The Giant Molecular Cloud Environments of Infrared Dark Clouds
We study Giant Molecular Cloud (GMC) environments surrounding 10 Infrared Dark Clouds (IRDCs), using emission from the Galactic Ring Survey. We measure physical properties of these IRDCs/GMCs on a range of scales extending to radii, , of 30 pc. By comparing different methods for defining cloud boundaries and for deriving mass surface densities and velocity dispersions, we settle on a preferred “CE,,G” method of “Connected Extraction” in position-velocity space plus Gaussian fitting to opacity-corrected line profiles for velocity dispersion and mass estimation. We examine how cloud definition affects measurements of the magnitude and direction of line-of-sight velocity gradients and velocity dispersions, including associated dependencies on size scale. CE,,G-defined GMCs show velocity dispersion versus size relations , which are consistent with the large-scale gradients being caused by turbulence. However, IRDCs have velocity dispersions that are moderately enhanced above those predicted by this scaling relation. We examine the dynamical state of the clouds finding mean virial parameters for GMCs and 1.6 for IRDCs, broadly consistent with models of magnetized virialized pressure-confined polytropic clouds, but potentially indicating that IRDCs have more disturbed kinematics. CE,,G-defined clouds exhibit a tight correlation of , with for GMCs and 1.3 for IRDCs (c.f., a value of 0.5 expected for a population of virialized clouds). We conclude that while GMCs show evidence for virialization over a range of scales, IRDCs may be moderately super virial. Alternatively, IRDCs could be virialized but have systematically different gas phase abundances, i.e., due to freeze-out, affecting mass estimations.
Subject headings:ISM: clouds - ISM: kinematics and dynamics - Stars: formation
Galactic star formation resides mostly within Giant Molecular Clouds (GMCs), conventionally defined to have masses and observed to extend up to several (e.g., Blitz, 1993; Williams et al., 2000; McKee & Ostriker, 2007). With typical mass surface densities of , GMCs have mean radial sizes of pc, assuming simple circular symmetry. However, GMCs are highly irregular and hierarchical structures. Their dense clumps can spawn stellar clusters and associations, creating the bulk of the Galactic field star population. The efficiency and rate of star formation from these clumps is relatively low, i.e., a few percent per local free-fall time (Zuckerman & Evans, 1974; Krumholz & Tan, 2007). This appears to be mostly because much of the GMC material is stable with respect to gravitational collapse, especially material below a threshold mag (e.g., Lada et al., 2010). Higher total star formation efficiencies, , appear to be possible in the star-forming clumps that form at least moderately bound clusters (Lada & Lada, 2003).
Different theoretical models of the processes that create star-forming clumps within GMCs or prevent overdensities in the bulk of the cloud are actively debated. These processes include the regulation of star formation and stabilization of gas by magnetic fields (McKee, 1989; Mouschovias, 2001) or turbulence (Krumholz & McKee, 2005; Padoan & Nordlund, 2011), and/or the initiation of star formation by discrete triggering events, such as converging atomic flows (e.g., Heitsch et al., 2006), cloud collisions (e.g., Tan, 2000) or stellar feedback (e.g., Samal et al., 2014).
Infrared dark clouds (IRDCs) are likely to be examples of early stage star-forming clumps (e.g., Perault et al., 1996; Egan et al., 1998; Carey et al., 2000; Rathborne et al., 2006; Butler & Tan, 2009; Peretto & Fuller, 2009; Battersby et al., 2010). Thus their study may help us understand the processes that initiate star formation in GMCs. There have been many investigations of the internal properties of IRDCs, including their temperatures (e.g., Pillai et al., 2006; Peretto et al., 2010; Ragan et al., 2011; Chira et al., 2013), mass surface density structure (e.g., Butler & Tan, 2009; Peretto & Fuller, 2009; Ragan et al., 2011; Butler & Tan, 2012; Kainulainen & Tan, 2013; Butler et al., 2014), kinematics (e.g., Henshaw et al., 2013; Jiménez-Serra et al., 2014) and dynamics (e.g., Hernandez & Tan, 2011; Hernandez et al., 2012), CO depletion (Fontani et al., 2006; Hernandez et al., 2011), chemistry (Sanhueza et al., 2013). See Tan et al. (2014) for a review.
However, there have been fewer studies connecting IRDCs to their larger-scale environments, such as the morphology, kinematics and dynamics of their parent clouds. Theories involving production of dense gas in shocks have been supported by detection of large-scale SiO emission along IRDCs (Jiménez-Serra et al., 2010; Nguyen-Lu’o’ng et al., 2013). However, these studies focus only on a few individual clouds and are still confined to a few-parsec scales in and around the filamentary molecular clouds.
Here we study the CO(1-0)-emitting gas in and around 10 well-studied IRDCs, utilizing data from the BU-FCRAO Galactic Ring Survey (GRS; Jackson et al., 2006). We consider a range of scales out to 30 pc projected radius, expected to encompass the potential GMC environment of the IRDC. While our study connects to scales typical of other large-sample GMC studies (e.g., Heyer et al., 2009; Roman-Duval et al., 2009, 2010), by focusing on just 10 regions we are able to investigate their kinematic properties in much greater detail. Other studies done on such a range of scales, from clump to GMCs, have been performed on nearby GMCs, such as Orion A (e.g., Shimajiri et al., 2011), Taurus (e.g., Goldsmith et al., 2008), and Perseus (Ridge et al., 2006; Foster et al., 2009; Kirk et al., 2010). However, these local GMCs do not seem to give rise to the more extreme range of star-forming clumps that is found in IRDCs.
The main focus of this paper is to “bridge the gap” between IRDC and GMC studies. The questions we aim to address include: Are IRDCs typically found within GMCs? Are IRDCs found within specific locations with respect to GMCs? Are IRDCs and their surrounding GMCs virialized, and does their degree of virialization vary as a function of cloud physical scale?
The IRDC/GMC sample is presented in §2. Methods for defining cloud boundaries and estimating masses and kinematic properties are described in §3. Results are presented in §4, including derived physical properties of the clouds (§4.1), the location of IRDCs within their respective GMCs (§4.2), and the kinematic and dynamical analysis of the IRDC and GMCs (§4.3). We conclude in §5.
2. The IRDC Sample and Molecular Line Data
We utilize the data of the GRS survey (Jackson et al., 2006), which covers , and (for ) and (for . The GRS had a spatial resolution of 46″, with 22″ sampling, and a spectral resolution of . The typical rms noise is K (or K with a main beam efficiency of 0.48).
Our selected IRDCs are the 10 clouds from Butler & Tan (2009, hereafter BT09) (see also Butler & Tan, 2012; Kainulainen & Tan, 2013). This sample, a subset of that of Rathborne et al. (2006), was chosen while considering the 8m (IRAC band 4) images from the Spitzer Galactic Legacy Mid-Plane Survey Extraordinaire (GLIMPSE; Benjamin et al., 2003). These IRDCs were selected for being relatively nearby, massive, dark (i.e., relatively high contrast against the surrounding diffuse emission), and surrounded by relatively smooth diffuse emission within the 8m GLIMPSE images. Characteristic sizes and boundaries for each IRDC were taken from Simon et al. (2006a), where ellipses were fitted based on extinction in MSX images. Although these fitted ellipses are not necessarily accurate of IRDC shapes, they provide a convenient measure of the approximate cloud structure. The catalog coordinates and sizes for the 10 IRDCs (including “single component” sub-classifications, labelled “-s”, see below) are listed in Table LABEL:tab1.
|IRDC||111IRDC coordinates and elliptical sizes adopted from Simon et al. (2006a)||111IRDC coordinates and elliptical sizes adopted from Simon et al. (2006a)||PA111IRDC coordinates and elliptical sizes adopted from Simon et al. (2006a)||222IRDC kinematic distances adopted from Rathborne et al. (2006)||333The excitation temperatures adopted from Roman-Duval et al. (2010)|
Kinematic distances were adopted from Rathborne et al. (2006) and Simon et al. (2006b), where the IRDC central velocity was matched morphologically between the mid-infrared (MIR) extinction estimated from MSX and the emission from the GRS. These distances were estimated assuming the rotation curve of Clemens (1985). We assume uncertainties of 20%, given the size of streaming motions of of clouds with trigonometric parallax measurements (e.g., Brunthaler et al., 2009; Reid et al., 2014).
3. Methods: Derivations of Molecular Cloud Properties
3.1. Definition of Cloud Boundaries
We first determined the cloud center-of-mass in position-velocity space. To do this, we co-added the spectra within an elliptical boundary defined by twice the and values, using the full GRS velocity range. Then, a velocity interval of was considered, centered on the peak of the emission profile. The cloud’s center-of-mass was evaluated for both optically thin and opacity corrected gas, ) and ), respectively. We then use this as the reference point around which to search for GMC-scale gas, out to a projected radius of 30 pc and over a new velocity range of . This range is wide enough to sample the velocities of GMC gas, even those that are potentially undergoing collisional interactions at velocities set by Galactic shear at the tidal radii of the clouds (Tan, 2000; Tasker & Tan, 2009). The GLIMPSE m and GRS integrated intensity maps centered on these locations are presented in Figures 1 and 2.
We examined emission within the km/s velocity range and within apertures of varying radii from the cloud center-of-mass coordinate. To explore how boundary definitions affect estimated physical properties, we used two different methods to select line emission associated with the cloud. First, “Simple Extraction (SE)” selected all the emission within radii , and 30 pc and km/s. Second, “Connected Extraction (CE)” defined a cloud as a connected structure in ---space with all cloud voxels required to be above a given threshold intensity: we defined a given voxel as “molecular cloud gas” if its (1-0) line intensity K (i.e., the GRS noise level). The CE search was also limited to within a 30 pc radius and of the IRDC center-of-mass. Position-velocity maps for each cloud, defined by both SE and CE, are shown in Figures 1 and 2.
Clouds defined by CE do not have simple radial sizes. Therefore, we estimated three circular boundaries, centered on the extracted cloud’s center-of-mass: (1) mass-weighted radius, , defined as mean projected radial distance of cloud mass from the center of mass; (2) areal radius, , defined by the total projected area , where is the total number of pixels subtended by the cloud and is the area of one image pixel; (3) half-mass radius, , defined as the radius from cloud center that contains half of the total mass.
For the IRDCs, we also selected “cloud” material via SE and CE using the spatial coordinates and elliptical boundaries from Simon et al. (2006a), along with the velocity intervals.
3.2. Column Densities and Masses from emission
We estimated the column density of each molecular cloud voxel, from their line emission assuming a partition function with a thermal distribution described by an excitation temperature via:
where is the partition function, is the Einstein coefficient, cm, and are the statistical weights of the lower and upper levels and is the optical depth of the line at frequency , i.e., at velocity . Each GRS voxel has a velocity width of . For linear molecules, the partition function is with where is the rotational quantum number and is the rotational constant. For (1-0) we have K.
Many studies of the physical properties of GMCs and IRDCs have accounted for line optical depth when estimating their physical properties (e.g., Heyer et al., 2009; Roman-Duval et al., 2010; Hernandez & Tan, 2011; Hernandez et al., 2011, 2012). In Hernandez & Tan (2011), we showed that optical depth correction factors can increase the column density by a factor of in the densest, sub-parsec scale clumps of IRDCs. However, for the more diffuse GMCs the optical depth correction factors are expected to be smaller. To gauge the importance of this effect, we carry out column density estimates for both the optically thin assumption and accounting for opacity corrections.
The optical depth is evaluated via
where is the brightness temperature at frequency , , and K is the background temperature. is derived from the antenna temperature, , via , where is the main beam efficiency ( for the GRS) and is the beam dilution factor of the emitting gas, which we assume to be unity due to the large scale extent of GMCs. Smaller scale structures are undoubtedly present, e.g., as revealed in the BT09 MIR extinction maps, but to gauge the effects of these on the CO emission requires higher resolution molecular line maps of the clouds. For , Equation (2) can be simplified to express by: . With this simplification, for an observed voxel and an assumed , the optically thin column density per voxel is given by combining Equations (1) and (2):
Early studies of IRDCs estimated typical gas kinetic temperatures K (Carey et al., 1998, 2000; Pillai et al., 2006). IRDC F was estimated to have a temperature of 19 K based on (1,1) and (2,2) VLA observations (Devine, 2009). However, as discussed below, CO excitation temperatures appear to be significantly lower.
In our previous study of IRDC H, we used IRAM 30m observations of (2-1) and (1-0) emission from around the IRDC filament to estimate a mean K (Hernandez et al., 2011, hereafter H11). Here, for a uniform analysis of the 10 IRDCs, we now use estimates from Roman-Duval et al. (2010, hereafter RD10). In their study of 580 molecular clouds, brightness temperatures from (1-0) emission line data (Univ. of Massachusetts-Stony Brook (UMSB) Galactic Plane Survey), were used to derive proxy excitation temperatures, assuming emission was optically thick and that and excitation temperatures are equal. Ultimately, RD10 cited a mean excitation temperature for all their molecular clouds based on all cloud voxels above . The RD10 clouds were extracted from the GRS data using a modified version of CLUMPFIND (Williams et al., 1994), which allowed for varying thresholds (i.e., contour increment and minimum brightness, see Rathborne et al. (2009) for details).
Eight of our IRDCs overlapped with at least one of their molecular clouds, and in these cases we adopted from the RD10 value from the overlapping cloud(s). For the remaining two IRDCs (D and E) we set K, similar to the mean values of 7.2 K of H11 and 6.32 K of RD10. Our adopted values of are listed in Table LABEL:tab1. For our column density estimates that assume optically thin CO emission, these values are assumed to be constant throughout the cloud. These temperatures are slightly lower (by a few K) than those used in previous studies (e.g., Simon et al. 2001, 2006b who assumed a fixed value of 10 K). However, the results from Heyer et al. (2009) indicate that CO gas throughout GMCs is mostly sub-thermally excited. Note that, in this optically thin limit, varying from 5 K to 10 K would change the derived column density of the cloud by only .
For the opacity-corrected case, the use of a single mean excitation temperature of relatively low value ( K) can lead to non-physical results in certain regions of the cloud. Equation 2 implies . Thus, for a given observed , will become undefined if (i.e., when ). For example, a voxel with an observed brightness temperature of K will have a numerically undefined opacity at K. H11 showed modest temperature variations were present in IRDC H, with a peak temperature of K within the densest clumps and K in the more diffuse gas within the filament envelope (see H11, Fig. 1). Hence, we expect that the voxels containing the largest brightness temperatures have excitation temperatures which are a few K larger than the constant values adopted from RD10.
To estimate the opacity-corrected column density, we first apply our adopted to estimate in each voxel. Then, for each voxel with an undefined , we specified a new excitation temperature of K, given the voxel’s observed . This revised excitation temperature allows us to estimate a real solution for throughout the cloud. After considering a range of possible temperature offsets from , we estimate that the uncertainty in in an individual voxel by this method is at a level of .
Ideally, excitation temperatures would be estimated locally from CO observations of the clouds, assuming that the CO line is optically thick. However, there are currently no other CO surveys that match both the resolution and spatial extent of the GRS. The widely used Columbia-CfA CO survey covers the whole Galactic plane, but with a low angular resolution of 8′ (Dame et al., 2001). The RD10 temperatures measurements are based on the UMSB survey, which has a 44″ spatial resolution with 3′ sampling.
For the clouds defined by SE extraction and with pc, we find that on average of the cloud voxels require higher temperatures than those listed in Table LABEL:tab1. Cloud D has the highest percentage, 1.5%, and for these voxels peaks at 16.4 K with a mean K. For the clouds defined by CE extraction and , we find that on average of voxels require higher excitation temperatures. Here, Cloud J has the highest percentage, 27%, with a peak of 25.0 K and a mean K.
where is the cloud distance, and are the angular sizes of the GRS pixels, and assuming a mass per H nucleus of . The total -derived cloud mass, , is simply the total mass of all cloud voxels within radius and velocity range . For clouds defined by SE, any pixels with total integrated intensities below were omitted from further analysis.
We estimate an uncertainty of in due to its intrinsic variation, which in addition to uncertainties due to intrinsic abundance variations, leads to uncertainty in the mass surface density (). We thus estimate 50% random errors in , after accounting for these uncertainties in and the cloud kinematic distance estimates (assumed to be ). However, we also anticipate that there could be global systematic uncertainties in (of the whole cloud sample) of up to a factor of 2, given the uncertainties in overall absolute abundance.
3.3. Cloud Kinematics
We used co-added column density velocity distributions (e.g., Fig. 3: Right) to determine the mean velocity, of each extracted cloud. The velocity dispersion was estimated using two standard methods: 1) the rms 1D velocity dispersion, ; 2) the width of a fitted Gaussian profile, . We estimate an mean uncertainty in of . Additionally, we used these Gaussian fitted profiles to estimate a Gaussian profile mass, , of each cloud. To visualize how and vary throughout the cloud, we show mass-weighted first and second moment maps of each GMC (e.g., Fig. 3).
Using the first moment maps, we derive the velocity gradients in each spatial direction. For example, the longitudinal velocity gradient, , was derived by first estimating the mean velocity at each longitudinal position, via a mass-weighted sum along the perpendicular () direction, then finding the best (mass-weighted) linear fit to these velocities. This method was repeated for . The magnitude of the total linear velocity gradient across the cloud, , and its position angle direction, , were then calculated. Note, we choose to use mass-weighted velocity gradients to prevent the results being unduly affected by tenuous wisps of cloud material. Also, in the context of interpreting velocity gradients as due to rotation and thus measuring rotational energies of the cloud (below), this mass-weighted gradient is the appropriate one to use.
Many studies of the dynamics of molecular clouds have interpreted total GMC linear velocity gradients as due to solid body rotation (e.g., Phillips, 1999; Rosolowsky et al., 2003; Imara & Blitz, 2011). However, the possibility remains that the identified single “cloud” actually consists of spatially independent structures. For a cloud undergoing solid body rotation, the line of sight velocity gradient in the plane of the sky, , is equal to the projected angular velocity, i.e., . The true angular velocity is , where is the angle between the rotation axis and the line of sight.
The position angle of the projected rotation axis of the cloud contains information that may constrain theories of GMC formation and evolution. For example, if a GMC forms rapidly from atomic gas in the Galactic plane, then the GMC rotation is expected to be prograde with respect to Galactic rotation (e.g., Tasker & Tan, 2009). If strong gravitational encounters and collisions are frequent between gravitationally bound GMCs (Tan 2000), then a more random set of orientations of the positional angles of projected rotation axes are expected, including both pro- and retrograde rotating clouds. For clouds observed in the Galactic plane, represents retrograde rotation and and represent prograde rotation.
We then also estimated the projected moment of inertia, , of each cloud using the sky projected rotation axis, defined by and the cloud center-of-mass coordinate, where is the shortest distance to the rotation axis. This allows estimation of the projected rotational energy of the cloud, .
|(pc)||( )||()||()||(pc)||()||(pc)||()||(km/s/pc)||()||( pc)||(erg)|
|(pc)||( )||()||()||(pc)||()||(pc)||()||(km/s/pc)||()||( pc)||(erg)|
|( )||()||()||(pc)||()||(km/s/pc)||()||( pc)||(erg)|
4.1. Molecular Cloud Physical Properties
Table LABEL:tab2 presents the derived physical properties for all 10 clouds. The global properties of each cloud were evaluated for data extracted out to seven different radii, shown on different lines in the table: 5, 10, 20, and 30 pc using SE; and out to mass weighted (), areal (), and half-mass () radii using CE. Each cloud definition case is noted in column 1. The center-of-mass coordinates of each cloud are listed in columns 2 and 3, evaluated for the optically thin and opacity-corrected cases. We note that the CE cloud definitions with , and share the same center. Mean velocities are reported in column 4. Columns 5 and 6 list the separations between the GMC centers and the IRDC centers in plane of sky distance and velocity offset, respectively, again for both optically thin and opacity-corrected methods. Column 7 lists the cloud radii (with , and cases having two values for the optical thin and opacity-corrected cases). Column 8 lists the three cloud mass estimates: optically thin mass (), opacity-corrected mass (), and the mass estimated from a gaussian fit of the co-added opacity corrected column density spectra (). Column 9 lists the three velocity dispersions measured: the standard deviations of the cloud’s optically thin and opacity corrected column density spectra (, ) and the dispersion estimated from the gaussian fitted opacity corrected column density spectra (). Column 10 presents the two velocity gradient magnitudes, one estimated from the optically thin column density weighted first-order moment map () and one estimated from the opacity-corrected column density weighted first-order moment map (). Column 11 presents the position angles of the projected cloud rotation axes estimated from each cloud’s angular momentum for both optically thin and opacity corrected masses ( and ). The projected moments of inertia estimated using the optically thin mass () and opacity corrected mass () are presented in column 12. Column 13 shows projected rotational energies, and , and column 14 shows the ratios of these to the gravitational energy, i.e., and . Finally, column 15 shows log of the virial parameters of the clouds using the three measures of mass, i.e., , , .
A visual overview of each GMC-scale region is given in Figure 3 and Figures 11-19. The left side of the figures display a series of images of the molecular gas out to a radius of 30 pc. The first row displays the GLIMPSE image and the mass surface density map, , generated using the velocity interval given solely by the IRDC. The IRDC elliptical boundary is also shown (Simon et al., 2006a). The subsequent rows display the maps assuming our defined velocity interval of , the column-density weighted mean velocity (first moment), the linear momentum map derived from each pixel’s column density and velocity with respect to the cloud center of mass, and the velocity dispersion map (second moment) for the SE cloud (Left column) and CE cloud (Right column). The right side of the figures display the optically thin and opacity-corrected co-added column density velocity spectra at various extraction radii for the SE and CE clouds and the SE IRDC.
Four of the clouds (D, F, G, and H) were found to contain multiple velocity components, based on the co-added column density spectra of their SE 60 pc diameter GMCs (e.g., Cloud D; Fig. 13). We used a series of integrated intensity maps to isolate the velocity range of the molecular gas associated with the IRDC (Fig. 4). In general, these velocity ranges span (Cloud D: ; Cloud F: ; Cloud G: ; Cloud H: ). To investigate how the cloud’s derived physical properties were affected by isolating only these velocity ranges, we include these four cases as separate entries in Table LABEL:tab2 (noted with “-s”). For all subsequent analyses, the “-s” values for these four clouds are used in place of those derived using the full velocity range of .
4.2. IRDC and GMC Morphologies
4.2.1 Description of Individual IRDCs/GMCs
Cloud A: We estimate a cloud areal radius of pc and a mass of . The GRS GMC study by RD10 identified five molecular clouds within the extent of Cloud A, with a total mass of (note, we scaled all masses of RD10 by a factor 0.49 to convert to our assumed /H abundance ratio; see §3.2). IRDC A is located about 20 pc north of its parent GMC’s highest clump, which appears bright at and is thus likely to be star-forming. The WISE H ii region catalog (Anderson et al., 2014) lists the W39 H ii region to be within of this clump’s center, and with a similar LSR velocity (). Another H ii region lies in a smaller dense clump at , with velocity and distance consistent with those of Cloud A. The isolated, simple emission profiles, along with the spatial connection between the IRDC and the H ii regions, demonstrated through CE, indicates that GMC A contains clumps at a variety of stages of star formation activity.
Cloud B: We estimate a radius of pc and a mass of . RD10 estimated a mass of and a smaller projected areal radius of 6.3 pc. The IRDC is located within the densest region of the GMC, which extends pc to the south-east of the IRDC and accounts for most of the cloud’s mass. Spatially, Cloud B is well defined and completely isolated at an extraction radius of 20 pc in CE. Although the cloud is isolated in velocity space, the column density profile indicates a low level of emission at neighboring velocities, both lower and higher. Although the GLIMPSE m emission map indicates a large bright emission structure to the west of the cloud, this feature is not at velocities consistent with Cloud B ().
Cloud C: We estimate a GMC radius of pc and a mass of . RD10 derived a radius of 20.9 pc and a mass of . IRDC C is located within the dense central region of its parent GMC. The integrated intensity maps indicate dense and clumpy structure throughout the cloud (Fig. 12). The velocity distribution shows a broad centrally-peaked profile, but with potential substructure. Additionally, there is a low level of emission at higher velocities. Butler et al. (2014) studied the IRDC C using near- and mid-infrared extinction mapping to probe the high-dynamic range of its mass surface density (down to mag), finding an IRDC mass of . This is one of the most massive, high- IRDCs known, with the potential to form a massive star cluster. Using the same IRDC elliptical boundary from Simon et al. (2006a), we estimate a mass of . Our slightly lower mass estimate may be the result of abundance uncertainties and/or CO depletion.
Cloud D: For the GMC at , we estimate a radius of pc and a mass of . RD10 identified two clouds within the Cloud D-s region with a total mass of . The IRDC is located in the dense central region of the GMC. However, the spectra show four cloud components within . Although the emission corresponding to the IRDC is evident within a radius of 5 pc, a lower-velocity gas component begins to dominate on scales out to about 20 pc. The position-velocity map (Fig. 1) shows there is a significant amount of emission at lower velocities, -, that extends largely over the northern region of the cloud. In the same velocity interval of the IRDC, , a substantial molecular gas clump contains the largest integrated peak intensity ( K ) at the western edge of the 30 pc extraction radius. Figure 4 suggests that this emission is only tenuously connected to the emission associated with the main GMC.
Cloud E: We estimate a GMC radius of pc and a mass of . No clouds within the Cloud E spatial and velocity ranges were listed in the RD10 catalog. The integrated emission map indicates that the GMC is highly substructured. The IRDC is contained within a small dense clump central to its parent GMC, but more massive clumps are seen on scales beyond 5 pc. CE finds an extended region of clumpy gas in the cloud’s south-west region, beginning at 10 pc and extending beyond the 30 pc boundary.
Cloud F: For the cloud in the velocity range , we estimate a radius of pc and a mass of , assuming a distance of 3.7 kpc (Rathborne et al., 2006). Note, Kurayama et al. (2011) estimated a parallax distance kpc, however this result has been called into question by Foster et al. (2012). RD10 found two GMCs associated with Cloud F: one cloud matches the location and velocity interval of Cloud F’s western massive clump, having a mass of and radius of 12.3 pc. The other molecular cloud is associated with the smaller eastern clump, for which they estimate a mass of . Our analysis recovers this larger total mass, , if we include all the emission within . The larger molecular cloud complex surrounding Cloud F was recently cataloged within the Giant Molecular Filament (GMF) study of Ragan et al. (2014). They found that their filament GMF38.1-32.4 is actually the projection of two overlapping filamentary structures. Cloud F is within the further filament (3.0-3.7 kpc), which has a total length of 232 pc. The GMF contains multiple infrared dark regions (identified in both GLIMPSE m and HiGAL m images) and several dense gas clumps (identified by the Wienen et al. (2012) and Bolocam Galactic Plane surveys), demonstrating that overall it is in an early phase of star formation. However, in Cloud F alone, there are two H ii regions within the two densest parts of the cloud with consistent LSR velocities and distances, at and (Anderson et al., 2014). While IRDC F, which is also highly filamentary, contains some of the highest column densities in the GMC, the highest clump is actually another feature extending west of the IRDC between 5 and 25 pc (Fig. 15). The velocity distribution of the GMC for material out to 30 pc indicates that there are four cloud components within our defined velocity interval