DISTINCT CHEMICAL REGIONS IN THE "PRESTELLAR" INFRARED DARK CLOUD G028.23−00.19

High-mass stars play a key role in the evolution of the energetics and chemistry of molecular clouds and galaxies. However, our knowledge of their formation is still poor, compared with that of low-mass stars. There are several factors that make the study of high-mass star formation more challenging than that of their low-mass counterparts. High-mass stars are rare and evolve quickly. Massive star-forming regions in the earliest stages of evolution are mostly located at large distances (≳3 kpc) and are heavily extincted by dust. The earliest well-understood phase of high-mass star formation is the hot molecular cores (e.g., Garay & Lizano 1999). However, at this stage, the high-mass stars have already begun burning hydrogen and reached the zero-age main sequence, disturbing and changing the kinematics and chemistry of the initially starless cocoon.

Prestellar cores represent the earliest stage of star formation. These starless cores are formed by dense, centrally concentrated parcels of gravitationally bound gas (e.g., André et al. 2009). Prestellar cores will eventually collapse and evolve to form stars. The prestellar phase in high-mass star formation has been very difficult to find. Initially, large infrared (IR)/millimeter surveys had poor angular resolution and sensitivity, producing a bias toward bright, evolved objects and a misidentification of starless candidates. However, with the advent of new space telescopes, such as Spitzer and Herschel, and submillimeter Galactic plane surveys, it has been possible to study in detail the best candidates for dense regions which will eventually form massive stars, the so-called Infrared Dark Clouds (IRDCs). IRDCs appear as dark silhouettes against the Galactic mid-IR background. Galactic plane surveys revealed thousands of IRDCs (Infrared Space Observatory, Perault et al. 1996; Midcourse Space Experiment, Egan et al. 1998; Simon et al. 2006; Spitzer, Peretto & Fuller 2009; Kim et al. 2010). The first studies characterizing them suggested that they were cold (<25 K), massive (∼10²–10⁴ M_☉), and dense (≳10⁵ cm⁻³) molecular clouds with high column densities (∼10²³ cm⁻²; Carey et al. 1998, 2000). More recent studies on IRDC clumps⁶ show that they have typical masses of ∼120 M_☉, sizes of ∼0.5 pc (Rathborne et al. 2006), dust temperatures that range between 16 and 52 K, and luminosities that range from ∼10 to 5×10⁴ L_☉ (Rathborne et al. 2010). NH₃ rotational temperatures are typically lower than dust temperatures and range from ∼10 to 20 K (Pillai et al. 2006; Sakai et al. 2008; Devine et al. 2011; Ragan et al. 2011).

Evidence of active high-mass star formation in IRDCs is inferred by the presence of ultracompact H ii regions (Battersby et al. 2010), hot cores (Rathborne et al. 2008), embedded 24 μm sources (Chambers et al. 2009), molecular outflows (Sanhueza et al. 2010; Wang et al. 2011), or maser emission (Wang et al. 2006; Chambers et al. 2009). IRDCs also host massive, cold IR-dark clumps that show similar physical properties to those clumps that host active high-mass star formation, except that the temperature and luminosity are lower. These characteristics suggest that massive, cold IR-dark clumps will form high-mass stars in the future. These objects are the best candidates for the most elusive and earliest phase of high-mass star formation, the "prestellar" or "starless" phase.

Although several studies have focused on IRDCs, the number of studies on starless candidates is small (e.g., Pillai et al. 2011). Zhang et al. (2009) studied in detail an IR quiescent clump, apparently starless, resolving it into five cores at 1.3 mm with Submillimeter Array (SMA). However, new SMA observations at 0.88 mm obtained by Wang et al. (2011) show that all cores are associated with molecular outflows. This suggests that this clump is in fact forming stars which have entered the accretion phase. Thus, it is crucial to evaluate the prestellar nature of a core/clump by using interferometric observations of molecular lines, and their starless nature should not be asserted just by the lack of a detection in the IR.

Little is known about the chemistry of IRDCs. A few works have focused on molecular line surveys of several IRDCs (Sakai et al. 2008, 2010; Vasyunina et al. 2011; Sanhueza et al. 2012). For instance, Sanhueza et al. (2012) find chemical variations between different evolutionary stages using a set of 10 different molecular lines at 3.3 mm. However, most of the effort related to chemistry in IRDCs has been focused on deuteration. Recent findings suggest that the deuterium fraction in high-mass star-forming regions resembles that seen in low-mass star-forming regions after the stars turn on, decreasing with time as the clumps evolve to warmer temperatures (Chen et al. 2011; Miettinen et al. 2011; Fontani et al. 2011; Sakai et al. 2012; Pineda & Teixeira 2013).

The internal kinematic structure of IRDCs has not yet been explored in depth. In two examples, IRDC G019.30+0.07 (Devine et al. 2011) and IRDC G035.39−00.33 (Jiménez-Serra et al. 2010; Henshaw et al. 2013), both IRDCs are composed of a few slightly different velocity components or "subclouds." It is still unclear what role these subclouds play in the earliest stages of high-mass star formation.

From the studies of Rathborne et al. (2010) and Sanhueza et al. (2012), we have identified an excellent candidate to study the initial conditions of massive star formation. As seen in Figure 1, IRDC G028.23−00.19 appears to be in a very early stage of evolution because it is dark at Spitzer/IRAC 3.6, 4.5, and 8.0 μm (Benjamin et al. 2003), Spitzer/MIPS 24 μm (Carey et al. 2009), and Herschel/PACS 70 μm (Molinari et al. 2010), except for a bright unrelated IR source superimposed against the northern part of the cloud. This source is an OH/IR star (Bowers & Knapp 1989), and is not a protostar. This foreground low-mass star, in a very late stage of evolution, is not associated with the IRDC. Indeed, its V_lsr is ∼52 km s⁻¹, whereas for IRDC G028.23−00.19 it is 80 km s⁻¹. Located in the center of the IRDC, the clump MM1 is one of the most massive, IR-quiescent clumps known, with a mass of ∼640 M_☉, as determined from 1.2 mm continuum emission (Rathborne et al. 2010) at a kinematical distance of 5.1 kpc (Sanhueza et al. 2012). By fitting the spectral energy distribution (SED), Rathborne et al. (2010) determined for MM1 a dust temperature of 22 K (see Section 4.1 for the updated values of the mass and the temperature of MM1). Using the Mopra telescope (38''), Sanhueza et al. (2012) detected weak molecular line emission of N₂H⁺, HNC, and HCO⁺ J = 1 → 0 toward this clump, consistent with cold gas. Battersby et al. (2010) observed this IRDC in 3.6 cm continuum emission with the Very Large Array (VLA) at 2''–3'' angular resolution and found no H ii regions associated with the cloud at σ_rms = 0.01 mJy beam⁻¹, which corresponds to a spectral type of B3 or later (Jackson & Kraemer 1999). Chambers et al. (2009) searched for, but failed to detect, 22.23 GHz H₂O and Class I 24.96 GHz CH₃OH masers with the Green Bank Telescope at ∼32'' (σ_rms≅0.05 K). Based on the results obtained from a large variety of observations, we suggest that IRDC G028.23−00.19 harbors the prime example of a starless or prestellar massive clump (MM1), being in a very early stage of evolution and showing no evidence of active star formation.

Zoom In Zoom Out Reset image size

To probe the chemistry, the internal structure, and the kinematics of IRDCs, interferometric observations are critical. In this paper, we present observations of IRDC G028.23−00.19 in nine different molecular lines obtained with the Combined Array for Research in Millimeter-wave Astronomy (CARMA). We aim to study the chemistry and internal structure of a massive star-forming region in a very early stage of evolution and test the starless nature of the cloud and, in particular, the clump MM1. This is the first study that explores the behavior of NH₂D, H¹³CO⁺, SiO, HN¹³C, C₂H, HCO⁺, HNC, N₂H⁺, and CH₃OH at ∼11'' angular resolution, studying both the chemistry and the structure of multiple subclouds traced by distinct, close velocity components (≲2 km s⁻¹) in a massive star-forming region that, apparently, has not yet formed high-mass stars.

Observations were carried out with CARMA, a 15-element interferometer consisting of nine 6.1 m and six 10.4 m antennas, during the CARMA Summer School 2011 July 27 in the compact E configuration. The projected baselines range from 8 to 66 m. The CARMA correlator has eight bands, each with an upper and lower sideband. One band was configured to have a bandwidth of 487.5 MHz, resulting in a total bandwidth of ∼1 GHz by including both sidebands, to observe continuum emission at 3.3 mm for calibration (91.878 GHz). The remaining seven bands were used to observe nine different molecular species (NH₂D, H¹³CO⁺, SiO, HN¹³C, C₂H, HCO⁺, and HNC in the lower sidebands; and N₂H⁺ and CH₃OH in the upper sidebands). For all seven bands, a bandwidth of 31 MHz (∼100 km s⁻¹) and a spectral resolution of 97 kHz (∼0.3 km s⁻¹) were used. Using natural weighting, the 1σ rms noise is 0.4 mJy beam⁻¹ for continuum emission, and ∼36.2 mJy beam⁻¹ (∼46.9 mK) per channel for molecular lines. The average of the major and minor axis of the final synthesized beam was 109, which corresponds to a physical size of ∼0.27 pc at the distance of 5.1 kpc. The typical rms noise, synthesized beam, and position angle (P.A.) for the continuum emission and each molecular line are given in Table 1.

Table 1. Summary of Observed Molecular Lines and Continuum Emission

Molecule/	Molecular	Frequency	Eu/k	Beam Size	P.A.	rms Noise
Continuum	Transition	(GHz)	(K)	('' × '')	(deg)	(mJy beam−1)	(mK)
NH2D	$J_{K_a,K_c}=1_{1,1} \rightarrow 1_{0,1}$	85.926260	20.68	15.3 × 8.4	−25.7	33.6	43.0
H13CO+	J = 1 → 0	86.754330	4.16	15.0 × 8.5	−26.9	31.1	39.8
SiO	J = 2 → 1	86.846998	6.25	15.0 × 8.4	−27.0	31.0	39.7
HN13C	J = 1 → 0	87.090859	4.18	14.5 × 8.6	−26.1	37.2	48.4
C2H	NJF = 1 (3/2) 2 → 0 (1/2) 1	87.316925	4.19	14.6 × 8.6	−26.3	36.4	46.6
HCO+	J = 1 → 0	89.188526	4.28	14.6 × 8.2	−27.5	33.3	42.9
HNC	J = 1 → 0	90.663574	4.35	14.4 × 8.0	−28.2	38.6	49.8
N2H+	JF1F = 112 → 012	93.171913	4.47	14.0 × 7.8	−28.2	40.5	52.7
	JF1F = 123 → 012	93.173772	4.47	14.0 × 7.8	−28.2	40.5	52.7
	JF1F = 101 → 012	93.176261	4.47	14.0 × 7.8	−28.2	40.5	52.7
CH3OH	JK = 2−1 → 1−1 E	96.739362	12.55	13.0 × 7.6	−26.0	44.5	58.8
	JK = 20 → 10 A	96.741375	6.97	13.0 × 7.6	−26.0	44.5	58.8
	JK = 20 → 10 E	96.744550	20.10	13.0 × 7.6	−26.0	44.5	58.8
3.3 mm	...	91.877525	...	14.2 × 7.9	−27.9	0.4	...

Notes. Rest frequencies and upper energy levels were adopted from the Splatalogue database (JPL, Pickett et al. 1998; CDMS, Müller et al. 2001, 2005), except for N₂H⁺ frequencies, which are from Daniel et al. (2006).

Download table as:  ASCIITypeset image

The system temperature varied from 190 to 250 K during the 7.1 hr track (5.4 hr on source). The receivers were tuned to a local oscillator frequency of 91.918 GHz. The phase center of the observations is (R.A., decl.)_J2000 = (18^h43^m3078, −04°13'1625). At the center frequency of 91.918 GHz, the primary beams or fields of view of the 6.1 and 10.4 m antennas are 135'' and 79'', respectively.

The data were reduced and imaged in a standard way using MIRIAD software. We used the quasar 1743-038 as a phase calibrator, the quasar 1751+096 as a bandpass calibrator, and Neptune as a flux calibrator. The uncertainty in the absolute flux is about 15%.

Figure 1 shows the Spitzer/IRAC 3.6, 4.5, and 8.0 μm image (left), Spitzer/MIPS 24 μm image (middle), and Herschel/PACS 70 μm image (right) overlaid with contours of the 1.2 continuum emission (11'' angular resolution). As can be seen from the IR images, the IRDC is completely dark, except for the foreground OH/IR star superimposed against the northern part of the cloud. The 1.2 mm continuum emission, from the IRAM 30 m telescope (Rathborne et al. 2006), matches the morphology of the IR extinction well. The massive clump MM1 is located at the center of the cloud.

Figure 2 shows the 3.3 mm continuum emission and integrated intensity maps of N₂H⁺ J = 1 → 0, H¹³CO⁺J = 1 → 0, and HN¹³C J = 1 → 0 in images and color contours (ranges of integration are given in the figure caption). For N₂H⁺, the transition with the lowest optical depth, the isolated hyperfine component JF₁F = 101 → 012 was used. White contours are the 1.2 mm continuum emission from the IRAM telescope. The 3.3 mm continuum emission from CARMA is more compact than the 1.2 mm emission. This is probably due to the low 3.3 mm flux at 12 K (see Section 4.1), the low sensitivity arising from using only ∼1 GHz of continuum bandwidth, and because the interferometer may be missing extended flux. In fact, comparing the measured flux at 3.3 mm with the expected 3.3 mm flux from the SED fitting done in Section 4.1, we estimate that the interferometer is only recovering about 40% of the total flux. The emission of N₂H⁺, H¹³CO⁺, and HN¹³C resembles the 1.2 mm emission from IRAM, although less extended. Using Mopra single-pointing observations from Sanhueza et al. (2012), we estimate that we recover all the flux for N₂H⁺. Except for N₂H⁺, HCO⁺, and HNC, we have no observations from single-dish telescopes to evaluate if any other molecular line suffers from missing extended flux. Based on the spatial extent of H¹³CO⁺, HN¹³C, NH₂D, C₂H, SiO, and CH₃OH emission, we suggest that zero-spacing problems are not important in these line observations and we recover most of the flux with the interferometer. On the other hand, using the Mopra single-pointing observations from Sanhueza et al. (2012), we estimate that ∼30% of the HCO⁺ and HNC flux is recovered by the interferometer. HCO⁺ and HNC integrated intensity maps are not presented here because their emission is substantially resolved out. Therefore, the distribution of these molecules is not discussed in the paper.

Zoom In Zoom Out Reset image size

Figure 3 shows the integrated intensity maps of NH₂D J = 1 → 1, C₂H J = 1 → 0, SiO J = 2 → 1, and CH₃OH J = 2 → 1 in images and color contours (ranges of integration are given in the figure caption). For NH₂D, the main component $J_{K_a,K_c}=1_{1,1} \rightarrow 1_{0,1}$ was used. A velocity range including all three transitions was used for CH₃OH. The emission of NH₂D and C₂H is compact and is mostly located in the densest part of the IRDC, the MM1 clump. As discussed in Section 4.4, the enhancement of these molecules is likely due to cold prestellar gas-phase chemistry. Emission from SiO and CH₃OH is highly unexpected because of the absence of IR signs of star formation (as discussed in more depth in Section 4.3). The SiO emission is located in the northern and southern regions of the IRDC, and it seems to be unassociated with the MM1 clump. An additional weak SiO component is located in the center-west part of the cloud. The CH₃OH emission is widespread over the full IRDC.

Zoom In Zoom Out Reset image size

Line center velocities, line widths, and intensity of the lines were determined by fitting Gaussian profiles. Molecular lines that show hyperfine structures were fitted using a multi-Gaussian function, with fixed frequency separation between hyperfine transitions. Gaussian fit parameters of molecular lines, in selected positions, are shown in Table 2.

Table 2. Gaussian Fit Parameters of Molecular Lines in Selected Positions

Molecule	Transition	TA	Vlsr	ΔV	Position
		(K)	(km s−1)	(km s−1)
NH2D	$J_{K_a,K_c}=1_{1,1} \rightarrow 1_{0,1}$	0.36 ± 0.03	79.64 ± 0.05	1.34 ± 0.13	MM1
H13CO+	J = 1 → 0	0.28 ± 0.02	79.42 ± 0.09	2.41 ± 0.21	MM1
HN13C	J = 1 → 0	0.28 ± 0.03	79.51 ± 0.09	1.89 ± 0.21	MM1
C2H	NJF = 1 (3/2) 2 → 0 (1/2) 1	0.18 ± 0.03	79.28 ± 0.15	1.82 ± 0.36	MM1
N2H+	JF1F = 101 → 012	0.66 ± 0.03	79.17 ± 0.04	1.85 ± 0.10	MM1
CH3OH	JK = 2−1 → 1−1 E	0.89 ± 0.03	79.32 ± 0.02	2.18 ± 0.09	MM1
	JK = 20 → 10 A	1.12 ± 0.03	79.32 ± 0.02	2.13 ± 0.07	MM1
	JK = 20 → 10 E	0.13 ± 0.04	79.32 ± 0.02	1.44 ± 0.47	MM1
SiO	J = 2 → 1	0.14 ± 0.02	81.79 ± 0.15	1.94 ± 0.34	Center-west
		0.26 ± 0.01	76.31 ± 0.15	6.26 ± 0.36	South
		0.13 ± 0.01	80.48 ± 0.33	7.74 ± 0.79	North

Download table as:  ASCIITypeset image

Multiple velocity components along a line of sight are not unusual in star-forming regions. However, it is not easy to separate two velocity components with a small offset of ≲2 km s⁻¹ in massive star-forming regions. In general, these kind of regions have line widths of 3–4 km s⁻¹ (e.g., Hoq et al. 2013). Two close velocity components can be separated only when the line emission has narrow profiles and both components are not fully blended. This can be done in regions containing quiescent gas and using high-density tracers. Following the discussion of Devine et al. (2011), we will refer to the structures defined by close velocity components as "subclouds." HN¹³C J = 1 → 0 and C₂H J = 1 → 0 show double-line profiles in the center-west part of the IRDC. The double-line profiles are actually two close velocity components separated by ∼2 km s⁻¹. Each individual component can be traced continuously from the overlapping region of the subclouds, where we see the double-line profiles, to regions where only one of the components is seen. Two velocity components are also seen in H¹³CO⁺, N₂H⁺, and CH₃OH; however, because these molecules have broader line emission, both components are blended. Figure 4 shows the HN¹³C center velocity map of each component obtained by fitting Gaussian profiles to the spectra. The map shows two spatially coherent structures. The blueshifted subcloud is mostly located in the southern-central part of the IRDC and has an average central velocity of 79.3 km s⁻¹. The redshifted subcloud is mostly located in the northern-central region of the IRDC and has an average central velocity of 80.7 km s⁻¹. The velocity difference between subclouds is therefore ∼1.4 km s⁻¹, similar to those found in other IRDCs, such as G019.30+0.07 or G035.39−00.33 (<2 km s⁻¹; see Devine et al. 2011; Henshaw et al. 2013). The massive MM1 clump is associated with the blueshifted subcloud. NH₂D emission is only detected in the blueshifted subcloud, while the narrow SiO component (described in the next paragraph) is only detected in the redshifted subcloud. All other lines are detected in both substructures.

Zoom In Zoom Out Reset image size

Figure 5 shows line width and center velocity maps of SiO across the full IRDC and spectra of SiO and CH₃OH in selected positions. Because in several positions over the IRDC the CH₃OH transitions are blended, Gaussian fits cannot be performed for all methanol spectra and thus the line width and center velocity maps are not useful. The line width map of the SiO line reveals the presence of two distinct SiO components defined by having different line widths and being located in different positions. The first SiO component has narrow line widths, ∼2 km s⁻¹, and peaks at the redshifted velocity component. It is also located in the center-west part of the IRDC, at about the same position where the subclouds overlap (see right panel in Figure 5). The second SiO component has broader line widths, ≳6 km s⁻¹, and is located in the southern and northern regions of the IRDC. CH₃OH emission is also broad in the same positions where SiO displays broad line profiles. Figure 5 also displays the SiO and CH₃OH spectra at the center position of MM1. As expected for quiescent molecular gas (Martin-Pintado et al. 1992; Requena-Torres et al. 2007), the SiO emission is not detected at this position. The CH₃OH J_K = 2₋₁ → 1₋₁ E and J_K = 2₀ → 1₀ A transitions (E_u/k of 12.6 and 7.0 K, respectively) are bright and are not blended. The higher-excitation CH₃OH J_K = 2₀ → 1₀ E transition (E_u/k = 20.1 K) is barely detected at 3σ level.

Zoom In Zoom Out Reset image size

Figure 6 shows all lines at the peak position of NH₂D (MM1). CH₃OH is the brightest line at this position. NH₂D, C₂H, HN¹³C, and H¹³CO⁺ lines show narrow profiles. SiO emission is not detected. Because the N₂H⁺ JF₁F = 123 → 012 and JF₁F = 112 → 012 transitions are saturated, the usually weakest JF₁F = 101 → 012 is the brightest line, suggesting a large optical depth for N₂H⁺ J = 1 → 0. HCO⁺ and HNC show only faint emission in MM1. Both HCO⁺ and HNC show profiles with absorption features that prevent their use in calculating physical parameters and make their interpretation difficult.

Zoom In Zoom Out Reset image size

4.1. Dust Temperature, H₂ Column Density, and Mass Determination

Taking advantage of the new Herschel Hi-GAL Survey (Molinari et al. 2010), we derive the dust temperature for MM1 by using the emission detected at 250, 350, and 500 μm, and the 1.2 mm emission from IRAM 30 m (Rathborne et al. 2006). We exclude from the analysis the 70 and 160 μm emission maps. At 70 μm the IRDC is seen in absorption and at 160 μm the emission from the IRDC is indistinguishable from the background/foreground emission.

After convolving the 250 μm, 350 μm, and 1.2 mm maps to the angular resolution of the 500 μm map (351), we subtract an extended component from the Herschel maps that represents the contribution from the background and foreground diffuse emission. Then, we fit the fluxes measured at the peak position of MM1 by using a single-temperature emission model, assuming optically thin emission, and dust opacities (κ_ν) for dust grains with thin ice mantles coagulated at 10⁵ cm⁻³ from Ossenkopf & Henning (1994). We defer the detailed description of the extended emission subtraction algorithm and the fitting procedure to an upcoming paper (A. Guzmán et al., in preparation). Using the measured fluxes at 250 μm (150.0 MJy sr⁻¹), 350 μm (610.8 MJy sr⁻¹), 500 μm (313.9 MJy sr⁻¹), and 1.2 mm (19.5 MJy sr⁻¹), we determine a dust temperature (T_dust) of 12 ± 2 K. We suggest that the main difference with the temperature of 22 K determined by Rathborne et al. (2010) is that they use a low value for the dust emissivity index (β = 1) and use the potentially optically thick Spitzer 85 and 95 μm emission in fitting the SED. The new dust temperature is more consistent with the CH₃OH temperature we determine by using the rotational diagram technique in Section 4.2 (12 K) and the NH₃ temperature (8 K) recently determined by Chira et al. (2013).

To calculate the H₂ column density and mass from the 1.2 mm continuum, we used the same procedure of Rathborne et al. (2006), assuming a gas-to-dust mass ratio of 100 and adopting for the dust opacity (κ_1.2  mm) a value of 1.0 cm² g⁻¹. Computed column densities at a given position are shown in Table 3. For the mass, we use the same integrated flux measure by Rathborne et al. (2010; 1.63 Jy) and the distance determined by Sanhueza et al. (2012; 5.1 kpc). We obtain a mass of 1520 M_☉ for the clump MM1.

Table 3. Derived Parameters in Selected Positions

Molecule	NMolecule	$N_{\rm H_2}$	Molecular	Position
	(cm−2)	(cm−2)	Abundance
NH2D	5.2(2.0) × 1013	3.9(1.4) × 1022	1.3(0.7) × 10−9	MM1
HCO+	4.5(0.6) × 1013	3.9(1.4) × 1022	1.2(0.5) × 10−9	MM1
HNC	5.7(0.9) × 1013	3.9(1.4) × 1022	1.5(0.6) × 10−9	MM1
C2H	1.6(0.4) × 1013	3.9(1.4) × 1022	4.1(1.8) × 10−10	MM1
N2H+	1.7(0.2) × 1013	3.9(1.4) × 1022	4.3(1.6) × 10−10	MM1
CH3OH	1.1(0.1) × 1014	3.9(1.4) × 1022	2.7(1.0) × 10−9	MM1
SiO	6.7(1.6) × 1011	1.6(0.5) × 1022	4.3(1.8) × 10−11	Center-west
	6.8(0.6) × 1012	1.8(0.7) × 1021	3.8(1.4) × 10−9	South
	8.1(1.2) × 1012	1.1(0.4) × 1022	7.5(2.7) × 10−10	North

Download table as:  ASCIITypeset image

4.2. Molecular Column Densities and Abundances

Assuming local thermodynamic equilibrium (LTE) conditions, beam-averaged column densities of SiO, H¹³CO⁺, HN¹³C, C₂H, and N₂H⁺ were calculated using the description of Sanhueza et al. (2012). Sanhueza et al. (2012) obtain column densities for the same set of lines and list all parameters needed for the calculation. For this work, we have assumed that all lines listed above are optically thin. However, there is evidence that the C₂H and N₂H⁺ emission lines may be moderately optically thick; their column densities should thus be treated as lower limits. HCO⁺ and HNC column densities are calculated from H¹³CO⁺ and HN¹³C, adopting a [¹²C/¹³C] isotopic abundance ratio of 50 (see discussion of Milam et al. 2005). Under LTE conditions, we assume an excitation temperature (T_ex) equal to the dust temperature determined in Section 4.1 (12 K).

For NH₂D, the optical depth of the main hyperfine component ($\tau _{_{\rm m}}$) was calculated from the ratio between the observed main beam brightness temperatures of the main component ($T_{\rm mb_{\rm m}}$) and the brightest satellite ($T_{\rm mb_{\rm s}}$), using the following relation (e.g., Sanhueza et al. 2012)

$$\begin{equation} \frac{1-e^{-\frac{1}{3}\tau _{\rm m}}}{1-e^{-\tau _{\rm m}}}=\frac{T_{\rm mb_{\rm s}}}{T_{\rm mb_{\rm m}}}, \end{equation} \tag{ 1 }$$

where we have used that the opacity ratio (r) between the main ($\tau _{_{\rm m}}$) and satellite ($\tau _{_{\rm s}}$) lines is $r=\tau _{_{\rm m}}/\tau _{_{\rm s}}=3$, which depends only on the transition moments. We obtained an optical depth of 0.7 for the main hyperfine component.

In LTE conditions (T_ex = T_dust), the column density is given by (e.g., Garden et al. 1991; Sanhueza et al. 2012)

$$\begin{equation} N=\frac{8\pi \nu ^3}{c^3R}\frac{Q_{\rm rot}}{g_{\rm u} A_{\rm ul}}\frac{\exp\, (E_{_ {l}}/kT_{\rm ex})}{[1-\exp\, (-h\nu /kT_{\rm ex})]}\int \tau \,dv, \end{equation} \tag{ 2 }$$

where τ_ν is the optical depth of the line, g_u is the statistical weight of the upper level, A_ul is the Einstein coefficient for spontaneous emission, $E_{_{l}}$ is the energy of the lower state, $Q_{\rm rot}=3.899+0.751\, T_{\rm ex}^{3/2}$ is the partition function for NH₂D (Busquet et al. 2010), ν is the transition frequency, and R = 1/2 is the relative intensity of the main hyperfine transition with respect to the others. R is only relevant for hyperfine transitions because it takes into account the satellite lines correcting by their relative opacities. It is equal to 1.0 for transitions without hyperfine structure. The values of g_u, A_ul, and $E_{_{l}}$ are given in Table 4. The beam-averaged column density is obtained by multiplying N by the filling factor, f. The filling factor can be found from

$$\begin{equation} T_{\rm mb} = f[J(T_{\rm ex}) - J(T_{\rm bg})](1 - e^{-\tau _{\nu }}), \end{equation} \tag{ 3 }$$

where J(T) is given by

$$\begin{equation} J(T) = \frac{h\nu }{k}\frac{1}{e^{h\nu /kT} - 1}. \end{equation} \tag{ 4 }$$

Table 4. Parameters Used for NH₂D and CH₃OH Column Density Calculations

Molecule	Transition	gu	gK	gI	Aul	El/k	μ2S
					(× 10−6 s−1)	(K)	(D2)
NH2D	$J_{K_a,K_c}=1_{1,1} \rightarrow 1_{0,1}$	15	...	...	5.8637	16.55	...
CH3OH	JK = 2−1 → 1−1 E	...	2	1	...	...	1.2134
	JK = 20 → 10 A	...	1	2	...	...	1.6170
	JK = 20 → 10 E	...	2	1	...	...	1.6166

Download table as:  ASCIITypeset image

The derived NH₂D column density and filling factor are 6.2×10¹⁴ cm⁻² and 0.08, resulting in a beam-averaged column density of 5.2×10¹³ cm⁻² in MM1.

The CH₃OH column density (N) was obtained by using the rotational diagram technique in the most general form (e.g., Girart et al. 2002). Assuming that all rotational levels are populated with the same excitation temperature (LTE), we have

$$\begin{equation} \log I= \log \frac{fN}{Q_{\rm rot}} - \frac{E_{\rm u}}{T}\log e - \log C_{T} - \log F_{T}, \end{equation} \tag{ 5 }$$

with Q_rot = 1.2327 T^1.5 (the rotational partition function for CH₃OH) and I equal to

$$\begin{equation} I = \frac{3k\int T_{\rm mb}dv}{8\pi ^3g_Kg_IS\mu ^2\nu }, \end{equation} \tag{ 6 }$$

where g_K and g_I are the K-level and nuclear spin degeneracies, μ is the electric dipole moment of the molecule, and S is the line strength. C_T and F_T are defined as

$$\begin{equation} C_{T} = \frac{\tau }{1 - e^{-\tau }}\quad {\rm and}\quad F_{T} = \frac{J(T_{\rm ex})}{J(T_{\rm ex}) - J(T_{\rm bg})}. \end{equation} \tag{ 7 }$$

The optical depth of the line is given by

$$\begin{equation} \tau = \frac{8\pi ^3g_Kg_I\mu ^2S\nu }{3h}\frac{N}{Q_{\rm rot}\Delta v}(e^{h\nu /kT} - 1)e^{-E_{\rm u}/kT}, \end{equation} \tag{ 8 }$$

where E_u is the energy of the upper state and Δv is the line width. Values for g_K, g_I, and μ²S are given in Table 4. A detailed description and derivation of the equations used in the rotational diagram technique are presented in A. Silva & Q. Zhang (in preparation).

Equation (5) is frequently used without the C_T and F_T terms, assuming that τ ≪ 1 and T_ex ≫ T_bg. However, in the cold clump MM1 these assumptions may not be valid. An additional condition for cold gas is that the E/A abundance ratio is expected to be 0.69 at low temperatures (∼10 K). This is because the ground state of E-methanol is 7.9 K above the ground state of A-methanol, which results in an overabundance of A-methanol at low temperatures (Friberg et al. 1988; Wirström et al. 2011). Thus, the integrated intensity of the transition J_K = 2₀ → 1₀ A was weighted by 0.69. Assuming that f (filling factor) equals 1.0, the remaining free parameters in Equation (5) are the beam-averaged column density and the rotational temperature of CH₃OH. By fitting the calculated integrated intensities, from Equation (5), to the observed integrated intensities of the three different methanol transitions, we can find the best solution for the beam-averaged column density and the rotational temperature. The fitting procedure was carried out in IDL using the MPFITFUN package (Markwardt 2009). At the peak position of MM1, the obtained beam-averaged column density and rotational temperature are 1.1×10¹⁴ cm⁻² and 12 K. The optical depth for the transition J_K = 2₀ → 1₀ A is 0.1.

To estimate the molecular abundances with respect to H₂, we take the ratio between the column density of a given molecule and the H₂ column density derived from the 1.2 mm dust continuum emission. We use the 1.2 mm continuum for deriving the H₂ column density, rather than the 3.3 mm continuum, because the emission is more extended, more closely matches the spatial extent of the molecular line emission, and is detected in the regions of interest (for example, where SiO is detected). Column densities and abundances for all molecules were calculated for the peak positions of NH₂D (center of MM1 clump). Additional SiO column densities and abundances were calculated for the positions marked in Figure 5 ("north," "center-west," and "south").

Uncertainties from the line fitting and dust temperature will propagate to the derived column densities and abundances. At the end, uncertainties in the molecular abundances range between ∼35% and ∼50%. All calculated column densities and abundances with their respective uncertainties are summarized in Table 3.

4.3. The Puzzling SiO and CH₃OH Emission

4.4. NH₂D and C₂H: Cold Gas Tracers

4.5. HCO⁺, H¹³CO⁺, HNC, HN¹³C, and N₂H⁺

4.6. The Role of the Subclouds

4.7. The "Prestellar" Nature of IRDC G028.23−00.19

We have observed the IRDC G028.23−00.19 in several molecular species (NH₂D, H¹³CO⁺, SiO, HN¹³C, C₂H, HCO⁺, HNC, N₂H⁺, and CH₃OH) and continuum emission at 3.3 mm using CARMA (11'' angular resolution). This IRDC is dark at Spitzer 3.6, 4.5, 8.0, and 24 μm and Herschel 70 μm. In its center, the IRDC hosts one of the most massive IR, quiescent clumps known (MM1). We have examined the spectral line observations and draw the following conclusions.

• 1.  
  Using the emission detected at 250, 350, and 500 μm with Herschel, and the 1.2 mm emission from IRAM 30 m, we updated the mass (1520 M_☉) and the dust temperature (12 K) of the central clump in IRDC G028.23−00.19.
• 2.  
  The low temperature, high NH₂D abundance, low HCO⁺ and HNC abundances, non-detection of SiO, narrow line widths, and absence of embedded IR sources in MM1 indicate that the clump still remains in a prestellar phase.
• 3.  
  Strong SiO components with broad line widths are detected in the northern and southern regions of the IRDC. The large line widths (6–9 km s⁻¹) and high abundances ((0.8–3.8) × 10⁻⁹) of SiO suggest that the mechanism releasing SiO from the dust grains into the gas phase could be molecular outflows from undetected intermediate-mass stars or clusters of low-mass stars deeply embedded in the IRDC. However, in the southern region where the SiO abundance peaks, there is no associated counterpart in the continuum.
• 4.  
  A weaker SiO component is detected in the center-west part of the IRDC. The narrow line widths (∼2 km s⁻¹) and low SiO abundances (4.3×10⁻¹¹) are consistent with either a "subcloud–subcloud" interaction or an unresolved population of a few low-mass stars. Higher angular resolution observations are needed to clearly establish the origin of the narrow SiO emission.
• 5.  
  We report unexpected widespread CH₃OH emission throughout the whole IRDC and the first detection of extended methanol emission in a massive prestellar clump. Because IR observations show no primary protostar embedded in the IRDC, we reject thermal evaporation of methanol from dust mantles as the production mechanism. CH₃OH emission with broad line widths supports the idea of molecular outflows in the southern and northern regions of the IRDC. The CH₃OH emission at the position of MM1 is rather narrow. We suggest that the most likely mechanism able to release methanol from the dust grains to the gas phase in such a cold region is the exothermicity of surface reactions (Garrod et al. 2006, 2007).
• 6.  
  C₂H has been suggested to be a PDR tracer. However, recent evidence suggests it could also trace cold, dense gas associated with earlier stages of star formation. Because the spatial distribution of C₂H emission resembles that of NH₂D, we suggest C₂H traces cold, dense gas in this IRDC.
• 7.  
  HN¹³C reveals that the IRDC is composed of two substructures ("subclouds") separated by 1.4 km s⁻¹. Remarkably, the subclouds overlap in the center-west region of the IRDC, exactly coincident with the narrow SiO component. One explanation for the narrow SiO component is that both subclouds may be interacting and producing low-velocity shocks that release small amounts of SiO to the gas phase. We speculate that IRDCs may be young molecular clouds that could still be in a stage where they are being assembled.
• 8.  
  In the densest part of the IRDC, the MM1 clump shows no signs of star formation. Notably, in other regions of the IRDC it appears that star formation activity may be occurring. Why star formation would begin in less dense regions and how the findings on IRDC G028.23−00.19 fit into current models of high-mass star formation are open questions that will be addressed in future investigations.

P.S. gratefully acknowledges the instructors of the CARMA Summer School 2011, especially for their enthusiasm and dedication to the professors John Carpenter, Nikolaus Volgenau, Melvyn Wright, Dick Plambeck, and Marc Pound. P.S. thanks Andrés Guzmán for his careful reading of the paper. We also thank the anonymous referee for helpful comments that improved the paper. P.S. and J.M.J. acknowledge funding support from NSF grant No. AST-0808001 and AST-1211844. I.J.-S. acknowledges the financial support from the People Programme (Marie Curie Actions) of the European Union's Seventh Framework Programme (FP7/2007–2013) under REA grant agreement number PIIF-GA-2011-301538. Support for CARMA construction was derived from the states of California, Illinois, and Maryland, the James S. McDonnell Foundation, the Gordon and Betty Moore Foundation, the Kenneth T. and Eileen L. Norris Foundation, the University of Chicago, the Associates of the California Institute of Technology, and the National Science Foundation. Ongoing CARMA development and operations are supported by the National Science Foundation under a cooperative agreement, and by the CARMA partner universities.