is not observed in any of the other regions (Figure 4). This feature appears to be contained to the area marked with the black “ × ” in Figure 2. This narrow-line velocity feature could be associated with NH₃ (3,3) maser emission and is discussed in detail in Section 3.2.2. Several of the profiles show “wings” in the main profile component, as annotated in the Shell panel of Figure 4, suggesting a secondary component offset in velocity. This secondary component is most visible in the Shell profile, but is also present in the Main Clump and Knot B (Figures 4 and 5).

Zoom In Zoom Out Reset image size

The hyperfine structure of NH₃ molecular emission has been used in many studies to extract valuable information on the physical properties (primarily temperature and optical depth/column density) of the regions under examination (e.g., GAS, KEYSTONE, RAMPS; R. K. Friesen et al. 2017; J. Keown et al. 2019; T. Hogge et al. 2018, respectively). However, due to the line blending seen here, such analysis is particularly complicated and may even be impossible (or so degenerate as to be inconclusive). In this publication we are only focusing on the gas morphology and kinematics and will therefore not be using the hyperfine lines in our analysis. A follow-up publication will investigate the hyperfine lines in more detail. Thus, we have chosen to use the (4,4) transition as the primary line from which to extract the basic velocity parameters (e.g., v_c, σ) of our observations.

Figure 5 shows the NH₃ (4,4) line profile for these regions (black histogram), fit with 1–2 Gaussian components (blue profiles) using the python program pyspeckit. The parameters for these fit components are shown in Table 2. Most of the spectra are best fit with two Gaussian components with the exception of Knot D, for which a single component is optimal. The solid green line, offset at −0.025 K in all panels, shows the residual of the best-fit line profile, shown as the red line in Figure 5. A secondary residual (blue dotted line) is shown for Knot D, for which there is slight improvement in the fit if a secondary component is included. This improvement in the residuals is best seen around 200 km s⁻¹. However, as this improvement is at the 1σ level of the noise, we will examine this source using the single Gaussian fit.

Table 2. Kinematics of the NH₃ (4,4) Transition

Parametera	Value
Main Clump	r  ∼  891 (3.46 pc)
Ac	64  ±  7 mK
vc	203.1  ±  0.4 km s−1
σ	5.2  ±  0.1 km s−1
Ac	52  ±  8 mK
vc	206.5  ±  0.2 km s−1
σ	2.8  ±  0.3 km s−1
Shell	r  ∼  1124 (4.36 pc)
Ac	68  ±  7 mK
vc	184.4  ±  0.1 km s−1
σ	5.7  ±  0.1 km s−1
Ac	12  ±  1 mK
vc	202.3  ±  1.5 km s−1
σ	10.0  ±  1.2 km s−1
Knot B	r  ∼  390 (1.51 pc)
Ac	207  ±  26 mK
vc	187.5  ±  0.2 km s−1
σ	3.7  ±  0.2 km s−1
Ac	31  ±  12 mK
vc	180.1  ±  4.5 km s−1
σ	5.8  ±  2.1 km s−1
Knot C	r  ∼  341 (1.32 pc)
Ac	224  ±  8 mK
vc	188.6  ±  0.1 km s−1
σ	4.4  ±  0.1 km s−1
Ac	47  ±  8 mK
vc	187.5  ±  0.2 km s−1
σ	1.7  ±  0.3 km s−1
Knot D	r  ∼  511 (1.98 pc)
Ac	154  ±  2 mK
vc	190.8  ±  0.1 km s−1
σ	5.0  ±  0.1 km s−1

Note. ^aA_c is the amplitude of the fit, v_c is the central velocity of the component, and σ is the velocity dispersion. The r parameter on the same line as each of the region names defines the radii value of that region. Region names are in bold.

Download table as:  ASCIITypeset image

Most of the measured velocity dispersion values are between 2 and 5 km s⁻¹ (Table 2). The largest velocity dispersion measured in our spectra is associated with the Shell (10.0  ±  1.2 km s⁻¹). Knot D is best fit with a broadest single component spectrum (5.0  ±  0.1 km s⁻¹). Knot C is the brightest of the five regions shown here. The offset in the secondary velocity component is the most extreme in the Shell and Knot B, with velocity offsets >5 km s⁻¹. These lower-intensity velocity components, discussed previously as “wings,” are annotated in these two panels of Figure 5.

Due to the hyperfine structure of the NH₃ molecule, analysis of the kinematics can be quite complex because of the satellite lines. These hyperfine lines become much more prominent at the lower transitions (e.g., Figure 4). Therefore, we will be using the NH₃ (3,3) line for the kinematic analysis for the following arguments. First, the emission in the (3,3) line is relatively bright when compared with the (1,1) line (see Figure 2). Second, hyperfine satellite lines in the (3,3) transition are relatively faint compared to the main component. These strong hyperfine lines in the (1,1) and (2,2) transitions are visible in Figure 4. Although the hyperfine lines in the (3,3) transition are visible in Figure 4, the ratio of the hyperfine lines to the main component is lower than the (1,1) and (2,2) transitions. Third, the larger velocity offset between the satellite lines and the main component allows for better isolation of the main component in velocity space. We note in later sections where the analysis could be influenced by the lower-level hyperfine emission. And lastly, while the HC₅N (9–8) transition would be an excellent candidate to use for kinematic analysis, due to the lack of hyperfine structure in the line profile, the relatively low SNR makes fitting the velocity components across the structure unfeasible. Although the NH₃ (4,4) line was used for the spectral line analysis, discussed earlier in the section and shown in Figure 5, the relatively faint nature of this transition on a pixel-by-pixel analysis (e.g., see the maximum intensity emission shown in Figure 3) makes this transition unfeasible to use in the “moment” analysis in the following section.

3.2.1. Central Velocity and Velocity Dispersion

In this section, we use the NH₃ (3,3) data to investigate the gas kinematics in the cloud by creating moment 0 (M₀, integrated intensity), moment 1 (M₁, central velocity), and moment 2 (M₂, velocity dispersion) maps. These moment maps are defined by the following equations:

and

where is the antenna corrected temperature of the data cube in kelvin, and v is the velocity in kilometers per second. These types of visualizations can illustrate velocity gradients across the source and can reveal regions with high velocity dispersion. However, care is needed when viewing these images—especially in cases where hyperfine lines are strong or multiple components are present. Figure 6 shows these three moment images, discussed above, for the NH₃ (3,3) line.

Zoom In Zoom Out Reset image size

All three moment maps shown in Figure 6, M₀–M₂, were created in CASA (CASA Team et al. 2022), using the immoments command. We used a robust lower limit cutoff of 150 mK (15σ) for all three moments to avoid fitting noise components in our integration. We used a velocity range of 146.4–257.8 km s⁻¹ when making these images.

The NH₃ (3,3) integrated emission (Moment 0; Figure 6, top) generally traces a distribution similar to that of the NH₃ (3,3) peak intensity (Figure 2, bottom). As noted previously in Section 3.1, the elongation observed in the peak emission in Knots B–D is also observed to be present in the integrated emission as well. The NH₃ (3,3) central velocity (Moment 1; Figure 6; middle panel) shows that velocities in the cloud generally range between 180 and 210 km s⁻¹. This velocity range is consistent with values expected for the midpoint of the dust lanes (∼200 km s⁻¹; see Figure 1(b)).

In general, the Nexus has slightly higher velocities, around 195−210 km s⁻¹, when compared to most of the emission in the Filament (180−195 km s⁻¹). This subtle velocity distribution (∼15 km s⁻¹) across the broad  ∼60 pc structure results in a large-scale velocity gradient across the M4.7–0.8 cloud of  ∼0.25 km s⁻¹ pc⁻¹, from east to west.

Within the Nexus region of the M4.7–0.8 cloud, there are no broad-scale velocity gradients. There are a few localized gradients of a few kilometers per second associated with emission across adjacent cloud clumps. South of the bulk emission in the Nexus is an elongated vertical clump that is labeled as the 230 km s⁻¹ cloud in the top panel of Figure 6. This feature is oversaturated in the M₁ map (Figure 6, middle panel) and is a distinct component along our line of sight. This feature is also observed to be a discrete component in position–velocity space (see Figure B2 and Appendix B.3 ).

The Filament shows a more complex velocity structure compared with the Nexus. Toward the “Gap” in the Filament (annotated in Figure 6, top) we observe the lowest velocity across the entire cloud, with velocity values extending down to 175 km s⁻¹. There is a generally increasing velocity gradient away from the “Gap” location in both directions along the Filament. Adjacent to the “Gap,” near the western rim of the “Shell,” we observe the highest velocity associated with the Filament (210–215 km s⁻¹), which we have labeled as the “high-velocity component.” In the Moment 1 map, in the region between the high-velocity component and the Gap, we observed a sharp jump in the central velocity across a few pixels. This type of discontinuity in the velocity distribution typically occurs when two components are located along the same line of sight. In this case, the central velocity in the fitting algorithm is shifted by bright emission in one component compared with the other (see Equation (2)) across a few pixels. This higher-velocity emission appears to be constrained to the western rim of the Shell, suggesting this feature is likely a secondary component along the line of sight. Furthermore, the high-velocity component is observed to be a distinct feature in the position–velocity diagram (Figure B2) and in channel maps (Figure B3)—both of which are discussed in Appendix B. This high-velocity component is likely the secondary profile observed in the Shell spectrum (10 km s⁻¹; Figure 5, Table 2). As the spectrum shown in Figure 5 is from the NH₃ (4,4), and therefore from the same NH₃ molecule shown in Figure 6, they are likely tracing the same velocity components.

The general velocity gradient in the Filament is clear, with an increasing velocity from Knot A (∼180 km s⁻¹) to Knot E (200 km s⁻¹).²² Knots B and C show a velocity gradient across the elongated knot substructure. In these two sources, we also observe a velocity gradient from the northeast to the southwest. In the northeast, the emission is higher, around 190 km s⁻¹, whereas in the southwest the emission lower, around 183–185 km s⁻¹. This results in a velocity gradient across the substructure of 2.5 km s⁻¹ pc⁻¹. Knot E is observed to have the highest velocity and appears distinct from its neighbor, Knot D.

The NH₃ (3,3) velocity dispersion (Moment 2; Figure 6, bottom) in the M4.7–0.8 cloud is generally around 2–10 km s⁻¹. The periphery of the cloud is observed to have lower velocity dispersion (≤2 km s⁻¹), whereas the interior of the cloud is observed to have higher velocity dispersion (∼6 km s⁻¹). There are a few clumps where the velocity dispersion exceeds 10 km s⁻¹. The high-velocity component is observed to be one of these regions. This high velocity dispersion is consistent with the spectral fitting results for the secondary component in the Shell (Table 2).

However, many of these other clumps with large velocity dispersion values have prominent hyperfine lines. Due to the approach in which CASA calculates the velocity dispersion, by using a simple, single Gaussian fit, satellite lines can result in a broadening of the velocity dispersion in regions with relatively bright hyperfine lines. While the hyperfine lines are suppressed in the (3,3) transition when compared with the (1,1) line (see Figure 4), regions where the hyperfine line is quite strong can result in high apparent velocity dispersion values in Moment 2 maps. Knot C is one of these regions which shows relatively high velocity dispersion in the Moment 2 map, but an examination of the (3,3) profile for this region reveals relatively strong hyperfine lines that are likely causing the high dispersion values in this plot. Here we take the presented dispersion values, in Figure 6, as illustrative and reserve a full analysis of the individual hyperfine component velocity dispersions for a future publication.

3.2.2. Detection of a New NH₃ (3,3) Maser

A narrow-line velocity feature is observed in the NH₃ (3,3) spectral cube at l = 453, b = −082 (see black “×” in Figure 2), and is clearly visible in Figure 4 (Knot B, NH₃ (3,3) panel), and Appendix B Figures B2 and B3. The location of this narrow-line emission feature suggests it is associated with the northwestern region of Knot B. Bright, narrow-line profiles in the NH₃ (3,3) are often indicative of maser emission (e.g., J. G. Mangum & A. Wootten 1994; K. E. Kraemer & J. M. Jackson 1995), supporting our identification of this feature as a new NH₃ (3,3) maser detection.

For this analysis, we will be using an un-smoothed version of the data set. This data set was sampled at the native resolution of 0.07 km s⁻¹, with no additional smoothing (see Section 2.3). This higher-resolution channel width is necessary to fully resolve and sample the narrow-line component associated with the maser emission. To better understand the nature of the source, we fit the NH₃ (3,3) spectrum, shown in Figure 4, using the python program pyspeckit (A. Ginsburg et al. 2022). This package allows us to account for the prominent hyperfine components of the NH₃ (3,3) transition when performing our fit. The raw NH₃ (3,3) spectrum for Knot B is shown in Figure 7 as a black histogram. We fit this profile with a two-component fit, using similar central velocity and velocity dispersion values previously measured in the NH₃ (4,4) data (see Table 2). We note that despite the higher spectral resolution of this data, compared with the spectrum used in Figure 5, we still recover the same components. The residuals from this two-component fit are shown as a dotted blue line at −0.25 K, in which the narrow-line feature is clearly detected. When adding an additional narrow-line component to our fitting algorithm, we can see that the residuals are greatly improved (solid green line at −0.25 K in Figure 7). The parameters of this three-component fit, with the combined fit shown as a red line in Figure 7, are presented in Table 3. The introduction of a fourth component, with a central velocity of 181 km s⁻¹ and a broad velocity dispersion of  ∼8 km s⁻¹, can slightly improve the residuals in the fit around 180 km s⁻¹. However, without additional evidence to support this fourth component, it is unclear whether this is an additional velocity component or the result of noise in the brightest emission channels. Higher-resolution observations are necessary to spatially disentangle these multiple components and determine whether a fourth component is indeed present in this region.

Zoom In Zoom Out Reset image size

Table 3. Kinematics of the NH₃ (3,3) Transition in Knot B

Parametera	Value
Main Component
vc	188.42  ±  0.02 km s−1
σ	3.05  ±  0.03 km s−1
Secondary Component
vc	181.88  ±  0.21 km s−1
σ	7.49  ±  0.09 km s−1
Maser Component
vc	193.16  ±  0.01 km s−1
σ	0.37  ±  0.01 km s−1

Note. ^av_c is the central velocity of the component, and σ is the velocity dispersion.

Download table as:  ASCIITypeset image

Figure 11 shows the dust temperature (top) and N(H₂) column density (bottom) from K. A. Marsh et al. (2017) for M4.7–0.8,²³ based on the Herschel 70, 160, 250, 350, and 500 μm measurements. The dust temperature in the cloud is observed to be between 19 and 21 K (Figure 11, top). The morphology of the cooler dust component (T_dust <  20.5 K) appears to trace the M4.7–0.8 cloud (black contours; see Figure 2). The dust in the region surrounding the NH₃ emission in the Nexus (∼20.25 K) is slightly cooler than the dust in the region around the Filament (∼20.75 K). This could be an indication that the environment around the two regions is different. The extended emission around the Nexus, discussed in Section 4, is also relatively cool and shows similar dust temperature values as the bulk M4.7–0.8 cloud. The region around the Filament is observed to be relatively warm compared with the colder interiors.

Zoom In Zoom Out Reset image size

The colder dust temperatures in the cloud (i.e., <20 K) are strongly correlated with the bright NH₃ (1,1) regions, located in the innermost regions of the clumps and knots. The coldest of these is Knot D, with a measured dust temperature of 19.2 K. Knot E is observed to be hotter than all other NH₃ peaks, with a temperature of 21 K. We note a slight positional offset between this hot dust core and the NH₃ (1,1) peak of  ∼10″. This offset is smaller than our GBT beam size (35″) and therefore could be caused by a pointing offset rather than a physical offset of the source. The emission surrounding Knot E appears to be slightly cooler (20.1 K) than the background emission (20.8 K), indicating this hot dust core could be embedded in a cooler envelope. We investigate this source, and the implication of an embedded core, in more detail in Section 5.3.2.

The N(H₂) column density in the cloud is around 0.5–2.0  ×  10²² cm⁻² (Figure 11, bottom). The highest column densities in the region trace the NH₃ emission morphology of M4.7–0.8. All of the knots located in the Filament show higher column densities in their interiors than the surrounding regions, with the highest column densities in M4.7–0.8 (1.82 × 10²² cm⁻²) associated with Knot D. Knot E also shows relatively high column densities (1.65 × 10²² cm⁻²), indicating that in addition to being quite compact (<10″; 0.4 pc), it is quite dense as well. We also note that these column density values are consistent with those in giant molecular clouds (GMCs; see Section 5.2).

The Nexus region of the cloud also shows moderately high column densities, with values around 1.0 × 10²² cm⁻². There are two main high column density clumps in this region, the Main Clump (as annotated in Figure 2) and a secondary clump located in the Eastern region of the cloud (l = 484, b = −080). This Eastern source does not appear to stand out in the NH₃ emission nor the HC₅N emission. However, this region does appear to be relatively bright in the 250 μm emission and in the Herschel short wavelength data sets (70 and 160 μm). This region of the cloud also appears to be quite large, extending well past the NH₃ (1,1) contours and appearing to correlate with the IR-dark region around the Nexus, observed in the Spitzer 8 μm data (Figure 9, bottom).

The Gap region in the Filament also exhibits relatively low column density, with only a faint bridge connecting the eastern and western sides of the Filament. The low column density regime appears to be associated with a large cavity that extends toward the south, below the filament, and carves a lower-density region north, above the Gap. This lower column density at the location of the Gap could be an indication that feature is produced by nearby stellar feedback that has disrupted the Filament (e.g., C.-P. Zhang et al. 2016; X. Chen et al. 2023).

Figure 12 shows the top-down geometry of the Milky Way Galactic Bar Dust Lanes from M. C. Sormani & A. T. Barnes (2019; see their Figure 2). The orientation and extent of the Galactic Bar Dust Lanes is marked with a dashed ellipse, with red and blue lines illustrating the near- and far-side Dust Lanes, respectively. The Galactic Bar Dust Lanes can roughly be divided into four main quadrants, labeled as Q1–Q4 in Figure 12. Quadrants 1 and 3 correspond to the near-side and far-side Dust Lanes, respectively. Quadrants 2 and 4 correspond to gas and dust that have “overshot” the GC (i.e., “spray”; T. Kim et al. 2024; see their Figure 1), from Quadrants 1 and 3, respectively, and are continuing to traverse the Bar. The gas and dust in these overshot regions can then interact with the approaching material on the opposing side to produce strong shocks from cloud–cloud collisions (M. C. Sormani et al. 2018a; M. C. Sormani 2021; S. R. Gramze et al. 2023). These shocks are visible as vertical features in position–velocity space (see Figure 1(b) and M. C. Sormani 2021).

Zoom In Zoom Out Reset image size

The midpoint cloud is located at a Galactic longitude of roughly l  ∼ 5° (angle ϕ in Figure 12). Based on the geometry shown in Figure 12, this suggests the cloud is located at (−0.59, −1.19), as marked by the yellow star. At this location, the cloud is calculated to be at a distance 7.035 kpc from the Sun. We also note the significantly low Galactic latitude of the cloud (b = −085), which suggests the cloud is below the midplane of the Milky Way. Using the geometry presented in A. A. Goodman et al. (2014; see their Figure 2), we calculate that M4.7–0.8 is located at 118.5 pc below the Sun, or 93.5 pc below the midplane (assuming the cloud has a distance of 7.0 kpc and the Sun has a positive z-height of 25 pc).

The cloud is also likely to be on the leading edge of the Dust Lanes. Orbital motions within the bar can create a shearing effect on the interstellar medium (ISM; E. Athanassoula 1992; E. Emsellem et al. 2015; F. Renaud et al. 2015). This strong shear can inhibit the formation of dense structures, such as clouds, which leads to low star formation within the bar, even in regions with relatively high densities. However, at the bar’s leading edge, supersonic turbulence, along with reduced shear and weaker tidal forces compared to the inner regions, promotes the formation and persistence of high-density structures (F. Renaud et al. 2015). The leading edges of the bar can also produce converging flows and large-scale shocks that promote the accumulation of dense gas into kiloparsec-scale structures (e.g., E. Athanassoula 1992).

5.1.1. Relative Location to Other Dust Lane Clouds

The leading edge of the Galactic Bar dust lanes tends to accumulate dense gas and promote the formation of overdense structures (E. Athanassoula 1992; F. Renaud et al. 2015). These overdense structures have the potential to form GMCs within the larger Dust Lane. The physical extent of the M4.7–0.8 cloud (∼60 pc) suggests this structure is likely a GMC within the leading edge of the near-side Dust Lane. GMCs are typically long (∼50–100 pc), dense (≳100 cm⁻³), massive (10⁴–10⁶ M_⊙) clouds, often observed to be stellar nurseries (see M. Chevance et al. 2023, for a review of GMCs).

The observed length of the M4.7–0.8 cloud is consistent with many other GMCs observed in the Milky Way and other local GMCs in nearby galaxies (e.g., M. Chevance et al. 2023). These local GMCs tend to be filamentary in nature and have been argued to be the “bones” of the Milky Way spiral arms (e.g., A. A. Goodman et al. 2014; C. Zucker et al. 2015). Many of these GMCs also tend to have localized regions of higher densities (i.e., knots) where the gas has been concentrated by its self gravity. The M4.7–0.8 cloud also shows these types of localized knots that are observed in the Filament (Knots A–E).

In general, the N(H₂) column densities in M4.7–0.8 (∼10²² cm⁻²; Figure 11, bottom) are similar to column density values measured in Disk GMCs (e.g., K. A. Marsh et al. 2017). However, when comparing our measured column density to values found in CMZ clouds, we find that CMZ clouds are an order of magnitude higher than those observed in M4.7–0.8 (∼10²³ cm⁻²; C. Battersby et al. 2020). This suggests that the column density in the cloud will either likely increase during accretion into the CMZ or will coalesce with other clouds in the CMZ to result in the higher values observed in the inner region.

Using the column density data from K. A. Marsh et al. (2017), shown in Figure 11 (bottom), we can calculate a total mass for the M4.7–0.8 GMC. We used the PPMAP-MASS jupyter notebook, which calculates a cloud mass from the K. A. Marsh et al. (2017) column density data set (S. R. Gramze et al. 2023).²⁵ This notebook first calculates an average column density over a specified region, which, in our case, uses the gray dashed box in Figure 11 (bottom). Using this average column density and area, with a set distance of 7.0 kpc (as calculated in Section 5.1), we estimate a cloud mass of 1.6 × 10⁵ M_⊙. We note, however, that this cloud mass is lower than previous mass estimates calculated using the T. M. Dame et al. (2001) CO data set (M. C. Sormani & A. T. Barnes 2019).

M. C. Sormani & A. T. Barnes (2019) calculated a mass estimate of this region of roughly 10⁵–10⁶ M_⊙ and used an X_CO factor of 2 × 10²⁰(N(H₂) cm⁻²)/(W(l, b) K km s⁻¹) for their mass estimates (see their Figure 3). While this X_CO factor, used in M. C. Sormani & A. T. Barnes (2019), represents the Galactic average, it is relatively high for CMZ clouds (A. D. Bolatto et al. 2013). Extragalactic studies of the CO-to-H₂ conversion factor illustrate that this value tends to be lower in the inner region of galaxies when compared to the disk (e.g., Y.-H. Teng et al. 2023; I.-D. Chiang et al. 2024).

In the S. R. Gramze et al. (2023) mass calculation of the G5 GMC, they calculated a mass of 2 × 10⁵ M_⊙—using an X_CO factor of 1.5 × 10¹⁹ cm⁻². However, in their PPMAP calculation, they are measuring a mass of 2 × 10⁴ M_⊙. This order-of-magnitude increase in the X_CO calculation, when compared with the PPMAP estimate, is similar to what we are observing in M4.7–0.8. This could be an indication that the X_CO factor could be different from values used in the Disk and CMZ. Despite the range of mass estimates for this cloud (10⁵–10⁶ M_⊙), which could be constrained in later publications, the estimated cloud mass falls well within the range for GMCs (10⁴–10⁶ M_⊙; e.g., M. Chevance et al. 2023)—supporting our argument that M4.7–0.8 is a GMC.

Based on the large-scale velocity gradient of the near-side of the Galactic Bar Dust Lanes, shown in Figure 1(b), we would expect an increasing velocity gradient from east to west across the cloud of  ∼+0.25 km s⁻¹ pc⁻¹. However, we instead observe a decreasing velocity gradient of −0.25 km s⁻¹ pc⁻¹ (see Section 3.2.1). We note that this is an average across the entire length of the GMC—where local velocity gradients in the Filament could be an order of magnitude higher (see Figure B2). This difference in the averaged velocity gradient across the cloud and the large-scale velocity gradient of the dust lanes could be an indication of some internal dynamics affecting the velocity structure of the GMC.

Velocity gradients across GMCs is not uncommon—as observed in E. Rosolowsky et al. (2003), who investigated the kinematics of 45 GMCs in M33 at 20 pc resolution (which would correspond to roughly 3 pixels beams⁻¹ across M4.7–0.8). They observed, both positive and negative, velocity gradients across the GMCs in their sample that tend to be around 0.1 km s⁻¹ pc⁻¹ (see their Figure 7)—indicating that our measure velocity gradient across the entire GMC is slightly more extreme than their average. In 40% of their sample GMCs, E. Rosolowsky et al. (2003) saw counterrotation when compared with the disk rotation of M33. If M4.7–0.8 is the result of a previous cloud–cloud collision, as hypothesized in our previous section, then the resulting material could “decelerate,” resulting in this apparent counterrotation, as has been observed in previous simulations (R. G. Tress et al. 2020; see their Figure 16). Follow-up CO observations at  <30″ resolution are needed to determine the kinematics of the extended region to ascertain whether the Filament is showing counterrotation relative to the bulk emission in the GMC. For this publication, we will limit our analysis to investigating the cloud kinematics and turbulence by using the spectral fitting information from Table 2 in the following section.

5.2.1. Elevated Turbulence in M4.7–0.8 When Compared to GMCs in the Disk

The CMZ is observed to have broader linewidths than the Galactic disk on similar size scales (e.g., R. Shetty et al. 2012). These broad lines are typically an order of magnitude greater than values measured in the disk. This increase is often attributed to turbulence; however, other possible scenarios include tidal shearing and feedback (e.g., J. D. Henshaw et al. 2023; see their Table 2). One method of understanding this turbulence is by plotting the measured linewidth against the physical radius of the region used to extract the line, also known as a Larson relationship (R. B. Larson 1981), defined as:

(see their Equation (1)). Here σ is the velocity dispersion in kilometers per second, and R is the radius of the cloud.

In Figure 13, we plot the R. B. Larson (1981) relationship for the cloud using the five spectra shown in Table 2 (shown as red “×” symbols) against other clouds in the CMZ (squares; T. Oka et al. 2001; J. Kauffmann et al. 2017; green and blue data points, respectively) and Galactic Disk GMCs (orange diamonds; M. Heyer et al. 2009). Data for these other clouds was pulled from J. S. Urquhart et al. (2015), who used NH₃ (1,1) and (2,2) observations to target high-mass star-forming clumps in the Disk, shown as purple triangles. We also show data from V. S. Veena et al. (2024), who targeted the Helix stream (pink star in Figure 12) as light-pink circles. We note that NH₃ (1,1) GBT observations of Sgr E exist (L. D. Anderson et al. 2020); however, these are single pointing spectra, and therefore, no radii information is available for this data set. We note, however, that the spectra were fit, and the measured values for Sgr E range from 1.3–40.3 km s⁻¹.

Zoom In Zoom Out Reset image size

We find that the measured velocity dispersion for the M4.7–0.8 cloud is fairly similar to values observed in the CMZ (green and blue squares), the Helix stream (pink circles), and high-mass star-forming regions (purple triangles) when compared with the Disk GMCs (orange diamonds). This similarity between the CMZ and Galactic Bar data points could be an indication that the turbulence observed in the CMZ is the result of infall to the gravitational potential. We note that the values measured in the midpoint cloud are fairly similar to those observed in the Helix stream (V. S. Veena et al. 2024). The Helix stream is argued to be “downstream” from the Midpoint and associated with material that has overshot the CMZ—indicating that it is likely located in Quadrant 2 (Figure 12). We note that this region is also “downstream” in the dust lanes from an “interaction location” (S. R. Gramze et al. 2023). Thus, the observed turbulence could be the result of a previous cloud–cloud collision. Results from M. Heyer et al. (2009) in Figure 13 represent GMCs in the Galactic Disk. Here we see that these GMCs have significantly lower velocity dispersions, while also presenting larger physical sizes, in comparison to our data and the results of V. S. Veena et al. (2024). This is not unexpected, due to the large (>10 pc) scale of the M. Heyer et al. (2009) sources. However, this is useful as a contrast. Further analysis of the structure of M4.7–0.8, with a focus on comparing cloud and clump properties to previously identified samples is beyond the scope of the current work and is planned for a follow-up investigation (L. K. Morgan et al. 2025, in preparation).

The higher linewidths observed in M4.7–0.8 are also consistent with the detection of the HC₅N (9–8) transition. M. Chevance et al. (2023), and references within, noted that clouds that are rich, with complex molecules often displaying broader linewidths when compared with other clouds in the ISM. The detection of the HC₅N (9–8) transition could be an indication that M4.7–0.8 is rich in other complex molecules, similar to other broad-line regions like the CMZ. Follow-up observations of additional complex molecules are currently underway with Atacama Large Millimeter/submillimeter Array (ALMA) to address this theory (N. Butterfield et al. 2025, in preparation).

Understanding star formation in barred spiral galaxies has been gaining momentum in recent years with the onset of high-resolution, high-sensitivity telescopes like ALMA and JWST (e.g., A. C. Phillips 1996; M. W. Regan et al. 1999; J. Neumann et al. 2019; K. Sheth et al. 2000; S. Verley et al. 2007; F. Renaud et al. 2015; A. Fraser-McKelvie et al. 2020; F. Maeda et al. 2023; L. Neumann et al. 2024). Bars within spiral galaxies are observed to contain abundant molecular material, which could potentially form stars (S. Díaz-García et al. 2021). This region of the galaxy also has a higher stellar density and is more metal-rich than the disk at similar radii (J. Neumann et al. 2024). These extragalactic studies of barred galaxies found that primarily low-mass galaxies host star formation in their bars and are typically usually on the leading edge of the bar (A. Fraser-McKelvie et al. 2020). Furthermore, these sites of star formation are regulated by shear forces, turbulence, and gas flows (F. Renaud et al. 2015; A. Fraser-McKelvie et al. 2020). In this section, we explore possible signs of star formation in the cloud by focusing on Knots B (Section 5.3.1) and E (Section 5.3.2) and investigate possible stellar feedback producing the Shell (Section 5.3.3).

5.3.1. Is There Star Formation in Knot B?

Knot B is observed to contain several tracers of possible star formation, including the new detection of an NH₃ (3,3) maser (Section 3.2.2). Analysis of the dust continuum emission shows compact emission southwest of the (3,3) maser, near the observed edge of the knot, in the 8, 12, 24, 70, and 870 μm bands (red box in Figures 8–10). Figure 14 shows a higher-resolution image of the 8 μm data in this compact emission. Overlaid are contours showing emission at 870 μm (blue; Figure 8), 70 μm (yellow; Figure 9, bottom), and 24 μm (red; Figure 10, top).

Zoom In Zoom Out Reset image size

The 8 μm observations show this source is extended (∼0.5 pc) and contains several bright point sources. The 24 μm emission is closely associated with the bright emission observed in the 8 μm. The 70 μm emission is concentrated toward the bright, extended 8 μm structure near the center of the region. The 870 μm emission appears to be marginally offset from this central source and slightly elongated—extending away from the main concentration of mid-IR emission.

The detection of relatively compact emission at mid-IR (8, 12, 24, and 70 μm) and in the submillimeter (870 μm), with no compact emission detected at far-IR wavelengths (160 and 250 μm), could indicate that this source is an H ii region. The detection in the 8, 12, 24, and 70 μm bands is likely associated with the hot dust component. The compact emission observed in the 870 μm band could be associated with the tail end of the “free–free” Bremsstrahlung emission from radio and millimeter wavelengths. Both of these emission mechanisms, free–free emission at submillimeter and hot dust at the mid-IR, are observed to be associated with H ii regions (e.g., C. G. Wynn-Williams 1974).

Furthermore, the detection of an NH₃ (3,3) maser may also support this argument of possible star formation in Knot B. NH₃ (3,3) masers have been previously observed toward regions of high-mass star formation (e.g., K. E. Kraemer & J. M. Jackson 1995; Q. Zhang & P. T. P. Ho 1995). K. E. Kraemer & J. M. Jackson (1995) observed NH₃ (3,3) masers near the ends of high-velocity molecular outflows as the results of shocks. The slight offset in the maser location could be due to the jets, from the massive star formation complex, interacting with the bulk emission in the filament.

While the mid-IR and submillimeter observations, combined with the maser detection, support this argument for an H ii region, we note that the source is lacking radio counterparts. We investigated the 2–3 GHz VLASS 3.0 data set (M. Lacy et al. 2020) for radio counterparts in the region and none were found above a 3σ level (<0.4 mJy beam⁻¹). The GLOSTAR C-band (6 GHz) survey was also checked, and no detection was found above a 3σ level in the B-configuration data of this survey either (<0.3 mJy beam⁻¹; A. Y. Yang et al. 2023). This nondetection at low radio frequencies could indicate that the turnover for the free–free emission in the radio spectrum is above this frequency range, or the emission is below the noise floor in the VLASS and GLOSTAR surveys. High-resolution radio and millimeter observations of this source are needed to determine (1) if this is indeed an H ii region by investigating the spectral energy distribution (SED) of the source, and (2) if the velocity of source is consistent with the GBT molecular emission, thereby confirming the association with the cloud.

5.3.2. Is Knot E a Free-floating Evaporating Gas Globule?

5.3.3. Is the Shell Evidence of Feedback?

Figure B1 shows the Moment 1 map for the NH₃ (3,3) emission in the Filament, constrained to the velocities in this region of the cloud (181–200 km s⁻¹). The velocity gradients in the knots are slightly more prominent in this constrained velocity range.

Zoom In Zoom Out Reset image size

Figure B2 shows a position–velocity diagram for the M4.7–0.8 molecular cloud. This plot was made in python using pvextractor,³⁰ with a slice width of 01625, centered at b = −081, and covering a Galactic longitude from l = 49 to 446. Most of the emission in the cloud is between 170 and 220 km s⁻¹ (solid black lines in Figure B2). The 230 km s⁻¹ cloud, discussed in Section 3.2.1 and annotated in Figures 6 and B2, is observed to be a distinct feature in position–velocity space.

Zoom In Zoom Out Reset image size

Between l = 49 and l = 47, the velocity in M4.7–0.8 is fairly constant—around 200 km s⁻¹ (dotted yellow line). Most of this emission is associated with the region we are defining as the “Nexus.” At l = 47, we observe a gradual decrease in the velocity from  ∼200 to  ∼ 180 km s⁻¹ (dotted–dashed line). This Δv (∼20 km s⁻¹) and Δl (∼0075,  ∼10 pc) results in a velocity gradient of  ∼2 km s⁻¹ pc⁻¹. Here, around l = 46–465, we observe a gap in the structure, where the emission is relatively faint. This feature is associated with the “Gap” region annotated in Figures 2 and 6. At lower Galactic longitudes, we observe a second velocity gradient between 46 and 45, with velocity increasing from 170 at l = 46 to 200 km s⁻¹ at l = 45 (dashed line), resulting in a velocity gradient of  ∼2 km s⁻¹ pc⁻¹. Both of the nonzero velocity gradients are associated with the feature we are identifying as the “Filament.”

In addition to the “Gap” between the two velocity gradients, we also observe a feature at a higher velocity: 210 km s⁻¹. This feature is associated with emission in the high-velocity component (as annotated in Figure 6, top), which shows high velocities and broad linewidths in that figure (middle and bottom panels, respectively). Lastly, we also observe a narrow velocity feature around l = 453 and v = 195 km s⁻¹. We have marked this feature as possible maser emission in Figure B2, and will discuss this emission feature in more detail in Section 3.2.2. The hyperfine structure of the NH₃ (3,3) line is also clearly visible in the Filament region of the cloud as faint emission that is  ±20.7 and  ±28.2 km s⁻¹ from the main, bright peak.

Figure B3 shows channels maps of the NH₃ (3,3) emission, from 170–220 km s⁻¹, for the M4.7–0.8 cloud. This velocity range contains all of the emission in M4.7–0.8, as observed in Figure B2, with the exception of the 230 km s⁻¹ cloud and a portion of the high-velocity component. Each panel was made by integrating the emission over a Δv = 5 km s⁻¹ velocity range, as specified in the bottom-left corner, essentially creating a Moment 0 map for each panel (Equation (1)).

The general, large-scale velocity gradient of 0.25 km s⁻¹ pc⁻¹, discussed in Section 3.2.1, is visible in this figure. Here, most of the higher-velocity emission (195–210 km s⁻¹) is on the eastern side of the panels in the region associated with the Nexus. The lower-velocity emission (170–195 km s⁻¹) is generally observed toward the western side of the panels in the region associated with the Filament. However, faint emission from the hyperfine lines is visible at  ±20.7 km s⁻¹ from the main emission features. In the 175–180 and 180–185 km s⁻¹ panels, the Shell feature is quite prominent and shows an increase in brightness along the leftmost rim. The high-velocity component, discussed in Section 3.2.1, is also visible in this figure and is the brightest emission in the filament in the 200–215 km s⁻¹ panels. The spatial location of the high-velocity component appears to coincide with an observed cavity visible in the 180–185 km s⁻¹ panel.

Zoom In Zoom Out Reset image size