The TOP-SCOPE Survey of Planck Galactic Cold Clumps: Survey Overview and Results of an Exemplar Source, PGCC G26.53+0.17

Planck was the third-generation mission to measure the anisotropy of the cosmic microwave background radiation, and it observed the entire sky in nine frequency bands (between 30 and 857 GHz). The high-frequency channels of Planck cover the peak of the thermal emission spectrum of dust colder than 14 K (Planck Collaboration et al. 2011b, 2016), indicating that Planck could probe the coldest parts of the ISM. The Planck team has cataloged 13,188 Planck galactic cold clumps (PGCCs), which are distributed across the whole sky, i.e., from the Galactic plane to high latitudes, following the spatial distribution of the main molecular cloud complexes. All 13188 PGCCs are overlaid on the Planck map in Figure 1.

Zoom In Zoom Out Reset image size

The catalog of PGCCs was generated using the CoCoCoDeT algorithm (Montier et al. 2010). The method uses spectral information to locate sources of cold dust emission. Each of the Planck 857, 545, and 353 GHz maps is compared to the IRAS 100 μm data. The 100 μm maps and the local average ratio of the Planck and IRAS bands represent the distribution of warm dust emission. The spectrum of this extended component is estimated from an annulus between 5' and 15' from the center position. This also sets an upper limit to the angular size of the detected clumps. Once the template of warm dust emission is subtracted from the Planck data, the residual maps of cold dust emission can be used for source detection. The initial catalogs were generated for the three Planck bands separately and were then merged, requiring a detection and a signal-to-noise ratio (S/N) greater than 4 in all three bands.

The properties of the PGCC catalog are described in Planck Collaboration et al. (2016). The source spectral energy distributions (SEDs; based on the four bands used in the source detection) give color temperatures 6–20 K with a median value of ∼14 K. About 42% of the sources have reliable distance estimates that were derived with methods such as 3D extinction mapping, kinematic distances, and association with known cloud complexes. The distance distribution extends from 0.1 to 10 kpc with a median of ∼0.4 kpc. Because of the low angular resolution of the data ($\sim 5^{\prime}$) and the wide distribution of distances, the PGCC catalog contains a heterogeneous set of objects from nearby low-mass clumps to distant massive clouds. This is reflected in the physical parameters where the object sizes range from 0.1 to over 10 pc and the masses from below 0.1 to over 10⁴ M_☉.

The PGCC catalog is the first homogeneous survey of cold and compact Galactic dust clouds that extends over the whole sky. The main common feature of the PGCC sources is the detection of a significant excess of cold dust emission. This is suggestive of high column densities, a fact that has also been corroborated by subsequent Herschel studies (Planck Collaboration et al. 2011b, 2016; Montillaud et al. 2015). The catalog covers a wide range of galactocentric distances (almost 0–15 kpc) and a wide range of environments from quiescent high-latitude clouds to active star-forming clouds and sites of potential triggered star formation. This makes the PGCC catalog a good starting point for many studies of the star formation process. Figure 2 shows the distributions of several parameters for the PGCC catalog sources.

Zoom In Zoom Out Reset image size

A large fraction of PGCCs seem to be quiescent, not affected by ongoing star-forming activity (Wu et al. 2012; Liu et al. 2014). Those sources are prime candidates for probing how prestellar cores form and evolve, as well as the initial stages of star formation across a wide variety of Galactic environments. The detection of gravitationally bound CO gas clumps (Liu et al. 2012a; Meng et al. 2013; Zhang et al. 2016) and dense molecular line tracers (Yuan et al. 2016; Tatematsu et al. 2017) inside PGCCs strongly suggests that many PGCCs have the ability to form stars. Moreover, their low level of CO gas depletion indicates that the natal clouds of PGCCs are still in the early stages of molecular cloud evolution (Liu et al. 2013b). A number of PGCCs were purposely followed up with observations with the Herschel satellite. These higher-resolution data revealed great variety in the morphology of the PGCCs (Planck Collaboration et al. 2011a; Juvela et al. 2012; Montillaud et al. 2015). The clumps typically have significant substructure and are often (but not always) associated with cloud filaments (Rivera-Ingraham et al. 2016). Further clues to the nature of the clumps are provided by the dust emission properties, the PGCCs showing particularly high values for submillimeter opacity and the spectral index β (Juvela et al. 2015a, 2015b). Prestellar cores and extremely young Class 0 objects have been detected in PGCCs (Liu et al. 2016b; Tatematsu et al. 2017), indicating that some PGCCs can be used to trace the initial conditions when star formation commences.

Because of the uniqueness and importance of PGCC sources to understanding the earliest stages of star formation, we have been conducting a series of surveys to characterize the physical and dynamical state of PGCCs. These surveys are described in the following sections.

"TOP" is a Key Science Program (KSP) of the Taeduk Radio Astronomy Observatory (TRAO). It is a survey of the J = 1–0 transitions of ¹²CO and ¹³CO toward ∼2000 PGCCs. Some interesting PGCCs were also mapped in the C¹⁸O (1–0) line. The "TOP" survey was initiated in 2015 December. The main aims of "TOP" are to find CO dense clumps; to study the universality and ubiquity of filamentary structures in the cold ISM; to study cloud evolution in conjunction with H i surveys; to investigate how CO abundances change with the evolutionary status of the cloud, and across a range of different environments; to investigate the roles of turbulence, magnetic fields, and gravity in structure formation; and to investigate the dynamical effect of stellar feedback and/or cloud–cloud collisions on star formation.

(1) The TRAO 14 m telescope

The Taeduk Radio Astronomy Observatory (TRAO⁷⁸ ) was established in 1986 October with the 14 m radio telescope located on the campus of the Korea Astronomy and Space Science Institute (KASI) in Daejeon, South Korea. The surface accuracy of the primary reflector is <130 μm rms, and the telescope pointing accuracy is better than 10'' rms. The main FWHM beam sizes for the ¹²CO (1–0) and ¹³CO (1–0) lines are 45'' and 47'', respectively. The new receiver system, SEQUOIA-TRAO, is equipped with high-performance 16-pixel MMIC preamplifiers in a 4 × 4 array, operating in the 85–115 GHz frequency range. The noise temperature of the receiver is 50–80 K over most of the band. The system temperature ranges from 150 K (86–110 GHz) to 450 K (115 GHz; ¹²CO). The second IF modules with a narrow band and the eight channels with four fast Fourier transform (FFT) spectrometers allow simultaneous observations at two frequencies within the 85–100 GHz or 100–115 GHz bands for all 16 pixels of the receiver. The backend system (an FFT spectrometer) provides 4096 × 2 channels with fine velocity resolution of 0.04 km s⁻¹ (15 kHz) per channel and full spectral bandwidth of 60 MHz (∼160 km s⁻¹ for ¹²CO).

(2) Source selection

The TOP survey has a target list that probes the early stages of star formation across a wide range of Galactic environments. The starting point is the PGCC catalog, which contains all cold dust clouds detected by Planck, in combination with the IRAS 100 μm data. Given the very high sensitivity of the Planck measurements, the PGCC catalog is currently the most complete unbiased all-sky catalog of cold clumps.

The TOP target list was constructed according to the following selection criteria. We excluded regions targeted as part of the previous Galactic plane (Moore et al. 2015) or Gould Belt surveys (Ward-Thompson et al. 2007) at the James Clerk Maxwell Telescope (JCMT⁷⁹ ) because these locations have been extensively observed by other telescopes (e.g., JCMT, NRO 45 m, and PMO 14 m) both by molecular line and by dust continuum emission observations. The TOP survey is therefore highly complementary to these previous surveys. The PGCCs in these excluded regions will also be investigated in our statistical studies using the archived data.

To ensure a representative study of the full PGCC population, the sources were divided into bins of Galactic longitude (every 30°), latitude (divided by $| b| =0^\circ ,4^\circ ,10^\circ$, and 90°), and distance (divided by d = 0, 200, 500, 1000, 2000, and 8000 pc and unknown), and targets were selected from each bin. Similarly, we cover the full range of source temperatures and column densities in Planck measurements. The sampling is weighted toward the very coldest sources. The sample includes 787 very cold sources for which the PGCC catalog only gives a temperature upper limit. To ensure a good detection rate, the sampling is weighted toward higher column densities. Within the constraints listed above, preference is given to regions covered previously by Herschel observations, which can further confirm the presence of compact sources and give more reliable estimates of their column densities. Figure 1 shows our target list of 2000 sources, sufficient to sample well the whole TRAO-visible sky and enable meaningful statistical studies of targets with different physical characteristics and in different environments.

About one-third of the 2000 sources have Herschel data at 70/100–500 μm. A large fraction of these data come from the Herschel Galactic Cold Cores survey (PI: M. Juvela), where the targets were also selected from the PGCC catalog, employing a random sampling similar to the one outlined above. The TOP sample covers widely different Galactic environments as shown in Figure 1. Among the 2000 sources, 219 are located in the Galactic plane with $| b| \lt 1^\circ$ and 154 at high latitudes with $| b| \gt 30^\circ$. Based on positional correlations, a sizable fraction of the targets in the Galactic plane are expected to be influenced by nearby H ii regions (Planck Collaboration et al. 2016). For example, 50 PGCCs in the λ Orionis complex were included in the TOP sample. The λ Orionis Complex containing the nearest giant H ii region has a moderately enhanced radiation field, and the molecular clouds therein seem to be greatly affected by stellar feedback (Goldsmith et al. 2016; Liu et al. 2016b). Among the 1181 sources with known distances in the TOP sample, 997 are within 2 kpc and 99 are beyond 4 kpc. Out of the 2000 sources, 753 have axial ratios larger than 2, suggestive of extended filamentary structures.

In Figure 2, we show the distributions of parameters for the TOP sample and the initial PGCC sample. In general, the TOP sample has similar distributions in longitude, latitude, and sizes to the initial PGCC sample. However, the TOP sample tends to have lower temperature and higher column density than the initial PGCC sample.

(3) Observation strategy

We first conduct single-pointing observations in the ¹²CO (1–0) and ¹³CO (1–0) lines simultaneously, to determine the systemic velocity of each target and also to find suitable reference positions for mapping. The on- and off-source integration time in the single-pointing survey is 30 s. For mapping observations, we applied the on-the-fly (OTF) mode to map the PGCCs in the ¹²CO (1–0) and ¹³CO (1–0) lines simultaneously. A small fraction (<10%) of interesting "TOP" targets will also be mapped in the C¹⁸O (1–0) line. The map size is set to $15^{\prime} \times 15^{\prime}$. Since the PGCC sources have average angular sizes of 8' (Planck Collaboration et al. 2016), the map size is large enough to well cover most PGCCs. We did, however, obtain larger (e.g., 30') OTF maps for some extended PGCCs. The FWHM beam sizes (${\theta }_{{\rm{B}}}$) for the ¹²CO (1–0) and ¹³CO (1–0) lines are 45'' and 47'', respectively. The main-beam efficiencies (${\eta }_{{\rm{B}}}$) for the ¹²CO (1–0) and ¹³CO (1–0) lines are 54% and 51%, respectively. The typical system temperatures for the ¹²CO (1–0) and ¹³CO (1–0) lines were ∼500 K and ∼250 K, respectively. We can achieve a typical sensitivity of ∼0.5 and 0.2 K in ${T}_{A}^{* }$ for the ¹²CO (1–0) and ¹³CO (1–0) lines, respectively, at a spectral resolution of 0.33 km s⁻¹ over the mapping field in ∼40 minutes of integration time. The OTF data are regridded into Class format files with a pixel size of 24'' with an OTFtool at TRAO. The OTF data were smoothed to 0.33 km s⁻¹ and baseline removed with Gildas/Class with a first- or third-order polynomial. Then the Class format files were converted to FITS format files for analysis with the MIRIAD and CASA packages.

The "SCOPE"⁸⁰ is one of the eight large programs at the JCMT of the East Asian Observatory (EAO) in the 2015 call. It is a survey of ∼1000 PGCCs at 850 μm continuum. The "SCOPE" project was launched in 2015 December. The main aims of the "SCOPE" survey are to obtain a census of dense clumps distributed in widely different environments, to study the roles of filaments in dense core formation, to investigate dust properties, and to detect rare populations of dense clumps (e.g., first hydrostatic cores; massive starless cores; pre-/proto-brown dwarfs; extremely cold cores).

(1) The SCUBA-2 bolometer camera at the JCMT 15 m telescope

With a diameter of 15 m, the JCMT is the largest single-dish astronomical telescope in the world designed specifically to operate in the submillimeter wavelength region of the spectrum. SCUBA-2 (Submillimetre Common-User Bolometer Array 2⁸¹ ) is an array operating simultaneously at 450 and 850 μm with a total of 5120 bolometers per wavelength (Holland et al. 2013). The instrument has a field of view of 45 arcmin². The instrument is cooled by three pulse tube coolers and a dilution refrigerator. The dilution fridge has a base temperature of about 50 mK. The focal plane that is the heat bath for the bolometers is temperature controlled at 75 mK. The main FWHM beam size of SCUBA-2 is 79 at 450 μm and 141 at 850 μm (Dempsey et al. 2013).

(2) Source selection

The 1000 "SCOPE" targets are selected from the 2000 PGCCs in the "TOP" survey sample. Preference is given to regions covered by Herschel observations or high column density ($\gt \,1\times {10}^{21}$ cm⁻² in Planck measurements) clumps. From our pilot study of 300 PGCCs (see Figure 3), we found that the dense core detection rates in SCUBA-2 observations drastically drop toward PGCCs with higher latitude and lower column density. Therefore, the "SCOPE" sample is further selected biased to high column density PGCCs in each parameter bin of the "TOP" survey sample. However, to ensure a good representation of the Galactic distribution of the full PGCC population, we also included many lower column density ($\gt 5\times {10}^{20}$ cm⁻² in Planck measurements) clumps at high latitudes as well. As shown in Figure 2, the SCOPE sample has similar distributions in most parameters (except column density, mass, and flux) to the TOP sample and the initial PGCC sample. The SCOPE sample tends to have larger column density and masses than the TOP sample and the initial PGCC sample. About half of the SCOPE sources have distances within 2 kpc. The high resolution of SCUBA-2 at 850 μm can easily resolve dense cores (with sizes of ∼0.1 pc) inside these PGCCs with distances smaller than ∼2 kpc.

Zoom In Zoom Out Reset image size

(3) Observation strategy

Since the PGCC sources have average angular sizes of 8' (Planck Collaboration et al. 2016), as noted above, the "SCOPE" observations were conducted primarily using the CV Daisy mode.⁸² The CV Daisy is designed for small compact sources providing a deep 3' (in diameter; the same as below) region in the center of the map but coverage out to beyond 12' (Bintley et al. 2014). All the SCUBA-2 850 μm continuum data were reduced using an iterative mapmaking technique (Chapin et al. 2013). Specifically, the data were all run with the same reduction tailored for compact sources, filtering out scales larger than 200'' on a 4'' pixel scale, for the first data release to the team. We also, however, tried different filtering and external masks in the data reduction for individual sources, which will be discussed below. The flux conversion factor (FCF) varies with time. A mean FCF of 554 Jy pW^–1 beam^–1 was used to convert data from pW to Jy beam^–1 in the pipeline for the first data release. The FCF is higher than the canonical value derived by Dempsey et al. (2013). This higher value reflects the impact of the data reduction technique and pixel size used by us. The observations were conducted under grade 3/4 weather conditions with 225 GHz opacities between 0.1 and 0.15. With 16 minutes of integration time per map, we reach an rms noise of ∼6–10 mJy beam⁻¹ in the central 3' region. The rms noise increases to 10–30 mJy beam⁻¹ out to 12', which is better than the sensitivity (50–70 mJy beam⁻¹) in the 870 μm continuum ATLASGAL survey (Contreras et al. 2013).

"SAMPLING⁸³ " is an ESO public survey to map up to 600 PGCCs in the J = 2–1 transition of ¹²CO and ¹³CO using the SMT 10 m telescope (Wang et al. 2018). The "SAMPLING" project was also launched in 2015 December. The typical map size in the "SAMPLING" survey is $5^{\prime} \times 5^{\prime}$ with an rms level of ∼0.2 K at a spectral resolution of 0.34 km s⁻¹. The beam size and main-beam efficiency are 36'' and 0.7, respectively. In conjunction with the "TOP-SCOPE" survey, the "SAMPLING" survey aims to resolve the clump structure, to derive the internal variations of column density and turbulent and chemical properties, and also to make a connection to Galactic structure. We observed the exemplar source, PGCC G26.53+0.17, in the ¹³CO (2–1) line with the SMT telescope. The observational parameters are summarized in Table 4 in Appendix A.

The Korean VLBI Network (KVN) is a three-element Very Long Baseline Interferometry (VLBI) network working at millimeter wavelengths. Three 21 m radio telescopes are located in Seoul, Ulsan, and Jeju Island, Korea. We use the single-dish mode of the three 21 m radio telescopes to observe the dense clumps detected in the "SCOPE" survey. The antenna pointing accuracy of the KVN telescopes is better than 5''. The KVN telescopes can operate at four frequency bands (i.e., 22, 44, 86, and 129 GHz) simultaneously. The beam sizes at the four bands are $\sim 126^{\prime\prime}$, $\sim 63^{\prime\prime}$, $\sim 32^{\prime\prime}$, and $\sim 23^{\prime\prime}$ at 22, 44, 86, and 129 GHz, respectively. The main-beam efficiencies are ∼50% at 22 and 44 GHz and ∼40% at 86 and 129 GHz.

A key science proposal for single-pointing molecular line observations of ∼1000 "SCOPE" dense clumps with KVN telescopes will be submitted in 2018. The main target lines are the 22 GHz water maser, 44 GHz Class I methanol maser, and other dense molecular lines, which are listed in Table 4 in Appendix A. The 22 GHz water maser and 44 GHz Class I methanol maser are indicators of outflow shocks. SiO thermal lines are also good tracers for outflow shocks or shocks induced by cloud–cloud collisions. Dense gas tracers (e.g., J = 1–0 transitions of ${{\rm{N}}}_{2}$H⁺, H¹³CO⁺, HN¹³C) can be used to determine the systemic velocities and the amount of turbulence in dense clumps. Optically thick lines (e.g., J = 1–0 transitions of HCN and HCO⁺) can be used to trace infall and outflow motions. ${{\rm{H}}}_{2}$CO${2}_{\mathrm{1,2}}-{1}_{\mathrm{1,1}}$ is a dense gas tracer and can also be used to reveal infall and outflow motions (Liu et al. 2016b). The deuteration of ${{\rm{H}}}_{2}$CO will be determined from observations of ${{\rm{H}}}_{2}$CO (${2}_{\mathrm{1,2}}-{1}_{\mathrm{1,1}}$) and HDCO (${2}_{\mathrm{0,2}}-{1}_{\mathrm{0,1}}$) lines. The [HDCO]/[${{\rm{H}}}_{2}$CO] ratios will be used to trace the early phase of dense core evolution (Kang et al. 2015). Pilot surveys of ∼200 "SCOPE" dense clumps with the KVN telescopes were conducted in 2016 (e.g., Liu et al. 2016b; S.-J. Kang et al., 2018, in preparation; H.-W. Yi et al. 2018, in preparation). In the single-pointing molecular line observations, it takes about 20 minutes (on+off) to achieve an rms level of <0.1 K in brightness temperature with dual polarization at a spectral resolution of ∼0.2 km s⁻¹ under normal weather conditions. The KVN observations of the exemplar source, PGCC G26.53+0.17, are summarized in Table 4 in Appendix A.

The KVN data are also reduced with the Gildas/Class package. All the scans were averaged to get the final averaged spectra, and the baselines of the averaged spectra were removed with a linear fit.

By using the 45 m telescope of Nobeyama Radio Observatory (NRO), we plan to carry out a comprehensive study of cores selected from the JCMT SCOPE survey to scrutinize the initial conditions of star formation in widely different environments, including massive star-forming regions. We will observe 100 cores in various environments in DNC, HN¹³C, ${{\rm{N}}}_{2}$D⁺, and cyclic-${{\rm{C}}}_{3}{{\rm{H}}}_{2}$ with receiver T70 in single-pointing mode. Among the 100 cores, 35 cores will also be mapped in ¹²CO, ¹³CO ,C¹⁸O, ${{\rm{N}}}_{2}$H⁺, ${\mathrm{HC}}_{3}{\rm{N}}$, and CCS with the FOREST receiver. The FWHM beam size of the NRO 45 m telescope at 86 GHz is $\sim 18\buildrel{\prime\prime}\over{.} 8$, and the main-beam efficiency is ∼54%. Adding 115 Orion cores already observed in single-pointing mode, we will have data for 215 single-pointing positions. By using DNC and ${{\rm{N}}}_{2}$D⁺ intensities and the chemical evolution factor (CEF) for DNC/HN¹³C, we will select the best targets of 35 cores for OTF mapping with the four-beam 2SB 2-polarization receiver FOREST. By applying CEF we developed (Tatematsu et al. 2017), we should be able to identify cores on the verge of massive star formation. From deuterium fractionation, we can identify the earliest protostellar phase. We will confirm/revise CEF by observing Orion cores having similar distances. The chemical nature of cores including deuterium fractionation will be investigated. Core stability will be investigated to see how star formation starts. Dynamics of parent filaments will be investigated to see whether there is accretion onto cores. Coefficients of the line width–size relation will be compared among different environments to see its meaning in star formation. Specific angular momentum will be compared among regions.

The large proposal requesting 350 hr over 2 yr since 2017 December was accepted for such follow-up observations. The NRO 45 m observations of the exemplar source, PGCC G26.53+0.17, have not yet been conducted. Therefore, no results from NRO 45 m observations of PGCC G26.53+0.17 will be presented in this paper. Some pilot studies with the NRO 45 m telescope toward SCUBA-2 dense cores in PGCCs were presented in Tatematsu et al. (2017).

The high-resolution SCUBA-2 850 μm data filtered out large-scale extended emission and thus are sensitive to the denser structures (filaments, cores, or clumps) in molecular clouds. The 850 μm data are also important in providing high-resolution information about the dust emission at a wavelength that lies in the regime between the far-infrared part (e.g., Herschel) and the millimeter part of the dust emission spectrum. One of the main goals of the "SCOPE" survey is to provide high-quality 850 μm data to constrain better the dust emission spectrum of dense structures in molecular clouds.

In this section, we fit the SEDs pixel by pixel with a modified blackbody function using four or five bands, combining SCUBA-2 measurements and fluxes extracted from Herschel maps that have been run through the SCUBA-2 pipeline. The algorithm of the SED fit is the same as in Section 6.1 in Appendix A. The Herschel and SCUBA-2 data with the subtraction of the local background were fitted with modified blackbody spectra with $\beta =1.8$ to estimate color correction factors and to derive estimates of the dust optical depth at a common spatial resolution of 40''. In Section 4.3, we fit the SEDs with free β to investigate the dust grain properties. The details of the SED fits using filtered Herschel and SCUBA-2 maps will be discussed in Juvela et al. (2017). In this paper, we performed four individual SCUBA-2 data reductions labeled R1, R2, R3, and R4 with different external masks and spatial filters (see Section 6.2 in Appendix A for details). R1 with a spatial filter of 200'' is very efficient at identifying dense clumps/cores and is used for creating dense core/clump catalogs in our first data release to the team. However, here we fit the SEDs with data sets obtained from two other data reduction methods (R2, R4), which use larger spatial filters (300'' and 600'') and thus can recover more flux. We applied two sets of SED fits (S1 and S2) using different data sets.

(1) S1:Herschel 160–500 μm data and SCUBA-2 850 μm data in the R2 data reduction (see Section 6.2 in Appendix A) were used. The extended emission larger than 300'' in the Herschel and SCUBA-2 maps was filtered out.

(2) S2:Herschel 160–500 μm data and SCUBA-2 850 μm data in the R4 data reduction (see Section 6.2 in Appendix A) were used. The extended emission larger than 600'' in the Herschel and SCUBA-2 maps was filtered out.

Figures 5(a) and (b) show the dust temperature (${T}_{d}$) and 250 μm optical depth (${\tau }_{250}$) maps from the two SED fits. The ${T}_{d}$ values across the filament ridge are around 10–14 K, except for the dense clumps located in the central region (∼17 K) and the southern end (∼19 K), where protostars have formed and heated their local surroundings. The less dense region to the east of the filament ridge has higher ${T}_{d}$ than the dense filament, indicating that the filament is externally heated from its eastern side.

Zoom In Zoom Out Reset image size

The ${T}_{d}$ and ${\tau }_{250}$ maps are quite different from SED fits with different effective spatial filters. In Figures 5(c)–(d), we compare the ${T}_{d}$ and ${\tau }_{250}$ values from SED fits S1 and S2 for pixels with S/N > 5 in the SCUBA-2 map from the R2 imaging scheme. The mean dust temperature from S1 and S2 is 13.7 and 14.3 K, respectively. In comparison, the mean ${\tau }_{250}$ (or column density N) in S1 and S2 is $6.5\times {10}^{-3}$ (or $N\,\sim 1.0\times {10}^{22}$ cm⁻²) and $8.3\times {10}^{-3}$ (or $N\sim 1.3\times {10}^{22}$ cm⁻²), respectively. In contrast to S1, S2 increases the ${T}_{d}$ in cold regions (with ${T}_{d}\lt 15$ K), whereas in warmer regions (with ${T}_{d}\gt 15$ K), the ${T}_{d}$ determined in S1 and S2 does not vary too much. S2 significantly increases ${\tau }_{250}$ by 0.003 more (or $N\sim 4.6\times {10}^{21}$ cm⁻²) than S1 in dense regions (with ${\tau }_{250}\sim 0.01$ or $N\sim 1.5\times {10}^{22}$ cm⁻² in the S2map). The peak column densities in S1 and S2 are $3.8\times {10}^{22}$ cm⁻² and $4.2\times {10}^{22}$ cm⁻², respectively. In less dense regions, S2 also significantly increases ${\tau }_{250}$, but not as much as in dense regions.

SED fits with large effective spatial filters (600''; Figure 5(b)) recovered more extended structure and thus more mass. The total mass of the G26 filament revealed in S2 is ∼6200 M_☉. Given the length ($L\sim 12$ pc) of the filament, the mean line mass (M/L) of the filament is ∼500 M_☉ pc⁻¹. For comparison, both the length and line mass of G26 are comparable to those ($L\sim 8$ pc; ${\text{}}M/L\sim 400$ M_☉ pc⁻¹) of the integral shaped filament in the Orion A cloud (Bally et al. 1987; Kainulainen et al. 2017).

The "SCOPE" survey aims to obtain a census of "all-sky" distributed dense clumps and cores. Since PGCC sources trace some of the coldest ISM in the Galaxy, an extensive survey of PGCC sources is able to provide us with a number of candidates of clumps and cores at their earliest evolutionary phases. In nearby clouds, we expect to discover a population of prestellar core candidates or extremely young (e.g., Class 0) protostellar objects (Liu et al. 2016b; Tatematsu et al. 2017). In Galactic plane PGCCs, we are particularly interested in searching for massive clumps or cores that may represent the initial conditions of high-mass star formation. Below, we demonstrate the source extraction process based on the G26 SCUBA-2 images. The properties of clumps identified in G26 will also be investigated.

Extraction of the dense clumps was done using the FELLWALKER (Berry 2015) source extraction algorithm, part of the Starlink CUPID package (Berry et al. 2007). The core of the FellWalker algorithm is a gradient-tracing scheme consisting of following many different paths of steepest ascent in order to reach a significant summit, each of which is associated with a clump (Berry et al. 2007). FellWalker is less dependent on specific parameter settings than CLUMPFIND (Berry et al. 2007). The source extraction process with FellWalker in the SCOPE survey is the same as that used by the JCMT Plane Survey, and details can be found in Moore et al. (2015) and Eden et al. (2017). A mask constructed above a threshold of 3σ (i.e., three times the pixel-to-pixel noise) in the S/N map is applied to the intensity map as input for the task CUPID:EXTRACTCLUMPS, which extracts the peak and integrated flux density values of the clumps. A further threshold for CUPID:FINDCLUMPS was the minimum number of contiguous pixels, which was set at 12, corresponding to the number of pixels expected to be found in an unresolved source with a peak S/N of 5σ, given a 14'' beam and 4'' pixels. The cores close to the map edges that are associated with artificial structures were further removed manually.

In total, 10 clumps were identified from the SCUBA-2 images. Those 10 clumps can be identified in at least two of the SCUBA-2 data reductions (R1, R2, R3, and R4; see Section 6.2 in Appendix A). Figure 6(a) presents the distributions of these 10 clumps on the 850 μm image from the R2 reduction. From checking Spitzer (see Figure 4) and the Herschel/PACS 70 μm data, we found that clumps "6" and "10" are associated with young stellar objects, while the others are likely starless. The parameters of these clumps are summarized in Table 1. The effective radius is defined as ${R}_{\mathrm{eff}}=\sqrt{{ab}}$, where a and b are the deconvolved FWHM sizes of the clump major and minor axes. The clump masses (M) are derived with Equation (4) in Appendix A with total fluxes of 850 μm continuum emission from the SCUBA-2 images. The radii and masses are listed in the seventh and 12th columns of Table 1, respectively. The particle number density (n) and column density (N) were calculated as $n=M/\tfrac{4}{3}\pi {R}_{\mathrm{eff}}^{3}\,\mu {{\rm{m}}}_{{\rm{H}}}$ and $N=M/\pi \,{R}_{\mathrm{eff}}^{2}\,\mu {{\rm{m}}}_{{\rm{H}}}$, respectively, where $\mu =2.37$ is the mean molecular weight per "free particle" (${{\rm{H}}}_{2}$ and He) and m_H is the atomic hydrogen mass.

Zoom In Zoom Out Reset image size

Table 1.  Parameters of Dense Clumps

Clump	R.A.	Decl.	${\theta }_{\mathrm{maj}}$	${\theta }_{\min }$	P.A.	${{\rm{R}}}_{\mathrm{eff}}$	${F}_{\mathrm{peak}}$	${F}_{\mathrm{int}}$	S/N	${T}_{d}$ a	M	n	N	${M}_{\mathrm{Jeans}}$
No.	(J2000)	(J2000)	(arcsec)	(arcsec)	(deg)	(pc)	(Jy beam−1)	(Jy)		(K)	(M☉)	(104 cm−3)	(1022 cm−2)	(M☉)
R1
1	18:37:12.00	−05:19:12.0	15	11	142	<0.26	0.10(0.01)	0.23(0.01)	7	10	89	>2.1	>2.2	<2
2	18:37:08.40	−05:20:27.6	19	9	108	<0.27	0.14(0.02)	0.35(0.02)	11	13	82	>1.7	>1.9	<3
3	18:37:09.36	−05:21:03.6	14	13	108	<0.27	0.17(0.02)	0.30(0.02)	13	13	70	>1.5	>1.6	<3
4	18:37:09.12	−05:21:54.0	12	10	191	<0.22	0.07(0.01)	0.10(0.01)	5	12	27	>1.1	>1.0	<4
5	18:37:07.44	−05:23:09.6	10	8	242	<0.18	0.25(0.03)	0.32(0.02)	20	13	75	>5.3	>3.9	<2
6	18:37:07.68	−05:23:56.4	29	20	133	0.40	1.35(0.14)	4.38(0.22)	103	17	654	4.2	7.0	3
7	18:37:02.64	−05:24:39.6	10	7	167	<0.17	0.11(0.01)	0.16(0.01)	9	13	37	>3.2	>2.2	<2
9	18:37:00.48	−05:25:40.8	27	13	90	0.25	0.14(0.01)	0.70(0.04)	11	14	144	3.8	3.9	2
10	18:36:57.84	−05:26:56.4	21	12	125	0.15	0.14(0.01)	0.46(0.02)	11	19	58	7.1	4.4	3
R2
1	18:37:12.00	−05:19:15.6	24	16	211	0.28	0.12(0.01)	0.44(0.02)	8	10	170	3.2	3.7	2
2	18:37:08.64	−05:20:20.4	22	11	112	0.13	0.16(0.02)	0.51(0.03)	10	13	119	22.5	12.0	1
3	18:37:09.36	−05:21:03.6	22	14	143	0.21	0.20(0.03)	0.53(0.03)	13	13	124	5.6	4.8	2
4	18:37:09.12	−05:21:54.0	25	14	155	0.25	0.11(0.01)	0.47(0.02)	7	12	127	3.4	3.5	2
5	18:37:07.44	−05:23:09.6	15	11	254	<0.26	0.31(0.04)	0.60(0.03)	20	13	140	>3.3	>3.5	<2
6	18:37:07.68	−05:23:56.4	39	27	147	0.60	1.40(0.12)	6.41(0.32)	88	17	957	1.8	4.5	5
7	18:37:02.64	−05:24:39.6	19	15	126	0.19	0.17(0.01)	0.71(0.04)	11	13	166	10.0	7.8	1
9	18:37:00.48	−05:25:40.8	22	17	247	0.27	0.17(0.02)	0.93(0.05)	10	14	191	4.0	4.5	2
10	18:36:57.60	−05:26:56.4	24	17	142	0.29	0.18(0.01)	0.91(0.05)	11	19	115	1.9	2.3	5
R3
3	18:37:09.36	−05:21:03.6	26	18	266	0.33	0.23(0.03)	1.20(0.06)	7	14	247	2.8	3.9	3
5	18:37:07.44	−05:23:09.6	22	13	230	0.19	0.38(0.05)	1.12(0.06)	12	14	230	13.9	10.9	1
6	18:37:07.68	−05:23:56.4	57	42	170	0.95	1.47(0.12)	11.44(0.57)	47	17	1708	0.8	3.2	7
7	18:37:02.64	−05:24:39.6	15	10	213	<0.25	0.21(0.02)	0.52(0.03)	7	14	107	>2.8	>2.9	<3
8	18:37:00.72	−05:25:01.2	32	26	209	0.51	0.22(0.02)	2.20(0.11)	7	14	452	1.4	3.0	4
10	18:36:57.60	−05:26:56.4	41	32	179	0.68	0.24(0.01)	2.42(0.12)	8	19	305	0.4	1.1	11
R4
1	18:37:12.24	−05:19:12.0	23	21	160	0.34	0.16(0.01)	1.04(0.05)	5	12	281	3.0	4.2	2
3	18:37:09.36	−05:21:03.6	31	23	103	0.46	0.26(0.03)	1.96(0.10)	9	14	403	1.7	3.2	4
4	18:37:09.12	−05:21:57.6	23	15	210	0.25	0.18(0.02)	0.96(0.05)	6	13	225	6.0	6.1	2
5	18:37:07.44	−05:23:09.6	25	14	223	0.25	0.40(0.05)	1.32(0.07)	13	14	271	7.2	7.4	2
6	18:37:07.68	−05:23:56.4	69	49	145	1.15	1.48(0.12)	13.87(0.69)	49	17	2071	0.6	2.7	8
7	18:37:02.64	−05:24:39.6	17	9	210	<0.25	0.21(0.02)	0.55(0.03)	7	14	113	>3.0	>3.1	<3
8	18:37:00.72	−05:25:01.2	38	27	217	0.59	0.23(0.02)	2.43(0.12)	8	14	499	1.0	2.4	5
10	18:36:57.60	−05:26:56.4	49	36	178	0.81	0.25(0.01)	3.60(0.18)	8	19	453	0.4	1.2	12

Note.

^a ${T}_{d}$ in R1 and R2 is derived from SED fit S1. ${T}_{d}$ in R3 and R4 is derived from SED fit S2.

Download table as:  ASCIITypeset image

Figure 6(b) compares the peak fluxes of the dense clumps in different imaging schemes. In general, applying larger spatial filters increase the peak fluxes particularly for less dense clumps. The peak flux of the most massive clump "6" does not change too much (<10%) in different imaging schemes. In contrast to R1, however, the peak fluxes of other less dense clumps increased by ∼20%−50% in R2 and by 50%−100% in R3 and R4. Imaging scheme R1 is used in the first data release of the SCOPE survey. R1 is very efficient to detect dense clumps but cannot well recover flux from extended structures. Therefore, we suggest to use external masks and larger spatial filtering in SCUBA-2 data reduction to recover more flux if Herschel data are available.

Figure 6(c) presents the mass–radius relation for the dense clumps. In general, larger scales kept by spatial filtering lead to larger masses and radii. The red dashed line shows a density threshold for high-mass star formation (Kauffmann & Pillai 2010). Except for clump "6," all the other clumps in R1 are below this threshold. Most clumps, however, will move above the empirical threshold when larger scales are kept by the spatial filtering, as in R2, R3, and R4. Most clumps (revealed in R2, R3, and R4) are above the predicted mass–radius relation for clumps undergoing quasi-isolated gravitational collapse in a turbulent medium (Li 2017). The clumps in G26, especially the dense and massive ones, are very likely bounded by gravity and have the ability to form high-mass stars.

We derived the thermal Jeans masses ${M}_{J}\,=0.877\,{M}_{\odot }{\left(\tfrac{T}{10K}\right)}^{3/2}{\left(\tfrac{n}{{10}^{5}{\mathrm{cm}}^{-3}}\right)}^{-1/2}$ following Wang et al. (2014). The resulting n, N, and M_J are listed in the last three columns of Table 1. The clump masses are also about one order of magnitude larger than the corresponding thermal Jeans masses. The clumps are all located along the filament with a mean separation of ∼1.3 pc, which is much larger than the thermal Jeans length (∼0.1 pc for T = 14 K and the mean volume density n = $6.2\times {10}^{4}$ cm⁻³ for clumps in R2), indicating that fragmentation in G26 is not governed by thermal instability but determined by other factors like turbulence or magnetic fields, as revealed in other IRDCs (e.g., Wang et al. 2011, 2014; Contreras et al. 2016).

The so-called "sausage instability" of a gas cylinder, proposed by Chandrasekhar & Fermi (1953) and promoted by many others (e.g., Stodólkiewicz 1963; Ostriker 1964; Inutsuka & Miyama 1992; Fiege & Pudritz 2000; Fischera & Martin 2012), has often been applied for studies of fragmentation in filamentary clouds (e.g., Jackson et al. 2010; Wang et al. 2011, 2014; Liu et al. 2017). In this scenario, an isothermal, hydrostatic, nonmagnetized, infinite filament becomes unstable and fragments into a chain of equally spaced fragments if its mass per unit length along the cylinder (or linear mass density) is close enough to the critical value so that perturbations in the filament have time to grow. In G26, the mean spacing and mass of clumps in G26 are ∼1.3 pc and ∼540 M_☉, respectively, which were derived for clumps in R4 reduction. The fragment mass in cylindrical fragmentation can be estimated as ${M}_{\mathrm{cl}}=575.3\,{M}_{\odot }{\left(\tfrac{\sigma }{1\mathrm{km}{{\rm{s}}}^{-1}}\right)}^{3}{\left(\tfrac{{n}_{c}}{{10}^{5}{\mathrm{cm}}^{-3}}\right)}^{-1/2};$ the typical spacing in cylindrical fragmentation is ${\lambda }_{\mathrm{cl}}=1.24\,\mathrm{pc}\left(\tfrac{\sigma }{1\,\mathrm{km}\,{{\rm{s}}}^{-1}}\right){\left(\tfrac{{n}_{c}}{{10}^{5}{\mathrm{cm}}^{-3}}\right)}^{-1/2}$ (Wang et al. 2014). σ is the 1D velocity dispersion and ${n}_{c}$ is the volume density at the center of the cylinder. Here we take ${n}_{c}$ as the mean number density ($6.2\times {10}^{4}$ cm⁻³) for clumps identified in R2 reduction because R2 filtered more extended emission than R4, and thus the mean number density for clumps in R2 is much closer to the volume density at the center of the cylinder. If we only consider thermal pressure (assuming T = 14 K), the σ, M_cl, and ${\lambda }_{\mathrm{cl}}$ are 0.22 km s⁻¹, 7.8 M_☉, and 0.35 pc, respectively. The M_cl and ${\lambda }_{\mathrm{cl}}$ are much smaller than the observed values in G26, indicating that thermal pressure alone cannot govern the fragmentation. However, if we also consider turbulent support, the σ, M_cl and ${\lambda }_{\mathrm{cl}}$ will become ∼0.94 km s⁻¹ (estimated from C¹⁸O line width; see Table 2), 600 M_☉, and 1.5 pc, respectively. The M_cl and ${\lambda }_{\mathrm{cl}}$ derived from turbulence-dominated cylindrical fragmentation are very consistent with the observed values (∼540 M_☉ and ∼1.3 pc), suggesting that the fragmentation in G26 from cloud (at ∼10 pc scale) to clumps (at ∼1 pc scale) is very likely dominated by turbulence in the absence of magnetic fields. Limited by the spatial resolution (∼0.3 pc at the distance of 4.2 kpc) and mass sensitivity (3σ mass sensitivity is ∼10 M_☉), we cannot study fragmentation from clump scale to core scale (at ∼0.1 pc scale) with the present data. Further fragmentation below the completeness and resolution limit of the present data is very likely to take place as seen in other IRDCs (Wang et al. 2011, 2014). The hierarchical fragmentation in G26 at all spatial scales can only be studied from higher spatial resolution interferometric observations.

Table 2.  Derived Parameters of Molecular Lines

Line	$\int {T}_{b}{dv}$ a	${V}_{\mathrm{LSR}}$ a	FWHMa	${T}_{\mathrm{mb}}$	${T}_{\mathrm{ex}}$ b	τb	Nb	Xc
	(K km s−1)	(km s−1)	(km s−1)	(K)	(K)		(cm−2)
C18O J = 1–0	5.05(0.25)	48.23(0.05)	2.22(0.12)	2.14	18.2	0.1	4.6E15	1.0E–07
${\mathrm{CH}}_{3}$OH ${7}_{0}-{6}_{1}A+$	0.48(0.06)	47.90(0.02)	0.39(0.04)	1.16
	0.64(0.07)	48.66(0.04)	0.79(0.11)	0.76
HCO+ J = 1–0	3.34(0.16)	50.95(0.04)	1.85(0.11)	1.70
	2.51(0.20)	47.06(0.11)	3.15(0.36)	0.75
H13CO+ J = 1–0	0.59(0.11)	49.78(0.12)	1.37(0.28)	0.40	3.9	0.5	1.9E13	4.2E–10
	0.86(0.12)	47.91(0.09)	1.46(0.23)	0.55	4.0	0.6	3.6E13	8.0E–10
${\mathrm{HC}}_{3}{\rm{N}}$ J = 10–9	2.43(0.18)	49.03(0.11)	3.07(0.27)	0.75	5.4	0.4	9.2E14	2.0E–08
${\mathrm{HC}}_{3}{\rm{N}}$ J = 14–13	1.18(0.19)	49.12(0.25)	3.07(0.55)	0.36	6.0	0.1
${\mathrm{HC}}_{3}{\rm{N}}$ J = 15–14	0.52(0.13)	49.29(0.31)	2.37(0.55)	0.21	6.6	0.1
SiO J = 2–1	1.57(0.40)	48.54(0.61)	7.28(1.23)	0.20	4.8	4.0	1.3E13	2.9E–10
	1.44(0.44)	37.16(2.29)	14.05(4.13)	0.10
SiO J = 3–2	3.12(0.52)	44.34(1.75)	22.01(2.79)	0.13
	0.63(0.24)	48.66(0.49)	4.44(1.45)	0.14	3.2	0.5	8.6E13	1.9E–09
HN13C J = 1–0	1.56(0.14)	49.06(0.14)	3.08(0.30)	0.48	3.5	1.0	2.5E14	5.6E–09
${{\rm{H}}}_{2}$CO ${2}_{\mathrm{1,2}}-{1}_{\mathrm{1,1}}$	2.35(0.10)	50.81(0.04)	2.38(0.13)	0.93
	1.32(0.09)	47.42(0.06)	1.87(0.17)	0.66
	${T}_{\mathrm{ex}}$ d	${V}_{\mathrm{LSR}}$ d	FWHMd	τ	${T}_{\mathrm{ex}}$	τ	N
CCH N = 1–0	4.4(0.3)	49.00(0.08)	3.02(0.20)	0.43(0.13)	4.5	0.4	1.5E14	3.3E–09
${{\rm{N}}}_{2}$H+ J = 1–0	5.58(0.59)	47.72(0.03)	0.58(0.03)	1.88(0.54)	4.1	2.2	2.4E13	5.3E–10
	5.53(0.79)	50.05(0.03)	0.57(0.04)	1.66(0.62)	4.0	2.0	2.0E13	4.4E–10

Notes.

^aFrom Gaussian fits with Gildas/Class. ^bThe values for H¹³CO⁺, ${\mathrm{HC}}_{3}{\rm{N}}$, and HN¹³C are derived by assuming a source size of 11''. The others are derived by assuming a filling factor of 1. ^c ${{\rm{N}}}_{{{\rm{H}}}_{2}}=4.5\times {10}^{22}$ cm⁻². ^dFrom hyperfine structure fits with Gildas/Class.

Download table as:  ASCIITypeset image

Observationally, dust properties (e.g., opacity) greatly affect the determination of density and temperature profiles of dense cores and therefore influence the estimation of core masses. The dust grain size distributions in molecular clouds and cores can change with time because of grain sputtering, shattering, coagulation, and growth (Ormel et al. 2009). Grain growth in the densest and coldest regions of interstellar clouds has been evidenced by an increase in the far-infrared opacity (Juvela et al. 2015a). Indeed, the dust spectral index (β) and its temperature (${T}_{d}$) dependence can provide additional information on the chemical composition, structure, and size distribution of interstellar dust grains (Juvela et al. 2011). This dust emissivity change has been attributed to the formation of fluffy aggregates in the dense medium, resulting from grain coagulation.

Recently, Juvela et al. (2015a) studied dust optical depths by comparing measurements of submillimeter dust emission and the reddening of the light of background stars in the near-infrared (NIR) for a sample of 116 PGCCs in Herschel fields. In their studies, there are indications that β increases toward the coldest regions and decreases strongly near internal heating sources. Although NIR reddening is an independent and reliable measure of column density, NIR extinction could not, however, trace the densest portion of the cloud, due to low resolution. Additionally, the Herschel 250 μm band is less sensitive to very cold dust and may completely miss the highest column density peak (see Figure 2 in Pagani et al. 2015). Herschel 500 μm data could trace cold dust well but have relatively low angular resolution (35''). Adding high-resolution photometric data at longer wavelengths (e.g., 850 μm from SCUBA-2) could help more accurately derive the dust spectral index, dust temperature, and column density (${{N}}_{{{\rm{H}}}_{2}}$) simultaneously (Sadavoy et al. 2013; Chen et al. 2016).

For PGCC sources having Herschel data in the "SCOPE" survey, we will determine the dust emissivity spectral index (β), dust temperature (${T}_{d}$), and column density (${{\rm{N}}}_{{{\rm{H}}}_{2}}$) simultaneously by combining 850 μm data with spatially filtered Herschel photometric data. Since the PGCC sources are at different stages of cloud evolution from starless clumps to protostellar cores and are located in different Galactic environments, we will search for variations of β that might be attributed to the evolutionary stage of the sources or to environmental factors, including the location within the Galaxy and stellar feedback. A pilot study of dust properties toward ∼100 PGCCs with Herschel data and SCUBA-2 data will be presented in Juvela et al. (2017). Below we investigate the dust properties of G26 as an example.

In Figure 7, we present the ${T}_{d}$, β, and ${\tau }_{250}$ maps of G26 from SED fits with free β and ${T}_{d}$. The extended emission larger than 600'' in the Herschel and SCUBA-2 maps (from imaging scheme R4; see Section 6.2 in Appendix A) was filtered out. In general, the β map is anticorrelated with the ${T}_{d}$ map. The most massive core, marked with a black plus sign, clearly shows higher temperature and lower β than its surroundings. The low β may indicate grain growth during core or star formation. Figures 8(a) and (b) show histograms of ${T}_{d}$ and β. Only the data points with S/N > 10 in the SCUBA-2 map, corresponding to the ridge region of the filament, are included in the statistics. In general, including 850 μm data tends to increase the ${T}_{d}$ by ∼1 K and lower β by ∼0.3. The dust temperature histograms show a high-temperature tail, which is caused by ongoing star formation. Interestingly, the β histograms can be well fitted with two normal distributions regardless of whether or not 850 μm data were included in the SED fits. Such a bimodal behavior in β distribution may suggest grain growth along the filament.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Figure 8(c) presents the correlation between ${T}_{d}$ and β. β is generally anticorrelated with ${T}_{d}$. Interestingly, the β versus ${T}_{d}$ correlations can be roughly depicted with three power laws, which again indicates the existence of different kinds of dust grains in the filament. Such an anticorrelation between ${T}_{d}$ and β was also found in all Perseus clumps (Chen et al. 2016). Chen et al. (2016) also found that these anticorrelations cannot be solely accounted for by anticorrelated ${T}_{d}$ and β uncertainties associated with SED fitting. The anticorrelation may be partially explained by the dust grains' intrinsic β dependency on temperature, but more likely by the sublimation of surface ice mantles, which can increase β when present on a dust grain (Chen et al. 2016). A thorough investigation of the correlation between ${T}_{d}$ and β is beyond the scope of this paper. We will systematically study this issue in the future with more "SCOPE" objects.

To understand better the star or core formation process, it is crucially important to study their parent molecular cloud properties. SCUBA-2 has better resolution and longer wavelengths than Herschel/SPIRE bands and thus is more suitable to trace the cold ISM. Due to variations of the atmosphere that mimic emission from extended astronomical objects, SCUBA-2 data are not well suited to capture the extended structures in molecular clouds. The loss of the filtered emission by SCUBA-2, however, can be somewhat corrected based on the Planck data. The combined Planck 353 GHz data and the SCUBA-2 850 μm data will be used to investigate the density distribution and structures at various cloud scales. Below we use G26 to demonstrate the combination of the Planck and SCUBA-2 data.

We combined the Planck 353 GHz image with the SCUBA-2 850 μm images in the R2 and R4 reductions following the same procedure as in Lin et al. (2016, 2017). In principle, the data combination was performed in the Fourier domain, yielding high-resolution (14'') combined data that have little or no loss of extended structure (Lin et al. 2016, 2017). The combined images are shown in Figure 9(a). In general, the two images show very similar morphology and flux density. From these combined images, we find that the dense filament resides in a cloud with very smooth structures. The cloud boundary is well enclosed by the 10% contours in Figure 9(a). Particularly, the dense filament is close to the cloud edge, indicating possible dynamical clues to its formation (e.g., external compression).

Zoom In Zoom Out Reset image size

Figure 9(b) presents the dust temperature and column density maps obtained from SED fits with Herschel/PACS 160 μm, SPIRE 250/350 μm and combined Planck+SCUBA-2 850 μm data in R2 reduction. The final dust temperature and column density maps have an angular resolution of 25''. The temperature map clearly reveals a temperature gradient from southeast to northwest.

To quantify the dense gas distributions systematically, we perform analyses of the column density probability distribution functions (N-PDF) following Lin et al. (2016, 2017). The natural logarithm of the ratio of column density and mean column density is $\eta ={ln}({N}_{{{\rm{H}}}_{2}}/\langle {N}_{{{\rm{H}}}_{2}}\rangle )$, and the normalization of the probability function is given by ${\int }_{-\infty }^{+\infty }p(\eta )d\eta ={\int }_{0}^{+\infty }p({N}_{{{\rm{H}}}_{2}})d({N}_{{{\rm{H}}}_{2}})=1$. Figure 9(c) shows the column density probability distribution functions (N-PDF). The N-PDF in molecular clouds is usually found to consist of a lognormal like part at low column density and a power-law-like part at high column density (Klessen 2000; Kritsuk et al. 2011; Schneider et al. 2012, 2015; Federrath & Klessen 2013; Lin et al. 2016, 2017). The N-PDF power-law tail is usually linked to the effects of gravity (Klessen 2000; Kritsuk et al. 2011; Federrath & Klessen 2013; Lin et al. 2016, 2017). We tried to fit the tail with a power law for the N-PDF of G26. Although the power-law fit is poor, its slope (−3.9) is comparable to those of IRDCs G28.34+0.06 (−3.9) and G14.225–0.506 (−4.1) but smaller than those of protoclusters (Lin et al. 2016, 2017), indicating that G26 is still at a very young evolutionary stage and is not greatly affected by star formation activities as in high-mass protoclusters.

To diagnose the characteristic spatial scales in G26, we determine the two-point correlation functions (2PT) of gas column density following Lin et al. (2016, 2017). Figure 9(d) shows the two-point correlation (2PT) function of column densities in the observed field. The correlation strength at a separation scale (lag) of l is calculated by

$$\begin{eqnarray}&&{S}_{{tr}}(l)=\displaystyle \frac{{\langle X({\boldsymbol{r}})X({\boldsymbol{r}}+{\boldsymbol{l}})\rangle }_{l}}{{\langle X({\boldsymbol{r}})X({\boldsymbol{r}})\rangle }_{l}},\end{eqnarray} \tag{ 1 }$$

where $X({\boldsymbol{r}})$ denotes the column density value at position ${\boldsymbol{r}}$, and the angle brackets are an average over all pairs of positions with a separation of l. The final form of the correlation function is normalized by the peak correlation strength to enable comparisons between different fields. For a more detailed description, see Lin et al. (2016, 2017).

In general, the 2PT function shows a smooth decay of correlation strengths over all spatial scales up to ∼15 pc. Such a flat profile in the 2PT correlation is also witnessed toward IRDCs like G11.11–0.12 (see Figure 5 in Lin et al. 2017) but is in contrast with most of the active OB-cluster-forming regions, which have a steep decay of correlation strength at small spatial scales (Lin et al. 2016). The flat 2PT correlation indicates that the column density distribution of G26 is very homogeneous over all spatial scales larger than 1 pc and the mass is less concentrated overall than in high-mass protoclusters.

The N-PDF and 2PT are powerful tools to investigate the cloud structure evolution and should be applied to other SCOPE fields.

The empirical power-law relation between the star formation rate (SFR), surface density (${{\rm{\Sigma }}}_{\mathrm{SFR}}$), and surface density of cold gas (${{\rm{\Sigma }}}_{\mathrm{gas}}$) pioneered in the works of Schmidt (1959) and Kennicutt (1998), the so-called Kennicutt–Schmidt (KS) law, is of great importance as an input for theoretical models of galaxy evolution. Recently, nearly linear correlations between star formation rates and line luminosities of dense molecular gas tracers (e.g., HCN and CS) have been found toward both Galactic dense clumps and galaxies (Gao & Solomon 2004; Wu et al. 2005, 2010; Lada et al. 2010; Zhang et al. 2014; Liu et al. 2016a; Stephens et al. 2016), strongly suggesting that star formation is mainly related to dense gas in molecular clouds. Therefore, it is important to evaluate the proportions of dense gas in molecular clouds. As discussed in Liu et al. (2016a), the relation between star formation rates and clump masses (traced by filtered continuum maps) is as tight as correlations between star formation rates and line luminosities of dense gas tracers (e.g., HCN, CS). Particularly, filtered continuum maps have advantages to trace the total dense gas masses rather than dense gas tracers (e.g., HCN, HCO⁺, CS), whose emissions are usually optically thick (Liu et al. 2016a). Since the filtered SCUBA-2 images intrinsically filter the large-scale diffuse gas and are most sensitive to the high volume densities (∼10⁴ cm⁻³ for G26) in clouds, the flux ratio between the SCUBA-2 data and the Planck+SCUBA-2 combined data is well representative of the dense gas fraction in the cold dust (${f}_{\mathrm{DG}}$), as also suggested in Csengeri et al. (2016).

Figures 10(a) and (c) show the dense gas fraction maps for G26. In general, the dense gas fraction increases with column density. In panels (b) and (d), we investigate the relationship between the dense gas fraction and flux density for the pixels with S/N > 5 in the SCUBA-2 images. The correlations can be well fitted with logarithmic functions toward high flux density points. The logarithmic function fits can also extend to low flux density points even though the low flux density points show much larger scatter. The ${f}_{\mathrm{DG}}$ in the image from R2 reduction range from 0.13 to 0.73 with a median value of 0.30, while the ${f}_{\mathrm{DG}}$ in the image from R4 reduction range from 0.17 to 0.75 with a median value of 0.37. The dense gas fractions determined from images with different spatial filters are slightly different but only vary less than ∼10%. It seems that only 30%–40% of the cloud gas in G26 is dense.

Zoom In Zoom Out Reset image size

4.6.1. Evidence for Large-scale Compression Flows

The origin of dense filaments in giant clouds still remains a puzzle to astronomers (André et al. 2014). Filaments have been well predicted in simulations of supersonic turbulence in the absence of gravity, which can produce hierarchical structure with a lognormal density distribution seen in observations (Vázquez-Semadeni 1994). Filaments in strongly magnetized turbulent clouds (e.g., B211/3, Musca; Cox et al. 2010; Palmeirim et al. 2013) are oriented preferentially perpendicular to the magnetic field lines, suggesting an important role of magnetic fields in filament formation. Filamentary structures in the Galaxy are preferentially aligned parallel to the Galactic midplane and therefore with the direction of large-scale Galactic magnetic field, suggesting a possible connection between large-scale Galactic dynamics and filament formation (Wang et al. 2015, 2016; Li et al. 2016). The formation of filaments by self-gravitational fragmentation of sheet-like clouds has been seen in simulations of 1D compression (e.g., by an expanding H ii region, an old supernova remnant, or the collision of two clouds; Inutsuka et al. 2015; Federrath 2016; Li et al. 2018). For example, the asymmetric column density profiles of filaments in the Pipe Nebula are most likely the result of large-scale compressive flows generated by the winds of the nearby Sco OB2 association (Peretto et al. 2012).

The "SCOPE" survey aims to resolve Planck cold clumps and reveal their substructures. We will assess the detection rate of filaments in Planck cold clumps in different environments (e.g., spiral arms, interarms, high latitude, expanding H ii regions, supernova remnants). In conjunction with molecular line data from the "TOP" and "SAMPLING" surveys, which provide kinematic information, we will be able to study the formation mechanisms of dense filaments in widely different environments. Below we present evidence of compression flows that may be responsible for the formation of the G26 filament.

As shown in the top panel of Figures 23 and 9(b), the temperature gradient on large scales is perpendicular to the dense filament G26. Figures 11(a)–(b) present maps of the first moment of ¹²CO (1–0) and ¹³CO (1–0), respectively. The maps reveal large-scale velocity gradients along the NW–SE direction across the whole map. The temperature gradient suggests an asymmetric heating source. Together with the velocity gradient, one can think of compression flows from the left.

Zoom In Zoom Out Reset image size

The origin of the large-scale compression flows is unclear. To the east of G26, we found an infrared bubble. Figure 12 shows a three-color composite Spitzer/IRAC image of the infrared bubble region overlaid with SCUBA-2 850 μm continuum emission in contours. As shown in Figure 13, the locations between the bubble and the filament have higher column density (∼1.6 times higher) and higher temperature (∼3 K higher) than the locations west of the filament, suggesting that a pressure gradient exists from the southeast to northwest. Also, the filament shows an asymmetric column density profile (panels (a) and (b)), again indicating that it may be compressed by external pressure from its southeast side. Panels (e) and (f) present the variance of the integrated intensity and intensity-weighted velocity of ¹²CO (1–0) and ¹³CO (1–0) line emission, respectively. Both ¹²CO (1–0) and ¹³CO (1–0) line emissions reveal a large-scale velocity gradient (∼0.16 km s⁻¹ pc⁻¹) perpendicular to the filament, again suggesting the existence of compression flows.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Note that the infrared bubble only has an extent of ∼5 pc and is ∼15 pc away in projection from the filament G26 (see Figures 23(a) and 26(a) in Appendix A). Hence, the bubble may not have the ability to generate the large-scale compression flows. Since the G26 filament is parallel to the Galactic plane (see Figure 4), the large-scale compression flow may originate from the ram pressure from the OB associations in the Galactic plane. In any case, the higher pressure to the southeast of G26 can continuously sweep up the interstellar gas to feed the dense filament.

Although the infrared bubble seems not to be responsible for the origin of the filament, it may have great impact on its surroundings. The clumps (shown in magenta contours in Figure 12) adjacent to its east seem to be compressed by the bubble and the star formation in those clumps may be triggered by the bubble as revealed near other infrared bubbles (e.g., Liu et al. 2012b, 2015, 2016). More detailed studies of those clumps adjacent to the bubble will be present in another work.

4.6.2. Collisions of Subfilaments?

Figure 14 presents channel maps of the J = 1–0 transitions of ¹²CO, ¹³CO, and C¹⁸O lines. In contrast to the 850 μm continuum emission, the molecular line emission reveals more complicated structures. Several velocity coherent subfilaments (F1 to F6) can be identified from the channel maps. The position angles of F1, F2, F3, F4, F5, and F6 are about 45°, 115°, 0°, 44°, 80°, and 20°, respectively. The subfilaments F3, F4, and F6 are also seen in SMT ¹³CO (2–1) channel maps (see Figure 15). Those subfilaments have noticeable differences in velocity because they emerge in different velocity channels.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

F4 and F6 form the main gas subfilament, most of whose emission is in the 47–48 km s⁻¹ channels. F1, which has the most redshited velocities among the subfilaments, is located to the northwest and only appears in the 51–53 km s⁻¹ channels. F1 seems to interact with the main filament (F6), as well as another sub-filament (F2). The interface between F1 and F6 shows much larger velocity dispersion, as revealed in the second-moment map of ¹³CO (1–0) emission (see the bottom panel of Figure 14). F5 has more blueshifted velocities than other subfilaments. It is connected to F4. The G26 filament as revealed by SCUBA-2 850 μm continuum has a curved (or "S"-type) shape, suggesting that it may be dynamically interacting with its surroundings (Wang et al. 2015, 2016). The continuum filament is mainly associated with the F3 and F4 gas subfilaments. Interestingly, F3 is clearly offset from the axis of the main gas filament (F6) and has redshifted velocities with respect to F6. F3 may be compressed owing to a collision between F1 and F6. The collision may have reshaped the continuum filament into a curved shape. F5 may also collide with F4 and reshape the southern arm of the continuum filament. With the detection of several subfilaments in G26 and because IRDCs are at the very early stages of star formation, we add further support to the idea that IRDCs are in a stage in which they are still being assembled (Jiménez-Serra et al. 2010; Sanhueza et al. 2013).

4.6.3. Filament Accretion

Figure 16 presents position–velocity diagrams of ¹³CO (2–1) line emission along subfilaments F3, F4, and F6. The peak emission of clump 6 is significantly blueshifted with respect to the systemic velocity and the main subfilament. As denoted by yellow dashed lines, the subfilaments connected to the central clump show clear velocity gradients with respect to the peak emission of the central clump. The velocity gradients are ∼0.4, ∼0.6, and 0.5 km s⁻¹ pc⁻¹ for "F3," "F4," and "F6," respectively, which are similar to the values found in other Galactic long filaments (Wang et al. 2016).

Zoom In Zoom Out Reset image size

One interpretation of velocity gradients is that they are caused by inflows along the subfilaments as also seen in other filamentary clouds (Kirk et al. 2013; Peretto et al. 2013; Liu et al. 2016c; Yuan et al. 2018). The velocity differences ($\delta V$) between the subfilaments and central clump become larger as they approach the central clump (i.e., $\delta V\propto \delta {R}^{-1}$, where $\delta R$ is the distance to the central clump), suggesting that the inflows along the subfilaments are likely driven by the gravity of the central clump MM6.

We can estimate the mass inflow rate (${\dot{M}}_{\parallel }$) along filaments following Kirk et al. (2013):

$$\begin{eqnarray}&&{\dot{M}}_{\parallel }=\displaystyle \frac{{\rm{\nabla }}{V}_{\parallel ,\mathrm{obs}}M}{\tan (\alpha )},\end{eqnarray} \tag{ 2 }$$

where ${\rm{\nabla }}{V}_{\parallel ,\mathrm{obs}}$, M, and α are the velocity gradient, mass, and inclination angle with respect to the plane of the sky of a filament, respectively. Since most gas is bounded in dense clumps, only a small fraction of free gas may be accreted along the filament. We estimated the mass of the free gas as the difference (∼1900 M_☉) between the total filament mass (∼6200 M_☉) and the sum of the clump masses (∼4300 ${M}_{\odot }$) from the R4 imaging scheme. We take a mean velocity gradient of $\sim {\rm{\nabla }}{V}_{\parallel ,\mathrm{obs}}=0.5$ km s⁻¹ pc⁻¹ to estimate the net mass inflow rate. Assuming $\alpha =45^\circ$, the inflow rate along the filaments to the central clump is ∼1 × 10⁻³ M_☉ yr⁻¹.

The CO molecular line data from the "TOP" and "SAMPLING" surveys will be used to study the gaseous CO properties (e.g., abundances, CO-to-${{\rm{H}}}_{2}$ conversion factors, CO depletion) of a large sample of PGCCs through joint analysis with the continuum emission data. The KVN survey is designed to observe dozens of dense gas lines toward dense clumps and cores detected in the "SCOPE" survey and characterize their chemical properties. Below, we will investigate the CO depletion of the G26 filament and the dense gas abundances in clump 6.

4.7.1. CO Depletion

Gaseous CO significantly freezes out onto grain surfaces when densities exceed ∼3 × 10⁴ cm⁻³ (Bacmann et al. 2002). As CO is a major destroyer of molecular ions, CO depletion leads to a change in the relative abundances of major charge carriers (e.g., ${{\rm{H}}}_{3}^{+}$, ${{\rm{N}}}_{2}$, H⁺, and HCO⁺; Bergin & Tafalla 2007; Caselli 2011). In addition, the abundances of the nitrogen hydrides and deuterated molecules are strongly enhanced, and these species are probing the gas where CO (and other carbon-bearing species) is depleted (Bergin & Tafalla 2007; Caselli 2011). Therefore, studies of CO depletion are very important for understanding the chemical processes in star formation. Through statistical studies of a sample of 674 PGCCs, Liu et al. (2013b) found that the CO abundance is strongly (anti)correlated to other physical parameters (e.g., dust temperature, dust emissivity spectral index, column density, volume density, and luminosity-to-mass ratio), suggesting that the gaseous CO abundance can be used as an evolutionary tracer for molecular clouds. Similarly, Giannetti et al. (2014) also found that less evolved ATLASGAL-selected high-mass clumps seem to show larger values for the CO depletion than their more evolved counterparts, and CO depletion increases for denser sources. However, both studies in Liu et al. (2013b) and Giannetti et al. (2014) used single-pointing molecular CO data, which limited the accurate determination of CO depletion in molecular clouds. The molecular CO mapping survey data in the TOP survey are more suitable for investigating how CO abundances or CO depletion varies inside individual molecular clouds and changes in different kinds of molecular clouds.

The top panel of Figure 17 presents the spectra of ¹³CO (1–0), ¹³CO (2–1), and C¹⁸O (1–0) at the continuum peak for G26. The ¹³CO (2–1) and C¹⁸O (1–0) emissions have a spatial distribution similar to that of the SCUBA-2 850 μm continuum as shown in the bottom panel of Figure 17.

Zoom In Zoom Out Reset image size

C¹⁸O (1–0) emission is optically thin in G26 (see Table 2). We derived C¹⁸O column densities following Liu et al. (2013b) by assuming that the excitation temperature of C¹⁸O (1–0) is equal to the dust temperature. The C¹⁸O abundance (${{\rm{X}}}_{{{\rm{C}}}^{18}{\rm{O}}}$) was derived by comparing the C¹⁸O column densities with the ${{\rm{H}}}_{2}$ column density map from the SED fits with Herschel data and Planck+SCUBA-2 data. The C¹⁸O column density map and ${{\rm{H}}}_{2}$ column density map were smoothed to the same resolution of $\sim 52^{\prime\prime}$. The CO gas depletion factor (${f}_{D}$) is defined as ${f}_{D}=\tfrac{\max ({X}_{{{\rm{C}}}^{18}{\rm{O}}})}{{X}_{{{\rm{C}}}^{18}{\rm{O}}}}$, where max(${{\rm{X}}}_{{{\rm{C}}}^{18}{\rm{O}}}$) is the maximum C¹⁸O abundance ($\sim 1.5\times {10}^{-7}$) across the map with C¹⁸O emission larger than 3σ. The CO gas depletion factor map is presented in Figure 18. The highest depletion (${f}_{D}\,\sim$5) occurs at the central massive clump ("6") owing to its largest column density and relatively low temperature. However, since MM6 hosts a relatively strong IR source (bright at 24 and 70 μm), the higher CO depletion means that the star formation process inside MM6 is really new and most of the clump gas is still cold. Due to the large beam of TRAO observations, we see most of the bulk, cold gas rather than the small, localized hot region, further suggesting that the IRDC G26 is still very young.

Zoom In Zoom Out Reset image size

The CO gas in the surroundings of the central clump MM6 is less depleted (${f}_{D}\sim 1.5\mbox{--}2.5$). The outskirts of the filament also show higher apparent CO depletion, which reflects the lower CO abundance therein. The low CO abundance in the outskirts suggests that CO gas is released from dust grains owing to external heating from interstellar radiation and then photodissociated.

4.7.2. Molecular Lines from KVN 21 m Single-pointing Observations

We observed the massive clump "6" in molecular lines that trace dense gas with the KVN 21 m single dishes to characterize its chemical properties. Figure 19 presents the spectra from KVN 21 m single-pointing observations. HCO⁺ J = 1–0 and ${{\rm{H}}}_{2}$CO ${2}_{\mathrm{1,2}}-{1}_{\mathrm{1,1}}$ lines show "red asymmetry profile" morphology with line wings, indicating that their emission is affected by outflows. H¹³CO⁺ J = 1–0 and ${{\rm{N}}}_{2}$H⁺ J = 1–0 spectra show two velocity components. Since the H¹³CO⁺ J = 1–0 and the isolated hyperfine components of ${{\rm{N}}}_{2}$H⁺ J = 1–0 lines are usually optically thin in IRDCs (Sanhueza et al. 2012), their two velocity components should not be the result of self-absorption. Indeed, the two velocity components may indicate the existence of converging flows inside the clump as resolved in other high-mass clumps (e.g., SDC 335, G10.6–0.4, G33.92+0.11, AFGL 5142; Liu et al. 2013a, 2015, 2016c; Peretto et al. 2013) or interacting subfilaments (e.g., G028.23–00.19; Sanhueza et al. 2013).

Zoom In Zoom Out Reset image size

We did not detect the 22 GHz water maser but detected the 44 GHz methanol maser. The 44 GHz methanol maser has two velocity components peaked at ∼47.9 and ∼48.7 km s⁻¹, respectively. The 44 GHz methanol maser is blueshifted with respect to the systemic velocity (∼49 km s⁻¹) and has much smaller line widths than other lines that trace denser gas.

SiO emission shows broad lines and two velocity components. The blueshifted component in SiO emission has much broader line widths than the central component, which may be caused by outflow shocks. CCH N = 1–0 and ${{\rm{N}}}_{2}$H⁺ (1–0) each have hyperfine structures. Their hyperfine lines were fitted assuming LTE to obtain their line widths and optical depths. HCN (1–0) shows a very complicated line profile, which is likely caused by its hyperfine structures and self-absorption. We did not fit the HCN (1–0) spectrum. The other lines were fitted with a Gaussian function. The line parameters are summarized in Table 2. The systemic velocity (∼49 km s⁻¹) is obtained from averaging the peak velocities of the optically thin lines (C¹⁸O, ${\mathrm{HC}}_{3}{\rm{N}}$, and HN¹³C). The HDCO ${2}_{\mathrm{0,2}}-{1}_{\mathrm{0,1}}$ line was not detected at an rms level of 0.06 K.

We estimated the column densities from molecular lines with the RADEX⁸⁴ radiation transfer code by fixing the kinetic temperature to 17 K, i.e., the dust temperature. For C¹⁸O, ${{\rm{C}}}_{2}$H, and ${{\rm{N}}}_{2}$H⁺, the ${{\rm{H}}}_{2}$ volume density of clump 6 is set to 6 × 10³ cm⁻³ from the R4 imaging scheme. For the other lines (H¹³CO⁺, HN¹³C, and ${\mathrm{HC}}_{3}$N), the ${{\rm{H}}}_{2}$ volume density of clump 6 is set as $1.8\times {10}^{4}$ cm⁻³ from the R2 imaging scheme since their effective excitation densities are about one order of magnitude higher than that of ${{\rm{N}}}_{2}$H⁺ (Shirley 2015).

As shown in Figure 20, with three transitions, the source size (${\theta }_{S}$) of the ${\mathrm{HC}}_{3}{\rm{N}}$-emitting area can be well determined, i.e., 11''. We applied the same filling factor $\left(f=\tfrac{{\theta }_{S}^{2}}{{\theta }_{S}^{2}+{\theta }_{\mathrm{beam}}^{2}}\right)$ for the H¹³CO⁺ and HN¹³C calculations. We assume a filling factor of 1 in the RADEX calculations for the other lines. The RADEX-derived excitation temperature, optical depth, column density, and abundance from each transition are listed in the last four columns of Table 2.

Zoom In Zoom Out Reset image size

As a major destroyer of ${{\rm{N}}}_{2}$H⁺, the absence of CO in the gas phase could enhance the abundance of ${{\rm{N}}}_{2}$H⁺. In warm regions, however, CO evaporation could destroy ${{\rm{N}}}_{2}$H⁺ and enhance the abundance of HCO⁺ by the following reaction (Lee et al. 2004; Busquet et al. 2011):

$$\begin{eqnarray}&&{{\rm{N}}}_{2}{{\rm{H}}}^{+}+\mathrm{CO}\to {\mathrm{HCO}}^{+}+{{\rm{N}}}_{2}.\end{eqnarray} \tag{ 3 }$$

Therefore, a small ${{\rm{N}}}_{2}$H⁺-to-H¹³CO⁺ abundance ratio may indicate that clumps are warm and thus chemically evolved. Feng et al. (2016) recently observed a sample of four 70 μm dark IRDC clumps, whose CO seems to be heavily depleted with depletion factors in the range of 14–50, i.e., about 3–10 times larger than that of G26 clump 6. They also derived high $[{{\rm{N}}}_{2}$ H⁺]/[H¹³CO⁺] abundance ratios (∼400–10,000) in the 70 μm dark young IRDC clumps, which are more than two orders of magnitudes larger than that of clump 6 in G26 (∼0.8). Therefore, clump 6 in G26 is more chemically evolved than those IRDC clumps. In addition, the abundance of ${\mathrm{HC}}_{3}$N, a hot core tracer, is about two orders of magnitude larger in clump 6 than in the 70 μm dark IRDC clumps, indicating that G26 clump 6 may be in a more chemically evolved phase.

The CO molecular line data from the "TOP" and "SAMPLING" mapping surveys can be used to study large-scale kinematics (e.g., collisions, filament accretion, outflows). The HCO⁺ J = 1–0 and ${{\rm{H}}}_{2}$CO ${2}_{\mathrm{1,2}}-{1}_{\mathrm{1,1}}$ lines in the KVN survey are also good tracers of kinematics like infall and outflows associated with dense clumps/cores. Below, we study the kinematics of G26 based on the ¹³CO (2–1) line emission from the "SAMPLING" survey and the HCO⁺ J = 1–0 and ${{\rm{H}}}_{2}$CO ${2}_{\mathrm{1,2}}-{1}_{\mathrm{1,1}}$ lines from KVN observations.

4.8.1. High-velocity Outflows

Outflow wings can be easily identified in the position–velocity diagrams in Figure 16. Figure 21 shows the integrated intensity maps of the outflow emission. Since the blue and red outflow lobes are well separated and have very high flow velocities; neither the very low (pole-on) inclination nor the very high (edge-on) inclination is likely. Therefore, we take 45° as the inclination angle for the outflows with respect to the line of sight. The radii of the blueshifted and redshifted lobes are ∼1.6 and ∼1.1 pc, respectively. The characteristic outflow velocities of the blueshifted and redshifted lobes are ∼5.1 and ∼3.3 km s⁻¹, respectively. Assuming that the excitation temperature of ¹³CO (2–1) line wing emission is 17 K, i.e., the dust temperature, the integrated intensity can be converted to the outflow masses following Garden et al. (1991). The ¹³CO abundance (7 × 10⁻⁷) is converted from C¹⁸O abundance (1 × 10⁻⁷) with a [¹³CO/C¹⁸O] abundance ratio (∼7) derived from empirical relations between molecular abundances and the galactocentric distance (Wilson & Rood 1994; Giannetti et al. 2014). The masses of the blueshifted and redshifted lobes are ∼67 and ∼97 M_☉, respectively. The momentum of the blueshifted and redshifted lobes is ∼342 and ∼314 M_☉ km s⁻¹, respectively. The energies of the blueshifted and redshifted lobes are ∼1.7 × 10⁴⁶ and ∼1.1 × 10⁴⁶ erg, respectively. The dynamical timescales $\left({t}_{\mathrm{dyn}}=\tfrac{2R}{{V}_{\mathrm{flow}}}\right)$ of the blueshifted and redshifted lobes are $6.2\,\times {10}^{5}\,\mathrm{yr}$ and $6.8\times {10}^{5}\,\mathrm{yr}$, respectively. The total outflow mass-loss rate is $\sim 2.5\times {10}^{-4}$ M_☉ yr⁻¹.

Zoom In Zoom Out Reset image size

4.8.2. Outflow-driven Expansion

The HCO⁺ J = 1–0 and ${{\rm{H}}}_{2}$CO ${2}_{\mathrm{1,2}}-{1}_{\mathrm{1,1}}$ lines show "red asymmetric profiles" morphologies, suggesting expansion. We apply a simple "two-layer model" to fit the lines. The model is described in Appendix B. The results are presented in Table 3. The expansion velocities inferred from the HCO⁺ J = 1–0 and ${{\rm{H}}}_{2}$CO ${2}_{\mathrm{1,2}}-{1}_{\mathrm{1,1}}$ lines are ∼0.15 and ∼0.06 km s⁻¹, respectively. Assuming that the whole clump (with M ∼ 2100 M_☉) is expanding with a velocity of 0.1 km s⁻¹, the total momentum and energy of the expanding clump are 210 M_☉ km s⁻¹ and ∼2.1 × 10⁴⁴ erg, respectively. Since it is unlikely that the whole clump is in expansion, these momentum and energy values should be treated as upper limits. Even in this extreme case, however, both the momentum and energy are smaller than those of the low-velocity outflows, indicating that the outflows have enough energy themselves to drive the envelope expansion.

Table 3.  Parameters of the Two-layer Model

Line	${\tau }_{0}$	Φ	${J}_{c}$	${J}_{f}$	${J}_{r}$	${V}_{\mathrm{cont}}$	σ	${V}_{\mathrm{out}}$
			(K)	(K)	(K)	(km s−1)	(km s−1)	(km s−1)
HCO+ J = 1–0	3.56	0.74	15.70	11.69	19.60	48.98	1.04	0.15
${{\rm{H}}}_{2}$CO ${2}_{\mathrm{1,2}}-{1}_{\mathrm{1,1}}$	2.21	0.79	15.44	12.03	19.54	49.07	0.93	0.06

Download table as:  ASCIITypeset image

The virial mass of the clump MM6 considering turbulent support can be derived as $\tfrac{{M}_{\mathrm{vir}}}{{M}_{\odot }}=210\left(\tfrac{R}{\mathrm{pc}}\right){\left(\tfrac{{\rm{\Delta }}V}{\mathrm{km}{{\rm{s}}}^{-1}}\right)}^{2}$ (MacLaren et al. 1988; Zhang et al. 2015), where R is the radius (1.15 pc) and ${\rm{\Delta }}V$ is the line width (2.22 km s⁻¹) of C¹⁸O (1–0). The virial mass is ∼1200 M_☉, which is much smaller than the clump mass (∼2100 ${M}_{\mathrm{Sun}}$), indicating that the whole clump may still be in collapse. Therefore, the expansion revealed by HCO⁺ J = 1–0 and ${{\rm{H}}}_{2}$CO ${2}_{\mathrm{1,2}}-{1}_{\mathrm{1,1}}$ lines should be very localized.

Figure 22 shows a cartoon to describe the PGCC G26 based on the results from the present data. Our findings of G26 are as follows: (i) the G26 filament may be formed as a result of large-scale compression flows evidenced by the temperature and velocity gradients across its natal cloud, (ii) subfilaments in G26 have different velocities and may interact with each other, (iii) the central massive clump where massive stars are forming may still accrete gas along subfilaments, and (iv) the entire massive clump may still be in global collapse while its inner part seems to be undergoing expansion due to outflow feedback.

Zoom In Zoom Out Reset image size

G26 was mapped with Herschel as part of the Hi-GAL project (Molinari et al. 2010). The observations include 70, 160, 250, 350, and 500 μm bands that were observed in parallel mode with 60'' s^–1 scanning speed. We use the level 2.5 maps available in the Herschel Science Archive,⁸⁵ SPIRE data (250–500 μm ) using extended source calibration, and the PACS maps (70, 160 μm ) made with the UNIMAP method (Piazzo et al. 2015). The resolutions of the original maps are approximately 12'', 15'', 183, 249, and 363, respectively. Given the fast scanning speed, the effective PACS beams are elongated.

Figure 23 presents the dust temperature and column density maps of PGCC G26.53+0.17 derived from Herschel/SPIRE data. The column density maps in Figure 23 are used as external masks in the below JCMT/SCUBA-2 data reduction. We fit the SEDs pixel by pixel with a modified blackbody function using the three Herschel/SPIRE bands. The modified blackbody function is

$$\begin{eqnarray}&&F(\nu )=F({\nu }_{0})\displaystyle \frac{B(\nu ,T)}{B({\nu }_{0},T)}{\left(\displaystyle \frac{\nu }{{\nu }_{0}}\right)}^{\beta }=\kappa (\nu )B(T)M/{d}^{2},\end{eqnarray} \tag{ 4 }$$

where $F(\nu )$ is the flux density at frequency ν, ${\nu }_{0}$ is a selected reference frequency, B stands for the Planck law, β is the opacity spectra index, M is the clump mass, and d is the cloud distance. The Herschel data were fitted with modified blackbody spectra with $\beta =1.8$ to estimate color correction factors and to derive estimates of the dust optical depth at a common spatial resolution of 40''. The values were also converted to estimates of hydrogen column density assuming a dust opacity of κ = 0.1(ν/1000 GHz)^β cm² g^–1 (Beckwith et al. 1990).

Zoom In Zoom Out Reset image size

The top panel of Figure 23 presents the dust temperature and column density maps derived from Herschel/SPIRE data without applying any local spatial filtering. A temperature gradient along the SE–NW direction is seen in the dust temperature map. The bottom panel in Figure 23 presents the column density map of the G26 filament derived from Herschel/SPIRE data with the extended emission larger than 300'' filtered out. The last closed contours in the column density maps outline the external masks used in the JCMT/SCUBA-2 data reduction in Section 6.2.

The CV Daisy observations of G26 were conducted on 2016 April 21. To test how well CV Daisy observations can recover faint and extended emission, we also conducted Pong1800 maps from 2017 May 13 to 23 for comparison. In contrast to the CV Daisy observations, the Pong1800 maps, which are designed for mapping larger fields, cover a circular area of ∼30' diameter with a uniform sensitivity and thus better recover extended emission.

The JCMT data reduction methods are similar to those used in the JCMT Gould Belt survey and JCMT transient survey (Ward-Thompson et al. 2007; Mairs et al. 2015, 2017; Herczeg et al., submitted). The data reduction procedure was performed using the iterative mapmaking technique makemap (Chapin et al. 2013) in the SMURF package of the Starlink software (Jenness et al. 2013; Currie et al. 2014). The user can also supply an external mask that surrounds the astronomical signal to constrain the solution derived by makemap, which, together with a larger spatial filter, can better recover faint and extended structure. Generally, the largest recoverable scales are $\sim 600^{\prime\prime}$ before atmospheric signal becomes significant. The details of JCMT data reduction can be found in Mairs et al. (2015, 2017). In this paper, we performed four individual data reductions labeled R1, R2, R3, and R4 with different external masks and spatial filters for comparison. The parameters of makemap used in the data reductions are the same as in Mairs et al. (2015).

(1) R1: We first run makemap with an effective spatial filter of 200'' and no external mask. Then an external mask was constructed from the final map with S/N > 3. The external mask was used to constrain the solution derived by makemap in a second run. Therefore, the structure was filtered to 200'' in the R1 data reduction. This data reduction is efficient to detect dense clumps and is used in the first data release of the "SCOPE" project. The top panel of Figure 24 shows the SCUBA-2 850 μm continuum emission in the R1 data reduction.

Zoom In Zoom Out Reset image size

(2) R2: An external mask from the Herschel data (the last closed contour in the bottom panel of Figure 23) was applied in the data reduction. The external mask corresponds to the contour of 3% of the peak column density value, which is the largest astronomical structure (see the blue contour in the bottom panel of Figure 24) that could be recovered in the CV Daisy map. Besides the external mask, an effective spatial filter of 300'' was applied in the data reduction with makemap. The bottom panel of Figure 24 shows the SCUBA-2 850 μm continuum emission in the R2 data reduction. Compared with R1, R2 better reveals not only dense clumps but also extended structures between them. The peak flux density (1.40 Jy beam⁻¹) in R2 is slightly larger than the value (1.35 Jy beam⁻¹) in R1.

(3) R3: Same as R1, but with a larger effective spatial filter of 600''. Figure 25(a) shows the SCUBA-2 850 μm continuum emission in the R3 data reduction. In contrast to R1 and R2, R3 recovers more extended emission.

Zoom In Zoom Out Reset image size

(4) R4: Same as R2, but with a larger effective spatial filter of 600''. The external mask was constructed from the Herschel data without any spatial filtering (the last closed contour in the top panel of Figure 23). The external mask corresponds to the contour of 15% of the peak column density value, which is the largest astronomical structure (see the blue contour in Figure 25(b)) that could be recovered in the CV Daisy map. Figure 25(b) shows the SCUBA-2 850 μm continuum emission in the R4 data reduction. Only the central coherent structure in the masked region is reliable. In general, R4 and R3 reveal similar structure. With the external mask from Herschel, makemap, however, recovers more extended and faint structures in R4 than in R3. Figure 25(c) reveals the extended emission in the difference image between R4 and R2. R4 recovers a more flattened structure with extended emission (∼50 mJy beam⁻¹) than R2.

In general, maps with external masks recover more flux. Therefore, we suggest to apply external masks in SCUBA-2 data reduction if Herschel data are available. Although R1 was used in the first data release of the SCOPE survey, R1 cannot recover the flux of the dense clumps as well as the other data reductions (R2, R3, and R4). R3 can recover most flux but detect the least number of dense clumps. Therefore, in this paper (except in Section 4.2), we mainly use the data reductions R2 and R4 in the further analysis. R2 is very efficient to reveal dense structures (e.g., dense clumps and filaments). In contrast, R4 recovers more extended structures surrounding the dense clumps.

To verify the detection of faint and extended structure in CV Daisy maps, we compare the Daisy and Pong1800 maps, both of which are obtained with the R4 imaging scheme. Figure 26(a) shows the Pong1800 map. In Figure 26(b), we compare Daisy and Pong1800 maps. In general, the Daisy map and the Pong1800 map show very consistent morphology, indicating that the Daisy map can recover large-scale structures as well as the Pong map. The flux recovered in the Daisy map is also linearly correlated with the flux in the Pong map as shown in Figure 26(c): ${\mathrm{Flux}}_{\mathrm{Pong}}=(0.883\pm 0.003){\mathrm{Flux}}_{\mathrm{Daisy}}$ $-\,(28.062\,\pm \,0.532);{R}^{2}\,=0.95$. The Pong1800 map will be discussed in detail in another paper.

Zoom In Zoom Out Reset image size