AN INFRARED THROUGH RADIO STUDY OF THE PROPERTIES AND EVOLUTION OF IRDC CLUMPS

In this paper, we investigate a sample of 17 IRDC clumps embedded within 8 IRDCs. These were selected from the MSX dark cloud catalog (Simon et al. 2006a), and from the previous work of Rathborne et al. (2006), who selected the 38 highest contrast IRDCs from the dark cloud catalog with known kinematic distances. These 38 IRDCs, therefore, are among the darkest and densest in the catalog by Simon et al. (2006a). From this subset, and based upon the work by Chambers et al. (2009), we selected mostly active clumps, with a few quiescent clumps for comparison. Our sample consists of nine active, four intermediate, and four quiescent clumps. We mostly selected clumps with signs of active star formation, motivated by our UC H ii region survey, as this would maximize our detection rate. We included a few quiescent clumps in our sample for comparison, although no UC H ii regions were expected in the quiescent clumps. Our sample of IRDC clumps is a subset of the darkest, most active IRDC clumps in the First Galactic quadrant (0 ⩽ ℓ ⩽ 90°). In a sample of 106 IRDC clumps, Chambers et al. (2009) find that 65% are quiescent and 35% are active clumps. We quantify the star-forming activity by four criteria: (1) an embedded 24 μm point source, (2) "green fuzzies" (Chambers et al. 2009), an extended 4.5 μm enhancement believed to be caused by shocked H₂, thus indicative of outflows (also known as EGOs; Cyganowski et al. 2008), (3) H₂O or CH₃OH maser emission, and (4) UC H ii regions. We assign the designation "active" to clumps which have three or four of these signs of star formation, which means that each "active" clump has either an H ii region or at least two outflow tracers and a 24 μm point source and is actively forming stars. The designation "intermediate" is for clumps which exhibit one or two signs of active star formation. These clumps, therefore, may have only shock/outflow signatures or a 24 μm point source. We reserve the designation "quiescent" for clumps with no signs of active star formation.

Selection of the clump positions from the sample discussed above required simply that there be an IR-dark feature, which can be near but not coincident with an IR-bright region, which has corresponding millimeter emission. The exact clump positions were selected by eye based on the above criteria and the overlap with existing data. While the exact clump positions were not selected by stringent thresholds, visual inspection of the images verifies that the sample accurately represents the IRDCs. The clump positions and the closest millimeter clump, as identified by Rathborne et al. (2006), are given in Table 1.

For this study, we examine a sample of 17 IRDC clumps within 8 IRDCs. We utilize existing and new data to determine the physical properties and evolutionary stages of these clumps. A summary of the data used is found in Table 2.

Table 2. Summary of Observations

Data	λ	Beam FWHM	1σ Sensitivity
GLIMPSE	3.6 μm	17	0.3 MJy/Sra
GLIMPSE	4.5 μm	17	0.3 MJy/Sr
GLIMPSE	5.8 μm	19	0.7 MJy/Sr
GLIMPSE	8.0 μm	19	0.6 MJy/Sr
MIPSGAL	24 μm	6''	0.67 mJy
HHT	1.12 mmb	28''	0.04 K
HHT	1.07 mmb	27''	0.04 K
GRS	2.86 mmb	46''	0.27 K
VLA	3.6 cm	28 × 24	0.05 mJy beam−1
IRAM	1.2 mm	11''	10 mJy beam−1
BPGS	1.1 mm	33''	30 mJy beam−1

Notes. ^aGLIMPSE raw 1σ surface brightness sensitivities from Reach et al. (2006). ^bHCO⁺ J = 3–2, N₂H⁺ J = 3–2, and ¹³CO J = 1–0, respectively. For HCO⁺ and N₂H⁺, 0.04 K is the 1σ sensitivity per 1.1 km s⁻¹ channel and for ¹³CO, 0.27 K is the 1σ sensitivity per 0.21 km s⁻¹ channel.

Download table as:  ASCIITypeset image

2.2.1. Mid-infrared Data

2.2.2. Millimeter Continuum Dust Emission

2.2.3. ¹³CO J = 1–0 Molecular Gas Tracer

2.2.4. VLA Radio Continuum Data

2.2.5. HCO⁺ and N₂H⁺ Dense Molecular Gas Tracers

CO is the second most abundant molecule in interstellar space. It is found nearly everywhere, including IRDCs. In these dense, cold environments, however, even the less abundant ¹³CO can become optically thick and can deplete onto dust grains (e.g., Tafalla et al. 2002). HCO⁺ and N₂H⁺ are less abundant than ¹³CO and therefore become optically thick at higher columns of H₂. The critical density is also much higher, 3.5 × 10⁶ cm⁻³ and 3.0 × 10⁶ cm⁻³ for the J = 3–2 transitions of HCO⁺ and N₂H⁺, respectively, compared with 1.9 × 10³ cm⁻³ for ¹³CO J = 1–0. Therefore, there is typically only one HCO⁺ or N₂H⁺ source along a given line of sight, which can then be confidently associated with the dense IRDC, as opposed to CO which is often seen at many velocities along a given line of sight through the Galactic plane. For this reason, we use the HCO⁺ velocity to determine the kinematic distance to each IRDC. A kinematic distance had already been determined to each of these objects using ¹³CO morphology matching by Simon et al. (2006b), however, we find that in the confused inner Galaxy, morphology matching can misassign distances. Two clouds of our sample of eight (G022.35+00.41 and G024.60+00.08) are spatially adjacent clumps along the line of sight at very different distances, as opposed to a contiguous object as found by Simon et al. (2006b).

Distances were estimated kinematically using the HCO⁺ clump velocities and the rotation curve of Reid et al. (2009), determined with trigonometric parallaxes and proper motions of masers using the Very Long Baseline Array and Japanese VERA project. In the First Galactic quadrant, a radial velocity can correspond to two possible distances, a problem known as the "Kinematic Distance Ambiguity." Since IRDCs are perceived as dark extinction features against the Galactic background, it is assumed that all IRDCs are located at the near kinematic distance. If we assume an average velocity scatter at any given distance of about 5% (about 5–10 km s⁻¹), this translates to an approximate distance uncertainty of 20%. The distances, line widths, and observed properties of the clumps are listed in Table 3.

The mid-IR extinction by IRDCs provides a powerful way to obtain an independent measure of their column density. We restrict our analysis of extinction to the 8 μm band of GLIMPSE, as this band is dominated by emission from the diffuse interstellar medium, rather than stellar sources. We apply the extinction mapping method put forth by Butler & Tan (2009). In brief, this method estimates the Galactic background and foreground emission, finds the optical depth required to produce the observed extinction feature, assumes a dust opacity, and finally extracts the gas mass surface density. Our application of this method is described below.

To accurately map the extinction, we need to estimate the intensity of radiation behind the cloud of interest, I_{ν 0}, and the radiation in front of the cloud, I_{ν 1}. Assuming no emission from the cloud itself at these wavelengths, we have

$$\begin{equation} I_{\nu 1} = I_{\nu 0} e^{-\tau _{\nu }}, \end{equation} \tag{ 2 }$$

where τ_ν = κ_ν Σ, κ_ν is the dust opacity, and Σ is the surface mass density. We adopt a value of κ_8 μm = 11.7 cm² g⁻¹, following the analysis of Butler & Tan (2009) (they adopt a value of 7.5 cm² g⁻¹ as they use a gas-to-dust ratio of 156, rather than the value of 100 used in this paper). This value is closest to the model of Ossenkopf & Henning (1994) of thin ice mantles that have undergone coagulation for 10⁵ yr at a density of $n_{\rm H_{2}}$ ∼ 10⁶ cm⁻³, which is a reasonable model for the cold, dense environment of an IRDC. This expression assumes that the dust opacity remains constant along the line of sight, which may not be true. However, most of the extinction arises from the coldest, densest parts of the cloud. In the less dense regions, the opacity may vary by up to a factor of two, but we assume that the majority of the extinction is caused by the coldest, densest regions of the cloud, which are well characterized by the adopted value of κ_8 μm.

Before proceeding with the background determination, we must first estimate the fraction of emission from dust in the foreground. We assume that the Galactic distribution of hot dust (in the Galactic midplane) is given by the Galactic surface density of OB associations (McKee & Williams 1997)

$$\begin{equation} \Sigma _{{\rm OB}} \propto e^{-R/{\rm H}_{R}} \end{equation} \tag{ 3 }$$

where R is the Galactocentric radius (in kpc) and H_R = 3.5 kpc is the Galactic radial scale length. We then integrate the column of dust from the Sun (Reid et al. 2009, at R₀ = 8.4 kpc from the Galactic center) to the cloud, and the column of dust along the same line of sight, from the Sun out to a Galactocentric radius of 16 kpc (beyond which there is a negligible contribution to the dust emission). The ratio of the column to the cloud over the column out to 16 kpc is called the "Foreground Intensity Ratio," f_fore; the fraction of emission along the line of sight that is from the foreground. Thus, the true radiation behind the cloud, I_{ν 0}, determined from that observed, I_{ν 0,obs}, is given by

$$\begin{equation} I_{\nu 0} = (1-f_{\rm fore}) I_{\nu 0, \rm obs}, \end{equation} \tag{ 4 }$$

and the true radiation in front of the cloud, I_{ν 1}, determined from that observed, I_{ν 1,obs} is given by

$$\begin{equation} I_{\nu 1} = I_{\nu 1, \rm obs}-f_{\rm fore} I_{\nu 0, \rm obs}. \end{equation} \tag{ 5 }$$

Butler & Tan (2009) note that f_fore is uncertain because of the inevitable small-scale variations in the Galactic background and the fact that the regions surrounding the IRDC are likely to be embedded in the same GMC that hosts the IRDC. This leads to a higher extinction of the integrated Galactic background, creating a tendency to underestimate f_fore and thus Σ.

An upper limit for f_fore is provided by the minimum flux at 8 μm divided by the average background. This value would be f_fore if the cloud were totally opaque at 8 μm. This upper limit is about 2 times higher in most cases, except in G028.37+00.07, where it is about the same. The remainder of the IRDCs may or may not be optically thick at 8 μm, however, the good correlation (see Figure 2) with BGPS 1.1 mm masses (which is almost certainly optically thin) is an indication that many of the IRDCs are still optically thin at 8 μm.

Zoom In Zoom Out Reset image size

For the purpose of estimating the Galactic background, we employ the more accurate small-scale median filter approach from Butler & Tan (2009). This approach is designed to capture the small-scale variations in the background, without altering the estimate by the darkness of the IRDC itself. Thus, we first exclude an ellipse approximating the size and shape of the IRDC from the filtering process (Figure 1, top). To smooth the data, we first remove the high-end tail of the distribution (all points greater than twice the mode) which is mostly due to stars. We then perform a circular spatial median filter at each point, i.e., we compute the median of all the data in a circle around a given point, and that becomes the smoothed data value at that point. A reasonable size for the filter was empirically determined to be a radius of one arcminute. This smoothed map is our estimate of the diffuse Galactic background. We then estimate the "background" inside the ellipse by taking the average of the smoothed data values from outside the ellipse, weighted by the inverse square of their distance from the point. This is depicted in the middle frame of Figure 1. This image is now our estimate of I_{ν 0,obs}, and our original image is I_{ν 1,obs}. Using Equation (6) we then create a surface density map, shown in Figure 1 in the bottom panel.

$$\begin{equation} \Sigma = -\frac{1}{\kappa _{\nu }} {\rm ln}\bigg [\frac{I_{\nu 1, \rm obs}-f_{\rm fore} I_{\nu 0, \rm obs}}{(1-f_{\rm fore}) I_{\nu 0, \rm obs}} \bigg ]. \end{equation} \tag{ 6 }$$

There is an additional correction that must be taken into account for quantities derived from extended sources in the IRAC data. Scattering from extended sources within the IRAC array systematically increases the measured surface brightness. For a uniformly distributed extended source, this scattering increases the measured surface brightness by about ∼30% at 8 μm (S. Carey 2010, private communication and http://ssc.spitzer.caltech.edu/irac/iracinstrumenthandbook/33/). This means that the measured background (I_{ν 0,obs}) is systematically brighter than the true background, which we will call T₀ for simplicity. Where s = 0.3, the scattering correction, the true background is given by T₀ = (1 + s)I_{ν 0,obs}. The measured foreground (I_{ν 1,obs}) is also contaminated by scattered light from the background, so the true foreground (T₁) is given by T₁ = I_{ν 1,obs} + sI_{ν 0,obs}. Plugging these values into Equation (6), simplifying, and writing the surface density in terms of our measured quantities, our final expression for the surface density is

$$\begin{equation} \Sigma = -\frac{1}{\kappa _{\nu }} \rm {ln}\bigg [\frac{(\rm {s} + 1)I_{\nu 1, \rm obs}-(\rm {s} + \rm {f})f_{\rm fore} I_{\nu 0, \rm obs}}{(1-f_{\rm fore})I_{\nu 0, \rm obs}} \bigg ]. \end{equation} \tag{ 7 }$$

Using distances determined from the HCO⁺ velocities and the Reid et al. (2009) rotation curve, we are then able to sum the surface density over the BGPS beam-sized apertures to obtain the mass (see Table 4 and Figure 2) in the clump. The extinction masses for the clumps range from 60 to 620 M_☉ and from 1200 to 26,020 M_☉ for the whole IRDC.

Obtaining mass estimates of IRDCs with molecular lines is often problematic due to high column densities and molecular freezeout onto dust grains. The cold dust emission, however, as observed in the millimeter, provides more reliable mass estimates as it is optically thin and does not deplete. We use both the 1.2 mm and 1.1 mm dust continuum emission from MAMBO and the BGPS, respectively, to achieve isothermal mass estimates using the expression

$$\begin{equation} M = \frac{S_{\nu } D^{2}}{\kappa _{\nu }B_{\nu }(T)}, \end{equation} \tag{ 8 }$$

where S_ν is the source flux density, D is the distance, κ_ν is the dust opacity, and B_ν(T) is the Planck function at dust temperature, T. We adopt a value of κ_1.1 mm = 0.0114 cm² g⁻¹ from Enoch et al. (2006) in which we have assumed a gas-to-dust ratio of 100. This opacity is consistent with the Ossenkopf & Henning (1994) model used for the 8 μm opacity in the extinction masses. The mass equation reduces to

$$\begin{equation} M_{1.1\, \rm mm} = 14.32 (e^{13.0/T} - 1) \Bigg (\frac{S_{\nu }}{1\, \rm Jy}\Bigg) \Bigg (\frac{D}{1\, \rm kpc}\Bigg)^{2} \,M_{\odot } \end{equation} \tag{ 9 }$$

$$\begin{equation} M_{1.2 \,\rm mm} = 20.82 (e^{12.0/T} - 1) \Bigg (\frac{S_{\nu }}{1\, \rm Jy}\Bigg) \Bigg (\frac{D}{1\, \rm kpc}\Bigg)^{2}\, M_{\odot }. \end{equation} \tag{ 10 }$$

We assume a dust temperature T = 15 K. We can also find the column density of H₂, as given by

$$\begin{equation} \hspace{-4pc}N(\mbox{H}_{2}) = \frac{S_{\nu } }{\Omega _{B}\kappa _{\nu }B_{\nu }(T) \mu _{{\rm H}_{2}}m_{{\rm H}}} \end{equation} \tag{ 11 }$$

$$\begin{equation} N(\mbox{H}_{2})_{1.1\, \rm mm}= 2.20\times 10^{22}(e^{13.0/T}-1)S_{\nu }\, \rm cm^{-2} \end{equation} \tag{ 12 }$$

$$\begin{equation} N(\mbox{H}_{2})_{1.2 \,\rm mm}= 3.20\times 10^{22}(e^{12.0/T}-1)S_{\nu }\, \rm cm^{-2} \end{equation} \tag{ 13 }$$

where Ω_B is the beam size, $\mu _{{\rm H}_{2}}$ is the mean molecular weight for which we adopt a value of $\mu _{{\rm H}_{2}}$ = 2.8 (Kauffmann et al. 2008). The dust continuum masses and column densities are listed in Table 4. The total BGPS 1.1 mm cloud masses range from 1200 to 13,000 M_☉ while the clump masses range from 80 to 1100 M_☉.

The molecular line width of these objects provides an independent measure of mass. The virial mass is based on the assumption that the object is gravitationally bound and in virial equilibrium. Therefore, the virial mass estimate, in combination with other mass estimates, may allow discussion of whether or not the clump is bound and virialized. For a density profile of ρ(r) = r^−α, the virial mass is given by

$$\begin{equation} M_{\rm vir} = 3 \Bigg [\frac{5-2\alpha }{3-\alpha }\Bigg ]\frac{R\sigma ^{2}}{G}, \end{equation} \tag{ 14 }$$

where σ is the line-of-sight velocity dispersion, R is the virial radius, and G is the gravitational constant. We adopt a spectral index of α = 1.8, which was found to be the mean in a sample of 31 massive star-forming regions by Mueller et al. (2002) with a standard deviation of 0.4. However, this spectral index was derived from a sample of "active" clumps, so this is an added source of uncertainty for quiescent clumps. We do not include a correction for ellipticity, as these clumps are compact enough to be well approximated by spheres. This expression reduces to

$$\begin{equation} M_{\rm vir} = 147 \Bigg (\frac{R}{1 \,\rm pc}\Bigg) \Bigg (\frac{\Delta v_{\rm fwhm}}{1\, \rm km\, s^{-1}}\Bigg)^{2}\, M_{\odot }, \end{equation} \tag{ 15 }$$

where R is the virial radius and Δv_fwhm is the line width at half maximum. In order to determine the best estimate for the virial radius, a two-dimensional circular Gaussian was fit to each BGPS clump. The Gaussian fit was then deconvolved from the BGPS beam. The effective radius (see Equation (1)) of the deconvolved Gaussian fit is the virial radius, given in Table 4. In this study, we assign a virial "radius" to a Gaussian beam which is likely convolved with a power-law distribution of cores within the beam. The meaning of the virial "radius" in this case is tricky as the deconvolved source size can vary by up to a factor of 2 depending on the slope of the power-law distribution of sources within the beam (Shirley et al. 2003, Figure 16). However, this method is consistent between the clumps in our sample and is a reasonable proxy for the size scale of the diffuse emission.

We use the FWHM of the HCO⁺ line for the virial masses presented here. Using the N₂H⁺ line width instead does not change the results qualitatively, as the two tracers provide consistent line widths across different activity levels. N₂H⁺ has been shown to be well correlated with submillimeter dust column in starless cores (Tafalla et al. 2002; Wyrowski et al. 2000), meaning that these virial masses are most likely based on the same material which is producing the dust continuum. The virial masses of the clumps were found to range from 330 to 4400 M_☉.

¹³CO is commonly used to estimate the column density and mass of molecular clouds in the Galaxy. However, due to depletion onto dust grains (Tafalla et al. 2002) and optical depth effects in cold, dense environments, ¹³CO is not the most reliable tracer of column in IRDCs. While there is good morphological agreement between the 8 μm extinction and the ¹³CO toward the dense clumps in IRDCs, Du & Yang (2008) show that ¹³CO is optically thick toward many IRDCs, so we caution the reader that these tracers are probably not probing to the same depth in dense IRDC clumps. Du & Yang (2008) also demonstrate that typical excitation temperatures of ¹³CO in IRDCs are of the order of ∼10 K. This indicates that the ¹³CO is at the very least tracing to some depth in the cold, dense region of the clump, however, the ¹³CO is also tracing the surrounding GMC and IRDC envelope. We include column density and mass estimates from ¹³CO as a secondary mass tracer, and as a comparison to other tracers. In the optically thin, thermalized limit, the ¹³CO column density is given by

$$\begin{equation} N(\mbox{H}_{2}) = \frac{8\pi k\nu ^{3}X_{^{13}\mbox{CO}}}{3c^{3}hB_{J}A_{10}} (1-e^{-h\nu /kT_{{\rm ex}}})^{-1}\int \! T_{{\rm mb}} \, dv \end{equation} \tag{ 16 }$$

where ν is the frequency of the ¹³CO J = 1–0 transition, A₁₀ is the Einstein A coefficient of ¹³CO from state J = 1 to J = 0, T_ex is the excitation temperature, B_J is the rotation constant, and $X_{^{13}{\rm CO}}$ is the fraction of ¹³CO to H₂. We adopt a value of ¹²CO/¹³CO of 58 from Lucas & Liszt (1998), and a value of ¹²CO/H₂ of 10⁻⁴, a value of 55.101038 GHz for B_J, and standard NIST values for all constants and spectral transition values. This expression then reduces to

$$\begin{equation} N(\mbox{H}_{2}) = 1.45 \times 10^{17} \frac{\int \! T_{{\rm mb}} \, dv}{1 - e^{-5.29/T}}\,\rm cm^{-2}. \end{equation} \tag{ 17 }$$

This column density can then be converted to a mass via

$$\begin{equation} M = N(\mbox{H}_{2})\mu _{{{\rm H}}_{2}}m_{{\rm H}}A, \end{equation} \tag{ 18 }$$

where $\mu _{{{\rm H}}_{2}}$ is the mean molecular weight of H₂, m_H is the mass of hydrogen, and A is the physical area over which the mass is summed. We adopt a value of $\mu _{{{\rm H}}_{2}}$ = 2.8 (Kauffmann et al. 2008). This expression reduces to

$$\begin{equation} M = 5.00\times 10^{-25} N(\mbox{H}_{2}) \Bigg (\frac{A}{1\, \rm arcsec^{2}}\Bigg) \Bigg (\frac{D}{1 \,\rm kpc}\Bigg)^{2}\, M_{\odot }, \end{equation} \tag{ 19 }$$

where A is the area in arcsec², over which the mass is summed, and D is the distance in kpc. The ¹³CO clump masses range from 70 to 610 M_☉.

For this study, we find that the systematic uncertainties far outweigh the statistical uncertainties. We present below our estimated systematic uncertainties and the method used to derive them.

We assign each variable in the mass equation (Equations (6), (9), (10), (15), and (19)) a fractional uncertainty, perform a Monte Carlo simulation randomly drawing the variables from a Gaussian distribution, and report the 68% confidence interval over which the mass varies. Since we have reason to believe that each of the variables is centered at our adopted value, we assign each variable a randomly populated Gaussian distribution, centered at the adopted value and with a width such that 99.7% of the points are within the lower and upper bounds of the fractional uncertainty. For example, we assign an uncertainty to the distance of 20%. Therefore, for each run, we randomly sample the distance value from a Gaussian which is centered at the adopted value and has a width such that 99.7% of the points lie between 0.8 and 1.2 times the adopted distance value. For each mass estimate, we ran at least 10⁷ points in the simulation. We note that several of the mass probability distributions below are asymmetric. To be consistent with the literature, we present the mass estimates at the adopted variable values (which is equal to the mean of the mass probability distribution), but note that this is not the same as the median or most probable value of the distribution. The 68% confidence intervals were drawn from the median of the distribution outward. We divide the high- and low-mass ends of the confidence interval by the mass at the peak of the distribution to achieve the asymmetric 68% confidence intervals.

The dominant sources of uncertainty in the extinction mass estimate are the background and foreground estimates, the scattering coefficient, as well as the distance, opacity, and f_fore. We find that typical fluctuations in the 8 μm emission near IRDCs, on the scale of the IRDCs, are around 40%. This background fluctuation is the best measure we have for how much the background behind the IRDC will vary from what we measure, so we assign a 40% uncertainty to the background estimate. In calculating the foreground uncertainty, we assign the same 40% uncertainty to the measured background and an additional 50% uncertainty to the value of f_fore. We include a 15% uncertainty in the value of the scattering coefficient as the color variation is about 5% and not being able to characterize the spatial dependency of the scattering has about a 10%–15% effect (S. Carey 2010, private communication). We also include a 20% uncertainty in the distance determination (as discussed in Section 3.1) and factor of two uncertainty in the value of the opacity as recommended by Ossenkopf & Henning (1994). The 68% confidence intervals for the extinction mass are from 80% to 2.2 times the quoted mass.

For the dust continuum (BGPS 1.1 mm and MAMBO 1.2 mm) mass estimates, the dominant uncertainties are the temperature, the flux density, the opacity, and the distance. For both the 1.1 mm and 1.2 mm data, we assign an uncertainty of 50% to the temperature (for most clumps the temperature is between 10–20 K; Dunham et al. 2010), 10% to the flux density (Aguirre et al. 2010), a factor of two uncertainty in the opacity (as above), and 20% to the distance. The 68% confidence intervals for both dust continuum masses are from 80% to 180% of the quoted mass.

The systematic uncertainty associated with the virial mass is difficult to characterize. The basic assumption that the clump is a gravitationally bound object in virial equilibrium introduces an inherent uncertainty that cannot be quantified. Another major uncertainty is the virial radius. The true radius of these clumps is in most cases unknown, so assigning an uncertainty is somewhat questionable. Our typical virial radii are around 0.4 pc and some interferometric millimeter studies of IRDC clumps (e.g., Beuther et al. 2005) find sizes of about 0.1 pc. Therefore, we assign an uncertainty of a factor of four to the virial radius. The line width is uncertain by about 20% for these data. These uncertainties give us a 68% confidence interval for the virial masses from 50% to 170% of the quoted value. In cases where the adopted virial radius is off by a wide margin, the virial mass will also be off by that same wide margin (as the calculated mass scales linearly with the adopted virial radius).

The quantifiable systematic uncertainties in the ¹³CO mass are the excitation temperature, integrated line intensity, and distance. However, there is a large uncertainty that we are unable to quantify associated with the optical depth, freezeout, excitation conditions, and variable abundance of ¹³CO in cold, dense clumps (e.g., Tafalla et al. 2002). These will have the effect of lowering the ¹³CO mass and column density estimates. We assign an uncertainty to the excitation temperature of 50% (as above), an uncertainty in the integrated intensity of 20%, and an uncertainty in the distance of 20%. We find that the 68% confidence interval for the ¹³CO masses is from 80% to 120% of the quoted mass, neglecting additional uncertainties introduced by optical depth and freezeout effects.

Given the systematic uncertainties, the agreement between the various mass tracers is reasonable, as shown in Table 4 and Figure 2. All of the mass tracers increase with increasing BGPS 1.1 mm dust continuum mass, except for extinction and ¹³CO at the highest BGPS 1.1 mm masses, as these are associated with UC H ii regions (see Table 5). The virial masses are uniformly higher than the BGPS 1.1 mm masses, the ratio of the two ranges from 1.1 to 14.4 with a median of 6.4 ± 4.2. A detailed comparison of the BGPS 1.1 mm mass with the virial mass is given in Section 4.3. The MAMBO 1.2 mm masses are also uniformly higher than the BGPS 1.1 mm masses, though not nearly as high as the virial masses. The ratio of MAMBO 1.2 mm to BGPS 1.1 mm masses ranges from 1.2 to 2.6 with a median of 1.8 ± 0.3. This offset is intrinsic to data taken with MAMBO at 1.2 mm versus Bolocam at 1.1 mm. The reason for this offset is being investigated by the BGPS team and will be discussed in Aguirre et al. (2010). The ¹³CO masses agree surprisingly well with the BGPS 1.1 mm masses, the ratio between the two ranges from 0.2 to 1.9 with a median of 0.8 ± 0.5. Recall, however, that the ¹³CO mass is taken over a 46'' beam rather than the 33'' effective beam size of the rest, so we are including more of the GMC and the IRDC envelope, and therefore more emission. This means that the true comparable ¹³CO mass would be generally lower than the BGPS 1.1 mm mass. The agreement between the two is best at lower BGPS 1.1 mm masses. At higher BGPS column there is likely a higher than assumed BGPS temperature, as well as optical depth effects or freezeout of ¹³CO at high column densities (Tafalla et al. 2002). The extinction mass and BGPS 1.1 mm mass agree well, except at the highest BGPS 1.1 mm masses (M_BGPS > 600 M_☉), where the sources are associated with UC H ii regions. The ratio of extinction to BGPS 1.1 mm mass ranges from 0.1 to 2.8 with a median of 1.1 ± 0.8. A more detailed comparison of the extinction mass to the BGPS 1.1 mm mass is given below.

The relationship between 8 μm extinction and 1.1 mm dust emission is reasonably good, and where it is not, there is good indication that warm gas/dust due to a hot clump is skewing both extinction and emission measures of mass. Recent dust temperature estimates from Rathborne et al. (2010) corroborate that active clumps are indeed warmer than quiescent/intermediate clumps. The agreement between 8 μm extinction and 1.1 mm dust emission masses is poor in the four highest BGPS 1.1 mm mass sources, which are associated with UC H ii regions. In these regions, the temperature is almost certainly warmer than the assumed 15 K, which contributes to the very large BGPS fluxes. Additionally, the UC H ii region illuminates polycyclic aromatic hydrocarbon (PAH) features in the 8 μm band, boosting the 8 μm flux and reducing the measured extinction mass.

Figure 3 shows the extinction column density versus BGPS 1.1 mm column density, and we note that while the column density derived from extinction and the BGPS 1.1 mm dust continuum generally agree, there are some peculiarities. Note that the clumps indicated in Table 4 are only a small portion of what is plotted in Figure 3. Figure 3 plots the entire IRDC, while the table only includes the subsections of the cloud chosen for this study as clumps. In some sources, the extinction flattens out as the 1.1 mm column increases, while in others the extinction column is higher than that 1.1 mm column. These indicate a hot clump (with imperfect temperature assumptions for the 1.1 mm dust emission and 8 μm contamination by PAH illumination) and either a cold clump or imperfect foreground/background estimation, respectively. For the individual clumps, the median of the ratio of extinction column to BGPS column is 1.1 ± 0.8 overall, 0.7 ± 0.9 for active clumps, 1.4 ± 0.6 for intermediate clumps, and 2.0 ± 0.9 for quiescent clumps. This is in agreement with our assessment that high extinction per 1.1 mm column is characteristic of quiescent clumps, low extinction per 1.1 mm column is characteristic of active clumps, however, there is still a lot of scatter. Though there are some exceptions, in general, dust extinction and emission are well correlated for quiescent/intermediate clumps, but often diverge for active clumps, particularly ones associated with UC H ii regions.

Zoom In Zoom Out Reset image size

Simple calculations show that the measured virial masses are uniformly higher than the masses measured through other techniques, such as the dust continuum emission. The median of the ratio of the virial mass to the BGPS 1.1 mm mass is 6.4 ± 4.2 and ranges from 1.1 to 14.4. Interferometric observations (Rathborne et al. 2007, 2008; Wang et al. 2008; Zhang et al. 2009; for more details on specific sources, please see the Appendix) of several of these sources find that most clumps fragment into smaller cores at higher resolution, with sizes of about 0.03 pc or less. Zhang et al. (2009) find that on the scale of ∼0.02 pc the virial and dust mass estimates agree quite well. On the scale of about 1 pc, we find that the virial masses are much higher than the dust masses. The virial radius required for our measured virial masses to match our measured BGPS 1.1 mm masses ranges from 0.02 to 0.25 pc with a mean of 0.09 ± 0.06 pc, similar to the interferometric size scales noted above. The information presented is consistent with the picture of small bound cores (about 0.01 pc in size), surrounded by a dense envelope (about 0.1 pc in size) which is also bound, engulfed in a diffuse envelope (size scale of about 1 pc). This dense envelope is where the majority of dense gas is emitting and therefore dominates the line width we measure, while the diffuse envelope extends out to about 1 pc and comprises the extended emission we see in the dust continuum. In this picture, the line widths we measure are primarily from a region that is smaller than the virial radius we assign to it, thus explaining the very large virial to dust mass ratio. Alternatively, if the material producing the large line widths and the dust continuum emission are the same, it is possible that something more than virialized motion (e.g., magnetic fields, turbulence) plays a role in the line widths or that the clumps are not gravitationally bound, and possibly ram pressure confined by converging flows (e.g., Heitsch et al. 2008). The question of whether these clouds, clumps, and cores are bound is of great importance in understanding cluster formation.

The ratio of HCO⁺ to N₂H⁺ in clumps of varying activity levels is of special interest, as it is thought that HCO⁺ will trace "hot" clumps while N₂H⁺ will trace "cold" clumps (Jørgensen et al. 2004). HCO⁺ is created by CO in the gas phase and will also deplete onto dust grains at cold temperatures, so it should be more abundant in hotter environments. N₂H⁺, on the other hand, is destroyed by CO in the gas phase and should be more abundant in colder environments. Figure 4 is a pixel-by-pixel comparison of N₂H⁺ to HCO⁺ integrated intensities, and the line is of slope one, representing equality. Throughout the majority of the clouds, the ratio of HCO⁺ to N₂H⁺ is greater than one. In a few cases (most notably G028.37+00.07), however, the N₂H⁺ integrated intensity is greater than the HCO⁺. This is due to a high column of gas and dust, in which HCO⁺ is optically thick, and self-absorbed which lessens the peak (e.g., Figure 15 (bottom), making N₂H⁺ brighter by comparison. The general trend is that HCO⁺ and N₂H⁺ are well correlated and that HCO⁺ is slightly brighter. In Figure 5, we see that our present sample shows no trend between clump activity and the ratio of HCO⁺ to N₂H⁺. It is important to note that we are not resolving individual cores, so adjacent cores with differing chemistry will result in an average, rather than resembling distinct core chemistries. Additionally, we are comparing line strengths, not abundances.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

While we see no evidence for a chemical differentiation between HCO⁺ and N₂H⁺ on a ∼1 pc scale between "active" and "quiescent" clumps, we note that active clumps are significantly brighter in HCO⁺ and N₂H⁺ than intermediate or quiescent clumps. The increased brightness of HCO⁺ and N₂H⁺ in active clumps could be the result of an increased column density, higher volume density, higher temperature, and/or a larger beam filling factor. We argue that an increase in column is not the primary effect as not all active clumps have higher dust column, yet all active clumps show an increased brightness in HCO⁺ and N₂H⁺. The change in filling factor is a possible explanation, however the HCO⁺ and N₂H⁺ maps show extended emission around all the clumps. Since at the typical density of an IRDC (n ∼ 10⁵ cm⁻³), we are approaching the critical densities of HCO⁺ and N₂H⁺ (see Daniel et al. 2007), we suggest that the active clumps may have begun to compress and become denser, increasing the intensity of HCO⁺ and N₂H⁺. This increase in density is also responsible for the increased optical depth of HCO⁺ in active clumps, as evidenced by the self-absorption of HCO⁺ in some active clumps. Recent dust temperature estimates from Rathborne et al. (2010) suggest that active clumps are warmer, which would also increase the observed brightness of HCO⁺ and N₂H⁺. We observe that HCO⁺ and N₂H⁺ are brighter in active than quiescent clumps and suggest that this is because the active clumps are either warmer, have higher volume densities, or both.

We also compare the HCO⁺ and N₂H⁺ peak line intensities with those of ¹³CO in Figure 6. As discussed in Section 3.5, while there is good evidence that ¹³CO is optically thick toward many of these clumps, we see that it at least traces to some depth in the cold, dense portion of the clump. However, the ¹³CO is almost certainly tracing some part of the GMC and IRDC envelope as well. Therefore, we include it as an optically thick comparison point for the dense gas tracers. The ¹³CO peak varies very little among the different clump types and shows no noticeable increase with column, again corroborating that it is optically thick. The HCO⁺ and N₂H⁺ peak temperatures increase somewhat with column density, but most notably with activity level. This means that the active clumps all have much smaller variation in line strength than the quiescent clumps (the line connecting the peak T_mb's is shorter). The variation in line strength among the three tracers can be quantified by the difference between the brightest peak line strength of the three tracers and the weakest peak line strength of the three tracers. The mean of the variation in line strength is 4.5 ± 0.8 for active clumps, 4.8 ± 1.1 for intermediate clumps, and 5.5 ± 1.8 for quiescent clumps. While the numbers are uncertain, the trend is clear, and this is a potentially useful tool for identifying active clumps.

Zoom In Zoom Out Reset image size

We detect radio continuum sources toward four of the clumps. Two of these are unresolved, optically thick UC H ii regions. One is a slightly resolved, optically thin UC H ii region. The final source is an unresolved, marginal detection which may be a UC H ii region. The peak fluxes of these regions are given in Table 5, as well as the limiting (3σ) flux in the clumps without detections. The properties of all the detections (outside or inside the clumps) are reported in Table 6. The sources are named based on the IRDC in which they were found (though they are not necessarily associated with the cloud) and in order of decreasing brightness (VLA1, VLA2, etc.). All the Gaussian-type sources were fit with two-dimensional elliptical Gaussians, and the fluxes reported are from the fits. The σ reported is the rms of the fit residuals. For the sources with non-Gaussian morphology, the fluxes were measured using a circular aperture and an outer annulus of equal area. The mean of the flux in the outer annulus is assumed to be the background level and is subtracted from the source flux, while the standard deviation in the outer annulus is our quoted σ. Some of these detections are more evolved H ii regions, peripheral to the IRDC. Others are point sources unassociated with any GLIMPSE emission. These sources may be unresolved H ii regions that happen to be in our field of view but unassociated with the mid-IR, or they may be other types of Galactic or extragalactic radio continuum sources. We restrict our analysis to the four radio continuum sources associated with the millimeter clump peaks, as these are almost certainly UC H ii regions in the IRDC.

The UC H ii region radio continuum flux is directly related to the number of Lyman continuum photons (Q) of the ionizing star if the region is optically thin. Through comparison with the 20 cm and 6 cm radio continuum data from the Multi-Array Galactic Plane Imaging Survey (MAGPIS; White et al. 2005), we find that G034.43+00.24: GLM2 (VLA1) is slightly resolved (D ∼ 0.14 pc) and optically thin at 3.6 cm. G034.43+00.24: GLM1 (VLA3) is a marginal detection in both our survey and the 6 cm survey by Shepherd et al. (2004). The measured flux densities at 3.6 and 6 cm are not consistent with the source being a UC H ii region. However, the fluxes reported are highly uncertain, and systematic errors could be large enough to account for this inconsistency if it is in fact a UC H ii region. Since an optically thick UC H ii region is the most likely explanation for a radio continuum source associated with star formation in an IRDC clump, we assume that this is the explanation for the radio continuum source. We use the 3.6 cm flux to calculate the Lyman continuum flux of an unresolved, optically thick, UC H ii region, but acknowledge that this assumption may be incorrect. For G024.33+00.11: GLM1 (VLA2) and G023.60+00.00: GLM1 (VLA4) the 6 cm upper limits are consistent with the source being an optically thick UC H ii region at 3.6 cm.

In the optically thin case (G034.43+00.24: GLM2 (VLA1)), we fit a thermal bremsstrahlung curve to the 3.6, 6, and 20 cm points. We find an emission measure, EM = 4.4 × 10⁶ cm⁻⁶ pc, a turnover frequency ν(τ = 1) = 1.3 GHz, and Q = 6.1 × 10⁴⁵ Lyman continuum photons per second. This translates to a B0.5 star (Vacca et al. 1996), in agreement with Shepherd et al. (2004).

For the unresolved, optically thick sources, we first derive a source radius

$$\begin{equation} r = \Bigg [\frac{S_{\nu }c^{2}}{2\nu ^{2}kT_{e}}4D^{2}\Bigg ]^{1/2}, \end{equation} \tag{ 20 }$$

where S_ν is the radio continuum flux, ν is the band-center frequency, T_e is the ionized gas temperature, and D is the distance. We assume a T_e = 8000 K ionized gas, typical of an H ii region. We conservatively assume that the observed frequency is the turnoff frequency (where τ = 1). This gives us a lower limit for the EM, electron density (n_e), and number of Lyman continuum photons (Q). If our observing frequency is the turnover frequency (ν) then τ = 1 in our expression for the EM (Wood & Churchwell 1989)

$$\begin{equation} \mbox{EM}\; ({\rm cm^{-6}}\, {\rm pc)} = \frac{\tau }{8.235\times 10^{-2}\, \alpha (\nu,T_{e})T_{e}^{-1.35}\nu ^{-2.1}}. \end{equation} \tag{ 21 }$$

The lower limit for the number density of electrons, which we take to be equal the number density of ions, is

$$\begin{equation} n_{e} = \sqrt{\frac{\mbox{EM}}{2r}}. \end{equation} \tag{ 22 }$$

The number of Lyman continuum photons is given by the Stromgren sphere equation,

$$\begin{equation} Q = \frac{4}{3}\pi r^{3}\alpha _{B}n_{e}^{2}, \end{equation} \tag{ 23 }$$

where α_B is the recombination rate coefficient, 3.1 × 10⁻¹³ cm³ s⁻¹ for T_e = 8000 K. The Stromgren sphere equation assumes that the H ii region is spherical, in equilibrium, that there is one electron per ion, and that each energetic photon ionizes an atom or molecule. Since the observed structure is clumpy, the Stromgren sphere approximation is probably not a very good one. Using this method for the two (possibly three) optically thick, unresolved UC H ii regions yields a log(Q) of 46.14, 46.55, and 46.15 and stellar types (Vacca et al. 1996) of B0, O9.5, and B0 for G023.60+00.00: GLM1 (VLA4), G024.33+00.11: GLM1 (VLA2), and G034.43+00.24: GLM1 (VLA3), respectively.

A major driving question in the study of IRDCs is the evolutionary sequence of the clumps and clouds: identifying the different evolutionary stages and their relative lifetimes. This requires an understanding of the star formation activity. We employ four different star formation tracers: (1) an embedded 24 μm point source, (2) "green fuzzies," regions of enhanced, extended 4.5 μm emission, indicative of shocks and outflows, (3) H₂O and class I CH₃OH maser emission, and finally (4)UC H ii regions, an unambiguous indication that a massive star has "turned on." Tracers of star formation activity in these clumps are summarized in Table 5.

Each of these tracers is sensitive to somewhat different environments. A 24 μm point source traces material accreting onto forming stars, the gravitational contraction of a young stellar object, or even the nuclear energy emitted once the star has "turned on." "Green fuzzies" (also called EGOs; Chambers et al. 2009; Cyganowski et al. 2008) are thought to arise from shock-excited spectral line emission (Marston et al. 2004; Noriega-Crespo et al. 2004) due to outflows from young stars. Many of the HCO⁺ and N₂H⁺ spectra (see Figures 11 (bottom)–26 (bottom)) show asymmetric line profiles commonly associated with outflows, many of which (e.g., G034.43+00.24 GLM3; Figure 26 (bottom)) are associated with "green fuzzies," further implicating "green fuzzies" as positive outflow tracers. H₂O (22.23 GHz) and CH₃OH (class I 24.96 GHz) masers are well-known signposts of star formation activity, likely formed in shocks and outflows. Finally, we include the presence of 3.6 cm radio continuum emission which when associated with a millimeter peak or mid-IR emission is a clear indicator of a UC H ii region.

Together, these tracers are sensitive to warm, embedded dust indicative of accretion, shocks and outflows, and thermal bremsstrahlung from newly formed H ii regions. While these are all excellent tracers of star formation, none are perfect, so we would not expect to see every tracer in every actively star-forming region. Viewing angle or dust obscuration could play a significant role in a non-detection. We assign the designation "active" to clumps that exhibit three or four signs of active star formation ("green fuzzy," 24 μm point source, UC H ii region, or maser emission). This ensures that each "active" clump has either an H ii region (unambiguously star forming) or at least two outflow tracers (maser emission or "green fuzzy") and a 24 μm point source. We assign the designation "quiescent" to clumps that exhibit no signs of active star formation. We reserve the title "intermediate" for clumps which exhibit one or two signs of active star formation. Intermediate clumps, therefore, may have only shock/outflow signatures or a 24 μm point source. These categorizations represent a crude separation between the quiet, massive-starless clumps and the active, star-forming clumps, not an evolutionary sequence. Evolutionary sequences are discussed in Section 5.

The detection of several UC H ii regions embedded in IRDC clumps unambiguously shows that some IRDCs are forming massive stars. We note that all the clumps associated with UC H ii regions have significantly brighter (≳1 Jy) 24 μm flux. All but one of the UC H ii regions are associated with a CH₃OH maser and two of the four have a "green fuzzy." The lack of "green fuzzies" and a CH₃OH maser in some sources may indicate that the outflow stage has ceased or could simply be related to an unfavorable viewing angle. We also note that the brightest UC H ii regions (G034.43+00.24: GLM2 and near the edge of G028.37+00.07: GLM4) are associated with "diffuse red clumps," regions of extended, enhanced 8 μm emission. The ionizing radiation from a young H ii region will excite PAH features in the 8 μm band. As the H ii region grows it will begin to clear some of the surrounding dust and gas and reveal itself as a "diffuse red clump." As the H ii region continues to evolve, there will be PAH destruction by the strong UV flux and it will no longer have enhanced 8 μm emission. The clumps which contain UC H ii regions, but are not classified as "red," are either in an earlier evolutionary phase or are perhaps still obscured by the dust at 8 μm. There are no "diffuse red clumps" in our sample which are not associated with UC H ii regions, so these are potentially very useful tracers of young H ii regions.

Chambers et al. (2009) discussed the use of "red clumps" (regions of enhanced 8 μm emission) as tracers of embedded H ii regions. Since an enhancement at 8 μm could presumably be caused by any significant UV flux and most high column density clumps are nearly optically thick at 8 μm, we suggest that only the more diffuse "red clumps" are necessarily associated with H ii regions. Only when the forming star has cleared out some dust obscuration and is no longer associated with a millimeter peak can it be identified as a "diffuse red clump." Other "red clumps" could be embedded UC H ii regions that have not cleared out their envelopes, or they could be PAH nebulae excited by main sequence or evolved AB stars.

Figure 7 shows the 24 μm flux (as measured by Chambers et al. 2009) versus the dust temperature (measured through spectral energy distribution (SED) fits in the sample of 100 clumps by Rathborne et al. 2010). Notice that the "red" clumps are associated with brighter 24 μm flux and that in general, the more star formation tracers, the warmer the clump is. This general trend indicates that these tracers combined are successfully measuring star formation activity. However, the large scatter may be due to viewing angle biases, misassigned star formation tracers, or the inadequacy of single-temperature SED fits.

Zoom In Zoom Out Reset image size

G022.35+00.41 (see Figures 10 and 11) is actually composed of two clumps along the line of sight, one at 3.6 kpc and the other at 4.8 kpc. At least one of the two clumps is truly an active star-forming region. G022.35+00.41: GLM1 shows a bright millimeter peak, masers, a "green fuzzy," and a 24 μm point source. Interferometric measurements at 1 and 3 mm on the IRAM Plateau de Bure Interferometer by Rathborne et al. (2007) show that GLM1 contains two cores, less than 0.026 pc each. The second clump (and associated filament) is quiescent and has diffuse millimeter emission. While the superposition of two clouds along the line of sight lessens the intrinsic column of either, this example shows us that in the confused inner Galaxy, it is not terribly uncommon to have two spatially coincident dense clumps.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

G023.60+00.00 (see Figures 12 and 13) is a particularly interesting example with regard to the comparison of 8 μm extinction and dust emission. The extinction masses of the two clumps are very close (100 and 120 M_☉), however, the BGPS 1.1 mm masses differ by almost a factor of two (140 and 80 M_☉). GLM1 is an active clump, so it is likely warmer and denser than GLM2, an intermediate clump. Also, there is a bright H ii region to the southwest corner of the image shown in Figure 12, which increases the foreground 8 μm emission. Also, the HCO⁺ and N₂H⁺ maps show only diffuse emission around GLM2 and a bright peak toward GLM1 (see Figure 13, top and middle), tracing the dense, hot gas associated with an active clump.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

At 2.9 × 1.8 pc (major × minor axes of the IRDC ellipse, see Figure 1), G023.60+00.00 is on the smaller side for an IRDC and shows a filamentary morphology. It is near two evolved H ii regions and a complicated web of mid-infrared bubbles, filamentary IRDCs, and photodissociation regions (PDRs; see Figure 13, bottom). The location of G023.60+00.00 is at 3.6 kpc, right amidst the active star formation in the molecular ring (Jackson et al. 2006).

G024.33+00.11 (see Figures 14 and 15) contains a dense, active clump and more diffuse filamentary structure. There is a ridge between dark 8 μm extinction and bright 8 μm emission at the southern edge of the cloud. At a distance of 5.9 kpc, this cloud is among the most distant in our sample. G024.33+00.11 was not mapped in HCO⁺ and N₂H⁺, but it would be a good candidate for a future mapping project, to see if the dense ridge is associated with the IRDC. Additionally, portions of the cloud are dark at 8 μm, but show very little millimeter emission.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

G024.33+00.11: GLM1 is a particularly good example of a dense, active clump. Interferometric measurements at 1 and 3 mm on the IRAM Plateau de Bure Interferometer by Rathborne et al. (2007) show that G024.33+00.11: GLM1 has a single core smaller than 0.035 pc. Figure 15 (bottom) exemplifies the increase of HCO⁺ and N₂H⁺ in an active clump, and the self-absorption of ¹³CO and HCO⁺ in a particularly dense environment. This is in contrast to the weak HCO⁺ and N₂H⁺ detections in G024.33+ 00.11: GLM2.

G024.60+00.08 (see Figures 16–18) is the second (of two) examples in our sample which is comprised of two unique clumps along the line of sight, rather than one contiguous object. The two clumps are spatially adjacent with only a slight overlap of diffuse emission, which is an astonishing coincidence. The HCO⁺ and N₂H⁺ maps confirm (see Figure 17) this chance alignment.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

At 6.0 kpc, G024.60+00.08: GLM1 is the most distant clump in our sample. The 8 μm extinction does not appear abnormally large, but given the amount of foreground emission between us and the clump, it has one of the highest extinction masses in our sample. G024.60+00.08: GLM2, on the other hand, is located in the molecular ring at 3.4 kpc. This clump shows self-absorption of HCO⁺, and an asymmetric line profile in ¹³CO, indicative of outflows. These clumps are both categorized as intermediate. Interferometric measurements at 1 and 3 mm on the IRAM Plateau de Bure Interferometer by Rathborne et al. (2007) show that G024.60+00.08: GLM1 contains five cores, less than 0.041 pc each, while GLM2 contains three cores, smaller than 0.024 pc each.

G028.23–00.19: GLM1 (see Figures 19 and 20) is the prime example of a starless IRDC clump. It is typical of an IRDC clump in size, mass, 8 μm extinction, and in having a relatively compact millimeter core. However, this clump has absolutely no signs of star formation. The detections of HCO⁺ and N₂H⁺ in this source are incredibly weak. The dust column is comparable to many other clumps with bright HCO⁺ and N₂H⁺, the primary difference being the lack of density and a heating source in G028.23–00.19: GLM1. The 8 μm point source just north of GLM1 may be an indication of nearby, recent star formation. Given the compact nature of GLM1, its high column density, and nearby star formation, it seems likely that it is a quiescent precursor to massive star formation.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

G028.37+00.07 (see Figures 21 and 22) is an extremely varied cloud. It was identified by Simon et al. (2006a) as the darkest (by contrast) cloud in the First Galactic quadrant. Due to a particularly bright background around the cloud at 8 μm (as discussed in Section 3.2) the extinction column density is overestimated as compared with the BGPS 1.1 mm column density. This cloud is the most massive (by all tracers) in our sample. G028.37+00.07 is composed of dense filaments and surrounded by an extremely bright mid-IR background, including nearby evolved H ii regions.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

G028.37+00.07 hosts a prime example of a "diffuse red clump." In Figure 21 (top), we see a "diffuse red clump," spatially separated from the bright millimeter peak. The H ii region producing this "red" clump is asymmetric. Away from the millimeter clump, the H ii region has expanded and become more diffuse and toward the millimeter clump the H ii region remains confined by the high density.

The structure of HCO⁺ and N₂H⁺ in G028.37+00.07 is a great example of the differing ratio of HCO⁺/N₂H⁺. The HCO⁺ and N₂H⁺ maps show a similar overall structure, but the intensities of the two tracers vary across different clumps. In HCO⁺, GLM2 is the brightest, and in N₂H⁺, GLM4 is by far the brightest. As pointed out previously, however, this is not due to chemical differentiation. Rather, the HCO⁺ and even the N₂H⁺ become optically thick (and self-absorbed) in the densest clump, GLM4. The self-absorption of HCO⁺ makes N₂H⁺ appear brighter by comparison.

Wang et al. (2008) and Zhang et al. (2009) observed G028.37+00.07: GLM2 and GLM4 (referred to as P1 and P2) with the VLA and Submillimeter Array (SMA), respectively. In GLM2, they find five cores along the filament with line widths of about 1.2 km s⁻¹ and temperature of about 15 K. In GLM4, they find two cores with line widths of about 4.3 km s⁻¹, temperatures of about 30 K, and a rich molecular spectra.

G028.53–00.25 (see Figures 23 and 24) is a curious example of a source with a very small extinction column as compared with the BGPS 1.1 mm column. This cloud either has an exceptionally low background or high foreground. The region surrounding G028.53–00.25 is mostly devoid of extended H ii regions, PDRs, and other strong emitters of diffuse 8 μm emission. Due to the paucity of nearby diffuse 8 μm emission, the background estimate is decreased, and the extinction column estimate too low.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

G028.53–00.25 is a fairly quiescent cloud with few indicators of star formation. The clump, GLM1, is identified as intermediate due to the presence of a "green fuzzy" (though a fairly compact, faint one) and 24 μm point source. GLM1 shows very weak HCO⁺ and N₂H⁺ emission (the OTF maps suffer from low signal to noise). G028.53–00.25: GLM1 was observed on the SMA by Rathborne et al. (2008) who found that GLM1 contains at least three compact (less than ∼0.06 pc) cores. This quiescent cloud could be the host of low-mass star formation, or it could be a transient density enhancement or a quiescent proto-cluster.

G034.43+00.24 (see Figures 25 and 26) is a stunning example of an active IRDC. This object is just northeast of the enormously bright H ii region at G34.3+0.2, RH85 (Reid & Ho 1985). This complex contains the bright H ii region, RH85, and the shells of supernova remnants. G034.43+00.24, along with a few other IRDCs not discussed in the paper, is extremely filamentary structures, streaming radially out of this bright H ii region complex, increasing in width with distance from the complex. This is an excellent source to study the triggered formation mechanism of IRDCs (e.g Redman et al. 2003).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

South of G034.43+00.24: GLM2 is long train of "green fuzzies," along the IRDC. Rathborne et al. (2008) observed G034.43+00.24: GLM1 with the SMA and found that GLM1 remains unresolved and contains a core smaller than the ∼0.03 pc beam. GLM2 itself is an excellent example of a "diffuse red clump," offset from the millimeter peak. The bright 8 μm emission nullifies the use of an extinction mass in this clump; the extinction mass is 60 M_☉ while the BGPS 1.1 mm mass is 780 M_☉. This clump is one of the most "massive" sources in our sample according to the BGPS 1.1 mm mass, which is almost certainly biased by a higher dust temperature than the assumed 15 K. This spectacular clump is an example where all of our conventional mass tracers fail. The HCO⁺ and N₂H⁺ are extremely bright in all the clumps, especially GLM1 and GLM2 (the brightest by far in our sample). There is also strong self-absorption in ¹³CO toward these clumps, further indication of the high density and temperatures of these active clumps.