G22: a hub-filament cloud

High-mass Star Formation through Filamentary Collapse and Clump-fed Accretion in G22

Jinghua Yuan11affiliationmark: , Jin-Zeng Li11affiliationmark: , Yuefang Wu22affiliationmark: , Simon P. Ellingsen33affiliationmark: , Christian Henkel44affiliationmark: 55affiliationmark: , Ke Wang66affiliationmark: , Tie Liu77affiliationmark: 88affiliationmark: ,
Hong-Li Liu99affiliationmark: , Annie Zavagno1010affiliationmark: , Zhiyuan Ren11affiliationmark: , Ya-Fang Huang11affiliationmark:
National Astronomical Observatories, Chinese Academy of Sciences, 20A Datun Road, Chaoyang District,
Beijing 100012, China; jhyuan@nao.cas.cn;
Department of Astronomy, Peking University, 100871 Beijing, China;
School of Physical Sciences, University of Tasmania, Hobart, Tasmania, Australia;
Max-Planck-Institut für, Radioastronomie, Auf dem Hügel 69, 53121 Bonn, Germany;
Astronomy Department, Faculty of Science, King Abdulaziz University, PO Box 80203, 21589 Jeddah, Saudi Arabia;
European Southern Observatory, Karl-Schwarzschild-Str. 2, D-85748 Garching bei München, Germany;
Korea Astronomy and Space Science Institute 776, Daedeokdae-ro, Yuseong-gu, Daejeon 34055, Korea;
East Asian Observatory, 660 N. A’ohōkū Place, Hilo, Hawaii 96720-2700, USA;
Department of Physics, The Chinese University of Hong Kong, Shatin, NT, Hong Kong SAR;
Aix Marseille Univ, CNRS, LAM, Laboratoire d’Astrophysique de Marseille, Marseille, France
Abstract

How mass is accumulated from cloud-scale down to individual stars is a key open question in understanding high-mass star formation. Here, we present the mass accumulation process in a hub-filament cloud G22 which is composed of four supercritical filaments. Velocity gradients detected along three filaments indicate that they are collapsing with a total mass infall rate of about 440  Myr, suggesting the hub mass would be doubled in six free-fall times, adding up to Myr. A fraction of the masses in the central clumps C1 and C2 can be accounted for through large-scale filamentary collapse. Ubiquitous blue profiles in HCO  and CO  spectra suggest a clump-scale collapse scenario in the most massive and densest clump C1. The estimated infall velocity and mass infall rate are 0.31 km s and  yr, respectively. In clump C1, a hot molecular core (SMA1) is revealed by the SMA observations and an outflow-driving high-mass protostar is located at the center of SMA1. The mass of the protostar is estimated to be  and it is still growing with an accretion rate of  yr. The coexistent infall in filaments, clump C1, and the central hot core in G22 suggests that pre-assembled mass reservoirs (i.e., high-mass starless cores) may not be required to form high-mass stars. In the course of high-mass star formation, the central protostar, the core, and the clump can simultaneously grow in mass via core-fed/disk accretion, clump-fed accretion, and filamentary/cloud collapse.

ISM: clouds – ISM: individual objects: G22 – ISM: kinematics and dynamics – stars: formation – stars: massive
slugcomment: Draft Version, August 3, 2019

1 Introduction

The process of mass accumulation in high-mass star formation has remained elusive for decades (Tan et al., 2014; Motte et al., 2017). In the turbulent core model, which is one of the two extensively debated scenarios, the final stellar mass is pre-assembled in the collapsing high-mass core (McKee & Tan, 2003). Otherwise, the competitive accretion model allows the mass reservoir, often referred to as the star-forming core, to keep growing in the course of high-mass star formation (Bonnell et al., 2001).

Prevalent filaments detected in our Galaxy are well-established as some of the main sites for star formation (André et al., 2010, 2014; Molinari et al., 2010a; Wang et al., 2014, 2015, 2016). Specifically, hub-filament systems are frequently reported forming high-mass stars (e.g., Hennemann et al., 2012; Liu et al., 2012, 2016; Peretto et al., 2013). Converging flows detected in several hub-filament systems channel gas to the junctions where star formation is often most active (e.g., Kirk et al., 2013; Peretto et al., 2014; Liu et al., 2016). However, how gas flows help individual star-forming cores grow in mass is still poorly understood.

High-mass clumps ( pc) are the objects that fragment into dense cores ( pc) which subsequently contract to form individual or bound systems of protostars. Clump-scale global collapse has been reported in case studies (Schneider et al., 2010; Csengeri et al., 2011; Ren et al., 2012; Liu et al., 2013; Peretto et al., 2013), based on which “clump-fed accretion” has been proposed as an alternative high-mass star formation mode (Tigé et al., 2017; Motte et al., 2017). It still remains unknown, how significant such clump-scale collapse may be in sustaining mass growth in individual cores where high-mass protostars have already formed.

Located in the inner Galactic Plane with and , the G22 cloud contains ten Spitzer infrared dark clouds (IRDCs) from Peretto & Fuller (2009). These IRDCs are mainly distributed in a hub-filament system. Several young stellar objects (Robitaille et al., 2008), an extended green object (EGO G022.04+00.22, Cyganowski et al., 2008), and methanol maser emission (Cyganowski et al., 2009) detected in G22 indicate active high-mass star formation therein. These features make G22 an ideal laboratory to investigate how mass is accumulated in the high-mass star formation process, from cloud-scale down to individual cores. Using the Bayesian Distance Calculator developed by Reid et al. (2016), the distance to G22 is estimated to be kpc based on its systemic velocity of 50.9 km s.

In this work, we carry out an extensive study of G22 based on observations with varying angular resolutions to reveal a promising high-mass star formation scenario where the central high-mass protostar, the core, and the clump grow in mass simultaneously. The data used in this work are described in Section 2. We present results about the large scale cloud and the SMA observations in Sections 3 and 4. More in-depth discussions and a summary of the findings are given in Sections 5 and 6.

2 Data

2.1 Galactic Plane Survey Data

We extracted multi-wavelength data covering the entire cloud from the GLIMPSE, MIPSGAL, Hi-GAL and ATLASGAL surveys. Using the IRAC instrument on board the Spitzer Space Telescope (Werner et al., 2004), the Galactic Legacy Infrared Mid-Plane Survey Extraordinaire (GLIMPSE, Benjamin et al., 2003; Churchwell et al., 2009) project surveyed the inner of the Galactic Plane at 3.6, 4.5, 5.8, and 8.0 µm with sensitivities of 0.2, 0.2, 0.4, and 0.4 mJy, respectively. The Galactic Plane Survey using the Multiband Infrared Photometer on Spitzer (MIPSGAL, Cyganowski et al., 2009) mapped an area comparable to GLIMPSE at longer infrared wavelengths. Version 3.0 of the MIPSGAL data provides mosaics from only the 24 µm band with an angular resolution of 6″ and a sensitivity of about 1.7 mJy. As a key project of the Herschel Space Observatory, the Herschel Infrared Galactic Plane Survey (Hi-GAL, Molinari et al., 2010b) mapped the entire Galactic plane with nominal (following the Galactic warp) at 70, 160, 250, 350, and 500 µm with angular resolutions of 102, 135, 181, 250, and 364. Using the LABOCA camera, the APEX Telescope Large Areas Survey of the Galaxy (ATLASGAL Schuller et al., 2009) mapped 420 square degrees of the Galactic plane between at 870 µm  with a 192 angular resolution. Cutouts of were retrieved from the above surveys to reveal the large scale morphology and global properties of the entire cloud. Note that we used the ATLASGAL-Planck combined data to better recover large scale features at 870 µm (Csengeri et al., 2016).

2.2 Single Dish Molecular Data

From the FCRAO Galactic Ring Survey (GRS, Jackson et al., 2006) we retrieved CO  line data of the cloud. The GRS CO  data have velocity and angular resolutions of 0.21 km s and 46″. The main beam efficiency () and the typical rms sensitivity are 0.48 and 0.13 K, respectively.

From the JCMT data archive, we extracted CO , C, and HCO  data which cover only the central . The CO  and C were simultaneously observed in August of 2007 with an angular resolution of 15″, a velocity resolution of 0.055 km s, and a rms noise of about 1.20 K. Observed in May of 2008, the HCO  data have an angular resolution of 20″, a velocity resolution of 0.55 km s, and a rms noise of about 0.18 K.

2.3 SMA Observations

Observations towards the central massive clump C1 using the Submillimeter Array (SMA) at 1.3 mm were performed on 2010 August in the compact-north configuration with eight antennae used. The data are publicly available in the SMA data archive (PI: Claudia Cyganowski). The GHz correlator was tuned to cover the frequency range 216.8–220.8 GHz in the lower sideband (LSB) and 218.8–232.8 GHz in the upper sideband (USB). The 4 GHz bandwidth in each sideband is allocated to 48 spectral windows, each of which consists of 128 channels with an uniform width of 812.5 kHz (1.1 km s). The projected baseline lengths ranged from 8 to 105 k, which indicates that the SMA observations are insensitive to smoothed structures larger than  .

3c454.3, J1733-130/J1911-201, and Callisto (Jupiter’s moon) were observed as bandpass, gain and flux density calibrators. The data were reduced and imaged using MIRIAD (Sault et al., 1995). GILDAS (Pety, 2005), APLpy111APLpy is an open-source plotting package for Python hosted at http://aplpy.github.com. and Astropy (Astropy Collaboration et al., 2013) packages were used for data visualization and analysis. The continuum data were constructed using line free channels from both sidebands, which resulted in a synthesized beam of and a rms noise of 3.2 mJy beam. For the line data, the continuum emission was subtracted by modeling an order one polynomial. We also performed self-calibration to the continuum data and applied the solutions to the line data. The typical noise in a single channel of the line data is about 80 mJy beam.

3 A Hub-Filament Cloud

Name Length Mass aaAveraged column density.
(pc) () ( pc) ( cm)
F1 2.91 357 123 4.59
F2 2.62 575 220 7.28
F3 3.02 485 161 7.72
F4 1.88 102 54 3.07
Table 1: Properties of filaments
Figure 1: Morphology of G022.04+0.22 in multi-wavelength bands. The blue and magenta crosses mark the most massive clumps C1 and C2 (see Figure 2 and Section 3.1) in the hub region. The skeletons of the four filaments are also shown in (e). The Hi-GAL images shown here are the unfiltered ones.

Cutouts at wavelengths from 3.6 to 870 µm of the G22 cloud are shown in Figure 1. Extinction features, which correspond to 10 infrared dark clouds identified by Peretto & Fuller (2009), dominate in the north at 8.0 µm and can be partly seen at 24 and 70 µm. Extended emission in the south at 8.0, 24, and 70 µm originates from an infrared dust bubble, MWP1G022027+002159 (Simpson et al., 2012). This infrared dust bubble is slightly farther from us but physically interacting with G22 (see Section 5.4).

As shown in Figure 1 (b)–(h), two clumps (C1 and C2, see bellow) dominate emission at wavelengths longer than 24 µm. These two clumps also represent the most active star-forming regions in G22. Detected at far-IR and sub-mm wavelengths, several filaments converge at clumps C1 and C2.

3.1 Dust Properties

Figure 2: (a) Column density map from SED fitting. The eight clumps, which have been identified based on this map and designated to be C1 to C8, are shown as open ellipses. Also labeled are the visually identified filaments F1 to F4. (b) Dust temperature map from SED fitting with column density overlaid as contours in levels of cm. Infrared dark clouds from Peretto & Fuller (2009) are marked as purple diamonds. The red stars represent young stellar object candidates from Robitaille et al. (2008). EGO G022.04+0.22 is shown as a filled triangle. The three blue open triangles represent molecular hydrogen emission objects from Ioannidis & Froebrich (2012). The cyan filled inverted triangle labels IRAS 18278-0936. The two green ellipses delineate the inner and outer rings of the infrared bubble MWP1G022027+002159 (Simpson et al., 2012). (c) H column density from SED fits (cyan contours) and SMA 1.3 mm continuum emission (magenta contours) overlaid on a IRAC three color image with emission at 8.0, 4.5, and 3.6 µm rendered in red, green, and blue, respectively. Cyan contours represent H column densities of cm. Magenta contour levels at the center (better visible in (d)) start from and increase with a step of , where mJy beam. (d) A close-up view of SMA 1.3 mm continuum. The single core is designated as SMA1. A filled star represents the MIR source SSTGLMC G022.0387+00.2222 (MIR1) from the GLIMPSE survey.

Column density and dust temperature maps have been obtained via fitting the spectral energy distribution (SED) to the multi-bands data. To reach an intermediate angular resolution we only used data at 160, 250, 350, and 870 µm to perform the fitting. The images were convolved to a common angular resolution of 25 which is essentially the poorest resolution of the considered wavelengths. The convolution kernels of Aniano et al. (2011) were used to take the different instrumental responding functions into account. For the ATLASGAL images, any uniform astronomical signal on spatial scales larger than 2.′5 has been filtered out together with atmospheric emission during the data reduction (Schuller et al., 2009). The filtering of Hi-GAL images was performed using the CUPID-findback algorithm of the Starlink suite. For further details on the algorithm please see the online document for findback222http://starlink.eao.hawaii.edu/starlink/findback.html. Following Yuan et al. (2017), we run the algorithm iteratively using a common filtering box of 2.′5 to get stable background images which were subtracted from the post-convolution data for Hi-GAL bands to remove large-scale structures.

Intensity as a function of wavelength for each pixel was obtained from the smoothed and background-removed far-IR to sub-mm image data and modeled to a modified blackbody.

(1)

where

(2)

Here, is the mean molecular weight adopted from Kauffmann et al. (2008), is the mass of a hydrogen atom, is the H column density, is the gas to dust ratio, is the dust opacity at 600 GHz for coagulated grains with thin ice mantles (Ossenkopf & Henning, 1994). The dust emissivity index was fixed to 1.75 (Ossenkopf & Henning, 1994; Planck Collaboration et al., 2014). Free parameters in this model are the dust temperature and column density .

The resultant column density and dust temperature maps are presented in Figure 2. We note that the dust temperatures will be underestimated as the modeling only considered emission at wavelengths not shorter than 160 µm, while some regions significantly emit at 70 µm. Uncertainties in and can also originate from the dust opacity, which is subject to a factor of two uncertainty (Ossenkopf & Henning, 1994). The dust emissivity index () can also largely influence the resultant parameters. An increase of 0.5 for would lead to a increase for and a decrease for , and would decrease by and might increase by if decreases by 0.5 (Yuan et al., 2017).

Clulmp R.A. Dec. PA FWHM aaFull width at half maximum: bbEquivalent radius: , where .
(″) (″) () (″) (pc) (K) ( cm) () (g cm) ( cm)
C1 183034.64 -93455.9 38.8 34.6 90.3 36.5 0.39 20.9 6.55 590 0.26 3.52
C2 183038.12 -93441.1 48.1 37.8 49.2 42.3 0.49 18.4 2.39 319 0.09 0.91
C3 183044.04 -93459.7 43.9 32.1 123.4 36.8 0.39 13.8 2.68 270 0.12 1.56
C4 183042.38 -93328.0 62.9 44.3 67.2 52.4 0.67 14.8 2.17 409 0.06 0.48
C5 183040.15 -93625.2 47.4 45.0 66.3 46.2 0.56 23.5 0.41 62 0.01 0.12
C6 183028.44 -93356.0 53.9 40.2 139.6 46.1 0.56 14.0 1.11 174 0.04 0.34
C7 183022.97 -93514.0 57.6 41.7 116.0 48.5 0.60 19.3 0.54 96 0.02 0.15
C8 183024.90 -93626.3 38.0 29.0 137.0 32.3 0.30 15.9 0.98 79 0.06 1.03
C9 183025.54 -93811.2 63.9 42.5 154.6 51.5 0.65 15.0 1.00 190 0.03 0.24
Table 2: Properties of the identified dense clumps

From the map, we have visually identified four filamentary structures which intersect at the center to constitute a hub-filament system. These four filaments are named as F1 to F4 and listed in Table 1 together with some physical parameters. Their loci are overlaid in Figure 2 (a). Nine peaks were identified as clumps and fitted to 2D Gaussian profiles using the 2D Fitting Tool of the CASA viewer. Their FK5 coordinates, and major and minor axes ( and ) are given in Table 2. These clumps are designated as C1 to C9 and delineated as open ellipses in Figure 2 (a). Interestingly, six of the nine clumps are located inside filaments with the most massive one (C1) in the intersecting hub.

Clump masses were estimated via integrating the column densities in the Gaussian ellipses. Source-averaged number densities and mass surface densities were calculated assuming a spherical morphology with a constant density profile and an equivalent radius of , where . The resultant equivalent radii, clump masses, number densities, and source-averaged surface densities are given in Table 2 together with peak column densities and source-averaged dust temperatures. These eight clumps have masses ranging from 62 to 590 , radii ranging from 0.30 to 0.67 pc. The total mass of all clumps is 2189 , holding about 47% of the cloud mass (5943 ).

3.2 Molecular Emission

Figure 3: Velocity-integrated intensity map of CO . The velocity interval for integration is [46, 59] km s. The contours start from 6.5 K km s and increase with a step of 1.0 K km s.

The velocity-integrated intensity map of CO  is shown in Figure 3. Although filamentary features are smoothed out due to the poor angular resolution, the strongest emission is still spatially coincident with the hub region.

Under local thermodaynamic equilibrium (LTE) conditions, the column density of a linear molecule can be expressed as

(3)

Here, is the permanent dipole moment. For the transition, the line strength , the degeneracy , where is the rotational quantum number of the lower state. We have followed McDowell (1988) to write the partition function as where is the rotational constant.

In the calculation, a single excitation temperature was assumed to be 20 K which is approximately the averaged dust temperature of the cloud. The H column density in each pixel was obtained by assuming a isotope ratio (estimated using Equation 4 of Pineda et al., 2013) and a canonical CO abundance of . The resultant column densities range from to cm with a mean of cm. Integrating the whole density map resulted in a total mass of 6117 . We note that CO  is assumed to be optically thin and the resultant cloud mass might be lower limit.

4 Results of SMA Observations

4.1 1.3 mm Continuum Emission

A zoom-in IRAC three color image of C1, the most massive clump in G22, is shown in Figure 2 (c) with and the SMA 1.3 mm continuum emission overlaid as cyan and magenta contours. With the current SMA angular resolution and sensitivity, only one 1.3 mm continuum core has been detected at the center and designated as SMA1.

A close-up view of SMA1 is shown in Figure 2 (d). Two-dimensional Gaussian fitting of the 1.3 mm continuum emission resulted in an ellipse with a major axis of , a minor axis of and a position angle of . The deconvolved major and minor axes are and , smaller than the synthesized beam (). The equivalent angular size of the core is , corresponding to a physical diameter of 0.034 pc ( AU). The flux density of the core is about 562 mJy with a peak intensity of 350 mJy beam.

Assuming the 1.3 mm continuum emission is optically thin, the core dust mass can be obtained via , where is the observed flux density, is the distance and cm g is the dust opacity (Ossenkopf & Henning, 1994). Using the rotational temperature of CHOH ( K, see Section 5.3.1) as the dust temperature, a core mass of is reached. The source averaged H number density is about cm. The beam-averaged column density was estimated to be about cm via .

4.2 Line Emission

Figure 4: SMA LSB and USB spectra toward the peak of SMA1. Lines with peaks no smaller than 3 (240 mJy beam) have been labeled.

Figure 4 shows the full band spectra towards the SMA1 peak. Cross-checking the rest frequencies with the Splatalogue molecular database, we have identified 103 emission lines with peaks where mJy beam. These transitions are from 19 species (26 isotopologues) including nitrogen-bearing species (e.g., HCN, CHCN), complex organic molecules (e.g., CHCHO, CHOCH), and classical tracers of bulk motions (e.g., CO, SiO).

We extracted spectra of all detected transitions towards the SMA1 peak and used the software to perform Gaussian profile fitting to get the line peak intensities, velocity centroids, linewidths, and velocity-integrated intensities. The systemic velocity of SMA1 was estimated to be km s via averaging line centers of transitions with km s and Jy beam.

Inspection of velocity-integrated intensity maps shows that lines from 19 molecules trace a single-core morphology in accord with the 1.3 mm continuum emission. On the other hand, lines of CO, CO, CO, HCO, SO, and SiO reveal extended structures either from the surrounding envelope or entrained outflows (see Section 4.3). Among the 15 detected CHOH lines, 13 trace a single-core spatially coincident with the 1.3 mm core. The other two methanol transitions show anomalous emission in the vicinity (see Section 4.4).

4.3 Molecular Outflows

Figure 5: Outflows revealed by CO (a), SiO (b), and HCO (c). Velocity intervals and noise levels of the red and blue lobes for each line are labeled in the bottom-right corner of each panel. For CO , the blue contours start from and increase by , the blue contours start from and increase by . For SiO , the blue contours start from and increase by , the blue contours start from and increase by . For HCO , the blue contours start from and increase by , the blue contours start from and increase . The 44 GHz Class I methanol masers from Cyganowski et al. (2009) are marked with crosses and color-coded according to their velocities. The green diamond and the black star show the locations of the 1.3 mm core SMA1 and embedded protostar SSTGLMC G022.0387+00.2222 (MIR1).
Lobe
(pc) (km s) () ( km s) ( km s) (10 yr) (10 yr)
B1 0.26 22.47 0.57 12.96 182.94 1.12 5.16
B2 0.27 19.63 1.23 24.22 300.82 1.33 9.25
R1 0.13 27.43 0.37 10.21 158.28 0.47 7.95
R2 0.26 28.18 0.38 10.71 170.64 0.89 4.29
Total 2.56 58.10 812.68 26.65
Table 3: Outflow Parameters

In multi-object surveys, previous works have detected outflow indicators towards EGO G022.04+0.22 (C1 in our work), including class I methanol masers (Cyganowski et al., 2009; Chen et al., 2011), “red-skewed” asymmetric line profiles of CO and HCO (Rygl et al., 2013), and broad SiO emission (Cyganowski et al., 2009). The existence of outflows in C1 is confirmed by the SMA observations presented in this work.

SMA CO shows very broad ( km s with respect to the systemic velocity) wing emission. Velocity-integrated intensities of the blue and red wings (see Figure 5 (a)) reveal two blue-shifted and two red-shifted lobes (B1, B2, R1 and B2). These outflow lobes can be well traced back to the SMA1 peak. The multiple lobes of the outflow also could be due to the presence of two, unresolved protostellar objects. The outflows are also seen in emission from SiO and HCO (see Figure 5 (b) and (c)). The GLIMPSE point source SSTGLMC G022.0387+0.2222 (MIR1, hereafter) which is an early-stage high-mass protostar (see Section 5.3), associated with SMA1 could be the driving source.

Copious 44 GHz Class I CHOH masers detected by Cyganowski et al. (2009) are shown in Figure 5 with crosses (“”). The spatial distribution of the 44 GHz masers shows four distinct groups. The southern group of masers are distributed mainly along a N-S line closely associated with the R1 lobe. The northeast group of masers is spatially coincident with the B2 lobe and the extended 4.5 µm emission. In the north, a crowd of masers are located in between the B1 and B2 lobes. There are also several maser spots on the edge of the B1 lobe. A small group of masers detected in the southwest are not consistent with any high-velocity CO emission. Noticeably, the velocities of maser spots generally agree with those of the associated outflow lobes, supporting their shock-driven origin. There is a reversal of maser velocities at the southern tip of the R1 lobe coincident with a change of outflow direction from southwards to the southeast. This intriguing feature could be due to the influence of the adjacent infrared bubble (see Section 5.4).

Physical parameters of the outflows have been calculated from the CO line wings under LTE conditions and assuming optically thin emission. The CO column density of each pixel in each velocity channel can be straightforwardly derived from Equation 3 as:

(4)

where is the velocity interval in km sand is the measured brightness temperature in K. In the calculation, the excitation temperature has been assumed to be approximately the source-averaged dust temperature of clump C1 which is 21 K. The outflow mass, momentum, energy, dynamical age, mass loss rate, and mechanical force have been obtained from:

(5)
(6)
(7)
(8)
(9)
(10)

Here, is the total solid angle that the flow subtends, is the flow velocity with respect to the systemic velocity, is the characteristic outflow velocity, and is the flow length.

The resultant outflow parameters are given in Table 3. The four lobes have dynamical ages ranging from to yr, comparable to those of other outflows in early high-mass star-forming regions, e.g., G240.31+0.07 ( yr, Qiu et al., 2009), G24.78+0.08 ( yr, Beltrán et al., 2011), G28.34+0.06 ( yr, Wang et al., 2011), G11.11-0.12 ( yr, Wang et al., 2014), and G9.62+0.19 MM6 ( yr, Liu et al., 2017). The relatively small dynamical ages are also consistent with the stellar age estimate from SED fitting ( yr, see Section 5.3.2).

4.4 Millimeter Methanol Maser Emission

Figure 6: (a)-(o): Velocity integrated intensities (contours) of all the 15 CHOH transitions with lowest upper energies overlaid on the 1.3 mm continuum image. Contours for each line start from with an increment of . The relevant transitions, velocity intervals for integration, and noise levels are labeled at the bottom individual panels. Panels (p)-(s) presents spectra of CHOH (), CHOH (), CHOH (), and CHOH () at core SMA1, and positions A, B and C labeled in (b) and (d).

Figures 6 (a)-(o) show the velocity integrated intensity maps of all the 15 detected transitions of CHOH with upper energies ranging from 40 K to 802 K. All CHOH transitions, with the exception of the (at 218.440 GHz with K) and (at 229.759 GHz with K), show single core morphologies. Emission from the and transitions reveals more complex structures. Figure 6 shows maps of the four CHOH transitions with lowest upper energies. Compared to the other methanol transitions, strong CHOH () and CHOH () emission can be detected to the north, south, and northeast of the 1.3 mm core SMA1, spatially coincident with Class I 44 GHz methanol masers, suggestive of maser emission in the CHOH () and CHOH () transitions.

CHOH () and CHOH () spectra towards core SMA1, and positions A, B and C marked in Figures 6 (b) and (d) are presented in Figures 6 (p)-(s). For comparison, spectra of CHOH () at 230.027 GHz with K and CHOH () at 216.946 GHz with K are also shown. CHOH () and CHOH () spectra at position A are narrow, resembling features of maser emission. Relatively broader spectra at positions B and C may be due to the blending of unresolved maser spots.

The peak intensities of CHOH () at positions SMA1, A, B, and C are 4.9, 3.5, 3.5, and 1.1 Jy beam. For CHOH (), the peak intensities at SMA1, A, B, and C are 4.1, 2.9, 2.5, and 1.4 Jy beam. The peak intensities of CHOH () at SMA1, A, B, and C are 1.4, 0.4, 0.2, and 0.1 Jy beam. The line ratio has previously been used as a discriminant between thermal and non-thermal emission, with ratios indicative of non-thermal (Slysh et al., 2002; Cyganowski et al., 2011). The ratios for SMA1, A, B, and C are 3.5, 8.8, 17.5, and 11.0, suggestive of non-thermal emission at positions A, B, and C. We can similarly use the line ratio to distinguish thermal from non-thermal emission. Under LTE and optically thin conditions, the thermal emission ratio of two CHOH transitions can be obtained from

(11)

With an assumed temperature of 21 K (similar to the dust temperature of clump C1) deviating from the 1.3 mm core, the thermal emission ratio will be approximately 3. Thus, we suggest a ratio threshold of 3 as a demarcation of thermal and non-thermal 218.440 GHz emission. The ratios for SMA1, A, B, and C are 2.9, 7.3, 12.5, and 14.0, supportive of non-thermal 218.440 GHz emission at positions A, B, and C. The CHOH () and CHOH () emission features at positions A, B, and C spatially coincide well with 44 GHz Class I CHOH masers. This agreement, along with large and ratios and narrow line widths (especially for position A), gives strong support to the interpretation of millimeter CHOH maser emission.

Maser emission of CHOH () at 229.759 GHz has been discovered by Slysh et al. (2002) in DR21(OH) and DR21 West, and reported in HH 80-81 (Qiu & Zhang, 2009), IRAS 05345+3157 (Fontani et al., 2009), NGC 7538 (Qiu et al., 2011), and EGOs G11.92-0.61, G18.67+0.03 and G19.01-0.03 (Cyganowski et al., 2011, 2012). However, maser emission of the CHOH () transition at 218.440 GHz has, to our knowledge, only been tentatively detected in DR21(OH) by Zapata et al. (2012) and in NGC 6334 I(N) by Hunter et al. (2014) prior to this work. Population inversion for the 218.440 GHz transition has been predicted in maser models (Voronkov et al., 2012). The transition and association with 44 GHz class I masers suggest that the 218.440 GHz CHOH masers detected in G22 are Class I type. The small number of known sources exhibiting maser emission in this transition is likely to be because few previous observations of high-mass star formation regions have covered this line.

5 Discussion

5.1 Global Collapse of the Cloud

Figure 7: (a) Probability distribution function (PDF) of . The blue solid line shows a power-law fit to the high-column density part of the PDF. Here, cm. A conversion factor of cm mag is used to convert column density to visual extinction. (b) Radial column density profile. Each point represents one pixel value in the map. The x-axis gives the distance to clump C1. The dashed line shows a power-law fit. The exponent and the exponent () of the underlying radial number density profile are also labeled.

5.1.1 Density Structure

The probability distribution function (PDF) of is shown in Figure 7 (a). The high-density end of the -PDF of G22 can be fitted to a power-law. Such -PDF profiles have been reported in many IRDCs and IR-bright clouds (e.g., Schneider et al., 2015b, and references therein ). Power-law tails in -PDFs also have been observed in numerical works (e.g., Federrath et al., 2008; Klessen & Hennebelle, 2010; Kritsuk et al., 2011). In both theoretical and observational studies, self-gravity has been suggested to be the dominant process in the formation of the power-law tail. Assuming a spherical symmetry, the power-law slope of the -PDF is related to the exponent of a radial density profile of and (Federrath & Klessen, 2013). An of about is consistent with self-gravity (Girichidis et al., 2014; Schneider et al., 2015a). Such self-gravity can be attributed to local free-fall of individual clumps/cores or global collapse.

Figure 7 (b) shows the radial column density profile which can be well fitted with a power-law . Here is the exponent of the underlying radial number density profile. The resultant is comparable to the from the power-law tail of the -PDF, consistent with gravitational collapse on large scales. This scenario is further supported by large scale velocity gradients discussed in the following section.

Figure 8: GRS CO  channel maps overlaid on the column density map from SED fits. The velocity interval of integration for each subplot covers 1 km s with the center value marked in the lower-right corner. CO Contours start from and increase with a step of . Here, K km s.
Figure 9: (a) Velocity centroids of CO  extracted along filaments overlaid on top of a column density map. Also labeled are the two most massive clumps (black “+” and small ellipses) in the hub region. (b) Spectra of JCMT/HCO  overlaid on the IRAC 8.0 µm image. (c) Spectra of JCMT/CO  overlaid on the SMA/CO outflows. The white diamond in panels (b) and (c) labels the 1.3 mm core SMA1. The large ellipse in (c) delineates clump C1.
Figure 10: (a) Line-of-sight velocity of CO  as a function of position from the potential well centers, i.e., clump C1 for F1, F2, and F4, and clump C2 for F3. (b) Velocity dispersion of CO  as a function of offset from the potential well centers.

5.1.2 Filamentary Collapse

Assuming cylindrical hydrostatic equilibrium, a critical mass per unit length has been proposed to be (Ostriker, 1964; Inutsuka & Miyama, 1997; Kirk et al., 2013; André et al., 2014). With a typical temperature of about 20 K (see Section 3.1 and Figure 2), the critical value is about 33  pc for the filaments in G22. All four filaments are supercritical with significantly larger than the critical value, consistent with the velocity measurements showing that gas is globally infalling in these filaments (see bellow). The mass per unit length for the four filaments ranges from 54 to 220  pc (see Table 1). The of filaments in G22 are comparable to those in DR21 where gas is channeled along the filaments onto the central ridge region (Hennemann et al., 2012).

GRS CO  channel maps are shown in Figure 8. Intriguingly, emission spatially associated with filament F1 mainly originates from gas with velocity smaller than 51 km s while the gas associated with filaments F2 and F3 is systematically red-skewed with velocities larger than 50 km s. The variation of velocity indicates bulk motions in these filaments.

Spectra of CO  have been extracted along the filaments with a step-size of half the angular resolution () of the GRS observations, and fitted to a one-dimensional Gaussian profile using the GILDAS/CLASS software. The resultant centroid velocities of the extracted spectra are shown in Figure 9 (a) as color-coded circles. The smoothly varying velocity along filaments F1, F2, and F3 is reminiscent of the situation in the SDC13 infrared dark clouds where longitudinal filamentary collapse has been reported (Peretto et al., 2014).

Alternative possible physical processes, which can lead to the observed velocity pattern, include collapse, rotation, filament collision, expansion, and wind-driven acceleration (Peretto et al., 2014). The observed velocity gradients along the filaments in G22 do not allow us to determine a common axis for rotation to take place unless the axis goes through the filament junction. However, such a scenario is unrealistic as differential rotation would tear apart these filaments (Peretto et al., 2014). No signs of enhanced linewidth in the interaction and bridging features (Benjamin et al., 2003; Gong et al., 2017) can be observed in the junction, excluding the possibility of filament collision. No H ii region is observed at the center of G22. And the outflows driven by MIR1 in clump C1 cannot explain the red-skewed velocity along F2 as the outflowing gas in that direction is blue-shifted (see Figure 5). Similar to the situation in SDC13 (Peretto et al., 2014), there also resides a mid-IR bright nebula to the southwest of G22 (see Figures 1). This mid-IR nebula corresponds to an infrared bubble MWP1G022027+002159 and is interacting with G22 (see Section 5.4). As MWP1G022027+002159 is located slightly behind G22 (see Section 5.4), its expansion would lead to systematic blue-shifted velocities in the interaction region, but the affected area is limited. The largely blue-skewed velocity in the most distant section of F1 and the red-skewed velocity in F2 cannot be explained by wind-driven expansion.

Figure 10 (a) shows the difference between the velocities of the filaments and the junction as a function of distance to the center. Monotonically increasing profiles for F1, F2, and F3 are consistent with those observed in other reported collapsing filaments (e.g., SDC13 and AFGL 5142 in Peretto et al., 2014; Liu et al., 2016). The estimated velocity gradients in F1, F2 and F3 are 0.36, 0.35 and 0.21 km s pc, which are comparable to those in SDC13 (, Peretto et al., 2014) , and smaller than those in The Serpens South cluster (, Kirk et al., 2013) and AFGL 5142 (, Liu et al., 2016). The small velocity gradients in G22 may indicate low filament inclination.

The velocity dispersion of the filaments as a function of distance to the center is shown in Figure 10 (b). For filaments F3 and F4, an increasing trend of velocity dispersion can be seen toward the center region. This increase could be due to additional kinetic energy converted from gravitational energy during the collapse of the filaments (Peretto et al., 2014). Such increase of velocity dispersion can also be observed toward the inner part of filament F1. In contrast, filament F2 has significantly larger velocity dispersions. As CO  could be optically thick, observations of more optically thin lines, such as NH, would help better reveal the velocity dispersions.

Following Kirk et al. (2013), we estimated the mass infall rates to be 132, 206 and 104  Myr in F1, F2, and F3. Here, we assumed an inclination angle of 45. Therefore, about 440  gas will be channeled to the junction regions in Myr if the current accretion rates are sustained. This high accretion will double the mass of the hub region in about 6 free-fall times where the free-fall time is about 0.31 Myr. Although two million years of high-accretion may not be plausible, according to models of the evolution of high-mass star, it is tenable to posit that a fraction of the masses in the central clumps C1 and and C2 have been assembled through large-scale filamentary collapse.

5.2 Clump-Fed Accretion in Clump C1

Figure 11: (a) JCMT HCO  , CO , and C spectra averaged over clump C1. (b) SMA CO , and C and DCN  spectra at the peak of SMA1.
Figure 12: Spectrum of HCO  averaged over clump C1. The HCO  spectra obtained from RATRAN modeling of a collapsing clump are shown as blue. Details about the modeling are given in Section 5.2. Here, we only show the models with reduced smaller than 3. The HCO abundance, infall velocity, and velocity dispersion are labeled for each model.

A virial parameter of for clump C1 has been estimated based on a velocity dispersion of km s measured from the JCMT C data. With an smaller than 1, C1 is likely dominated by gravity and potentially collapsing. Magnetic fields might provide additional support against gravity. Following Pillai et al. (2011), we estimated that the magnetic field strength necessary to virialize C1 is G, which is comparable to estimates for the field in clouds at similar densities (Crutcher, 2012; Li et al., 2014). Note that the virial parameter may have been overestimated, as any systematic motions, such as infall and outflow, could significantly increase the velocity-dispersion estimate.

Figure 9 (b) and (c) show JCMT HCO  and CO  spectra of clump C1 overlaid on the IRAC/8.0 µm image and SMA CO outflows. All spectra with significant emission () for both HCO  and CO  show blue-skewed self-absorbed profiles, with the exception of some spectra at the center (for both lines) and the north (for CO ). Such blue profiles could originate from collapsing or rotating clouds with an inward increasing temperature profile. The different line profiles in the center could be due to the inclusion of emission from the outflow from SMA1. Although we cannot exclude the contribution of rotation, inward motions must occur to explain the dominance of blue profiles in both lines, especially in the clump-confined region (see the large circle in Figure 9 (c)). This is further supported by the blue-skewed clump-averaged spectra shown in Figure 11 (a). Intriguingly, the absorption dips of the source averaged HCO and CO spectra are both red-skewed with respect to the systemic velocity. This feature is consistent with a global collapse scenario as suggested in Jin et al. (2016).

We estimated the infall velocity by fitting the clump-averaged HCO spectrum using the RATRAN 1D Monte Carlo radiative transfer code (Hogerheijde & van der Tak, 2000). The input parameters include the clump size, density profile, kinematic temperature profile, HCO fractional abundance (),and infall velocity. We deduced a power-law density profile () based on the mass ( ) and radius ( pc) of C1. Following Peretto et al. (2013), we used a constant temperature which is approximately the source-averaged dust temperature of C1 (21 K, see Table 2). We ran a grid of 880 models with varying HCO abundance in the range , infall velocity in the range [] km s, and velocity dispersion in the range km s. A reduced parameter was calculated for the averaged spectrum of each model. Note that only the central 4 km s portion was considered in calculating reduced as the emission from the outflows has a large impact on the outer channels of the HCO spectrum. Fifty-six models with were considered to be good fits, here . Figure 12 shows the averaged spectra of HCO  of models with . From weighted parameters of all models with good fits, we derived the infall velocity (), velocity dispersion (), and HCO abundance to be km s, km s, and , respectively. We also investigated the affect of varying the radius of the collapse and found that the HCO  can be well modeled only for pc. This is consistent with the observed blue-skewed CO  spectra beyond the extent of clump C1 (see Figure 9 (c)) and further supports a clump-scale collapse scenario.

We estimated the current mass infall rate using to be about  yr. Here, is assumed. The mass infall rate is about five times smaller than the free-fall accretion rate (  yr) but still significantly larger than the mass outflow rate of SMA1, suggesting that the central hot core can keep growing in mass via clump-fed accretion. Using a velocity dispersion of 0.87 km s measured from the JCMT C, we estimate the crossing time of C1 to be 0.43 Myr which is about 2.5 free-fall times ( Myr). In spite of a difference of about a factor 2, the crossing and free-fall times agree order-of-magnitude wise considering they are just approximate values. This further supports a dynamical global collapse scenario for clump C1.

5.3 SMA1: A Collapsing Hot Molecular Core

Figure 13: Rotation diagrams of CHOH (a) and CHCN (b).

Most of the 103 molecular lines detected towards the SMA1 peak are from complex organic molecules. This, together with the small size (0.034 pc) and high density ( cm), suggests that SMA1 is a hot molecular core. The physical properties of hot cores are characterized by a small source size ( pc), a high density ( cm), and warm gas/dust temperature ( K) (Kurtz et al., 2000; van der Tak, 2004).

5.3.1 High Gas Temperature

Fifteen CHOH transitions with ranging from 40 to 802 K and eight CHCN transitions with ranging from 69 to 419 K have been reliably detected toward SMA1. Assuming that these lines are optically thin and SMA1 is in LTE, beam-averaged temperatures and column densities of CHOH and CHCN can be obtained using the rotational diagram method (Goldsmith & Langer, 1999). The column density () of the upper state of a transition can be expressed as a function of its upper state energy(),

(12)

where

(13)

Here, is the upper state degeneracy, is the beam-averaged total column density of a species, is the partition function at a given temperature, is the rotation temperature, is the Einstein A-coefficient for the transition, and is the velocity integrated intensity of a specific line.

The partition function of CHOH and CHCN can be expressed as,

(14)

The constants of and were obtained by fitting the JPL partition function values at 9.375 to 300 K.

With multiple transitions observed, which have upper state energies spanning through a large range, the column density and rotation temperature of a specific molecule can be obtained via a least-square fitting to as a linear function of .

The fitting results are shown in Figure 13. A rotation temperature of K and a column density of cm have been attained for CHOH. The resulting temperature and column density for CHCN are K and cm. The gas temperatures of SMA1 are consistent with those of hot cores in our Galaxy (120–480 K, Hernández-Hernández et al., 2014).

5.3.2 MIR1: an Embedded High-Mass Protostar

Figure 14: Spectral energy distribution of MIR1 (SSTGLMC G022.0387+00.2222). Photometric data from GLIMPSE, MIPSGAL, Hi-GAL, and ATLASGAL were fitted using YSO models of Robitaille et al. (2006, 2007). The black line shows the best fit, and the gray lines show subsequent good fits with .
Figure 15: (a) Velocity integrated intensity of CHOH (A+) overlaid on the 1.3 mm continuum image. Contours start from with a increasing step of . (b) Spectrum of CHOH (A+) at the SMA1 peak. The vertical dashed line marks the systemic velocity (50.4 km s) of SMA1.

Near the peak of SMA1, there is a mid-IR point source (SSTGLMC G022.0387+0.2222, MIR1 for short). Photometric data collected from the GLIMPSE, MIPSGAL, Hi-GAL, and ATLASGAL surveys have been fitted to models of YSOs developed by Robitaille et al. (2006, 2007). The resultant spectral energy distribution of MIR1 is shown in Figure 14. The good fits suggest that MIR1 is a Stage 0/I type high-mass prototar with a stellar mass of , a total luminosity of , an envelope accretion rate of  yr, and an age of yr.

5.3.3 Possible Core-fed Accretion

As shown in Figure 11 (b), blue profile and inverse P-cygni profile have been observed in SMA CO  and C spectra. A similar blue-skewed profile is also seen in CHOH (A+) (see Figure 15). Although the self-absorption in these lines can be partially due to the short spacing issues related to interferometer observations, blue-skewed profiles still could be tracing some dynamical motions. Additionally, the inverse P-cygni profile of C cannot be fully explained by short spacing problem, and has been frequently interpreted as a convincing tracer of infall motions in other interferometer observations (Wu et al., 2009; Liu et al., 2011b, a, 2013; Qiu et al., 2011, 2012). And the short spacing problem would filter structures larger than 20″and cannot significantly alter the profile of CHOH (A+) as this line mainly traces the dense core (see Figure 15 (a)). Careful inspection of moment 1 maps of all transitions shows there is no detectable rotation in SMA1 with current SMA observations. This suggests that the blue-skewed profiles could be due to infall motions. Blue profiles of CO  based on SMA observations have also been interpreted as tracing infall in other star-forming regions (e.g., Lee et al., 2006; Wu et al., 2009; Zhang et al., 2014).

Using the model of Myers et al. (1996), an infall velocity of about 0.16 km s was estimated. Assuming a density profile of , we obtain a mass infall rate of  yr which is a quarter of the free-fall accretion rate (  yr) and comparable to mass accretion rates of high-mass star-forming cores in numerical studies (Keto, 2007; Kuiper et al., 2010, 2011) as well as in interferometric observations (e.g., Liu et al., 2013; Wu et al., 2014). Assuming that about half of the gas flow can be successfully channeled onto the central protostar, MIR1 would grow to an O9 star in about yr if such accretion rate is sustained.

In summary, the gas in G22 is being channeled into the central hub region through global filamentary collapse. The most massive clump C1 at the center is collapsing and feeding the embedded hot molecular core SMA1. The high-mass protostar MIR1 in SMA1 is also gaining mass from SMA1 where infall motions have been tentatively detected.

5.4 Influence from the Adjacent Bubble

The infrared bubble MWP1G022027+0022159 has been observed in the south of G22 (see Figures 1 and 2 (b)). The dark lane observed to the near-IR suggests the bubble is behind G22. This is also supported by the changes of velocities from red to blue for masers associated with the outflow lobe R1 (Figure 5). The reversal of maser velocities and large intensity gradients of R1 suggest that the bubble might be strongly interacting with G22 and modifying the direction of the outflowing gas.

As the bubble is slightly behind G22, its expansion is apt to push the gas in G22 towards us. And the interacting region would have a higher temperature and more blue-shifted velocities. Thus, any existing self-absorption due to the bubble would result in line profiles with a higher red peak (i.e., red profiles). The ubiquitous blue profiles in C1 cannot originate from the the influence of the IR bubble. However, an expansion of the bubble would introduce a blue wing to the spectrum of HCO  as shown in Figures 11 and 12.

6 Summary

We have presented a detailed study of a hub-filament system in the G22 cloud. The main findings of this work are summarized as follows.

  1. The G22 cloud is composed of four supercritical filaments with mass per unit length ranging from 54 to 220  pc. Velocity gradients along three filaments revealed by CO  indicates they are collapsing and channeling gas towards the junction. A total mass infall rate of 440  Myr suggests that the hub mass would be doubled in 6 free-fall times. Although such a high accretion may not be sustained for almost two million years, it is tenable to posit that a fraction of the mass in the central clumps C1 and C2 have been built through large-scale filamentary collapse.

  2. The most massive and densest clump C1 at the junction is globally collapsing. With a virial parameter , C1 is supercritical unless strong magnetic fields can provide additional support. Prevalent blue-profiles of HCO  and CO  spectra support a scenario of clump-scale collapse. An infall velocity of 0.31 km s was estimated via modeling the clump-averaged HCO spectrum using the RATRAN code. The inferred mass infall rate is about  yr.

  3. Embedded in clump C1, a hot molecular core (SMA1) has been revealed by the SMA observations. More than 100 lines from 19 species (26 isotopologues including complex organic molecules) have been detected. The gas temperature was estimated to be higher than 200 K via fitting the rotation diagrams of CHOH and CHCN. The detected outflows and the Class-I like SED of the embedded mid-infrared source suggest that high-mass star formation is taking place in SMA1. The high-mass star-forming hot core SMA1 may be still growing in mass via clump-fed accretion. This is supported by the high mass infall rate of C1 which is significantly larger than the mass loss rate of the outflows.

  4. A high-mass protostar (MIR1) is located at the center of SMA1 and could be the driving source of the observed outflows. The mass of MIR1 is estimated to be in the range []  and may be still growing via core-fed accretion which is supported by the detected infall features in SMA1.

  5. Anomalous methanol emission of CHOH () at 229.759 GHz and CHOH () at 218.440 GHz was detected in three positions around SMA1. Large and ratios suggest maser emission of these two methanol transitions.

The coexistence of infall through filaments, clumps and the central core has revealed continuous mass growth from large to small scales. This suggest that pre-assembled mass reservoirs may not be indispensable to form high-mass stars. In the process of high-mass star formation, the masses of the central protostar, the core, and the clump can simultaneously grow via core-fed/disk accretion, clump-fed accretion and, filamentary/global collapse.

We are grateful to the anonymous referee for the constructive comments that helped us improve this paper. Prof. Neal J. Evans II and Dr. Kee-Tae Kim are also acknowledged for useful discussions. This work is supported by the National Natural Science Foundation of China through grants of 11503035, 11573036, 11373009, 11433008 and 11403040, 11403041, the International S&T Cooperation Program of China through the grant of 2010DFA02710. Tie Liu is supported by the EACOA fellowship. KW is supported by grant WA3628-1/1 through the DFG priority program 1573 “Physics of the Interstellar Medium”.

References

  • André et al. (2014) André, P., Di Francesco, J., Ward-Thompson, D., et al. 2014, Protostars and Planets VI, 27
  • André et al. (2010) André, P., Men’shchikov, A., Bontemps, S., et al. 2010, A&A, 518, L102
  • Aniano et al. (2011) Aniano, G., Draine, B. T., Gordon, K. D., & Sandstrom, K. 2011, PASP, 123, 1218
  • Astropy Collaboration et al. (2013) Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al. 2013, A&A, 558, A33
  • Beltrán et al. (2011) Beltrán, M. T., Cesaroni, R., Zhang, Q., et al. 2011, A&A, 532, A91
  • Benjamin et al. (2003) Benjamin, R. A., Churchwell, E., Babler, B. L., et al. 2003, PASP, 115, 953
  • Bonnell et al. (2001) Bonnell, I. A., Bate, M. R., Clarke, C. J., & Pringle, J. E. 2001, MNRAS, 323, 785
  • Chen et al. (2011) Chen, X., Ellingsen, S. P., Shen, Z.-Q., Titmarsh, A., & Gan, C.-G. 2011, ApJS, 196, 9
  • Churchwell et al. (2009) Churchwell, E., Babler, B. L., Meade, M. R., et al. 2009, PASP, 121, 213
  • Crutcher (2012) Crutcher, R. M. 2012, ARA&A, 50, 29
  • Csengeri et al. (2011) Csengeri, T., Bontemps, S., Schneider, N., Motte, F., & Dib, S. 2011, A&A, 527, A135
  • Csengeri et al. (2016) Csengeri, T., Weiss, A., Wyrowski, F., et al. 2016, A&A, 585, A104
  • Cyganowski et al. (2009) Cyganowski, C. J., Brogan, C. L., Hunter, T. R., & Churchwell, E. 2009, ApJ, 702, 1615
  • Cyganowski et al. (2011) Cyganowski, C. J., Brogan, C. L., Hunter, T. R., Churchwell, E., & Zhang, Q. 2011, ApJ, 729, 124
  • Cyganowski et al. (2012) Cyganowski, C. J., Brogan, C. L., Hunter, T. R., et al. 2012, ApJ, 760, L20
  • Cyganowski et al. (2008) Cyganowski, C. J., Whitney, B. A., Holden, E., et al. 2008, AJ, 136, 2391
  • Federrath & Klessen (2013) Federrath, C., & Klessen, R. S. 2013, ApJ, 763, 51
  • Federrath et al. (2008) Federrath, C., Klessen, R. S., & Schmidt, W. 2008, ApJ, 688, L79
  • Fontani et al. (2009) Fontani, F., Zhang, Q., Caselli, P., & Bourke, T. L. 2009, A&A, 499, 233
  • Girichidis et al. (2014) Girichidis, P., Konstandin, L., Whitworth, A. P., & Klessen, R. S. 2014, ApJ, 781, 91
  • Goldsmith & Langer (1999) Goldsmith, P. F., & Langer, W. D. 1999, ApJ, 517, 209
  • Gong et al. (2017) Gong, Y., Fang, M., Mao, R., et al. 2017, ApJ, 835, L14
  • Hennemann et al. (2012) Hennemann, M., Motte, F., Schneider, N., et al. 2012, A&A, 543, L3
  • Hernández-Hernández et al. (2014) Hernández-Hernández, V., Zapata, L., Kurtz, S., & Garay, G. 2014, ApJ, 786, 38
  • Hogerheijde & van der Tak (2000) Hogerheijde, M. R., & van der Tak, F. F. S. 2000, A&A, 362, 697
  • Hunter et al. (2014) Hunter, T. R., Brogan, C. L., Cyganowski, C. J., & Young, K. H. 2014, ApJ, 788, 187
  • Inutsuka & Miyama (1997) Inutsuka, S.-i., & Miyama, S. M. 1997, ApJ, 480, 681
  • Ioannidis & Froebrich (2012) Ioannidis, G., & Froebrich, D. 2012, MNRAS, 421, 3257
  • Jackson et al. (2006) Jackson, J. M., Rathborne, J. M., Shah, R. Y., et al. 2006, ApJS, 163, 145
  • Jin et al. (2016) Jin, M., Lee, J.-E., Kim, K.-T., & Evans, II, N. J. 2016, ApJS, 225, 21
  • Kauffmann et al. (2008) Kauffmann, J., Bertoldi, F., Bourke, T. L., Evans, II, N. J., & Lee, C. W. 2008, A&A, 487, 993
  • Keto (2007) Keto, E. 2007, ApJ, 666, 976
  • Kirk et al. (2013) Kirk, H., Myers, P. C., Bourke, T. L., et al. 2013, ApJ, 766, 115
  • Klessen & Hennebelle (2010) Klessen, R. S., & Hennebelle, P. 2010, A&A, 520, A17
  • Kritsuk et al. (2011) Kritsuk, A. G., Norman, M. L., & Wagner, R. 2011, ApJ, 727, L20
  • Kuiper et al. (2010) Kuiper, R., Klahr, H., Beuther, H., & Henning, T. 2010, ApJ, 722, 1556
  • Kuiper et al. (2011) —. 2011, ApJ, 732, 20
  • Kurtz et al. (2000) Kurtz, S., Cesaroni, R., Churchwell, E., Hofner, P., & Walmsley, C. M. 2000, Protostars and Planets IV, 299
  • Lee et al. (2006) Lee, C.-F., Ho, P. T. P., Beuther, H., et al. 2006, ApJ, 639, 292
  • Li et al. (2014) Li, H.-B., Goodman, A., Sridharan, T. K., et al. 2014, Protostars and Planets VI, 101
  • Liu et al. (2012) Liu, H. B., Quintana-Lacaci, G., Wang, K., et al. 2012, ApJ, 745, 61
  • Liu et al. (2011a) Liu, T., Wu, Y., Liu, S.-Y., et al. 2011a, ApJ, 730, 102
  • Liu et al. (2013) Liu, T., Wu, Y., Wu, J., Qin, S.-L., & Zhang, H. 2013, MNRAS, 436, 1335
  • Liu et al. (2011b) Liu, T., Wu, Y., Zhang, Q., et al. 2011b, ApJ, 728, 91
  • Liu et al. (2016) Liu, T., Zhang, Q., Kim, K.-T., et al. 2016, ApJ, 824, 31
  • Liu et al. (2017) Liu, T., Lacy, J., Li, P. S., et al. 2017, ArXiv e-prints, arXiv:1705.04907
  • McDowell (1988) McDowell, R. S. 1988, J. Chem. Phys., 88, 356
  • McKee & Tan (2003) McKee, C. F., & Tan, J. C. 2003, ApJ, 585, 850
  • Molinari et al. (2010a) Molinari, S., Swinyard, B., Bally, J., et al. 2010a, A&A, 518, L100
  • Molinari et al. (2010b) —. 2010b, PASP, 122, 314
  • Motte et al. (2017) Motte, F., Bontemps, S., & Louvet, F. 2017, ArXiv e-prints, arXiv:1706.00118
  • Myers et al. (1996) Myers, P. C., Mardones, D., Tafalla, M., Williams, J. P., & Wilner, D. J. 1996, ApJ, 465, L133
  • Ossenkopf & Henning (1994) Ossenkopf, V., & Henning, T. 1994, A&A, 291, 943
  • Ostriker (1964) Ostriker, J. 1964, ApJ, 140, 1056
  • Peretto & Fuller (2009) Peretto, N., & Fuller, G. A. 2009, A&A, 505, 405
  • Peretto et al. (2013) Peretto, N., Fuller, G. A., Duarte-Cabral, A., et al. 2013, A&A, 555, A112
  • Peretto et al. (2014) Peretto, N., Fuller, G. A., André, P., et al. 2014, A&A, 561, A83
  • Pety (2005) Pety, J. 2005, in SF2A-2005: Semaine de l’Astrophysique Francaise, ed. F. Casoli, T. Contini, J. M. Hameury, & L. Pagani, 721
  • Pillai et al. (2011) Pillai, T., Kauffmann, J., Wyrowski, F., et al. 2011, A&A, 530, A118
  • Pineda et al. (2013) Pineda, J. L., Langer, W. D., Velusamy, T., & Goldsmith, P. F. 2013, A&A, 554, A103
  • Planck Collaboration et al. (2014) Planck Collaboration, Ade, P. A. R., Aghanim, N., et al. 2014, A&A, 564, A45
  • Qiu & Zhang (2009) Qiu, K., & Zhang, Q. 2009, ApJ, 702, L66
  • Qiu et al. (2012) Qiu, K., Zhang, Q., Beuther, H., & Fallscheer, C. 2012, ApJ, 756, 170
  • Qiu et al. (2011) Qiu, K., Zhang, Q., & Menten, K. M. 2011, ApJ, 728, 6
  • Qiu et al. (2009) Qiu, K., Zhang, Q., Wu, J., & Chen, H.-R. 2009, ApJ, 696, 66
  • Reid et al. (2016) Reid, M. J., Dame, T. M., Menten, K. M., & Brunthaler, A. 2016, ApJ, 823, 77
  • Ren et al. (2012) Ren, Z., Wu, Y., Zhu, M., et al. 2012, MNRAS, 422, 1098
  • Robitaille et al. (2007) Robitaille, T. P., Whitney, B. A., Indebetouw, R., & Wood, K. 2007, ApJS, 169, 328
  • Robitaille et al. (2006) Robitaille, T. P., Whitney, B. A., Indebetouw, R., Wood, K., & Denzmore, P. 2006, ApJS, 167, 256
  • Robitaille et al. (2008) Robitaille, T. P., Meade, M. R., Babler, B. L., et al. 2008, AJ, 136, 2413
  • Rygl et al. (2013) Rygl, K. L. J., Wyrowski, F., Schuller, F., & Menten, K. M. 2013, A&A, 549, A5
  • Sault et al. (1995) Sault, R. J., Teuben, P. J., & Wright, M. C. H. 1995, in Astronomical Society of the Pacific Conference Series, Vol. 77, Astronomical Data Analysis Software and Systems IV, ed. R. A. Shaw, H. E. Payne, & J. J. E. Hayes, 433
  • Schneider et al. (2010) Schneider, N., Csengeri, T., Bontemps, S., et al. 2010, A&A, 520, A49
  • Schneider et al. (2015a) Schneider, N., Ossenkopf, V., Csengeri, T., et al. 2015a, A&A, 575, A79
  • Schneider et al. (2015b) Schneider, N., Csengeri, T., Klessen, R. S., et al. 2015b, A&A, 578, A29
  • Schuller et al. (2009) Schuller, F., Menten, K. M., Contreras, Y., et al. 2009, A&A, 504, 415
  • Simpson et al. (2012) Simpson, R. J., Povich, M. S., Kendrew, S., et al. 2012, MNRAS, 424, 2442
  • Slysh et al. (2002) Slysh, V. I., Kalenskiĭ, S. V., & Val’tts, I. E. 2002, Astronomy Reports, 46, 49
  • Tan et al. (2014) Tan, J. C., Beltrán, M. T., Caselli, P., et al. 2014, Protostars and Planets VI, 149
  • Tigé et al. (2017) Tigé, J., Motte, F., Russeil, D., et al. 2017, ArXiv e-prints, arXiv:1703.09839
  • van der Tak (2004) van der Tak, F. F. S. 2004, in IAU Symposium, Vol. 221, Star Formation at High Angular Resolution, ed. M. G. Burton, R. Jayawardhana, & T. L. Bourke, 59
  • Voronkov et al. (2012) Voronkov, M. A., Caswell, J. L., Ellingsen, S. P., et al. 2012, in IAU Symposium, Vol. 287, IAU Symposium, ed. R. S. Booth, W. H. T. Vlemmings, & E. M. L. Humphreys, 433–440
  • Wang et al. (2016) Wang, K., Testi, L., Burkert, A., et al. 2016, ApJS, 226, 9
  • Wang et al. (2015) Wang, K., Testi, L., Ginsburg, A., et al. 2015, MNRAS, 450, 4043
  • Wang et al. (2011) Wang, K., Zhang, Q., Wu, Y., & Zhang, H. 2011, ApJ, 735, 64
  • Wang et al. (2014) Wang, K., Zhang, Q., Testi, L., et al. 2014, MNRAS, 439, 3275
  • Werner et al. (2004) Werner, M. W., Roellig, T. L., Low, F. J., et al. 2004, ApJS, 154, 1
  • Wu et al. (2014) Wu, Y., Liu, T., & Qin, S.-L. 2014, ApJ, 791, 123
  • Wu et al. (2009) Wu, Y., Qin, S.-L., Guan, X., et al. 2009, ApJ, 697, L116
  • Yuan et al. (2017) Yuan, J., Wu, Y., Ellingsen, S. P., et al. 2017, ApJS, 231, 11
  • Zapata et al. (2012) Zapata, L. A., Loinard, L., Su, Y.-N., et al. 2012, ApJ, 744, 86
  • Zhang et al. (2014) Zhang, C.-P., Wang, J.-J., Xu, J.-L., Wyrowski, F., & Menten, K. M. 2014, ApJ, 784, 107
Figure 16: Spectra of CO 1-0 along filaments.The red line in each panel shows the one-dimensional Gaussian fitting result.
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
""
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
   
Add comment
Cancel
Loading ...
174663
This is a comment super asjknd jkasnjk adsnkj
Upvote
Downvote
""
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters
Submit
Cancel

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test
Test description