The Lifetimes of Phases in High-Mass Star-Forming Regions

Transcription

1 The Lifetimes of Phases in High-Mass Star-Forming Regions Cara Battersby 1 & John Bally 2 ABSTRACT High-mass stars are born in dense molecular clumps, but the nature of this process is a subject of much debate. Whether the clusters which form high-mass stars are born through slow, hydrostatic, global contraction of dense molecular clumps or quickly and dynamically, is a key question. We investigate the lifetimes of phases of high-mass star formation in dense molecular clumps by comparing the relative fraction in each phase in a 2 2 degree region centered at [l, b] = [30, 0 ]. Of all the regions capable of forming high-mass stars on 1 pc scales, based on a surface density threshold, the starless phase lasts approximately Myr (70%) and the star-forming phase lasts approximately Myr (30%). The starless/star-forming lifetime is the estimated time during which an entire 1 pc region exceeds a critical surface density and appears devoid/full of massive stars (as traced by heated dust or UV-excited 8 micron emission). Regions need to have accumulated enough material to exceed the critical surface density to be considered in this phase and also can not have significantly cleared the environment through feedback. The relative lifetimes of the starless vs. star-forming phases are robustly determined, while the absolute lifetime determinations from the maser and UCHII region lifetimes are less precise and will benefit from larger-scale, higher resolution analyses. The 6.7 GHz Class II CH 3 OH masers tend to be found at the intersection of starless and star-forming High Surface Density (HiSD) regions, indicating that they exist for a short time when a high-mass star turns on. Subject headings: ISM: kinematics and dynamics dust, extinction HII regions radio emission lines stars: formation 1. Introduction High-mass stars dominate the life cycle of gas and dust in our Galaxy, driving the dynamics in the interstellar medium and lighting up the universe around us. The transition 1 Harvard-Smithsonian Center for Astrophysics 2 Center for Astrophysics and Space Astronomy University of Colorado Boulder

2 2 of Giant Molecular Clouds (GMCs) into high-mass stars and stellar clusters is fundamental to our understanding of the global process of star formation, the origin of the Initial Mass Function (IMF), planet formation in star clusters, and star formation rates in nearby galaxies. We have yet to fully understand the formation, early evolution, and lifetimes of high-mass star-forming regions and this quest remains a cornerstone of modern astrophysics. Whether star clusters and high-mass stars form as the result of slow, equilibrium global collapse of clumps (e.g., Tan et al. 2006) over several free fall times or if they collapse quickly on the order of a free-fall time (e.g., Elmegreen 2000, 2007; Hartmann & Burkert 2007), perhaps through large scale accretion along filaments (Myers 2009), remains an open question. To study the formation, early evolution, and lifetimes of high-mass star-forming regions, we investigate the earliest phases of high-mass star formation. Young, embedded clusters form from high-density (n(h cm 3 ) clumps, pc radii, temperatures of K, and masses 10 2 to 10 3 M (Lada & Lada 2003). The precursors to these clusters to are expected to be colder than the clumps actively forming stars. Hence, searches for proto-cluster-forming regions typically target the longer wavelengths where the dust continuum peaks in cold, dense clumps (e.g., BGPS, ATLASGAL, Hi-GAL, Aguirre et al. 2011; Schuller et al. 2009; Molinari et al. 2010). These clumps can also be seen in silhouette as Infrared Dark Clouds (IRDCs, Egan et al. 1998; Perault et al. 1996; Omont et al. 2003) absorbing the diffuse background Galactic mid-ir light. Due to their cold temperatures (T < 20 K, Pillai et al. 2006) and high densities (> 10 5 cm 3 ) the highest-mass IRDCs (M > 10 3 M ) are commonly cited as the birthplaces for high-mass stars and stellar clusters (e.g., Rathborne et al. 2006; Beuther et al. 2007; Parsons et al. 2009; Battersby et al. 2010), yet not all IRDCs will form high-mass stars (Kauffmann & Pillai 2010). Additionally, the selection of IRDCs is inherently biased becasue their identification requires being on the near-side of a bright-mid-ir background. Battersby et al. (2011) showed that the physical properties of dust continuum clumps (temperature and column density) can be used to distinguish the pre-star-forming and actively star-forming populations, independent of whether or not they are dark in the mid-ir. Unlike other similar analyses (e.g., Chambers et al. 2009; Miettinen 2012; Wilcock et al. 2012; Tackenberg et al. 2012; Parsons et al. 2009, discussed in 3.5), all pixels above a surface density threshold are analyzed and binned into either actively star-forming or quiescent categories. Instead of IRDCs, this analysis targets all pixels above the threshold column density determined from the sub-mm dust continuum flux and temperature derived from the Herschel Hi-GAL data in the HiSD regions.

3 3 2. Methods The Galactic plane centered at Galactic longitude 30 contains one of the largest concentrations of dense gas and dust in the Milky Way. Located near the end of the Galactic bar and the start of the Scutum-Centaurus spiral arm, this region contains the massive W43 star forming complex at a distance of 5.5 kpc (Zhang et al. 2014) and hundreds of massive clumps and molecular clouds with more than 80% of the emission having V LSR = 80 to 120 km s 1 implying a kinematic distance between 5 and 9 kpc (Carlhoff et al. 2013). Data from the Herschel Infrared Galactic Plane Survey (Molinari et al. 2010, Hi-GAL), the Spitzer legacy project GLIMPSE (Benjamin et al. 2003), and surveys for 6.7 GHz CH 3 OH masers (Pestalozzi et al. 2005) were used to investigate a 2 2 field centered at [l,b] = [30, 0 ]. All data were convolved to a common resolution of 25. Dust temperatures and column densities were determined from modified blackbody fits to background-subtracted Hi-GAL data from 160 to 500 µmusing methods described in Battersby et al. (2011). The iterative background subtraction removed diffuse Galactic cirrus component whose contribution (N(H 2 ) < cm 2 ) is not included in the discussion of the dense region below Battersby et al. (2011). High-mass stars tend to form in regions with surface densities Σ 1 g cm 2 (Krumholz & McKee 2008), corresponds to a column density N(H 2 ) cm 2. At distances of several kpc or more, most cores are highly beam-diluted in a 25 beam. To derive a realistic high-mass star-forming column density threshold for cores beam-diluted by a 25 beam, consider a spherical core with a constant column density Σ = 1 g cm 2 for r < r f and a power-law drop off for r > r f n(r) = n f (r/r f ) p (1) where p is the density power-law exponent. The Mueller et al. (2002) study of 51 high-mass star-forming cores found a best-fit central core radius, r f 1000 AU, and density power-law index, p = 1.8. This model implies an H 2 central density of n f = cm 3, which integrated over r f = 1000 AU, corresponds to the theoretical surface density threshold for forming high-mass stars of Σ = 1 g cm 2 and 0.35 M contained within r f. Integration of this model core along the line of sight and convolution with a 25 beam results in a beam-diluted column density threshold appropriate for the analysis of the Hi- GAL based column density maps. At typical distances of 5 and 9 kpc toward the l = 30 field (Ellsworth-Bowers et al., in prep.) a 25 beam subtends 0.7 and 1.1 pc and the beam-diluted column densities of the model cores would be N(H 2 ) = 0.8 and cm 2, respectively. Pixels above this column density threshold are referred to as High-Surface Density Regions (HiSD regions). The mass enclosed in the beam area ranges from 63 to 83 M, sufficient to

4 4 form a 20 M star with 30% efficiency. The identification of starless vs. star-forming HiSD regions is based on the approach from Battersby et al. (2011) and uses the HiSD region temperature and emission at 8 µm to determine whether it is starless or star-forming. Pixels above the column density threshold for forming high-mass stars (HiSD regions) are identified. Next, each HiSD region is tagged as starless or star-forming based on the absence or presence of 8 µm emission. From the relative fraction of HiSD regions capable of forming high-mass stars in the starless vs. starforming phase, we determine the relative lifetimes of each phase. Above our column density threshold, we also search for 6.7 GHz Class II CH 3 OH masers from the unbiased searches by Szymczak et al. (2002) and Ellingsen (1996), as described in 2.2. Using the relative fraction of HiSD regions containing 6.7 GHz masers and their absolute lifetime of 35,000 years from the literature (van der Walt 2005), we determine a range of absolute lifetimes for each phase. Figure 1 demonstrates that the relative lifetimes are insensitive to variations in the threshold column density from 0.3 to cm 2 (corresponding to distances of 11 and 3 kpc for the model core). Thus, two cutoffs are used to determine relative lifetimes: N(H 2 ) = cm 2 ) dubbed the generous and N(H 2 ) = cm 2 ) dubbed the conservative cutoff Starless vs. Star-Forming Battersby et al. (2011) found that Herschel dust continuum sources can be separated out into starless and star-forming based on the absence pr presence of 24 µm sources, 8 µm PAH, or excess 4.5 µm emission in Spitzer data (Cyganowski et al. 2008, 2011b,a), masers, or HII regions. Dust temperature were found to increase with star formation activity. In this Letter, dust temperature distributions are combined with each HiSD region s mid-ir star-formation signatures to determine weather the region is starless or star-forming. This study is only sensitive to the star-forming signatures of high-mass stars. Therefore, the term starless refers only to the absence of high-mass stars forming; the region may support active low-mass star formation. Pixels above the column density threshold described above are classified as either starless of star-forming, based on its temperature and signature at 8 µm. A HiSD region can be mid- IR-dark (an IRDC absorbing the background 8 µm light), mid-ir-neutral (no signature at 8 µm), or mid-ir-bright (from UV excited PAH emission at 8 µm). Above the column density thresholds, mid-ir-dark HiSD regions are starless, while the mid-ir-bright HiSD regions are associated with high-mass star formation. Above the generous and conservative thresholds

5 Starless Generous Starless Conservative Starry Generous Starry Conservative Maser Percent % 30% Number of Pixels Column Density x [cm 2 ] Fig. 1. The relative fraction of pixels which are starless (70%) vs. star-forming (30%) is mostly insensitive to column density cutoff over a reasonable range of thresholds (N(H 2 ) cm 2 ) in a 25 beam. The left y-axis and dashed lines show the percentage of pixels which are starless (cyan and green using the generous and conservative identification methods respectively) vs. star-forming (magenta and orange using the generous and conservative identification methods respectively) or which contain a maser (red) as a function of the column density threshold selected. The right y-axis and solid lines show the column density distribution of the same populations.

6 6 (N(H 2 ) = 0.4 and cm 2, respectively), a Gaussian fit to the mid-ir-dark and mid-ir-bright temperature distributions are used tto classify all HiSD regions as starless or star-forming. As with the column densities, two different cutoffs based on the temperature distributions are used, a generous and conservative cutoff corresponding to 3 and 2 σ cuts, respectively. A HiSD region is classified as starless if it: 1) is mid-ir-dark and falls within 2 (or 3 for the generous cutoff) σ of the mid-ir-dark temperature distribution (Conservative: T = K, Generous: T = 9-32 K) or 2) is mid-ir-neutral and falls within the same distribution, with an upper limit at 25 K so as not to overlap with the star-forming population. Conversely, a HiSD region is classified as star-forming if it: 1) is mid-ir-bright and falls within 2 (or 3 for the generous cutoff) σ of the mid-ir-bright temperature distribution (Conservative: K, Generous: K) or 2) is mid-ir-neutral and falls within the same distribution, with a lower limit at 25 K so as not to overlap with the starless population Associated Methanol Maser While H 2 O, OH, and SiO masers are associated with young and post main-sequence stars, methanol masers tend to be only found in regions of massive star formation. Thus, 6.7 GHz Class II CH 3 OH masers were selected from the unbiased Galactic plane searches by Szymczak et al. (2002) and Ellingsen (1996) as compiled by Pestalozzi et al. (2005). The sizes of the methanol masers are defined by a circle with a radius determined by the positional accuracy of the observations (30 and 36 respectively for Szymczak et al. 2002; Ellingsen 1996). While methanol maser emission comes from very small areas of the sky (e.g., Walsh et al. 1998; Minier et al. 2001), they are often clustered. Szymczak et al. (2002); Ellingsen (1996) find multiple maser spots towards the majority of sources, so these methanol maser sizes are meant to represent the extent of the star-forming region. Future higher sensitivity and resolution unbiased searches (e.g., the 6 GHz multi-beam maser survey, Green et al. 2009) will certainly improve this characterization. 3. Lifetimes 3.1. Discussion of Uncertainties Several assumptions were made in determining the relative lifetimes of the starless and star-forming stages for high-mass star-forming regions: i) The sample is complete and unbiased in time and space. ii) Each HiSD region represents a region which is forming or

7 Galactic longitude Galactic latitude Temperature 8 micron emission Column Density Galactic latitude Galactic latitude Starless Starry Methanol Maser Generous Galactic longitude Starless Starry Methanol Maser Conservative Galactic longitude Fig. 2. Depiction of starless (white contour), star-forming (black contour), and maser (green contour) associated HiSD regions plotted on a three-color image in which red is the temperature, blue is the column density, and green is the 8 µm emission. Predominantly blue regions in this map are starless (cold, high column density, and 8 µm dark or neutral), while predominantly white/purplish regions are star-forming (warm, high column density, and 8µm bright). Red/yellow regions are warm but have low column densities. The generous HiSD region identifications (N(H2 ) > cm 2, 3σ temperature distribution) are depicted on the left while the conservative HiSD region identifications (N(H2 ) > cm 2, 2σ temperature distribution) are depicted on the right. Both identifications yield relative lifetimes of approximately 70% starless and 30% star-forming, though the associated relative maser lifetime is 2% on the left and 4% on the right. The sharp edges in red and blue are the source masks in the temperature and column density maps (see Battersby et al. 2011, for details).

8 Galactic latitude Galactic latitude Generous Galactic longitude Conservative Galactic longitude Fig. 3. Depiction of starless (white contour), star-forming (black contour), and maser (green contour) associated HiSD regions plotted on a three-color image in which red is the temperature, blue is the column density, and green is the 8 µm emission. Predominantly blue regions in this map are starless (cold, high column density, and 8 µm dark or neutral), while predominantly white/purplish regions are star-forming (warm, high column density, and 8µm bright). Red/yellow regions are warm but have low column densities. The generous HiSD region identifications (N(H2 ) > cm 2, 3σ temperature distribution) are depicted on the left while the conservative HiSD region identifications (N(H2 ) > cm 2, 2σ temperature distribution) are depicted on the right. Both identifications yield relative lifetimes of approximately 70% starless and 30% star-forming, though the associated relative maser lifetime is 2% on the left and 4% on the right. The sharp edges in red and blue are the source masks in the temperature and column density maps (see Battersby et al. 2011, for details).

9 9 will form a high-mass star. iii) The HiSD regions lifetimes don t depend on mass. iv) The star formation rate is constant as a function of time. v) Starless and star-forming regions occupy similar areas on the sky (number of pixels). vi) The signatures at 8 µm (mid-ir bright or dark) and their associated temperature distributions are good indicators of the presence or absence of a high-mass star. vii) All pixels above the threshold column density are beam-diluted dense cores and none are trace beam-filling low-surface density gas. While these assumptions are generally reasonable, many are highly uncertain. Assumption (i) is reasonable if there has been no large-scale triggering event. However,the entire study field lies in the densest part of the Scutum arm. The major HII regions such as W43 may have been triggered by older generations of stars, and in turn may be triggering the current populations of forming young massive stars and clusters at its periphery. However, the conditions in this field may be typical of a spiral arm environment. The argument for assumption (ii) presented above suggests that the majority of HiSD regions (with characteristic density profiles and at typical distances) have the ability to form highmass stars, but there are always exceptions and outliers which will break this assumption. This assumption could be improved with distance (and hence mass and size) determinations. Assumption (iii) is necessary at this time, but could potentially be removed by a careful separation of sources into mass bins when distances are determined. Assumption (iv) is reasonable over a sufficiently large sample (similar to assumption (i)) and over these relatively short Myr timescales. Since the column density threshold is the same for starless and star-forming HiSD regions, and there is little variation above that threshold, both should represent equal regions capable of forming high-mass stars, meaning that assumption (v) is reasonable. Various studies (e.g., Battersby et al. 2010; Chambers et al. 2009; Rathborne et al. 2006) argue in favor of assumption (vi), but more sensitive and higher resolution studies will continue to shed light on the validity of this assumption Observed Relative Lifetimes In both the conservative and generous cases, the relative fraction of pixels in the starless phase (percent of total pixels) is 70% vs. the relative fraction in the star-forming phase of 30%. These percentages are robust over a range in the cutoffs as shown in Figure 1. Slight changes in the temperature distributions (for example, including all HiSD regions down to 0 K in the starless and up to 100 K in the star-forming case) also have a negligible effect. Under the assumptions discussed in 3.1, high-mass star-forming clumps spend about 70% of their lives in the starless phase and 30% in the actively star-forming phase.

10 Maser association and Lifetimes An absolute lifetime for the starless and star forming phases can be anchored to the duration of the 6.7 GHz Class II CH 3 OH masers estimated of have lifetimes of 35,000 years (van der Walt 2005) using the same unbiased surveys used to identify the maser locations (Szymczak et al. 2002; Ellingsen 1996). This estimate is based on extrapolating the number of masers detected in these surveys to a Milky Way total and using an IMF and global Milky Way star formation rate to estimate the lifetime of the masers observed in these surveys, a method similar to that used by Tackenberg et al. (2012) to determine the absolute lifetimes of starless clumps. This method includes the fact that not every high-mass star-forming region will necessarily go through a maser phase. The absolute lifetime derived using CH 3 OH masers is controversial, but will likely be improved upon in future studies (e.g., the 6 GHz multi beam maser survey, Green et al. 2009). While the fraction of starless vs. star-forming HiSD regions is insensitive to the column density cuts, the fraction of HiSD regions associated with methanol masers increases as a function of column density (see Figure 1). In the generous and conservative cuts, the methanol maser fraction is 2% and 4%, respectively. Therefore, in the generous identification, the starless lifetime is 1.2 Myr and the star-forming lifetime is 0.5 Myr for a total dust clump lifetime of 1.7 Myr. In the conservative case, the starless lifetime is 0.6 Myr while the star-forming lifetime is 0.3 Myr for a total dust clump lifetime of 0.9 Myr. These starless lifetimes correspond to roughly a free fall time for typical clump sizes and densities. The 6.7 GHz CH 3 OH masers are nearly always found near the intersection of starless and star-forming HiSD region distributions. The 6.7 GHz CH 3 OH masers exist for a short time right when high-mass stars turn on UCHII Region Association and Lifetimes The HiSD regions tend to be associated with UCHII regions as reported by Wood & Churchwell (1989b). Wood & Churchwell (1989b,a), determine that the lifetimes of UCHII regions are longer than expected based on the expected expansion ration of D-type ionization fronts as HII regions evolve toward pressure equilibrium. They estimate that O stars spend about 10-20% of their main-sequence lifetime indie molecular clouds as UCHII regions. If we assume an O6 star, the main sequence lifetime is about years (Maeder & Meynet 1987). If the star spends 15% of this lifetime in the UCHII region phase, the corresponding lifetime of a UCHII region is about years. The remaining link between the absolute and relative lifetimes is the fraction of starry

11 11 pixels associated with UCHII regions, particularly O stars. While recent studies (e.g. Anderson et al. 2011) show more complete samples of HII regions, Wood & Churchwell (1989b) look for only the brightest UCHII regions, dense regions containing massive stars. More sensitive studies (e.g. Anderson et al. 2011) show HII regions over wider evolutionary stages, after much of the dense gas cocoon has been dispersed. Wood & Churchwell (1989b) searched 3 regions in the l = 30 field for UCHII regions finding them toward 2. These three regions were all classified as starry in our study. The Wood & Churchwell (1989b) survey found UCHII regions toward 2/3 starry regions surveyed. Since the 8 µm emission is indicative of UV excitation of PAH molecules, the starry pixels show warmer dust temperatures, and Bania et al. (2010) show that nearly all GLIMPSE bubbles are associated with UCHII regions, we estimate that % of our starry pixels are associated with an UCHII region. This corresponds to total lifetimes of 2.4 Myr (50% of starry pixels have UCHII regions) and 1.2 Myr (100% of starry pixels have UCHII regions). The absolute lifetimes of the starless phase then would be Myr and the starry phase would be Myr Comparison with Other Lifetime Estimates Previous lifetime estimates, based primarily on mid-ir emission signatures at 24 µm toward samples of IRDCs, found relative starless fractions between about 30-80% and extrapolate these to absolute starless lifetimes ranging from years and up to years. Chambers et al. (2009) using a sample of 106 IRDC cores find 65% starless and 35% with 24 µm emission and EGOs (or 82% starless and 18% star-forming if only those cores which contain 8 µm emission are considered to be star-forming). They extrapolate this to an absolute starless lifetime of years assuming a representative YSO accretion timescale of years (Zinnecker & Yorke 2007) for the star-forming phase. Miettinen (2012) found a starless / star-forming ratio of 44% to 56% using LABOCA, and extrapolate this in the same way to an absolute lifetime of years. Wilcock et al. (2012) using Hi-GAL find 18% starless, 15% with emission at 24 µm, and 67% with emission at 8 µm to derive an absolute starless lifetime of years. Tackenberg et al. (2012) target their search toward starless clumps in the ATLASGAL survey and derive a lifetime of the starless phase for the most high-mass clumps of years based on an extrapolated total number of starless clumps in the Milky Way and a Galactic star formation rate. Parsons et al. (2009) derive 33% starless and 67% with 24 µm emission using SCUBA targeted toward IRDCs, finding a starless lifetime of 10 3 to 10 4 years. Dunham et al. (2011) looked for mid-ir star formation signatures toward Bolocam Galactic Plane Survey (BGPS Aguirre et al. 2011) clumps and found that 56% are starless, or when accounting for chance alignment, 80% are starless. Peretto & Fuller (2009) found a starless fraction between 80%- 32% toward IRDCs,

12 12 based on their lack of association with 24 µm point sources. Combining chemical models with observations toward 59 high-mass star-forming regions, Gerner et al. (2014) found an IRDC lifetime of 10 4 years, 6 and years for high-mass protostellar objects and hot molecular core phases respectively, and 10 4 years for the UCHII region stage. Many previous lifetime estimates for high-mass star forming regions targeted IRDCs and used emission at 24 µm as the indicator of star formation. Our analysis includes all HiSD regions above a column density threshold. 8 µm emission is used to indicate high-mass star formation; 24 µm emission may turn on earlier (e.g., Battersby et al. 2010) but may not indicate high-mass star formation since it does not require UV-excited PAH emission. Previous analyses calculated relative fractions of clumps or cores, defined in various ways and with arbitrary sizes. Each clump or core is denoted as starless or star-forming; a single 24 µm point source would classify the entire clump as star-forming. In this study, individual pixels are used. Therefore a higher starless fraction is not surprising. Maser lifetimes anchor the absolute lifetimes. If a star-forming lifetime of years is used, (representative YSO accretion timescale, Zinnecker & Yorke 2007), the starless lifetime is reduced to years. 4. Conclusion Using column densities and dust temperature distributions derived from Hi-GAL, Spitzer 8 µm star formation signatures, and unbiased surveys for 6.7 GHz Class II CH 3 OH masers, the relative lifetimes of the starless and star-forming phases for high-mass star-forming regions in a 2 2 field centered at [l, b] = [30, 0 ] are determined. HiSD regions capable of forming high-mass stars are identified by their large column density in the dust continuum. They spend about 70% of their lifetimes in the starless phase and 30% in the star-forming phase or embedded. Starless refers only to a lack of high-mass stars and was determined by the temperature (cold, roughly 25 K, see 2.1 for details) and signature at 8 µm (dark or neutral). Star-forming refers to active high-mass star formation as indicated by warmer dust temperatures (> 25 K, roughly, see 2.1 for details) and emission at 8 µm. This relative fraction is determined robustly over a variety of reasonable cutoff parameters. Absolute lifetimes for the two phases are ached to the duration of methanol masers (35,000 years) determined from van der Walt (2005). Two column density cutoffs suggest a starless lifetimes of 0.6 to 1.2 Myr (70%) and a star-forming lifetimes of 0.3 to 0.5 Myr (30%) for high-mass star-forming regions identified in the dust continuum. These starless lifetimes correspond to roughly a free fall times at typical clump densities (high density cores embedded in these clumps will have much shorter free fall times). Using an estimate of the

13 13 absolute lifetimes for UCHII regions and estimating their association with regions defined as star-forming, we find a starless lifetime of Myr (70%) and a star-forming lifetime of Myr (30%). Together, these methods of anchoring the relative lifetimes to absolute lifetimes produce similar results with a starless lifetime ranging from Myr (70%) and the star-forming lifetime of molecular clumps lasting about Myr (30%). The 6.7 GHz CH 3 OH masers appear at the intersection between starless and star-forming HiSD regions. This indicates that these masers exist for a short period of time as a high-mass star turns on, and that 6.7 GHz masers trace the earliest phase of high-mass star formation. We thank H. Beuther, J. Tackenberg, A. Ginsburg, and J. Tan (and others?? XXXX) for helpful conversations regarding this work. This work has made use of ds9 and the Goddard Space Flight Centers IDL Astronomy Library. Data processing and map production of the Herschel data has been possible thanks to generous support from the Italian Space Agency via contract I/038/080/0. Data presented in this paper were also analyzed using The Herschel interactive processing environment (HIPE), a joint development by the Herschel Science Ground Segment Consortium, consisting of ESA, the NASA Herschel Science Center, and the HIFI, PACS, and SPIRE consortia. This work was supported by NASA through an award issued by JPL/Caltech via NASA Grant # REFERENCES Aguirre, J. E., Ginsburg, A. G., Dunham, M. K., et al. 2011, ApJS, 192, 4 Anderson, L. D., Bania, T. M., Balser, D. S., & Rood, R. T. 2011, ApJS, 194, 32 Bania, T. M., Anderson, L. D., Balser, D. S., & Rood, R. T. 2010, ApJ, 718, L106 Battersby, C., Bally, J., Jackson, J. M., et al. 2010, ApJ, 721, 222 Battersby, C., Bally, J., Ginsburg, A., et al. 2011, A&A, 535, A128 Benjamin, R. A., Churchwell, E., Babler, B. L., et al. 2003, PASP, 115, 953 Beuther, H., Walsh, A. J., Thorwirth, S., et al. 2007, A&A, 466, 989 Carlhoff, P., Nguyen Luong, Q., Schilke, P., et al. 2013, A&A, 560, A24 Chambers, E. T., Jackson, J. M., Rathborne, J. M., & Simon, R. 2009, ApJS, 181, 360 Cyganowski, C. J., Brogan, C. L., Hunter, T. R., & Churchwell, E. 2011a, ApJ, 743, 56

14 14 Cyganowski, C. J., Brogan, C. L., Hunter, T. R., Churchwell, E., & Zhang, Q. 2011b, ApJ, 729, 124 Cyganowski, C. J., Whitney, B. A., Holden, E., et al. 2008, AJ, 136, 2391 Dunham, M. K., Robitaille, T. P., Evans, II, N. J., et al. 2011, ApJ, 731, 90 Egan, M. P., Shipman, R. F., Price, S. D., et al. 1998, ApJ, 494, L199 Ellingsen, S. 1996, PhD thesis, Physics Department, University of Tasmania, GPO Box 252C, Hobart 7001, Australia Elmegreen, B. G. 2000, ApJ, 530, , ApJ, 668, 1064 Gerner, T., Beuther, H., Semenov, D., et al. 2014, A&A, 563, A97 Green, J. A., Caswell, J. L., Fuller, G. A., et al. 2009, MNRAS, 392, 783 Hartmann, L., & Burkert, A. 2007, ApJ, 654, 988 Kauffmann, J., & Pillai, T. 2010, ApJ, 723, L7 Krumholz, M. R., & McKee, C. F. 2008, Nature, 451, 1082 Lada, C. J., & Lada, E. A. 2003, ARA&A, 41, 57 Maeder, A., & Meynet, G. 1987, A&A, 182, 243 Miettinen, O. 2012, A&A, 542, A101 Minier, V., Conway, J. E., & Booth, R. S. 2001, A&A, 369, 278 Molinari, S., Swinyard, B., Bally, J., et al. 2010, A&A, 518, L100 Mueller, K. E., Shirley, Y. L., Evans, II, N. J., & Jacobson, H. R. 2002, ApJS, 143, 469 Myers, P. C. 2009, ApJ, 700, 1609 Omont, A., Gilmore, G. F., Alard, C., et al. 2003, A&A, 403, 975 Parsons, H., Thompson, M. A., & Chrysostomou, A. 2009, MNRAS, 399, 1506 Perault, M., Omont, A., Simon, G., et al. 1996, A&A, 315, L165 Peretto, N., & Fuller, G. A. 2009, A&A, 505, 405

15 15 Pestalozzi, M. R., Minier, V., & Booth, R. S. 2005, A&A, 432, 737 Pillai, T., Wyrowski, F., Carey, S. J., & Menten, K. M. 2006, A&A, 450, 569 Rathborne, J. M., Jackson, J. M., & Simon, R. 2006, ApJ, 641, 389 Schuller, F., Menten, K. M., Contreras, Y., et al. 2009, A&A, 504, 415 Szymczak, M., Kus, A. J., Hrynek, G., Kěpa, A., & Pazderski, E. 2002, A&A, 392, 277 Tackenberg, J., Beuther, H., Henning, T., et al. 2012, A&A, 540, A113 Tan, J. C., Krumholz, M. R., & McKee, C. F. 2006, ApJ, 641, L121 van der Walt, J. 2005, MNRAS, 360, 153 Walsh, A. J., Burton, M. G., Hyland, A. R., & Robinson, G. 1998, MNRAS, 301, 640 Wilcock, L. A., Ward-Thompson, D., Kirk, J. M., et al. 2012, MNRAS, 422, 1071 Wood, D. O. S., & Churchwell, E. 1989a, ApJ, 340, b, ApJS, 69, 831 Zhang, B., Moscadelli, L., Sato, M., et al. 2014, ApJ, 781, 89 Zinnecker, H., & Yorke, H. W. 2007, ARA&A, 45, 481 This preprint was prepared with the AAS L A TEX macros v5.2.