© ESO, 2016

1. Introduction

In contrast to low-mass stars, the process by which massive stars form remains poorly understood. The characterization of the initial conditions of massive star and star cluster formation is indeed challenging: massive stars are less common and have much shorter lifetimes than low-mass stars. The formation of massive stars and star clusters is believed to start in cold and dense molecular structures. When viewed against the bright Galactic mid-infrared (IR) background, these clouds are called infrared dark clouds (IRDCs). While some show evidence for ongoing star formation, others appear to be starless. Thus, IRDCs represent the best objects in which to study the initial conditions for massive star and star cluster formation. This has been verified by single-dish observations in low resolution (e.g., Pillai et al. 2006; Chambers et al. 2009) and interferometric observations in high resolution (e.g., Wang et al. 2011, 2012, 2014; Longmore et al. 2013; Tan et al. 2013; Zhang et al. 2015).

Some of the early studies of IRDCs focused on the characterization of the global physical properties of IRDCs using Spitzer 8 μm data, i.e., based on extinction maps (Rathborne et al. 2006, 2008; Peretto et al. 2010; Butler & Tan 2009). However, such studies have a statistical bias because IRDCs, by selection, only represent the population of dense clouds projected onto the bright background of the inner Galaxy, as they need a favorable location to absorb background infrared radiation. Not all the dark patches are real star-forming objects. IRDC catalogs based only on mid-infrared data indeed overestimate the number of real sources by a factor of ~2 (Jackson et al. 2008). In addition, Wilcock et al. (2012) studied 3171 IRDC candidates within the l = 300−330° region of the Hi-GAL survey. Only 38% of these candidates were bright at the Herschel wavelengths, and therefore associated with cold cloud structures. The other mid-infrared IRDC candidates are simply minima in the mid-infrared background.

An unbiased selection criterion to identify the whole Galactic population of massive prestar and precluster forming clouds is based on the emission properties of dust in the far-IR and submillimeter. The APEX Telescope Large Area Survey of the Galaxy (ATLASGAL; Schuller et al. 2009) and the CSO Bolocam Galactic Plane Survey (BGPS; Aguirre et al. 2011) have provided catalogs of submillimeter sources throughout the inner Galaxy. The outer Galaxy is however out of the coverage for these systematic surveys.

For the first time, the Planck satellite has provided an inventory of the cold condensations throughout the Galaxy. The Planck survey covered the submillimeter-to-millimeter wavelength range with unprecedented sensitivity, furnishing the first all sky catalog of cold objects. This is thus a perfect database to identify the coldest structures in the whole Galaxy, offering the opportunity to search for massive star-forming clumps in the outer Galaxy also and to investigate the dependence of massive star formation with location in the Galactic disk. While in extragalactic studies only the average star formation properties can be studied (Kennicutt & Evans 2012), a more detailed analysis can be performed in the Galaxy toward different environments (e.g., the Galactic center, spiral arms, and outer Galaxy). The angular resolution of the Planck observations is however scarce for a thorough study of these objects, and higher angular resolution observations are needed to obtain detailed information about the temperature, column density structure, and star formation content of these cold objects.

In this paper we present the analysis of the Herschel observations of 48 Planck sources that fall within the Hi-GAL (Molinari et al. 2010; Molinari 2011, 2012) survey area. We use the Planck ECC catalog, which provides continuum fluxes for over 900 Planck sources across a broad range of wavelengths (from 350 to 850 μm), to systematically analyze the distribution of cold clumps in the Galactic plane. The higher spatial resolution of Herschel PACS and SPIRE (Poglitsch et al. 2010; Griffin et al. 2010) makes it possible to examine the internal structure of these Planck sources and allows us to identify their high column density peaks and sample their deeply embedded sources with angular resolutions of ~5″− 36″. In Sect. 2 we present the sample and datasets used in this paper, including the distance estimate for our sample. In Sect. 3 we describe the analysis methods and results based on the Hi-GAL and Planck data. Finally, in Sect. 4 we discuss our results.

2. Data and sample selection

2.1. Planck catalogs of Galactic cold core objects

The Planck satellite (Tauber et al. 2010) has provided an all-sky submillimetre and millimetre survey covering the wavelengths around and longward of the intensity maximum of cold dust emission. While ν²B_ν(T = 10 K) peaks close to 300 μm, the coldest dust with a temperature of T ~ 6 K shows its maximum emission close to 500 μm. Combined with far-infrared data such as the IRAS survey, the data enable accurate determination of both the dust temperature and spectral index. The Early Cold Cores¹ Catalogue (ECC; Planck Collaboration XXIII 2011) is part of the Planck Early Release Compact Source Catalogue (ERCSC) and the ECC forms a subset of the full Cold Core Catalogue of Planck Objects (C3PO; more than 10 000 objects have been detected; Planck Collaboration XXIII 2011). The ECC, with a total of 915 objects over the sky, contains only the most secure detections (signal-to-noise ratio (S/N) > 15) with color temperatures below 14 K.

The Planck Catalogue of Galactic Cold Clumps (PGCC; Planck Collaboration 2015) is the last release of Planck clump data tables. It contains 13188 Galactic sources with a temperature ranging from 5.8 K to 20 K. Forty-two percent of these sources have available distance estimates. A few ECC sources are not part of the PGCC catalog because they do not satisfy the compactness criterion of the PGCC catalog, i.e., they are slightly more extended compared to the other sources. We note however that the old ECC and new PGCC data are in good agreement for the selected 48 clumps. Indeed, the derived dust temperatures, spectral indices, and clump sizes differ by less than 6%, 1.5%, and 2%, respectively.

2.2. Planck ECC sources in the Hi-GAL survey

We selected the ECC sources that were also covered by the Hi-GAL survey to build up our sample. Hi-GAL was one of the Herschel Open Time Key Projects (Molinari et al. 2010) and, with its extensions (in OT1 and Hi-GAL2pi; Molinari 2012, 2011), this survey mapped the entire Galactic Plane of the Milky Way (−1°<b< + 1°, following the Galactic warp) in five bands with the PACS instrument at 70 and 160 μm and with the SPIRE instrument at 250, 350, and 500 μm with spatial resolutions of 5″, 13″, 18″, 25″ and 36″, respectively.

We selected the 48 ECC sources that were also covered by the Hi-GAL survey (see Table A.1). These are all the ECC objects that fall within the Herschel Hi-GAL survey area and all five Herschel band images are available for each source. Three sources from our sample are not part of the latest PGCC catalog owing to their lower level of compactness (G001.64-00.07, G201.13+00.31, and G296.52-01.19), but we included these sources in our analysis. Based on their physical properties, there are no significant differences between these three rejected ECC clumps and the rest of the ECCs within the Hi-GAL region (see later in Sect. 3.3).

The main parameters associated with the Planck clumps are shown in Table A.1. The mass calculation based on Planck data is described in Sect. 3.1.

2.3. Distance determination of ECC objects

Distance determination is needed for the mass and size calculation of the ECC clumps. For this purpose, association with known IRDCs and molecular line follow up survey data were used. Table B.1 shows all the available estimated distances for the ECC objects and the final adopted distances. The data come from different sources:

• 1.

  Kinematic distance based on the CO J = 1−0 survey data of the PurpleMountain Observatory by Wuet al. (2012). COJ = 1−0 emission is detected in all ECC clumps except one.Among these sources, 28 clumps belong to our Hi-GAL ECCsample and have velocity and distance estimates in the Wuet al. (2012) survey. Half of them(15 sources) show more than one velocity components. In thosecases, we used the velocity component with the brightest lineintensity to estimate the distance of the ECC clump.

• 2.

  Kinematic distance based on the MALT90 (Millimetre Astronomy Legacy Team 90 GHz, ATNF Mopra 22-m telescope; Jackson et al. 2013) survey data. Six ECC clumps were covered by this survey. For the clumps with more than one molecular line detected (i.e., N₂H⁺, ¹³CO, and ¹²CO), we adopted the kinematic distance derived from the species with higher critical density.

• 3.

  Kinematic distance based on our own APEX observations (Zahorecz et al., in prep.). APEX observations were performed toward three ECC clumps as part of the E-093.C-0866A-2014 project. Again, the kinematic distance adopted for these clumps was that associated with the emission from the species with higher critical density.

• 4.

  Kinematic distance based on the CfA CO survey (Dame et al. 2001) data. Twenty-four objects out of our sample of 48 Planck clumps show a single or dominating velocity component in the CfA ¹²CO survey data. In Fig. 1, we show some examples of the CfA CO survey spectra measured for the ECC sources that do not have any distance estimate from the Purple Mountain Observatory survey.

• 5.

  Associations with known IRDCs. Six ECC clumps are associated with known IRDC objects. Their kinematic distances were obtained by Parsons et al. (2009) based on the Galactic Ring Survey data.

• 6.

  Distance from the Planck PGCC catalog. This distance determination available in the PGCC catalog was obtained using several methods: kinematic distance estimates, optical extinction based on SDSS DR7, near-IR extinction toward IRDCs, and near-IR extinction. In case of the near-IR extinction methods, negative values indicated upper limits.

The distance flags for the ECC clumps considered in our sample are indicated in Table A.1. Once the molecular line velocity information was collected for our sample, we estimated the kinematic distance of the ECC clumps using the rotation curve of the Galaxy from Reid et al. (2014). We employed the publicly available Fortran code from Reid et al. (2009) to relate the radial velocity to the source’s Galactrocentric radius. In the inner Galaxy, this Galactocentric radius corresponds to two distances along the line of sight: the near and far kinematic distances. The near kinematic distance was adopted if kinematic distance estimation alone was available for the source. For the sources in the outer Galaxy, the distance solution is unique.

We note that in 12 cases, both kinematic- and extinction-based distance estimates are available. The average difference between these two methods is 10%, although for two objects they differ by more than 50%. When an extinction-based distance estimate was available, we used that as the adopted distance (see Table A.1) since this method reproduces maser parallax distances (with very low level of uncertainties) better than kinematics distances (see Foster et al. 2012). Otherwise, we adopted the kinematic distance estimates obtained for the remaining ECC clumps.

Calculations of mass, size, and H₂ column density were performed only for the 40 clumps with available distance estimations. The distances are between 0.1 kpc and 8.1 kpc, the angular sizes lie between 4′− 15′, which give us spatial scales in the range of 0.5 pc−29 pc.

We note that the distance estimates for the majority (92%) of our ECC clumps are likely correct within a factor of 2, and so our mass estimates also lie within a factor of 4. Only three sources (i.e. 8% of the sample) may have distance discrepancies as large as a factor of 5, implying a factor of 25 uncertainty in the calculated mass. Therefore, our results on the mass distribution of ECC clumps as a function of Galactic longitude and Galactocentric distance are not expected to change substantially because the errors in the distance determination.

2.4. Location within the Galaxy

Once the distance of the selected ECC clumps has been estimated, we can determine the location where these objects fall within the Galaxy. Figure 2 shows the Galactic distribution of the selected clumps. The figure shows that 1/3 and 2/3 of these clumps are located in the inner and outer part of the Galaxy, respectively, between Galactic longitudes of 278° to 89° and 144° to 262°. Based on our estimated distances, 60% of these clumps fall on top of the Galactic spiral arms as defined in Reid et al. (2014).

3. Methods and results

3.1. Mass determination based on Planck data

From the fluxes measured by Planck for every cold clump and provided in the ECC catalog, we can estimate their individual masses using $\mathit{M}\mathrm{=}\frac{{\mathit{S}}_{\mathit{\nu }}{\mathit{D}}^{\mathrm{2}}}{{\mathit{\kappa }}_{\mathit{\nu }}{\mathit{B}}_{\mathit{\nu }}\mathrm{\left(}\mathit{T}\mathrm{\right)}}\mathit{,}$(1)where S_ν is the integrated flux density, D is the distance, κ_ν is the dust opacity, and B_ν(T) is the Planck function for dust temperature at T. The dust opacity was adopted by Planck Collaboration XXIII (2011) as${\mathit{\kappa }}_{\mathit{\nu }}\mathrm{=}\mathrm{0.1}\mathrm{\times }{\left(\frac{\mathit{\nu }}{\mathrm{1}\mathrm{THz}}\right)}^{\mathit{\beta }}\mathrm{c}{\mathrm{m}}^{\mathrm{2}}\mathit{/}\mathrm{g}\mathit{.}$(2)For this calculation, we used the integrated flux density at the frequency ν = 857 GHz from the ECC catalog, the derived core temperature, T_core, and the emissivity spectral index, β_core. We note that the Planck fluxes at ν = 857 GHz were used for the computation because this frequency is the closest one to the frequency reference (i.e., 1 THz) for the formula of the dust opacity (see Eq. (2)), so that the impact on the spectral index β remains small compared to the uncertainty in κ_ν. In addition, the 857 GHz band also has the best S/N of all Planck observations and dust emission at this frequency is optically thin (see details in Planck Collaboration XXIII 2011).

The average calculated mass based on the Planck data is 1.2 × 10⁶M_⊙ with a minimum and maximum value of 10 M_⊙ for G089.62+02.16 and 2.7 × 10⁷M_⊙ for G000.65-00.01, respectively. This average value is biased by a few very massive sources within the sample. The 25th, 50th, and 75th percentile of the data are 1.6 × 10²M_⊙, 1.7 × 10³M_⊙, and 2.2 × 10⁴M_⊙, respectively.

3.2. Clump categories based on Hi-Gal 70 μm data

In this section, we evaluate the level of star and star cluster formation within the Planck cold clumps observed by Herschel with the 70 μm Hi-GAL images. We note that the 70 μm images have the best angular resolution within the Hi-GAL data and therefore provide information about the point-like sources found within the clumps and likely associated with protostellar objects. Also it is an excellent probe of the population of intermediate- and high-mass young stellar objects in star-forming regions (see, e.g., Ragan et al. 2012; Veneziani et al. 2013). Planck clumps without 70 μm sources are the best candidates to study the earliest phases of star formation, since the absence of 70 μm emission likely indicates that protostellar activity has affected its surrounding environment very little.

In order to investigate whether there are significant differences in the physical properties (i.e., dust temperature, mass, and maximum H₂ column density) between clumps with and without star formation, we divided the sample of Planck cold clumps into two different categories: clumps with no sign of star formation (Category I, 11 sources) and clumps with active star formation (Category II, 37 sources).

We probe the star formation activity by checking the presence or absence of 70 μm point sources within the defined clump boundaries (see Sect. 3.3). From our classification of clumps with or without 70 μm point-sources, we find that 11 clumps (23% of the sample of 48 clumps) do not show any star formation activity. The criterion may be distant dependent because of the sensitivity of the 70 μm observations. To check for this possible bias, we plotted the distribution of ECC clumps with and without 70 μm sources as a function of distance (see Fig. 3). We find that sources without 70 μm sources are not, on average, further away than sources with 70 μm emission. On the contrary, most of the cores with no 70 μm sources lie within ~4.5 kpc from the Sun with only one exception, i.e., G354.81+00.35 at ~6 kpc.

Figure 4 shows examples of Category I (G146.71+2.05) and Category II (G319.35+0.87) clumps. Herschel 70, 250, and 500 μm images are shown with angular resolution of 5″, 18″, and 36″, respectively. A similar set of figures is available for all the 48 sources in Appendix A.

To investigate the possibility of chance alignments, we estimated the expected number of point sources within the clump regions. For this, we identified the sources present within 2 × 2 sq. deg Hi-GAL 70 μm tiles using SExtractor (Bertin & Arnouts 1996). Except for the ECC clumps located in the inner 20 deg of the Galaxy, the expected number of point sources due to chance alignment is less than 2 (median values of 0.02 for | l−180° | < 90° and of 1.8 for 90°< | l−180° | < 170°), which implies that the 70 μm sources detected in the ECC clumps genuinely trace the star formation activity within the clumps. For the ECC clumps in the Galactic center, we note however that the number of expected sources by chance alignment is higher than for the ECC clumps in the Galactic disk (median value of 78 for | l−180° | > 170°). This is related to the H₂ column density value chosen to define the clump boundary (3 × 10²¹ cm^-2, see Sect. 3.3), which covers larger areas. In any case, even for these Galactic center ECC clumps, the number of 70 μm sources detected in the Hi-GAL images is, on average, higher than the expected number of sources by chance alignment.

From the integrated flux of the faintest detected 70 μm point sources, we can estimate the luminosity (and mass) limit below which no source is detected in the Hi-GAL images. The faintest detected sources in the ECC clumps have fluxes F₇₀ ~ 45 mJy. In order to make the conversion from 70 μm flux to luminosity, we use the following correlation (see Dunham et al. 2008): ${\mathit{L}}_{\mathrm{int}}\mathrm{=}\mathrm{3.3}\mathrm{\times }{\mathrm{10}}^{\mathrm{8}}{\mathit{F}}_{\mathrm{70}}^{\mathrm{0.94}}{\left(\frac{\mathit{d}}{\mathrm{0.14}\mathrm{kpc}}\right)}^{\mathrm{2}}{\mathit{L}}_{\mathrm{\odot }}\mathit{,}$(3)where F₇₀ is in [ergs cm^-2 s^-1] and is normalized to 0.14 kpc and d is the distance in kpc. The faintest sources in our ECC clumps (with F₇₀ = 45 mJy) have luminosities 0.2−0.7−1.5−2.6 L_⊙ at 1−2−3−4 kpc. For distances larger than 3 kpc, only high-mass protostars are luminous enough to be visible.

The all-sky YSO candidate catalogs of Tóth et al. (2014) and Marton et al. (2016) can be used to search for low- and intermediate-mass star formation. We find that four of our ECC clumps without 70 μm point sources show AKARI or WISE YSOs. This implies that, although these clumps have not formed massive protostars yet, they show a certain level of low-mass star formation. From this, the total number of clumps without any star formation activity (low mass and high mass) within their clump boundaries is reduced to seven, which represents 15% of the whole sample. Since we are mostly interested in the potential of the ECC clumps to form massive protostars, we hereafter focus on clump classification based on the presence or absence of 70 μm sources alone.

3.3. Physical properties derived from Hi-GAL data

The higher angular resolution of the Herschel Hi-GAL observations allows us to investigate not only the internal physical structure (compactness and fragmentation) of the Planck cold clumps, but also the small-scale variations in the dust temperature and H₂ column density distribution of these clumps. These higher angular resolution data also provide more accurate estimates of the fraction of the mass locked into the densest parts of the clumps, i.e., into the regions that form massive stars and star clusters.

Dust temperature (T_dust) and H₂ column density maps (N(H₂)) were derived by fitting the SED constructed based on the Herschel PACS 160 μm and SPIRE 250−500 μm data pixel by pixel (see the detailed description of the method in Wang et al. 2015). As a first step, we smoothed the maps to the SPIRE 500 μm data resolution (i.e., to 36′′) and we performed background subtraction. See Appendix C for a detailed description of the background subtraction methods we tested. As an example, Fig. 5 shows the 70 μm images (left) and calculated column density (middle) and dust temperature (right) maps for ECC G146.71+02.05 and G319.35+0.87.

To derive the mass, average temperature, and size of the clumps, we defined the boundaries of the clumps on the background-subtracted H₂ column density maps. The peak column density values span 2 order of magnitudes, so we used two different column density thresholds for the source definition: 3 × 10²¹ cm^-2 and 10²² cm^-2. These values correspond to extinctions of ~3 mag and ~9 mag, respectively (see Güver & Özel 2009). These thresholds are shown in Fig. 5 in red (for N(H₂) = 10²² cm^-2) and yellow contours (for N(H₂) = 3 × 10²¹ cm^-2). The different column density thresholds give us information about the fraction of mass associated with the diffuse structure of the clump and the densest parts of the clumps, where star formation is expected to occur.

We note that for one clump, G249.23-01.64, the derived peak column density is smaller than 3 × 10²¹ cm^-2, and therefore its average dust temperature, mass, and size, within the 3 × 10²¹ cm^-2 and 10²² cm^-2 contours, could not be calculated (see Table A.1). We do not detect 70 μm sources in G249.23-01.64. The peak column density is smaller than 10²² cm^-2 for 24 clumps; for 12 of these clumps we have multiple subclumps. In this case, the total mass of the clump was calculated as the addition of the subclump masses. For four sources in the Galactic centre region, the peak column density is larger than 10²² cm^-2, but the masses, sizes, and temperatures within the contours we used were not calculated because the size within these column density limits is not comparable with the size of the ECC. The region within the column density limit of 5−10 × 10²² cm^-2 is comparable with the size of the ECC objects.

In order to provide a characteristic dust temperature for the ECC clumps based on the Hi-GAL data, a column density weighted temperature was calculated within the clump boundary for every object. In this way, the derived temperature is dominated by the temperature of the dense material. The average temperature among all ECC clumps is 13.9 K based on the Herschel images, while the average dust temperature based on the Planck data differs by 1.5−3.0 K with respect to the Herschel one. This difference is discussed in more detail in the next section.

In Table A.1, we report all the calculated average temperatures, sizes, and masses values based on the two column density limits of 10²² cm^-2 and 3 × 10²¹ cm^-2.

3.4. Distribution of the ECC physical properties with Galactic longitude and Galactocentric distance

After calculating the average H₂ column densities, dust temperatures, and masses of the ECC clumps from the Herschel data, we can investigate whether variations in these properties are observed depending on the location of the ECC clumps within the Galaxy.

In Fig. 6, we show the maximum H₂ column density distribution of the ECC clumps as a function of Galactic longitude and Galactocentric distance. ECC clumps toward the Galactic center region show the highest peak column density values. The peak column density decreases with Galactocentric distance. There is a 3 order of magnitude difference between the peak N(H₂) values in the Galactic center region and in the outer part of the Galaxy. From Fig. 6, we also find that H₂ peak column densities N(H₂) > 9 × 10²³ cm^-2 are only detected toward the Central Molecular Zone region. No high column density clumps with N(H₂) > 10²³ cm^-2 are found in the outer Galaxy. There is no clear trend between Category I and Category II clumps and their location within the Galaxy. These two types of clumps are found indistinctly across the Galaxy.

Figure 7 shows the column density weighted temperature, T_weighted, distribution of the ECC clumps as a function of Galactic longitude and Galactocentric distance. Figure 7 also shows similar trends as the distribution of the maximum column density values. ECC objects in the Galactic center show higher temperatures (by more than 5 K difference). For the clumps with Galactocentric distances larger than 6 kpc, there is no significant difference in the calculated temperature between clumps with (in red) and without (in black) 70 μm sources. For Galactocentric distances shorter than 6 kpc, we have too few sources to draw conclusions. The weighted temperature outside the molecular ring is ~13 K, and values larger than 14.5 K are only detected within the molecular ring.

3.5. Comparison of Planck and Hi-GAL temperatures

In Fig. 8, we make the comparison between the clump dust temperature calculated from Planck data, T_ECC, and the column density weighted average dust temperature derived from the Herschel images, T_weighted. The Planck and Herschel based temperatures agree within ±3 K, except for the warmer sources in the Galactic center region, where our column density criteria select a much larger region than the Planck clump size and includes warmer regions. The average temperatures in the inner Galaxy are 10.4 K and 15.4 K based on the Planck and Herschel data, respectively. The same values are 11.8 K and 12.5 K in the outer part of our Galaxy. We note that there is one outlier clump in the outer Galaxy without any 70 μm emission (G201.26+00.46), for which the Planck based temperature is higher by ~5 K with respect to the average temperature derived from Herschel data. A bright 70 μm-emission region with high dust temperatures is visible toward the southwest from the clump. The Planck ECC size is larger than the denser region used in the Herschel-based calculation, which causes lower weighted temperature for the Herschel data. This comparison shows that the Planck ECC temperatures are, in most cases, a reasonable approximation for the average temperatures of the high column density regions of the cores. Temperatures derived from the integrated Herschel fluxes within the two column density contours that we used in the paper lead to the same conclusion.

4. Discussion

In this section, we evaluate the potential of forming massive stars and star clusters by the selected Planck cold clumps and determine their evolutionary stage. We accomplished this using the column density maps and bolometric luminosities based on the Herschel observations. In Sect. 4.1 we investigate the fraction of mass in the densest parts of the clumps compared to their surrounding, more diffuse regions, while in Sect. 4.2 we identify the clumps with ongoing massive star formation as probed by 70 μm emission. Kauffmann & Pillai (2010) showed that regions forming massive stars are, at a given radius, more massive than the limit mass (m_lim[r] = 870 M_⊙[r/pc]^1.33) and thus, we evaluate the fraction of Planck cold clumps with the potential to form massive stars and star clusters.

4.1. Mass distribution

As mentioned in Sect. 3.3, the clump masses and sizes were calculated from the Hi-GAL column density maps considering two different H₂ column density thresholds of 3 × 10²¹ and 10²² cm^-2. The comparison of the masses contained within these levels gives us information about the fraction of gas mass locked into the densest parts of the clumps and the total mass enclosed within the surrounding region. This parameter may be important since it is possible to increase the total mass of the dense core through the accretion of this lower column density material.

Figure 9 shows the mass distribution of the sources as a function of Galactocentric distance. The top panel shows the total mass associated with a column density limit of 3 × 10²¹ cm^-2, while the bottom panel shows the mass contained within the high column density gas, N(H₂) > 10²² cm^-2. Solid lines show the range of masses and Galactocentric distances for the sources with large distance discrepancies. No sources with mass above 10⁴M_⊙ were detected at a Galactocentric distance larger than 6 kpc. All the densest clumps with N(H₂) > 10²² cm^-2 are found within 9 kpc to the Galactic center.

Figure 10 shows the size and mass ratio for the two column density thresholds considered, 3 × 10²¹ cm^-2 and 10²² cm^-2, and the mass ratio as a function of Galactocentric distance. The ECC clumps with a Galactocentric distance of >9 kpc do not have high column density gas in our sample. The fraction of the mass in the densest parts of the nearby Planck clumps is lower than that of the more massive clumps closer to the Galactic center. We also note that the densest regions tend to be more compact (i.e., smaller in spatial extent) for the clumps in the outer Galaxy than for the clumps in the inner Galaxy. This is likely a consequence of the higher overall H₂ column densities found in the inner Galaxy (by more than a factor of 10) with respect to the outer Galaxy.

4.2. Mass-size relation

Kauffmann & Pillai (2010) found an empirical mass-size threshold above which massive star formation can occur using solar neighborhood clouds devoid of massive star formation. Regions forming massive stars must thus be, at a given radius, more massive than the limit mass (m_lim[r] = 870 M_⊙[r/pc]^1.33). In Fig. 11, we show the mass-size diagram for the selected Planck cold clumps based on their Herschel data. Solid lines indicate the range of masses and radii for the sources with large distance discrepancies. They are all located below the empirical mass-size threshold of massive star formation. Twenty-five percent of the Planck ECC clumps in the Galactic Plane with Herschel-based derived masses have the potential to form massive stars since these masses lie above the mass limit proposed by Kauffmann & Pillai (2010). These ECC clumps are the most massive and also the largest in size among our sample, which is consistent with the idea that massive stars tend to form in the largest, most massive clumps (Wang et al. 2009). Their distances are very typical for the sample, between 1.1 kpc and 6.04 kpc. We note that three out of these clumps do not have 70 μm point sources (G006.96+00.89, G009.79+00.87, and G354.81+00.35), which makes them good candidates to study the earliest stages of star formation. Two of the sources above the mass limit are in the outer Galaxy (G224.27-00.82 and G224.47-00.65), but they contain 70 μm point sources already, which confirms their ability to form stars.

In addition, we can investigate the possible evolution of the mass of the clumps by calculating the ECC clump masses and sizes contained within five different H₂ column density thresholds: 10²¹, 3 × 10²¹, 5 × 10²¹, 8 × 10²¹, and 10²² cm^-2. The bottom panel of Fig. 11 shows the mass-size relationship. The lines represent the evolution of the calculated values within the different thresholds. As shown in Fig. 11, the lines are almost parallel to the mass-size threshold. This demonstrates that the number of sources with the potential to form high-mass stars does not depend on the selected column density threshold.

We stress that two clumps are found in the Outer Galaxy that fulfill the criteria of high-mass star formation. Their masses and sizes based on the Planck catalog information also fulfill this criteria. Three clumps without 70 μm emission are found above the high-mass star formation threshold in the inner Galaxy. These clumps represent good candidates to study the earliest phases of massive star and star cluster formation.

5. Summary

We investigate the basic physical properties of 48 Planck ECC clumps in the Galactic plane using Herschel Hi-GAL data. Starless (~22%) and star-forming (~78%) clumps were identified in the inner and outer part of the Galaxy. We calculated their masses and sizes based on background subtracted Herschel images, and we characterized their evolutionary stage based on the presence, or absence, of point-like 70 μm emission (an excellent probe of massive star formation). No clear difference was found in the mass and temperature of starless and star-forming clumps based on our data. Planck clumps located in the Galactic center show higher column densities and average dust temperatures than those located in the outer Galaxy. We identified five particularly interesting objects in the Galactic plane that are good candidates for higher resolution continuum and molecular line studies. Three of these objects (G006.96+00.89, G009.79+00.87, and G354.81+00.35) are located in the inner Galaxy, fulfill the criteria of massive star and star cluster formation, and show no 70 μm emission. These objects are thus excellent templates to study the earliest phases of massive star and star cluster formation. Two objects with 70 μm emission (G224.27-00.82 and G224.47-00.65) were found in the outer part of the Galaxy, which also fulfill the massive star formation criteria; with a follow-up study of these objects, the properties of massive star-forming clumps can be compared directly in the inner and outer part of the Galaxy.

---

Acknowledgments

This work has made use of the APLpy plotting package (https://aplpy.github.io/) and the agpy code package (https://pypi.python.org/pypi/agpy/0.1.4). K.W. acknowledges the support from Deutsche Forschungsgemeinschaft (DFG) grant WA3628-1/1 through priority program 1573 (“Physics of the Interstellar Medium”). I.J.-S. acknowledges the financial support received from the People Programme (Marie Curie Actions) of the European Union’s Seventh Framework Programme (FP7/2007–2013) under REA grant agreement PIIF-GA-2011-301538 and from the STFC through an Ernest Rutherford Fellowship (proposal number ST/L004801/1). L.V.T. and S.Z. acknowledges support by the OTKA grants NN-111016 and K101393.

References

Appendix A: Herschel images of the clumps

In Figs. A.1−A.8, we show the Herschel 70, 250 and 500 μm images of the investigated Planck ECC objects (intensity scale in units of MJy/sr) with a resolution of 5″, 18″, and 36″. Black circles are centered at the positions of the Planck clumps and their sizes correspond to the derived major axis of the clumps.

Appendix B: Available distance estimates for the clumps

Table B.1 shows all the available estimated distances for the ECC objects, see the detailed description in Sect. 2.3.

Appendix C: Background subtraction

For the precise determination of clump masses and sizes, we have to subtract the background and foreground emission in the observed images. Different methods are available to perform this task: (i) the Fourier transform (FT) method, which separates the large-scale and small-scale structures (Wang et al. 2015); (ii) to use a reference, relatively emission-free region to derive a constant background value (Juvela et al. 2012); and iii) to perform the Gaussian fitting of the Galactic background along the Galactic latitude (Battersby et al. 2011).

For the background subtraction, we used 2° × 2° maps around the central positions of the Planck cold clumps. These maps are much larger than the structures of interest. We smoothed all intensity images to the same resolution of 5′ and used the smoothed map as background. In contrast to methods (ii) and (iii), this technique allows spatial variations in the background values across the maps. After background subtraction, the derived column densities are 15−20% lower and the dust temperatures are 1−3 K lower. We compared our method with the FT method introduced by Wang et al. (2015). Our column density values are in good agreement with the FT method, and the difference is on average only ~10%. We caution, however, that this may be partly because of the specific properties of the power spectrum of the Hi-GAL images in large tiles, as both works have found; the same may not apply for other images with a different power spectrum characterization. Calculated dust temperature and column density maps based on the background subtracted maps are shown in Appendix D.

Appendix D: T_dust and N(H₂) maps

Figures D.1−D.8 shows the Herschel 70 μm images ([MJy/sr], left panels), the calculated column density ([cm^-2], middle panels) in logarithmic scale, and the dust temperature maps ([K], right panels) for the ECC sources with a resolution of 36″. The calculation was based on the background subtracted Herschel 160−500 μm images. Black circles show the location (central position) and size (major axis given in the ECC catalog) of the Planck clumps. Yellow and red contour levels refer to the H₂ column density thresholds of 3 × 10²¹ and 10²² cm^-2, respectively.

All Tables

All Figures