This site uses cookies. By continuing to use this site you agree to our use of cookies. To find out more, see our Privacy and Cookies policy.

Deep, Broadband Spectral Line Surveys of Molecule-rich Interstellar Clouds

, , , , , , , , , , , and

Published 2017 August 22 © 2017. The American Astronomical Society. All rights reserved.
, , Citation Susanna L. Widicus Weaver et al 2017 ApJS 232 3 DOI 10.3847/1538-4365/aa8098

Download Article PDF
DownloadArticle ePub

You need an eReader or compatible software to experience the benefits of the ePub3 file format.

0067-0049/232/1/3

Abstract

Spectral line surveys are an indispensable tool for exploring the physical and chemical evolution of astrophysical environments due to the vast amount of data that can be obtained in a relatively short amount of time. We present deep, broadband spectral line surveys of 30 interstellar clouds using two broadband λ = 1.3 mm receivers at the Caltech Submillimeter Observatory. This information can be used to probe the influence of physical environment on molecular complexity. We observed a wide variety of sources to examine the relative abundances of organic molecules as they relate to the physical properties of the source (i.e., temperature, density, dynamics, etc.). The spectra are highly sensitive, with noise levels ≤25 mK at a velocity resolution of ∼0.35 km s−1. In the initial analysis presented here, column densities and rotational temperatures have been determined for the molecular species that contribute significantly to the spectral line density in this wavelength regime. We present these results and discuss their implications for complex molecule formation in the interstellar medium.

Export citation and abstract BibTeX RIS

1. Introduction

Nearly 200 molecules have been detected in interstellar and circumstellar environments (Müller et al. 2001, 2005; CDMS 2017), and a great fraction of these are complex organic molecules (COMs). Here, following the convention of Herbst & van Dishoeck (2009), we define a COM as a molecule containing six or more atoms. Laboratory, observational, and modeling studies have revealed that very little is understood about the chemical and physical mechanisms driving complex molecule formation in the interstellar medium (ISM). Calculations (Horn et al. 2004) and laboratory results (Geppert et al. 2006) have shown that the textbook chemical mechanisms thought to lead to complex organics in the ISM—ion–molecule and electron recombination reactions—are actually inefficient processes for the formation of most complex molecules. Even the ubiquitous molecule methanol presents a challenge. Ion–molecule reactions account for only a small percentage of the methanol observed in dense clouds, and neutral methanol comprises only ∼3% of the dissociative recombination products from protonated methanol (Geppert et al. 2006).

Conversely, chemical modeling suggests that reactions of simple radicals on interstellar grain surfaces may be integral to COM formation (Garrod et al. 2008; Laas et al. 2011), and that grain-surface reactions are the dominant formation mechanism for many of the most abundant COMs under typical interstellar cloud conditions. Based on this type of chemistry, many other previously undetected COMs are also predicted to be present in high abundance in hot cores. Modeling studies must be compared with observations to test their robustness. Unfortunately, the subset of sources for which extensive COM inventories have been compiled is quite small. Therefore, the modeling studies have primarily focused on comparisons with the Sagittarius B2(N) hot molecular core, which has the highest number of detected molecules of any sightline studied to date (Snyder 2006), but may not be representative of all sources.

Before a full understanding of interstellar organic chemistry can be obtained, we must test these models against observations to gain a comprehensive view of the possible chemical and physical parameter space. Previously, molecular observations largely consisted of targeted searches focusing on a few lines of only the most abundant molecules, or on molecules that were high-priority targets. Unbiased, self-consistent, broadband spectral line surveys, however, are more effective in determining the true composition of interstellar clouds as they allow the acquisition of spectral data over large frequency ranges and sample many transitions of each molecule. It has already been shown through the initial results reported from the Atacama Large Millimeter/Submillimeter Array (ALMA) that more COMs will be detected in all interstellar environments as spectral sensitivity, spectral bandwidth, and spatial resolution continue to increase. Therefore, such broadband surveys in the millimeter band will guide observations at higher frequencies with ALMA and other submillimeter and far-IR telescopes by identifying the sources that have the richest chemical inventory.

We present here the first results from our molecular line survey observations using two broadband $\lambda =1.3\,\mathrm{mm}$ receivers at the Caltech Submillimeter Observatory (CSO). The goal of this program is to conduct deep, broadband spectral line surveys of interstellar clouds to probe the influence of the physical environment on molecular complexity. We are observing a wide variety of sources in an attempt to correlate the relative abundances of organic molecules with the physical properties of the source (i.e., temperature, density, dynamics, etc.). Our broader research goal is to improve astrochemical models to the point where accurate predictions of complex molecular inventory can be made based on the physical and chemical environment of a given source. The information gained from our observations will serve as a benchmark for these astrochemical models, and therefore will hold the promise of significantly advancing our understanding of interstellar chemical processes. We report here on the details of the observations, the spectral analysis process, and the initial results for the molecular species that contribute significantly to the spectral line density in this spectral regime.

2. Observations and Data Reduction

The observations were performed with the 10.4 m Leighton Telescope of the CSO10 between 2007 September and 2013 June. The nominal source positions and values assumed for vlsr are listed in Table 1. Typical system temperatures (Tsys) during the observations were <400 K; Tsys was higher during periods of high opacity, but did not exceed 1100 K.

Table 1.  Sources, Positions, and Velocities Used in the Observations, and Source Physical Parameters

Source Name Source Type α δ vlsr Distance Luminosity Mass (Gas and Dust) References
    (J2000) (J2000) (km s−1) (pc) (${L}_{\odot }$) (${M}_{\odot }$)  
W3(H2O) Hot Core 2h27m04fs61 $+61^\circ 52$'25 0'' −47.0 2040 $\gt {10}^{5}$ 26 1, 2
L1448 MM-1 Class 0 + Outflow 3h25m38fs80 $+30^\circ 44$'05 0'' 0.0 232 11.6 $\lt 2$ 3–5
NGC 1333 IRAS 2A Hot Corino 3h28m55fs40 $+31^\circ 14$'35 0'' 7.8 235 20.7 0.12 4, 6, 7
NGC 1333 IRAS 4A Hot Corino 3h29m10fs30 $+31^\circ 13$'31 0'' 6.8 235 7.7 1.1 4, 6, 7
NGC 1333 IRAS 4B Hot Corino 3h29m11fs99 $+31^\circ 13$'08 9'' 5.0 235 7.7 0.56 4, 6, 7
B1-b Class 0 3h33m20fs80 $+31^\circ 07$'40 0'' 0.39 230 $\lt 3$ ∼1.5 8
Orion-KL Hot Core 5h35m14fs16 $-05^\circ 22$'21 5'' 8.0 500 $\gt {10}^{5}$ 10 9, 10
NGC 2264 Hot Core 6h41m12fs00 $+9^\circ 29$'09 0'' 7.6 800 71 − 102 48 11–13
NGC 6334–29 Class 0 17h19m57fs00 $-35^\circ 57$'51 0'' −5.0 1.7 k $1.7\times {10}^{5}$ $\lt 200$ 14–16
NGC 6334–38 Class 0 17h20m18fs00 $-35^\circ 54$'42 0'' −5.0 1.7 k $\lt 2\times {10}^{5}$ $\lt 175$ 14–16
NGC 6334–43 Class 0 17h20m23fs00 $-35^\circ 54$'55 0'' −2.6 1.7 k $\lt 2\times {10}^{5}$ $\lt 150$ 14–16
NGC 6334-I(N) Class 0 17h20m55fs00 $-35^\circ 45$'40 0'' −2.6 1.7 k $5\times {10}^{4}$ $\lt 500$ 14, 15, 17
Sgr B2(N-LMH) Hot Core 17h47m19fs89 $-28^\circ 22$'19 3'' 64 8.5 k $6\times {10}^{6}$ $\gt 2000$ 9
GCM+0.693-0.027 Shocked Region 17h47m21fs86 $-28^\circ 21$'27 1'' 68.0 a a a  
GAL 10.47+00.03 Hot Core 18h08m38fs40 $-19^\circ 51$'51 8'' 67.8 10.8 k $\sim 3\times {10}^{5}$ 2200 9, 18, 19
GAL 12.21–0.10 H ii Region 18h12m39fs70 $-18^\circ 24$'20 9'' 24.0 13.5 k $\lt {10}^{6}$ $\lt 14000$ 18, 20, 21
GAL 12.91–00.26 Hot Core 18h14m39fs00 $-17^\circ 52$'0. 0'' 37.5 13.6 k 105 $\gt 70$ 18, 22, 23
HH 80–81 Outflow 18h19m12fs30 $-20^\circ 47$'27 5'' 12.2 1.9 k $\sim 2.4\times {10}^{6}$ $\gt 10$ 24, 25
GAL 19.61–0.23 Hot Core 18h27m37fs99 $-11^\circ 56$'42 0'' 40.0 12.0 k $\lt 2\times {10}^{5}$ 450 9, 18, 20
G24.33+00.11 MM1 Hot Core 18h35m08fs14 $-7^\circ 35$'01 1'' 113.4 6.3 k $\sim 5\times {10}^{4}$ 26 26
GAL 24.78+0.08 Hot Core 18h36m12fs60 $-7^\circ 12$'11 0'' 111.0 6.1 k $\lt 2\times {10}^{5}$ 250 18, 21, 27
GAL 31.41+0.31 Hot Core 18h47m34fs61 $-1^\circ 12$'42 8'' 97.0 7.9 k $\sim 2\times {10}^{5}$ ∼2000 28, 29
GAL 034.3+00.2 Hot Core 18h53m18fs54 $+1^\circ 14$'57 9'' 58.0 3.7 k $\sim 9\times {10}^{4}$ 7000 30, 31
GAL 45.47+0.05 Hot Core 19h14m25fs60 $+11^\circ 09$'26 0'' 62.0 8.0 k $\sim 1\times {10}^{6}$ 250 9, 32
W51 Hot Core 19h23m43fs50 $+14^\circ 30$'34 0'' 55.0 5.4 k $\lt 5\times {10}^{5}$ ∼250 21, 33, 34
GAL 75.78+0.34 H ii Region 20h21m44fs09 $+37^\circ 26$'39 8'' 4.0 3.83 $\lt 3\times {10}^{4}$ $\lt 100$ 35, 36
W75N Hot Core 20h38m36fs60 $+42^\circ 37$'32 0'' 10.0 1.7 k $\lt 5\times {10}^{4}$ $\lt 2$ 20, 37, 38
DR21(OH) Hot Core 20h39m01fs10 $+42^\circ 22$'49 1'' −3.0 3.0 k $5\times {10}^{4}$ 350 9
L1157 B1 Class 0 + Outflow 20h39m10fs20 $+68^\circ 01$'11 5'' 2.7 325 5.8 0.2 7, 39, 40
NGC 7538 Hot Core 23h13m45fs70 $+61^\circ 28$'21 0'' −57.0 2.8 k $2\times {10}^{4}$ $\lt 2000$ 41–43

Note.

aThis source has not been well-characterized.

References. (1) Qin et al. (2015), (2) Wyrowski et al. (1999b), (3) Hirota et al. (2011), (4) van Dishoeck et al. (2011), (5) Tobin et al. (2015), (6) Hirota et al. (2008), (7) Jørgensen et al. (2007), (8) Öberg et al. (2010), (9) Kurtz et al. (2000), (10) McKee & Tan (2003), (11) Raimond (1966), (12) Cunningham et al. (2016), (13) Ward-Thompson et al. (2000), (14) Neckel (1978), (15) Loughran et al. (1986), (16) Matthews et al. (2008), (17) Gezari (1982), (18) Pandian et al. (2008), (19) Cesaroni et al. (2010), (20) Wu et al. (2010), (21) Ginsburg et al. (2012), (22) Faúndez et al. (2004), (23) Galván-Madrid et al. (2010), (24) Maud et al. (2015), (25) Fernández-López et al. (2011), (26) Rathborne et al. (2011), (27) Beltrán et al. (2004), (28) Le Teuff et al. (2000), (29) Cesaroni et al. (2010), (30) Wood & Churchwell (1989), (31) Mookerjea et al. (2007), (32) Hernández-Hernández et al. (2014), (33) Sato et al. (2010), (34) Shi et al. (2010) (35) Sánchez-Monge et al. (2013), (36) Pipher et al. (1979), (37) Schneider et al. (2006), (38) Minh et al. (2010), (39) Straizys et al. (1992), (40) Shirley et al. (2000), (41) Crampton et al. (1978), (42) Willner (1976), (43) Werner et al. (1979).

Download table as:  ASCIITypeset image

Observations were performed using two sets of receivers and spectrometers. The first was a prototype 230 GHz wideband receiver (Rice et al. 2003; Kaul et al. 2004; Sumner 2011), which was used with the facility acousto-optical spectrometer (AOS; Horn et al. 1999). The receiver has a 12 GHz bandwidth, but the total IF bandwidth for the observations was limited by the AOS to 4 GHz, with a channel width of ∼0.65 MHz. The second receiver was the facility 230 GHz wideband receiver (Kooi et al. 2007), used with the facility Fast Fourier Transform Spectrometer (Klein et al. 2006). The receiver is a dual-frequency, tunerless wide-IF bandwidth (4–8 GHz) system with a fully synthesized LO; the total IF bandwidth for the observations was 4 GHz, with a channel width of ∼0.27 MHz, which is equivalent to a velocity resolution of ∼0.35 km s−1.

All collected spectra were double sideband (DSB). A set of rest frequencies in the range 223.192–251.192 GHz was chosen, with each rest frequency setting being separated by 4 GHz. Each frequency point was sampled with a minimum redundancy of six to ensure sufficient sampling for deconvolution of the DSB spectra; most frequencies were sampled with a redundancy of eight. To achieve this redundancy, four IF offsets (4.254, 5.268, 6.754, and 7.795 GHz) were applied to each rest frequency, and DSB spectra were collected at each of these frequency settings. Additional IF offsets of 4.753, 6.283, 5.767, and 7.269 GHz were applied to the two lowest rest frequency settings. The number of DSB spectra acquired for each source varied, depending on weather conditions and the amount of available observing time.

Spectral intensities for the raw data were calibrated using the standard chopper wheel calibration method and placed on the atmosphere-corrected temperature scale, ${T}_{a}^{* }$. The chopper settings used were either 70'' ± 8'' of throw with a chopping frequency of 1.123 Hz, or 90'' ± 8'' of throw with a chopping frequency of 1.1 Hz. The estimated percent calibration error in antenna temperature is <20%, based on previous experience with this hardware. Integration times were adjusted based on the Tsys determined after tuning the receiver to a given frequency setting (i.e., for one set of rest frequency and IF values), such that the noise level achieved for a given frequency setting was ∼30 mK. Pointing was verified on average every two hours, and pointing offsets were consistent to within ≤5'' over the course of a given night. As an independent verification of the pointing accuracy, each spectrum was checked against previous spectra recorded for consistency in spectral line intensity. The full-width-half-power beam size at 230 GHz was determined to be 33farcs4 for the prototype receiver, and 35farcs5 for the facility receiver.

Spectral cleaning and deconvolution were performed using the CLASS software package (versions jun13a and jul14a) included in the GILDAS suite of programs (Institut de Radioastronomie Millimétrique, Grenoble, France; GILDAS 2017). Baselines were fitted and removed from each DSB spectrum using a linear function for most sources. A small number of the sources required more complicated baseline fitting via boxcar averaging or interpolating a low-order polynomial. In cases where the line density precluded a reliable baseline fit, a constant offset was subtracted from the ${T}_{a}^{* }$ scale. This offset was determined through comparison to the baseline in neighboring spectral windows. In general, the baselines were adjusted until they were as close as possible to zero. The process was checked for reproducibility, and the noise that was introduced was comparable to or less than the overall noise level of the spectra. In addition to baseline subtraction, some spurious noise features were present in the spectra, but their identification was straightforward because they were significantly narrower than the spectral line widths. Such noise features were removed from the spectra by blanking the affected channels prior to processing. After baseline subtraction and removal of noise features, the spectra were resampled at a uniform channel spacing of 1 MHz. The deconvolution was performed using the standard routine included in CLASS. An initial deconvolution was performed assuming no gain variations between the upper and lower sidebands. This was then used as an initial guess to constrain a second deconvolution allowing for varying gains between sidebands. Due to the nature of the algorithm used, extremely strong lines in the spectrum can produce ghosts visible above the final residual noise. Therefore, the CO line at 230.5 GHz and other strong spectral features with intensities above 2 K were masked during deconvolution and later stitched back into the spectrum to prevent the introduction of spurious features. After deconvolution was completed, it was found that a few strong features in the image sideband produced spectral artifacts appearing as ghosts, that manifest as negative features in the deconvolved spectrum. These features were blanked in the deconvolved spectrum. After deconvolution, all intensities were corrected to the main beam temperature scale, Tmb, using the relation ${T}_{\mathrm{mb}}={T}_{a}^{* }/{\eta }_{\mathrm{mb}}$. Here, ${\eta }_{\mathrm{mb}}$ is the main beam efficiency, which was determined through observation of planets to be 57% ± 10% for both receivers. The root-mean-squared noise in the deconvolved spectra was determined to be ≤25 mK on the Tmb scale.

3. Line Identification and Analysis

The deconvolved spectra are broadband, highly sensitive, chemically complex, and line-blended in many cases. Detailed spectral analysis using the traditional Boltzmann diagram approach was therefore challenging. Such analyses are often unreliable because they are conducted one molecule at a time using a small set of lines for which line widths are determined independently and do not account for possible line blending. We have therefore developed a new program that analyzes all of the lines from one molecule across the entire broadband spectrum simultaneously. This Global Optimization and Broadband Analysis Software for Interstellar Chemistry (GOBASIC) is described in detail elsewhere (Rad et al. 2016). The performance of this software has been previously benchmarked (Rad et al. 2016) against both synthetic spectra and through a comparison between our Orion-KL data set and the results of Blake et al. (1986), and has been found to reliably model line surveys such as those considered here. In short, the software assumes local thermodynamic equilibrium (LTE) and Gaussian line shapes, and conducts a simultaneous, multimolecule, multicomponent, broadband spectral analysis. It is worthwhile to note that the rotational temperatures of molecules from this LTE analysis are not guaranteed to be the true gas kinetic temperatures. Rather, it only means that the molecular line emission can be modeled by a single excitation temperature under Boltzmann statistics. We used published line surveys for Sgr B2(N) (Nummelin et al. 1998a) and Orion-KL (Sutton et al. 1985; Blake et al. 1986, 1987) to guide the initial list of molecules that were included for the analysis of this data set. We imported spectral line catalogs from the Cologne Database for Molecular Spectroscopy (CDMS; Müller et al. 2005) and the Jet Propulsion Laboratory (JPL) Spectral Line Catalog (Pickett et al. 1998). Unless combined in the JPL/CDMS catalog files, individual vibrational states were treated as unique molecular components. In the case where both databases contained information for a given molecule, we used the catalog that had most recently been revised and updated according to the online documentation provided by the databases.

The partition function interpolation required for such an analysis is not trivial for many molecules. In order to match the line intensities tabulated in the JPL/CDMS catalogs, we used the tabulated partition function values provided by these databases along with the molecular catalogs. These partition function values are determined from direct summations of the Boltzmann factors of all accessible energy levels in the catalog. These values are provided in the documentation for each molecule at several specific temperature points spanning 10 to 300 K. Blake et al. (1986) showed that for either linear or nonlinear molecules, their pure rotational partition function ${Q}_{\mathrm{rot}}(T)$ is always proportional to T or ${T}^{3/2}$, respectively. Unfortunately, this pure rotational model did not always describe the tabulated partition function values provided by the databases. This was a problem in particular for molecules that possess significant internal motions, and for catalogs (and thus partition function values) merging multiple vibrational states. Therefore, we adopted the general functional form

Equation (1)

for partition function interpolation. This functional form consists of two terms: a pure ${Q}_{\mathrm{rot}}$ term and a ${Q}_{\mathrm{rot}}$ combined with a Boltzmann factor for a vibrational state. When holding part of the four coefficients to specific values, this functional form reduces to simpler cases. For example, a pure ${Q}_{\mathrm{rot}}$ term is derived when $\gamma =\epsilon =0$ and β = 1 or 3/2. . For each molecule to be analyzed, we minimized the coefficients to be fitted in Equation (1), only adding additional terms if necessary to obtain a reliable fit. Since the partition function values span several orders of magnitude, the residual may not be reliable to evaluate the goodness of the interpolation. Instead, we examined the range of the relative deviation between the interpolated values and the tabulated values. This relative deviation is defined as $\delta ={Q}_{\mathrm{interpolation}}/{Q}_{\mathrm{catalog}}-1$. For the molecules included in our analysis, this interpolation method works well in general, except for a few molecules: CH3COCH3, H2CCO, H2CO, and CH3SH, whose maximum relative deviation exceeds 10%. We believe this level of deviation is acceptable because of the 20% uncertainty in the flux calibration in addition to the fact that the deviation is much smaller than in the other models reported (Friedel et al. 2005; Peng et al. 2013). The line catalogs, their databases and versions, the coefficients for the partition function interpolation, and the relative deviation range of our interpolation are summarized in Table 2.

Table 2.  Summary of the Information of Line Catalogs

Molecule Formula Database Version Partition Function Interpolation
        α β γ epsilon δ(%)a
Acetaldehyde CH3CHO JPL 2012 Jan 17.1065 1.5 0.5489 −256.2868 [−0.51 0.27]
Acetone CH3COCH3 JPL 2008 Mar 15274.522 1.5 0.0165 −527.2334 [−12.48 27.87]
Carbon Monosulfide CS CDMS 2008 Oct 0.8510 1 0 0.3957 [−0.10 0.06]
Carbon Monosulfide 13CS CDMS 2004 Jan 1.8034 1 0 0.3631 [−0.07 0.04]
Carbon Monosulfide 13C34S CDMS 2004 Jan 0.9171 1 0 0.3593 [−0.04 0.03]
Carbon Monosulfide C33S CDMS 2004 Jan 3.4325 1 0 0.3845 [−0.04 0.03]
Carbon Monosulfide C34S CDMS 2004 Jan 0.8650 1 0 0.3790 [−0.07 0.04]
Carbon Monoxide Ion CO+ CDMS 2013 Aug 0.7105 1 0 0.9603 [−4.04 0.63]
Carbonyl Sulfide OCS CDMS 2005 Nov 3.4361 1 0 0 [−0.75 0.21]
Carbonyl Sulfide OC34S CDMS 2005 Apr 3.5220 1 0 0 [−0.73 0.21]
Cyanide Radical CN (2Σ+) CDMS 2005 May 2.2040 1 0 0.9449 [−0.26 0.33]
Cyclopropenylidene c-HCCCH CDMS 2012 May 2.4549 1.5 0 0.2384 [−0.48 0.30]
Diazenylium N2D+ CDMS 2009 Mar 4.9476 1 0 0 [−4.80 1.51]
Dimethyl Ether CH3OCH3 CDMS 2011 Nov 91.2742 1.5 0 0 [−1.80 4.00]
Ethanol C2H5OH JPL 2008 Apr 6.6674 1.5 0.4972 −59.8213 [−0.11 0.17]
Ethenone (Ketene) H2CCO CDMS 2003 Nov 1.0186 1.5 1 −3.8746 [−8.45 11.59]
Ethyl Cyanide C2H5CN JPL 2009 Jul 7.1989 1.5 0 0 [−0.38 0.17]
Ethynyl CCH CDMS 2010 Jul 1.9078 1 0 0.7175 [−0.13 0.16]
Formaldehyde H2CO CDMS 2005 Aug 0.2791 1.5 1 −2.1456 [−9.43 14.90]
Formaldehyde HDCO CDMS 2000 Feb 0.3694 1.5 0 0.7692 [−2.54 3.62]
Formaldehyde H213CO JPL 2005 Oct 0.2903 1.5 1 −2.7718 [−4.11 2.85]
Formamide NH2CHO CDMS 2003 Apr 1.8945 1.5 0 0 [−4.55 1.15]
Formic Acid trans-HCOOH CDMS 2003 Jul 1.7189 1.5 0 0 [−1.20 0.38]
Formyl Cation HCO+ CDMS 2007 Oct 0.4671 1 0 0.7344 [−0.14 0.17]
Formyl Cation HC18O+ JPL 1983 Dec 0.4894 1 0 0.6847 [−0.01 0.02]
Formyl Cation H13CO+ CDMS 2007 Oct 0.4802 1 0 0.7143 [−0.12 0.15]
Formyl Radical HCO JPL 1983 Aug 0.5697 1.5 0 1.8171 [−2.27 4.02]
Hydrogen Cyanide HCN CDMS 2007 May 1.4357 1 0 0 [−5.69 1.55]
Hydrogen Cyanide H13CN CDMS 2014 Nov 1.4489 1 0 0.6870 [−0.12 0.05]
Hydrogen Cyanide HC15N CDMS 2005 Aug 0.4892 1 0 0 [−2.64 0.78]
Hydrogen Isocyanide HN13C JPL 1979 Dec 0.4889 1 0 0 [−5.20 1.95]
Hydrogen Sulfide H2S CDMS 2008 Oct 0.1015 1.5 0 0 [−5.25 2.55]
Hydrogen Sulfide HDS CDMS 2008 Oct 0.0797 1.5 0 2.7203 [−6.86 8.96]
Isocyanic Acid HNCO CDMS 2009 May 0.4451 1.5242 0 2.8006 [−7.77 5.17]
Isocyanide DNC CDMS 2009 Sep 0.5460 1 0 0.6251 [−0.09 0.11]
Methanethiol CH3SH CDMS 2012 Jan 46.0687 1.7072 0.1085 −1244.7174 [−16.44 29.21]
Methanol CH3OH JPL 2010 Mar 1.8680 1.7336 0.2229 −1005.6869 [−3.17 3.14]
Methanol CH2DOH JPL 2012 Jun 0.5941 1.7821 0 0 [−4.97 5.51]
Methanol 13CH3OH CDMS 2002 Dec 1.2810 1.7314 0.3372 −929.5959 [−3.12 3.13]
Methenamine CH2NH JPL 1981 Jan 1.1497 1.5 0 0 [−3.52 1.19]
Methyl Cyanide CH3CN CDMS 2016 Nov 1.9286 1.5 0 1.2952 [−1.94 2.11]
Methyl Cyanide CH3CN (${v}_{8}=1$) CDMS 2016 Nov 0.1464 2.1176 0 −481.7178 [−10.57 5.61]
Methyl Cyanide ${\mathrm{CH}}_{3}^{13}\,\mathrm{CN}$ CDMS 2009 Dec 1.9251 1.5 0 1.2743 [−1.28 1.79]
Methyl Cyanide 13CH3CN CDMS 2009 Dec 2.0506 1.5 0 0 [−10.77 2.30]
Methyl Formate HCOOCH3 JPL 2009 Apr 25.2869 1.5 0.9900 −190.5726 [−0.30 0.22]
Methylamine CH3NH2 JPL 2009 Jan 23.9534 1.5 0 0 [−0.20 0.23]
Nitric Oxide NO JPL 2010 Oct 0.9345 1.2435 0 6.2186 [−4.16 3.89]
Nitrogen Sulfide NS JPL 1995 Feb 2.6639 1.1673 0 3.1202 [−4.17 4.36]
Nitrogen Sulfide N34S CDMS 2011 Aug 3.9120 1.0988 0 1.6296 [−6.92 8.19]
Propyne CH3CCH CDMS 2008 Aug 1.0301 1.5 1 2.4731 [−1.02 1.57]
Propyne CH3CCD CDMS 2002 Apr 1.1330 1.5 1 2.4383 [−1.02 1.57]
Propyne CH2DCCH CDMS 2002 Apr 1.8964 1.5 0 0 [−0.91 0.29]
Propynenitrile HCCCN CDMS 2000 Oct 4.6073 1 0 0 [−2.07 0.53]
Propynenitrile HCCCN (${v}_{7}=1$) CDMS 2000 Oct 9.1357 1 0 −321.0257 [−0.01 0.00]
Propynenitrile H13CCCN JPL 1979 Dec 14.2113 1 0 0 [−0.53 0.20]
Propynenitrile HC13CCN JPL 1979 Dec 13.8318 1 0 0 [−0.54 0.21]
Propynenitrile HCC13CN JPL 1979 Dec 13.8295 1 0 0 [−0.56 0.21]
Silicon Monoxide SiO JPL 1984 Jan 0.9693 1 0 0 [−2.69 1.25]
Sulfur Dioxide SO2 CDMS 2005 Jul 1.1492 1.5 0 0 [−3.94 1.01]
Sulfur Dioxide 33SO2 JPL 1996 Nov 4.6045 1.5 0 0 [−0.95 0.30]
Sulfur Dioxide 34SO2 JPL 1996 Nov 1.1608 1.5 0 0 [−0.95 0.30]
Sulfur Monoxide SO CDMS 1998 Aug 1.4598 1 1 −16.7494 [−1.19 1.95]
Sulfur Monoxide 33SO CDMS 1998 Aug 5.8908 1 1 −16.7632 [−0.89 2.16]
Sulfur Monoxide 34SO CDMS 1998 Aug 1.4851 1 1 −16.7767 [−0.83 2.36]
Sulfur Monoxide Ion SO+ JPL 1996 Dec 1.9142 1 0 0.0914 [−8.00 5.28]
Thioformaldehyde H2CS CDMS 2008 Feb 0.5848 1.5 1 −3.2965 [−9.84 13.46]
Thioformaldehyde HDCS CDMS 2004 May 0.7709 1.5 0 0 [−1.73 0.55]
Thioformaldehyde H2C34S CDMS 2006 Feb 0.6004 1.5 1 −3.6703 [−4.28 3.12]
Thiomethylium HCS+ CDMS 2003 Jun 0.9871 1 0 0 [−2.55 0.90]
Vinyl Cyanide C2H3CN JPL 2009 Oct 29.1521 1.5 0.5212 −396.4232 [−0.31 0.21]
Water HDO JPL 1989 Oct 0.0273 1.5 0 4.7166 [−2.85 5.14]

Note. The molecules are sorted alphabetically by their chemical names. The coefficients for the interpolated partition function (Equation (1)) and the range of the interpolation deviation are listed.

aDeviation between the interpolated Q(T) values and the catalog tabulated Q values. Listed as a range of $[{\delta }_{\min },{\delta }_{\max }]$ in units of percent.

Download table as:  ASCIITypeset images: 1 2

The values for the source velocity, vlsr, were set to those listed in Table 1 based on information available in the literature for previous molecular observations in each source. The column density (${N}_{{\rm{T}}}$), rotational temperature (${T}_{\mathrm{rot}}$), line width (characterized as the FWHM), and velocity shift ($\delta v$) with respect to the source vlsr were determined iteratively for each molecule using the procedures outlined in Rad et al. (2016). In brief, four steps were followed for each source. First, spectral lines from CO, 13CO, C17O, and C18O were flagged in GOBASIC because of their large intensity and non-Gaussian line shapes. Data in the frequency ranges corresponding to these four lines were omitted from analysis. Second, strong spectral features arising from ubiquitous small molecules, such as CS, CN, CCH, and HCN, were fitted with arbitrary temperature and column density in GOBASIC to achieve only a best-fit model for the emission. These molecules only present a few lines in the observed frequency range, and therefore the information is not sufficient to perform a full LTE analysis. Nonetheless, their flux was properly accounted for in the analysis by using the arbitrarily fitted parameters. After that, each target molecule was analyzed using GOBASIC with the inclusion of the results from the second step. All fits were constrained within a reasonable parameter space for the search grid, where the temperature, column density, velocity shift, and line widths were held to a range of physically meaningful values; this shortened the analysis time by limiting the size of the search grid. In this initial fitting step, we determined which molecules were unambiguously detected in a source, and obtained a coarse estimate of their LTE parameters. In the final step, all detected target molecules were optimized simultaneously in GOBASIC, using their LTE parameters obtained from the third step as the initial guess for the final analysis. The convergence requirements for the change of rms of the residual and the change of parameter vector were both set to $\lt {10}^{-6}$ in GOBASIC. Caution was taken to make sure the final results did not hit the boundaries of the searching parameter space, with the exception of those cases that are specified below. Spectral lines with $\gt 1\,\mathrm{MHz}$ uncertainty in the catalog were discarded in our analysis to avoid unreliable assignments. Not all sources listed in Table 1 have interferometric data available in the 1.3 mm band. As a consequence, for consistency, beam dilution effects were not accounted for in our analysis, despite the ability of GOBASIC to do so. Uncertainties in the final parameters were determined from the quality of the fit as described by Rad et al. (2016), using 10−3 as the grid size for its numerical gradient calculation.

4. Results

The baseline-subtracted, deconvolved spectra for the 30 star-forming regions are shown in Figure 1. The spectra are plotted with intensity in units of antenna temperature, and sorted from high flux on the top to low flux on the bottom. The spectral coverage is 220–267 GHz in general, with individual variations up to 10 GHz depending on the specific spectral coverage obtained for that source. The Orion-KL survey is particularly short relative to the others, because it was used only as a benchmark to test receiver operation and deconvolution reliability. Given the extensive literature related to Orion-KL, no new astrochemical information was expected to be gained from its analysis. Likewise, although Sgr B2(N) is included here, a full analysis was not performed, because it has been so extensively studied by other groups (see Mehringer et al. 1997; Miao & Snyder 1997; Nummelin et al. 1998a, 2000; Hollis et al. 2000, 2006; Martín-Pintado et al. 2001; Remijan et al. 2002; Snyder et al. 2002; Friedel et al. 2004; Apponi et al. 2006; Halfen et al. 2006, 2011, 2013; Leigh et al. 2007; Belloche et al. 2008, 2009, 2013; Neill et al. 2014).

Figure 1.

Figure 1. Baseline-subtracted, deconvolved spectra for the 30 observed sources. Spectral baselines are offset on the Tb scale for clarity. The data used to create this figure are available.

Standard image High-resolution image

The CO isotopologue lines, including CO at 230.5 GHz, 13CO at 220.4 GHz, C17O at 224.7 GHz, and C18O at 219.6 GHz, are observed in all sources. Other strong spectral features that appear in most sources include the H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line at 225.7 GHz, the CN $N=2\to 1$ hyperfine lines at 226 GHz, the CH3CN $J=13\to 12$ Q-branch at 239 GHz, the CS $J=5\to 4$ line at 244.9 GHz, the CH3OH $J=5\to 4$ Q-branch around 242 GHz and ${K}_{a}=3\to 2$ R-branch around 252 GHz, the CCH $N=3\to 2$ hyperfine lines at 262.1 GHz, and the HCN $J=3\to 2$ line at 265.9 GHz. Because of the source-to-source variation of the spectral coverage, molecules such as HCN and HCO+ are undetected in some sources simply because their lines are outside of the observed spectral window. This non-detection due to the limited spectral range is distinguished from the non-detection due to weak spectral intensity, which impacts the interpretation of the chemistry in these sources.

Here, because of the vast nature of the data set, we will focus the discussion on the molecules that can be reliably fit using LTE, rather than considering all possible molecular detections in the sources. A reliable LTE analysis requires the detection of sufficient unblended molecular lines that span a wide range of upper-state energy. This criterion prohibits us from reporting the abundance and temperature information for a few important COM precursors, such as H2CO, from our analysis. Notable detections of small molecules will be mentioned, and the full analysis results can be found in Table 3, which can be retrieved in machine-readable format. In Table 3, each detected molecule is listed with four parameters and their associated uncertainties: the column density ${N}_{{\rm{T}}}$ in cm−2, the rotational temperature ${T}_{\mathrm{rot}}$ in K, the FWHM in km s−1, and the velocity shift with respect to ${v}_{\mathrm{lsr}},\delta v$, in km s−1. For the molecules with insufficient information for LTE analysis, arbitrary ${N}_{{\rm{T}}}$ and ${T}_{\mathrm{rot}}$ may be assigned, and their associated uncertainties may be unphysical.

Table 3.  Detected Molecules in the 30 Sources and Their LTE Parameters and Associated $1\sigma $ Uncertainties

Source Molecule ${N}_{{\rm{T}}}$ (cm−2) ${T}_{\mathrm{rot}}$ (K) FWHM (km s−1) $\delta v$ (km s−1)
B1-b CH3OH 50 ± 9 × 1012 12.1 ± 1.6 1.43 ± 0.09 6.50 ± 0.05
B1-b SO 46 ± 7 × 1012 9.9 ± 0.5 1.16 ± 0.04 6.66 ± 0.02
DR21(OH) CH3CCH 124 ± 4 × 1013 36.8 ± 1.0 5.0 ± 0.2 0.12 ± 0.07
DR21(OH) CH3CHO 8 ± 1 × 1013 139 ± 24 5.5 ± 0.4 0.2 ± 0.1
DR21(OH) CH3CN 231 ± 5 × 1011 56.7 ± 2.1 5.4 ± 0.2 0.63 ± 0.09
DR21(OH) CH3OH 176 ± 2 × 1013 18.2 ± 0.3 5.63 ± 0.09 0.34 ± 0.02
DR21(OH) CH3OH 152 ± 3 × 1013 92.8 ± 2.4 5.9 ± 0.2 0.78 ± 0.04
DR21(OH) CH3OCH3 35 ± 2 × 1013 60.0 ± 4.9 4.1 ± 0.4 0.3 ± 0.2
DR21(OH) H2CCO 30 ± 3 × 1012 58.0 ± 9.7 4.8 ± 0.8 0.4 ± 0.2
DR21(OH) H2CS 148 ± 2 × 1012 53.7 ± 1.6 5.59 ± 0.09 0.56 ± 0.04
DR21(OH) HCCCN 10 ± 2 × 1012 79 ± 12 6.0 ± 0.3 0.4 ± 0.1
DR21(OH) HCOOCH3 107 ± 7 × 1012 81 ± 26 4.1 ± 0.4 0.2 ± 0.2
DR21(OH) HDCO 12 ± 6 × 1012 79 ± 42 4.8 ± 0.6 0.6 ± 0.2
DR21(OH) OCS 4 ± 1 × 1014 42.3 ± 8.5 5.4 ± 0.4 0.6 ± 0.1
DR21(OH) SO 431 ± 2 × 1012 36.2 ± 0.5 6.5 ± 0.0 0.92 ± 0.01
DR21(OH) SO2 288 ± 7 × 1012 75.6 ± 2.6 7.7 ± 0.2 1.55 ± 0.06
DR21(OH) 33SOa N/A N/A 5.8 ± N/A 1.9 ± N/A
DR21(OH) 34SOa N/A N/A 6.3 ± N/A 1.4 ± N/A

Note. The four LTE parameters are the column density ${N}_{{\rm{T}}}$ in cm−2, the rotational temperature ${T}_{\mathrm{rot}}$ in K, the FWHM in km s−1, and the velocity shift with respect to ${v}_{\mathrm{lsr}},\delta v$, in km s−1. $1\sigma $ uncertainties are reported. Only molecules with more than three unblended lines at different upper-state energies can be fitted to a reliable ${T}_{\mathrm{rot}}$. Other small molecules are fitted with arbitrary ${N}_{{\rm{T}}}$ and ${T}_{\mathrm{rot}}$, and their associated uncertainties may be unphysically large.

aSpectral lines of this molecule are detected, but they are insufficient to perform a reliable LTE analysis. Only the FWHM and $\delta v$ are reported here, without their unphysical uncertainties. The full set of determined parameters for reproducing our best-fit spectra are given in the machine-readable table.

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as:  DataTypeset image

Figure 2 shows a sample frequency window, in which the molecular components fitted in the global LTE analysis are plotted.

Figure 2.

Figure 2.

Sample frequency window of the molecular components fitted in the global LTE analysis. The CH3OH line at 250510 MHz is non-LTE in some of the sources. (The complete figure set (29 images) is available.)

Standard image High-resolution image

Due to the complexity of the physical environment in most of these sources, molecules have a wide range of rotational temperatures and FWHMs. This is to be expected given that many of these sources have been shown through interferometric studies to have multiple kinematic components (Minh et al. 2010; Furuya et al. 2011; Friedel & Widicus Weaver 2012). Therefore, no upper-limit calculations were performed on undetected molecules. Instead, we simply report a non-detection. It is important to mention that these non-detections can be interpreted in two ways. The insufficient spectral line intensity could come from a low abundance (or non-presence) of the carrier molecule in that source, or from beam dilution if the carrier molecule's spatial distribution is compact compared to the single-dish beam.

These results can be compared to those previously reported, listed below for each source included in our data set. Due to the large number of sources, we do not provide here a comprehensive literature review of all previous observations as this is not within the scope of this manuscript. Instead, we focus on the molecules included in our analysis, especially COMs. We also do not attempt a quantitative comparison to previous results from the literature. Many of the previous observations were conducted in other wavelength ranges and therefore probed other temperature regimes and hence spatial scales. Likewise, previous observations used telescopes with different beam sizes, and there is little or no spatial scale information available for most of these sources to enable a proper beam dilution correction. Therefore, a direct, quantitative comparison is impossible. Nonetheless, we can look at overall trends in the molecules detected as well as their abundances and temperatures and draw conclusions from these results. An overview of the pertinent literature for each source and a comparison to our results are included below.

4.1. B1-b

B1-b is a dark molecular cloud with an embedded prestellar core; we observed this source at the "core" position as designated by Öberg et al. (2010). A cold complex chemistry was discovered in B1-b via the detections of CH3OH, HCOOCH3, CH3CHO, trans-HCOOH, C4H, HNCO, and the H2C3O isomers (Sakai et al. 2009a; Öberg et al. 2010; López-Sepulcre et al. 2015; Cuadrado et al. 2016; Loison et al. 2016), and tentatively CH3OCH3 (Öberg et al. 2010), as well as CH3O (Cernicharo et al. 2012). These previous detections were all reported at levels well below the noise level obtained in our observations. Therefore, in our survey, CH3OH was the only COM detected. Several other diatomic and triatomic molecules were detected here, including CN, CS, and HCN. The H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was also detected. Dark molecular clouds are known to have deuterium enrichment (Bergin & Tafalla 2007). We detected N2D+, DNC, and HDCO in B1-b, but not HDO. Other commonly detected deuterated molecules, such as DCN, DCO+, and H2D+, do not have spectral lines within the spectral coverage of our survey.

4.2. GCM+0.693-0.027

GCM+0.693-0.027, located in the Sgr B2 complex, is a shocked region where COMs such as CH3OH, C2H5OH, CH3OCH3, HCOOCH3, and trans-HCOOH have previously been detected (Requena-Torres et al. 2006). The line shape is non-Gaussian and asymmetric, with strong absorption features at the higher frequency edge of the lines. GOBASIC can only treat Gaussian line shapes; therefore, the fit did not perfectly describe the source. Nevertheless, we could fit CH3OH, CH3CCH, CH2NH, H2CS, and OCS with some confidence. The temperatures of these molecules range from 15 to 63 K. The spectral lines are extremely wide, with FWHMs from 18 to 24 km s−1. The velocities are also not consistent. OCS has a blueshift of 0.6 km s−1, whereas all of the other molecules display redshifts. CH3OH and CH3CCH are redshifted by 4 km s−1, followed by H2CS at 3 km s−1, and then CH2NH at 2 km s−1. The H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was detected and fit with one Gaussian profile.

4.3. HH 80–81

HH 80–81 is a massive bipolar outflow where CH3OH, CH3CN, H2CO, SO2, SO, HCO+, and NH3 were observed (Martí et al. 1993; Gómez et al. 2003; Qiu & Zhang 2009). In our survey, we reliably fit CH3OH, CH3CCH, H2CS, SO, and SO2. The molecules are cold, with the hottest one being SO2 at 47 K. The other four molecules are within 15–25 K. The spectral lines are also narrow, with FWHMs from 1.5 to 3.1 km s−1. The velocities are consistent within 1 km s−1. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with two Gaussian components.

4.4. L1157

The L1157 region is a bipolar outflow from the central star L1157 MM. We observed the B1 position in the outflow based on the interferometric study of Arce et al. (2008). Millimeter single-dish and interferometry studies on small molecules (CS, SiO, etc.), CH3OH, and CH3CN have revealed the outflow structure and molecular component (Avery & Chiao 1996; Codella et al. 2009; Flower et al. 2010; Gómez-Ruiz et al. 2013; Burkhardt et al. 2016). Additional observations were conducted in the submillimeter by the CHESS project on Herschel (Codella et al. 2010, 2012; Lefloch et al. 2010, 2012; Benedettini et al. 2012; Busquet et al. 2014). Bachiller & Gutiérrez (1997) and Bachiller et al. (2001) reported spectral lines from CH3OH, H2CO, SiO, HCN, CN, CS, SO, and SO2 in the 1.3 mm band using the IRAM 30 m telescope. Another line survey reported the detection of HCOOCH3, CH3CN, HCOOH, and C2H5OH in its B1 shock region, which is in the blueshifted lobe of L1157 (Arce et al. 2008).

In our survey of the B1 region, we confirmed the detection of small molecules reported by Bachiller & Gutiérrez (1997) and Bachiller et al. (2001), except for the SO2 lines, which were too weak to be observed. No spectral features, however, were found for any of the COMs reported by Arce et al. (2008). This is likely due to beam dilution, as the COM detections showed small spatial scales relative to the CSO beam. The spectral line shape in L1157 is asymmetric. We fit two velocity components to methanol and SO, one with a blueshift of 3.7–3.9 km s−1 and a wider FWHM of 6 km s−1, another one with a blueshift of 1 km s−1 and a narrower FWHM of 3–3.6 km s−1. The narrower component corresponds to the north region, where the lines are redshifted from the gas extending farther out from the star. The wider component corresponds to the south region toward the B1 and B2 blueshifted lobes. The molecules are extremely cold, with rotational temperatures below 15 K, and do not perfectly follow a Boltzmann intensity profile. Nevertheless, we can estimate the methanol column density to be $1.2\times {10}^{15}$ cm−2, and the SO column density to be $1.8\times {10}^{14}$ cm−2. The column density ratio between methanol and SO agrees well with that reported by Bachiller & Gutiérrez (1997). The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with two Gaussian components.

4.5. L1448 MM-1

The L1448 MM-1 source is a bipolar outflow. We observed the C(N) position. This source has been studied previously in the millimeter band (Bachiller et al. 1991; Gregersen et al. 1997; Maret et al. 2004; Jørgensen et al. 2005b, 2007; Maret et al. 2005; Requena-Torres et al. 2007). In these surveys, HCO+, H13CO+, HN13C, SiO, H2CO, and CH3OH were detected. Parise et al. (2006) studied the deuterated chemistry in detail using the HDCO, D2CO, CH2DOH, and CH3OD lines. In the 3 mm band, the long carbon chain molecule C4H was also observed (Sakai et al. 2009a). We confirmed the presence of CH3OH and H2CO lines in the 1.3 mm band. In addition to CH3OH and H2CO, we detected CS, CN, and HCN. We observed a high D/H ratio, indicated by strong lines from DNC, N2D+, and HDCO.

The spectral lines are much narrower, compared to the outflow of L1157. An FWHM of 1.5 km s−1 was found for CS, and line widths were narrower for the other detected molecules. CH3OH has a rotational temperature of 17.7 K, slightly warmer than L1157. We only observed one main velocity component, which is redshifted from the central star rest velocity by 5–6 km s−1. The H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was detected and fit with one Gaussian profile.

4.6. NGC 6334

NGC 6334, also called the Cat's Paw nebula, is a H ii region that contains several massive protostars (Straw et al. 1989). Although it has been studied in the infrared and millimeter for its star formation processes (e.g., Persi et al. 2000; Hunter et al. 2006; Brogan et al. 2009), only a few molecular surveys are available for this source (Bisschop et al. 2007; Thorwirth et al. 2007; Walsh et al. 2010). In these surveys, COMs such as CH3OH, 13CH3OH, CH3CCH, CH3CN, ${\mathrm{CH}}_{3}^{13}\,\mathrm{CN}$, C2H5OH, HCOOCH3, CH3OCH3, NH2CHO, C2H5CN, and trans-HCOOH were detected. CH3OH masers in the millimeter were also reported (Slysh et al. 2002).

We pointed to four Class 0 star-forming regions, NGC 6334–29, NGC 6334–38, NGC 6334–43, and NGC 6334-I(N), in this nebula. In general, rich chemistry was found in NGC 6334–29 and NGC 6334-I(N), whereas NGC 6334–38 and NGC 6334–43 have simpler chemistry.

In NGC 6334–29, we were able to fit 17 molecules with reliable temperatures, in addition to another 25 small molecules. For O-containing molecules, we detected CH3OH, 13CH3OH, CH3OCH3, H2CCO, HCOOCH3, and CH3CHO. For N-containing molecules, we detected CH3CN, 13CH3CN, HCCCN (ground state and ${v}_{7}=1$), NH2CHO, CH2NH, and C2H5CN. For S-containing molecules, we detected H2CS, OCS, SO, SO2, and 34SO2. CH3CCH is also abundant. CH3OH has a cold component at 19 K and a warm component at 128 K. HCOOCH3 is the warmest molecule detected, with a temperature of 285 K. The S-containing molecules in NGC 6334–29 are kinematically different from the rest of the COMs. Instead of having a Gaussian line shape with FWHMs of 4–6 km s−1, S-containing molecules show broadening in their line shapes. We modeled this with a convolution of two Gaussian components, a narrow one of $\,4$ km s−1, and a wide one of 8–13.5 km s−1. The two components have similar temperatures for each molecule. SO2, however, does not show these two components. Instead, it shows a cold component at 13 K, and a warm component at 138 K, similar to CH3OH.

NGC 6334-I(N) is in an early cluster formation stage with molecular outflows but rather weak mid-IR emission (Megeath & Tieftrunk 1999). It contains seven submillimeter continuum sources in a roughly 15'' × 15'' region, with an estimated total gas mass of 39–145 ${M}_{\odot }$ (Hunter et al. 2006; Brogan et al. 2009). Walsh et al. (2010) acquired a low-resolution map toward a 5' × 5' region around NGC 6334-I(N) and found the peak fluxes of CH3OH, CH3CN, H2CS, HCCCN, OCS, and SO on the source site. Spectral lines were also reported by Brogan et al. (2009) in the 1.3 mm band from their SMA maps. Detected molecules included CH3OH, HCOOCH3, C2H3CN, CH3OCH3, H2CO, and HCCCN.

In our survey of NGC 6334-I(N), we were able to fit 15 molecules with reliable temperatures, in addition to another 21 small molecules. For O-bearing molecules, we only detected CH3OH, 13CH3OH, CH3OCH3, CH3CHO, and H2CCO. For N-bearing molecules, we only detected CH3CN, CH2NH, and HCCCN. For S-bearing molecules, we only detected H2CS, OCS, SO, and SO2. CH3CCH is also abundant. Interestingly, despite the rich chemistry in NGC 6334-I(N), all molecules are cold. The temperature range is only 10–35 K, a temperature range more typical for prestellar cores. The reason for NGC 6334-I(N) having such a rich COM inventory at such a low temperature is not clear; it is possible there is an outflow driving shocks. The FWHMs are in the range of 3–9 km s−1. We observed an asymmetric line shape, similar to NGC 6334–29, for SO, but not for other S-bearing molecules. Instead, CH3OH in NGC 6334-I(N) shows a similar asymmetric line shape, with a narrow component at 3.4 km s−1 and a wider component at 6.9 km s−1. The temperature difference between these two CH3OH components is only 4 K. The velocities range from 0.5 km s−1 redshift to 2.2 km s−1 blueshift.

In NGC 6334–38, we were able to fit nine molecules with reliable temperatures, in addition to another 26 small molecules. For O-bearing molecules, we only found CH3OH and H2CCO. For N-bearing molecules, we only found CH3CN and HCCCN. For S-bearing molecules, we only found OCS and SO. CH3CCH is also present. All molecules have temperatures within the range of 20–50 K, much colder than NGC 6334–29. The lines are 3–5 km s−1 wide in FWHM, with a consistent redshift of 1.5–2.6 km s−1 with respect to the source's vlsr.

In NGC 6334–43, we were able to fit seven molecules with reliable temperatures, in addition to 18 other small molecules. For O-bearing molecules, we only found CH3OH and H2CCO. For N-bearing molecules, we only found CH3CN and HCCCN. For S-bearing molecules, we only found OCS and SO. CH3CCH is also present. All molecules have temperatures within the range of 10–35 K, even colder than NGC 6334–38. The lines are 3–4 km s−1 wide in FWHM, with a consistent redshift of 1.9–2.6 km s−1 with respect to the source vlsr.

An asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was found in all four regions in NGC 6334. In NGC 6334–29, −38, and −43, it was modeled with two Gaussian components; in NGC 6334-I(N), it was modeled with three Gaussian components.

4.7. NGC 1333

NGC 1333 is a fragmented molecular cloud with multiple low-mass protostars embedded. Intensive studies have focused on this low-mass star-forming region (Blake et al. 1995; Gregersen et al. 1997; Bachiller et al. 1998; Stark et al. 1999; Knee & Sandell 2000; Bottinelli et al. 2004, 2007; Jørgensen et al. 2004, 2005a, 2005b, 2007; Maret et al. 2004, 2005; Wakelam et al. 2005; Parise et al. 2006; Plume et al. 2007; Choi et al. 2010; Sakai et al. 2012, 2006; Coutens et al. 2013, 2015; Lyo et al. 2014; Taquet et al. 2015; De Simone et al. 2017). Numerous COMs, such as CH3CN, CH3OH, CH2DOH, 13CH3OH, trans-HCOOH, HCOOCH3, CH3OCH3, C2H5OH, C2H5CN, and CH2OCHO, were detected in the IRAS 2 and IRAS 4A regions in this cloud (Blake et al. 1995; Bachiller et al. 1998; Bottinelli et al. 2004, 2007; Jørgensen et al. 2004, 2005a, 2005b, 2007; Maret et al. 2005; Parise et al. 2006; Sakai et al. 2012; Coutens et al. 2015; Taquet et al. 2015; De Simone et al. 2017). (CH2OH)2 was also detected in IRAS 2 (Coutens et al. 2015). The IRAS 4B core has fewer COMs reported: CH3CN, CH3OH, CH2DOH, CH3OD, HCOOCH3, and CH2OHCHO (Jørgensen et al. 2005b; Maret et al. 2005; Parise et al. 2006; Sakai et al. 2006, 2012; De Simone et al.v2017).

In our survey, we observed three positions in NGC 1333: the IRAS 2A, 4A, and 4B regions. We also conducted observations on NGC 1333 IRAS 2B, but found after the observations were complete that the incorrect source position was used. Our survey revealed a relatively simple chemical composition of these sources, which is not surprising given that the COMs previously detected in these sources using single-dish observations were detected at signal intensities of ≤10 mK, which is our noise level (see Bottinelli et al. 2007 for an example). In the three regions observed, only CO, CS, cold CH3OH, and the H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line were detected. We could only reliably fit the cold CH3OH, which is 15–25 K, and has a non-Gaussian, asymmetric line shape. The total spectral flux of 4A and 4B is higher than that of 2A. In 4A and 4B, it is likely that they contain a warmer CH3OH component, indicated by weak spectral features around the warm CH3OH ${K}_{a}=3\to 2$ R-branch near 252 GHz. However, the signal-to-noise ratio is too low to generate a reliable LTE fit. In addition, two CH3OH lines, ${8}_{-\mathrm{1,8}}\to {7}_{\mathrm{0,7}}({E}_{\mathrm{low}}\,=78\ {\rm{K}})$ at 229.76 GHz, and ${11}_{\mathrm{0,11}}\to {10}_{\mathrm{1,10}}({E}_{\mathrm{low}}=141\ {\rm{K}})$ at 250.51 GHz, present inconsistently high intensities which indicates the non-LTE environment of these sources. The H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with a single Gaussian in 2A, whereas 4A and 4B were modeled using two Gaussian components.

The weak intensity of COM lines in NCG 1333 sources is likely due to their small source sizes, which result in a heavy beam dilution effect. The warmer lines from other COMs are likely below the noise level. To study these hot corinos, telescopes with smaller beam size and higher spatial resolution are required.

4.8. GAL 75.78+0.34

GAL 75.58+0.34 is a massive star-forming region and a hypercompact H ii region (Kurtz 2005; Bisschop et al. 2007; Riffel & Lüdke 2010). CH3OH, 13CH3OH, CH3CN, CH3CCH, and SO2 were detected in a 1.3 and 0.89 mm line survey (Hatchell et al. 1998b). CH3OH and CH3CN lines were also studied by Pankonin et al. (2001) and Sánchez-Monge et al. (2013). Bisschop et al. (2007) detected H2CO, CH3OH, 13CH3OH, H2CCO, CH3OCH3, HCOOCH3, trans-HCOOH, CH3CCH, CH3CN, and NH2CHO in this source with temperatures between 40 and 190 K. We reliably fit CH3OH, CH3CCH, CH3CN, H2CCO, H2CS, HCCCN, OCS, SO, and SO2. CH3OH has a cold component at 15 K and a warm component at 160 K, and is the hottest molecule in this source. Other than the warm CH3OH, the remaining molecules are below 100 K. Comparing GAL 75.78+0.34 and GAL 45.47+0.05, the former seems to be at a later stage in the star formation process, because of the presence of warm CH3OH and CH3CN (92 K). The FWHMs are 3–6 km s−1, and the velocities are between −4 and −3 km s−1. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with two Gaussian components.

4.9. GAL 12.21–0.10

GAL 12.21–0.10 is a hot core embedded in an ultracompact H ii region (de la Fuente et al. 2010). A few line surveys in the 1.3 and 0.89 mm bands have been reported toward this source, where several COMs were detected (Hatchell et al. 1998a, 1998b; Pankonin et al. 2001). These COMs include CH3OH and its isotopologue 13CH3OH, CH3CN, and CH3CCH. Here, we present the detection of a handful of COMs at moderate temperatures, supporting the existence of a hot core. We reliably fit CH3OH, CH3OCH3, CH3CCH, CH3CN, H2CCO, H2CS, HCCCN, OCS, SO, and SO2. CH3OH has two temperatures, a cold component at 16 K and a warm component at 162 K. For other molecules, instead of clustering into cold and warm groups, their temperatures range widely from 20 to 150 K. CH3OCH3 and SO2 are warm, whereas CH3CCH and SO are cold. The FWHMs of these molecular lines are around 7–10 km s−1, and the velocities are consistent within −1 to 1 km s−1. The H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was detected and fit with one Gaussian profile.

4.10. NGC 2264

NGC 2264 is a star-forming region with multiple millimeter continuum sources (Ward-Thompson et al. 2000). We pointed at the NGC 2264 C-MM3 source, where the peak CH3OH intensity is located (Sakai et al. 2007; Saruwatari et al. 2011), and a rich chemistry was observed by combining extensive Nobeyama 45 m dish (4 and 3 mm) and ASTE 10 m dish (0.89 mm; Watanabe et al. 2015). CH3OH, 13CH3OH, CH3CCH, HC5N, CH3CN, CH3CHO, H2CCO, CH3OCH3, HCOOCH3, and NH2CHO lines were detected in this survey. A 1.3 mm SMA map, however, reveals a C-MM4 core that has much higher continuum emission and more molecules than C-MM3 (Cunningham et al. 2016). C-MM4 is about 30'' away from C-MM3.

In our survey, we detected abundant CH3OH, along with CH3CCH, H2CO, HDCO, H2CS, and SO. More complex COMs, such as CH3CHO, CH3OCH3, and HCOOCH3, were not found. A consistent velocity component of 0–1 km s−1 with 4–5 km s−1 FWHM can be fit to the above molecules. The rotational temperature ranges from 20 to 50 K.

HCOOCH3 was reported as a detection in NGC 2264 MMS3 (Sakai et al. 2007) in the 3 mm band using the Nobeyama 45 m radio telescope. The A/E lines from the ${8}_{\mathrm{1,8}}\to {7}_{\mathrm{1,7}}$ transition of HCOOCH3 were observed at a main beam temperature intensity of 60–80 mK. The signal-to-noise ratio of that detection was only 2–3. In our 1.3 mm band survey, no HCOOCH3 was detected. The peak-to-peak noise level of our survey is ∼50 mK. Considering a CSO beam ($30^{\prime\prime} $) that is larger than the Nobeyama beam ($19^{\prime\prime} $) and a higher frequency band, which in general means lines arising from warmer gas, it is likely that the HCOOCH3 lines are below our detection limit. The H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was detected and fit with one Gaussian profile.

4.11. NGC 7538

NGC 7538 is a high-mass star-forming region with massive protostars and outflows (Scoville et al. 1986; Kraus et al. 2006), the IRS 1 core being one of the protostars (Buizer & Minier 2005). H2CO, CH3OH, and CH3CN were detected in the millimeter bands (Slysh et al. 1999; Kalenskii et al. 2000a, 2000b; van der Tak et al. 2000; Wirström et al. 2011), and one C2H5OH line was reported in the centimeter (Menten et al. 1986). Ikeda et al. (2001) detected c-C2H4O and CH3CHO in this source. Bisschop et al. (2007) detected H2CO, CH3OH, 13CH3OH, H2CCO, HCOOCH3, C2H5OH, CH3CHO, trans-HCOOH, CH3CCH, and NH2CHO, and gave upper limits for CH3CN, C2H5CN, and CH3OCH3. Among these molecules, the hottest ones are CH3OH at 156 K, C2H5OH and NH2CHO, both at 164 K, and HCOOCH3 at 134 K. The rest of the molecules are below 100 K.

In our survey, we detected CH3OH, H2CCO, CH3CCH, CH3CN, H2CS, HCCCN, SO, and SO2. CH3OH has two components, a cold one at 24 K and a warm one at 257 K. CH3CN is the other warm molecule in this source, with a temperature of 244 K. Due to the low signal-to-noise ratio of the CH3CN lines, its temperature is subject to a large uncertainty. Nonetheless, the value 244 K can be treated as an upper limit. The remaining molecules are between 13 and 66 K, which agrees with the result from Bisschop et al. (2007). The FWHMs are from 3 to 4.5 km s−1. All of the lines are blueshifted within the range of 0–2.4 km s−1. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with two Gaussian components.

4.12. GAL 45.47+0.05

GAL 45.47+0.05 is an ultracompact H ii region (Cesaroni et al. 1992; Hofner et al. 1999) with outflows (Ortega et al. 2012). Remijan et al. (2004a) reported the detection of CH3CN in the 3 mm band using BIMA, and the CH3CN was further mapped by Hernández-Hernández et al. (2014). Other COMs, such as CH3OH, 13CH3OH, CH3CCH, CH3OCH3, C2H5CN, and C2H3CN, were reported in several millimeter-band line surveys (Hatchell et al. 1998b; Fontani et al. 2007).

In our 1.3 mm band survey, we fit six molecules with reliable temperatures: CH3OH, CH3CCH, H2CS, HNCO, SO, and SO2. Among them, only two are COMs. We did not detect CH3CN. The gas temperatures are low with SO2 being the hottest one at 44 K. The FWHMs are between 4 and 6 km s−1. The velocity shifts with respect to vlsr span between 0 and 2.3 km s−1. The H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was detected and fit with one Gaussian profile.

4.13. GAL 12.91–00.26

GAL 12.91–00.26 is a massive hot core. CH3OH, CH3CN, CH3OCH3, HCOOCH3, and C2H5CN lines were reported by several line surveys across the millimeter bands (Slysh et al. 1999; van der Tak et al. 2000; Purcell et al. 2006; Reiter et al. 2011; He et al. 2012). Its rich chemistry was reported by Bisschop et al. (2007) using IRAM and JCMT. In their study, they detected CH3OH, 13CH3OH, C2H5OH, CH3OCH3, HCOOCH3, H2CO, NH2CHO, and CH3CN at elevated temperatures. In addition to these hot COMs, they also detected H2CCO, trans-HCOOH, and CH3CCH at cold temperatures. C2H5CN and CH3CHO were reported as upper limits.

In our survey, we reliably fit CH3OH, 13CH3OH, CH3OCH3, CH3CCH, CH3CN, CH2NH, H2CCO, H2CO, HCCCN, HCOOCH3, OCS, SO, and SO2. We also detected lines from the ${v}_{7}=1$ vibrational excited state of HCCCN. CH3OH has two temperature components, a cold one at 14 K and a warm one at 178 K. Similarly, CH3OCH3, CH3CN, HCCCN, and HCOOCH3 are warm, with temperatures above 100 K. The rest of the molecules, especially CH3CCH and CH2NH, are cold. This result agrees with that of Bisschop et al. (2007). The FWHM of these molecules are about 3–6 km s−1. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with two Gaussian components.

4.14. GAL 19.61–0.23

GAL 19.61–0.23 is identified as a hot molecular core by multiple line surveys. In the 3 mm band, CH3OH, CH3CN, C2H5CN, HCOOCH, and trans-HCOOH were detected (Remijan et al. 2004a; Furuya et al. 2005). In the 2 and 1.3 mm bands, C2H5CN, C2H3CN, CH3OCH3, CH3CN, HCOOCH3, CH2NH and CH2OHCHO (glycoaldehyde) were reported (Fontani et al. 2007; Calcutt et al. 2014; Favre et al. 2014; Suzuki et al. 2016). Wu et al. (2009) and Furuya et al. (2011) mapped the CH3CN emission in the 890 μm band. Qin et al. (2010) compiled a list of molecules detected from a line survey in a 2 GHz window in the SMA 890 μm band. CH3OH, 13CH3OH, CH3OCH3, HCOOCH3, C2H5OH, CO(CH2OH)2, NH2CHO, and CH3CN were reported with multiple unblended spectral lines observed. CH3CHO, CH2NH, 13CN3CN, ${\mathrm{CH}}_{3}^{13}\,\mathrm{CN}$, C2H5CN, and trans-HCOOH were reported with either only one detected spectral line, or with multiple lines that are blended with other molecular lines.

Our 1.3 mm broadband line survey supports the rich chemistry discovered by previous line surveys. We reliably fit CH3OH, 13CH3OH, CH3OCH3, CH3CCH, CH3CN, C2H3CN, C2H5CN, H2CCO, H2CS, HCCCN, HCOOCH3, HNCO, OCS, SO, and SO2. However, we did not find a reliable detection of CH3CHO, CH2NH, 13CN3CN, ${\mathrm{CH}}_{3}^{13}\,\mathrm{CN}$, C2H5OH, and NH2CHO. CO(CH2OH)2 was not included in the list of molecules that were analyzed in this source. CH3OH has two components, a cold one at 14 K and a warm one at 177 K. This result is different from Qin et al. (2010), who fit a single temperature of 151 K in their Boltzmann diagram despite the fact that a cold CH3OH line has a significantly higher $\mathrm{ln}({N}_{{\rm{u}}})/{g}_{{\rm{u}}}$ value than the rest of the warm CH3OH lines. For most of the other detected molecules, our analysis shows a lower rotational temperature with smaller uncertainties. This is because the global analysis of GOBASIC is less likely to overpredict the spectral intensity from a single molecule in a blended line. Still, the trend is similar to that shown by Qin et al. (2010). N-bearing molecules, including CH3CN, C2H5CN, and HNCO, are found to be at higher temperatures. The FWHM of these lines spans a wide range, from 4 to 14 km ${{\rm{s}}}^{-1}$. COMs are generally found to have FWHMs of 7–9 km ${{\rm{s}}}^{-1}$, with the exception of 13CH3OH, which is 4.6 km ${{\rm{s}}}^{-1}$. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with two Gaussian components.

4.15. G24.33+00.11 MM1

G24.33+0.14 is listed as an outflow candidate in the Extended Green Objects catalog (Cyganowski et al. 2008). It is a high-mass young stellar object that has not yet formed a compact H ii region. This stellar core was mapped in the 890 μm band by the SMA using molecular lines from CH3CN, 13CH3CN, CH3OH, CH3OCH3, SO, and 34SO, and a few other COMs such as C2H5OH and NH2CHO were tentatively detected (Rathborne et al. 2011). We pointed to the stellar core G24.33+00.11 MM1 at $\alpha =18$h35m08fs14, $\delta =-07^\circ 35$'01''1, which is offset from the nominal source position of α = 18h35m08fs1, δ = −07°35'04''. The flux of G24.33+00.11 MM1 is weaker compared to that of other hot cores we observed. Nevertheless, it is one of the sources displaying the richest chemistry. In addition to the commonly detected CH3OH, CH3OCH3, CH3CCH, CH3CN, H2CCO, H2CS, HCCCN, HCOOCH3, OCS, and SO, we reliably fit C2H5CN, C2H5OH, CH2NH, 13CH3OH, and CH3CN (${v}_{8}=1$). Interestingly, we did not detect SO2. The hottest molecules are CH3CN (302 K), warm CH3OH (233 K), CH3OCH3 (200 K), and HCOOCH3 (124 K). CH3OH also has a cold component at 14 K. Because we treated vibrational states as individual molecules, the CH3CN (${v}_{8}=1$) temperature is 150 K, lower than the vibrational ground state. The rest of the COMs are below 100 K. The FWHMs of the lines are 3–5 km s−1. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with two Gaussian components.

4.16. GAL 24.78+0.08

GAL 24.78+0.08 is an embedded cluster with high-mass protostars (Furuya et al. 2002). It possesses a rotation toroid (Beltrán et al. 2005) and gas infall accompanied by outflows (Cesaroni et al. 2003; Beltrán et al. 2006, 2011). In these observations, CH3OH, H2CCO, CH3OCH3, C2H5CN, C2H3CN, C2H5OH, and HCOOCH3 lines were reported, and CH3CN and ${\mathrm{CH}}_{3}^{13}\,\mathrm{CN}$ were used as kinematic tracers. Molecular lines from CH3OH, CH3OD, CH3CN, HCOOCH3, and C2H5CN were also reported in other surveys in the centimeter and millimeter bands (Purcell et al. 2006; Vig et al. 2008; Ratajczak et al. 2011; Calcutt et al. 2014; Favre et al. 2014). Bisschop et al. (2007) reported the detection of H2CO, CH3OH, C2H5OH, NH2CHO, CH3CN, C2H5CN, HCOOCH3, CH3OCH3, H2CCO, CH3CCH, and trans-HCOOH. The rotational temperature of these molecules ranges from 39 to 261 K. Specifically, glycolaldehyde (CH2OHCHO) was reported as a detection in this source (Beltrán et al. 2005; Calcutt et al. 2014).

In our analysis, we reliably fit CH3OH, 13CH3OH, CH3OCH3, CH2NH, CH3CCH, CH3CN, H2CS, HCCCN, HCOOCH3, OCS, SO, and SO2. CH3OH has a cold component at 17 K and a warm component at 141 K. We also detected the HCCCN ${v}_{7}=1$ vibrational excited state. The temperatures of CH3OH, CH3OCH3, HCOOCH3, HCCCN (${v}_{7}=1$), and CH3CN are above 100 K. CH3CN is the hottest molecule, with a temperature of 204 K. This temperature of CH3CN is slightly lower than the 261 K reported by Bisschop et al. (2007), but higher than the 132 K reported by Beltrán et al. (2005), who combined the lines from the normal species, the 13C isotopologue, and the ${v}_{8}=1$ vibrational state in their analysis. The FWHMs of the lines are 5–9 km s−1. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with two Gaussian components.

4.17. DR21(OH)

DR21(OH), located in the Cygnus X complex, is a giant, fragmented molecular cloud with multiple emission cores (Minh et al. 2012; Girart et al. 2013). Small molecules such as CS, OCS, H2CO, HDCO, H2CS, and c-C3H2 (cyclopropenylidene) have been observed in the centimeter band toward this cloud (Angerhofer et al. 1978; Goldsmith & Linke 1981; Madden et al. 1989; Chandler et al. 1993; Richardson et al. 1994; Mehringer et al. 1995; Wootten & Mangum 2009). Numerous COMs, such as CH3OH, CH3CN, CH3CHO, c-C2H4O, CH3CCH, CH3OCH3, C2H5CN, and CH2NH, were detected in multiple line surveys in the 3, 2, 1.3, and 0.89 mm bands (Kalenskii et al. 1997, 2000a, 2000b; Slysh et al. 1999, 2002; van der Tak et al. 2000; Ikeda et al. 2001; Alakoz et al. 2002; Charnley 2004; Kalenskii & Johansson 2010a; Wirström et al. 2011; Zapata et al. 2012; Girart et al. 2013; Suzuki et al. 2016). Kalenskii & Johansson (2010a) found low to moderate rotational temperatures, 9–56 K, for most molecules detected. The evidence for the existence of an embedded hot core is the hot rotational temperature of CH3OH at 252 K and SO2 at 186 K.

Our results agree well with those of Kalenskii & Johansson (2010a) and Minh et al. (2012) in the type of molecules detected. In brief, two methanol temperature components were observed, indicated by the coexistence of the cold methanol branch at 242 GHz and the warm methanol branch between 250 and 253 GHz. The temperature of the warm methanol, however, is 93 K and significantly lower than the value reported by Kalenskii & Johansson (2010a). It is possible that heavy beam dilution diminished the weaker lines of the methanol branch into the noise, which led to our lower estimation of the rotational temperature. Other detected COMs include CH3CN, CH3CCH, CH3CHO, CH3OCH3, H2CCO, HCCCN, and HCOOCH3. No rigorous detections of trans-HCOOH, NH2CHO, and C2H3CN were found, despite being reported by Kalenskii & Johansson (2010a). The two tentative detections reported by Kalenskii & Johansson (2010a), CH2NH and C2H5OH, were also not found. The LTE temperatures of the COMs we detected are all below 150 K, with CH3CHO being the warmest one at 139 K. The FWHMs of these molecular lines are consistent within the range of 4–7 km s−1. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with three Gaussian components.

Multiple methanol masers have been observed in the DR21 complex (Plambeck & Menten 1990; Araya et al. 2009; Fish et al. 2011). Similarly, we also observed non-LTE methanol emission at 229.76 GHz (${8}_{-\mathrm{1,8}}\to {7}_{\mathrm{0,7}}$) and 250.5 GHz (${11}_{\mathrm{0,11}}\to {10}_{\mathrm{1,10}},{A}^{+}$), as the total flux of these two methanol lines considerably exceeds the LTE emission at 93 K. These two lines may be associated with the same physical environment that drives the methanol masers at lower frequencies. The rest of the molecular lines can be modeled with the LTE model.

4.18. W75N

W75N is an ultracompact H ii region associated with high-mass star formation (Persi et al. 2003). Watson et al. (2002) mapped CH3CN, and more recently, three compact cores, MM1, MM2, and MM3, were found in the 1.3 mm continuum map (Minh et al. 2010). The largest core, MM1, is about 3'' × 6'', and all three cores are covered by the CSO beam. The MM1 core was further resolved into two emission peaks, MM1a and MM1b, which present different chemical compositions (Minh et al. 2010). SO, SO2, HCCCN, NH2CHO, and HCOOH were detected in MM1a, whereas MM1b exhibits more COMs such as CH3OH, HCOOCH3, and C2H5OH. Other studies also reported COMs such as CH3OH, CH3CN, C2H5OH, HCOOCH3, and trans-HCOOH (Menten et al. 1986; Pankonin et al. 2001; Minier & Booth 2002; Remijan et al. 2004a; Reiter et al. 2011).

We fitted CH3OH, H2CCO, CH3CCH, CH3CN, HCCCN, H2CS, OCS, SO, and SO2. Two CH3OH temperature components are found, a cold one at 19 K and a warm one at 207 K. CH3CN and SO2 are above 100 K, and the rest of the molecules are below 100 K. The widths of the lines show two categories: CH3CCH, cold CH3OH, H2CS, and SO are narrow, between 4.2 and 5.5 km s−1; CH3CN, warm CH3OH, H2CCO, HCCCN, OCS, and SO2 are wide, between 7.4 and 10.2 km s−1. The velocities are consistent within 1 km s−1. No detection was found for HCOOCH3, C2H5OH, NH2CHO, and HCOOH, likely due to the heavy beam dilution effect. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with two Gaussian components.

4.19. W51

W51 has two molecular cores, e1 and e2, and two extended continuum emission regions, IRS 1 and IRS 2 (Ho & Young 1996). We pointed to the center of e1, but also covered the entire e2 core with our CSO beam. Edges from IRS 1 may also be included in the beam. A spectral line survey of the e1 and e2 cores in the 3 mm band was reported by Kalenskii & Johansson (2010b), who detected COMs such as CH3OH, CH3CN, CH3OCHO, C2H5OH, CH3OCH3, CH3CH2CN, CH3COCH3, and C2H5OOCH. They found that O-bearing molecules dominate the e2 core whereas N-bearing molecules dominate the e1 core. Other observations across the millimeter bands have detected multiple COMs, including CH3OH, CH3CN, CH3CCH, HCOOCH3 (g.s. and vt = 1), HCOO13CH3, C2H5OH, CH3CHO, c-C2H4O, trans-HCOOH, CH3COOH, C2H5CN (g.s., vt = 1, and vb = 1), C2H3CN, and, tentatively, (CH2OH)2 (Dickens et al. 1997; Slysh et al. 1999; Kalenskii et al. 2000a, 2000b, 2002; Ikeda et al. 2001; Alakoz et al. 2002; Minier & Booth 2002; Remijan et al. 2002, 2008; Demyk et al. 2008; Jordan et al. 2008; Wirström et al. 2011; Favre et al. 2014; Hernández-Hernández et al. 2014; Lykke et al. 2015; Suzuki et al. 2016). The rotational temperatures of CH3OH, CH3CN, C2H5OH, and C2H3CN were derived to be above 150 K, whereas those of CH3CHO and HCOOCH3 were below 100 K (Ikeda et al. 2001; Remijan et al. 2002, 2004b).

We detected CH3OH, 13CH3OH, CH3OCH3, HCOOCH3, H2CCO, CH3CCH, CH3CN, CH2NH, HCCCN, H2CS, OCS, SO, and SO2. Due to limited spectral sampling, the noise level above 259.5 GHz is noticeably higher than the rest of the spectrum. To avoid the unrealistic temperature simulations of large COMs from overfitting the noise in this spectral region, we truncated the molecular catalogs of CH3OCH3, HCOOCH3, and H2CCO at 259.5 GHz. Visual examinations confirmed that fits better matched the actual observed lines after this adjustment. The temperatures of the molecules in this source span a wide range from 14 to 472 K. The 472 K HCOOCH3, in contrast to the result of Ikeda et al. (2001), appears to be much hotter than the rest of the molecules. Warm CH3OH and CH3CN are above 200 K, and CH3OCH3, 13CH3OH, and SO2 are above 100 K. The CH3CN rotational temperature, 226 K, is slightly higher than that found by Remijan et al. (2002), which is not surprising because the 3 and 2 mm lines of CH3CN showed a larger intensity scattering in the analysis of Remijan et al. (2002), likely due to the inconsistency of beam sizes. The cold CH3OH component is at 23 K. The FWHMs of the molecules range from 7.5 to 11.4 km s−1, and the velocities range from 1 to 4 km s−1, redshifted. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with two Gaussian components.

4.20. W3(H2O)

W3 is a massive star-forming region with a hot core, W3(H2O), indicated by the H2O maser (Alcolea et al. 1993; Wilner et al. 1999), and an ultracompact H ii region, W3(OH), which is indicated by the OH maser (Wyrowski et al. 1997). The OH maser is located 6'' to the west of W3(H2O) (Turner & Welch 1984; Wilner et al. 1995). CH3OH masers are also present in this region (Slysh et al. 1999). Our CSO beam covered both sources, but was centered on W3(H2O).

Both cores have a rich inventory of COMs, revealed by extensive line surveys across the millimeter bands (Helmich et al. 1994; Helmich & van Dishoeck 1997; Kalenskii et al. 1997, 2000a; Wyrowski et al. 1999a; van der Tak et al. 2000; Ikeda et al. 2001; Alakoz et al. 2002; Charnley 2004; Sutton et al. 2004; Kim et al. 2006; Bisschop et al. 2007; Fontani et al. 2007; Ratajczak et al. 2011; Reiter et al. 2011; Li et al. 2012; Gerner et al. 2014; Hernández-Hernández et al. 2014; Ligterink et al. 2015; Qin et al. 2015). Detected COMs include CH3OH, 13CH3OH, C2H5OH, NH2CHO, CH3CN, ${\mathrm{CH}}_{3}^{13}\,\mathrm{CN}$, C2H5CN, C2H3CN, HCOOCH3, CH3OCH3, H2CCO, CH3CHO, CH3CCH, and trans-HCOOH, as well as ground state and vibrationally excited HCCCN. The rotational temperature of most COMs ranges between 100 and 250 K (Helmich & van Dishoeck 1997; Fontani et al. 2007; Qin et al. 2015), and CH3OH has two temperature components (Qin et al. 2015). Cold CH3OH, CH3CCH, C2H5OH, and CH3CHO lines were also observed at low frequency (3 mm and centimeter; Menten et al. 1986; Alakoz et al. 2002; Charnley 2004; Kim et al. 2006).

CH3CN lines in the 1.3 mm band have been imaged to reveal the binary rotational structure of W3(H2O) (Chen et al. 2006). The CH3CN lines in W3(H2O) have wide velocity range from −55 to −43 km s−1 (i.e., −8 to 4 km s−1 with respect to our vlsr), whereas the CH3CN lines in W3(OH) are more concentrated in the velocity channels from −48 to −43 km s−1 (−1 to 4 km s−1 with respect to our vlsr).

In our survey, we fit CH3OH, 13CH3OH, CH3OCH3, H2CCO, HCOOCH3, CH3CCH, CH3CN, C2H5CN, HCCCN (both ground state and ${v}_{7}=1$), H2CS, OCS, SO, and SO2. The line shapes of the warm O-bearing molecules, including CH3OH, 13CH3OH, CH3OCH3, and HCOOCH3, are asymmetric, and can be modeled with two velocity components. For each of these molecules, one component is redshifted from 1.7 to 2.5 km s−1 with a narrower FWHM between 4 and 5.8 km s−1, while another component is blueshifted from 0 to 3.2 km s−1 with a wider FWHM between 5.4 and 9.7 km s−1. Comparing this kinematic information to Chen et al. (2006), it is likely that the wide, blueshifted components arise from the W3(H2O) region, and the redshifted, narrow components arise from the W3(OH) region. We found that the narrow components of CH3OH, 13CH3OH, CH3OCH3, and HCOOCH3 are all hot, with rotational temperatures of 100–162 K. CH3OH and CH3OCH3 from the W3(H2O) region, on the other hand, have both hot (150–156 K) and cold (17–23 K) components. The SO2 line profile is more complicated; it can be modeled with three components, each of which has a distinct rotational temperature (17, 70, and 148 K), velocity, and line width. It is difficult to attribute these components definitively to either the W3(H2O) or the W3(OH) regions. These temperatures are comparable to the results from Helmich et al. (1994) and Bisschop et al. (2007). The temperatures of N-bearing molecules are slightly colder than the warm O-bearing molecules, ranging between 83 and 152 K. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with three Gaussian components.

4.21. GAL 034.3+00.2

GAL 034.3+00.2 is a high-mass, chemically rich star-forming region with ultracompact H ii regions (Benson & Johnston 1984) and embedded hot cores (Garay & Rodriguez 1990; Carral & Welch 1992). Numerous small molecules are present in this region (Macdonald et al. 1996), and multiple COMs have also been detected across the millimeter bands. The detected COMs include common interstellar O-bearing and N-bearing molecules such as CH3OH, CH3CCH, CH3CN, C2H3CN, C2H5CN, NH2CHO, C2H5OH, CH3OCH3, HCOOCH3, HCOOCH3, trans-HCOOH, CH3CHO, and CH3COOH (MacDonald et al. 1995; Millar et al. 1995; Macdonald et al. 1996; Mehringer & Snyder 1996; Dickens et al. 1997; Hatchell et al. 1998b; Nummelin et al. 1998b; Slysh et al. 1999; Watt & Mundy 1999; Kalenskii et al. 2000a, 2000b; Kim et al. 2000, 2001; Ikeda et al. 2001; Alakoz et al. 2002; Remijan et al. 2003; Charnley 2004; Mookerjea et al. 2007; Reiter et al. 2011; Rosero et al. 2013; Lykke et al. 2015; Fu & Lin 2016; Suzuki et al. 2016). Istopologues such as CH2DCN and 13CH3OH are also observed (Gerin et al. 1992; Macdonald et al. 1996; Hatchell et al. 1998b).

GAL 034.3+00.2, is an incredibly rich source in chemistry. We were able to derive reliable rotational temperatures and densities for 22 molecules, in addition to the detection of spectral lines from 22 other small molecules and their isotopologues. In addition to the commonly detected CH3OH, CH3OCH3, CH3CCH, CH3CN, H2CCO, H2CS, HCCCN, HCOOCH3, OCS, SO, and SO2, we found C2H3CN, C2H5CN, C2H5OH, CH3CHO, CH3COCH3, NH2CHO, CH2NH, HCCCN (${v}_{7}=1$), and trans-HCOOH. 13CH3OH and 13CH3CN are both detected. We did not analyze CH3COOH because the available molecular catalog was incomplete. CH3OH was fitted to two components, a cold one at 16 K and a warm one at 190 K. On the other hand, one component at 155 K was sufficient to describe 13CH3OH. Other COMs have temperatures spanning 25 to 160 K. Surprisingly, CH3OH, instead of CH3CN, is the hottest molecule in GAL 034.3+00.2. The FWHMs are between 4 and 8 km s−1. The velocities are consistent in the range of −1 and 1 km s−1. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with three Gaussian components.

4.22. GAL 31.41+0.31

GAL 31.41+0.31 is a high-mass star-forming region (Cesaroni et al. 1994) with embedded hot cores (Osorio et al. 2009; Cesaroni et al. 2011) and outflows (Maxia et al. 2001; Beltrán et al. 2004; Araya et al. 2008; Moscadelli et al. 2013). CH3CN was used as a tracer for the structure and kinematics and extensively mapped in the millimeter (Olmi et al. 1995; Walmsley et al. 1995; Olmi et al. 1996a, 1996b; Pankonin et al. 2001; Beltrán et al. 2004, 2005; Araya et al. 2005; Purcell et al. 2006; Hernández-Hernández et al. 2014). Numerous COMs have been detected in this region via multiple millimeter and submillimeter line surveys. The detected COMs include CH3OH, 13CH3OH, CH3CN, CH3CCH, CH3OCH3, HCOOCH3, C2H5OH, CH3CHO, c-C2H4O, trans-HCOOH, C2H3CN, C2H5CN, and CH3COCH3 (Hatchell et al. 1998b; Nummelin et al. 1998b; Ikeda et al. 2001; Fontani et al. 2007; Araya et al. 2008; Jordan et al. 2008; Beltrán et al. 2009; Reiter et al. 2011; Isokoski et al. 2013; Calcutt et al. 2014; Ligterink et al. 2015; Suzuki et al. 2016). It is one of the sources where CH2OHCHO (glycolaldehyde), a sign of complex chemistry, was reported (Beltrán et al. 2009; Beltràn et al. 2010; Calcutt et al. 2014).

GAL 31.41+0.31 is one of the chemically richest sources in our list. We were able to derive reliable rotational temperature and density for 22 molecules, in addition to the detection of spectral lines from 19 other small molecules and their isotopologues. In addition to the commonly detected CH3OH, CH3OCH3, CH3CCH, CH3CN, H2CCO, H2CS, HCCCN, HCOOCH3, OCS, SO, and SO2, we found C2H3CN, C2H5CN, C2H5OH, CH3CHO, CH3COCH3, NH2CHO, CH2NH, and CH3SH. 13C isotopologues of CH3OH and CH3CN are also detected. The 13CH3OH:CH3OH ratio is 0.16, whereas the 13CH3CN:CH3CN ratio is 0.05. The temperatures of the COMs span 20 to 320 K. Both 13CH3OH and CH3OH have two components, a cold one at ∼40 K and a warm one above 200 K. The temperature of 13CH3CN is lower than that of CH3CN, with the former at 105 K and the latter at 321 K. Additionally, CH3CN (${v}_{8}=1$) is detected at 180 K. The FWHMs of the lines are between 5 and 10 km s−1. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with three Gaussian components.

4.23. GAL 10.47+00.03

GAL 10.47+00.03 is an ultracompact H ii region with an embedded hot core (Cesaroni et al. 1992; Pascucci et al. 2004). It is surprising that GAL 10.47+00.03 has chemical complexity comparable to that of Orion-KL. Hot cores have been identified in this UC H ii region (Pascucci et al. 2004), and methyl cyanide (CH3CN) (Olmi et al. 1993, 1995,1996a, 1996b; Araya et al. 2005; Hernández-Hernández et al. 2014) has been used to map this source. The molecular line surveys were conducted by a few groups (Hatchell et al. 1998b; Nummelin et al. 1998b; Ikeda et al. 2001; Pankonin et al. 2001; Hatchell & van der Tak 2003; Purcell et al. 2006; Fontani et al. 2007; Jordan et al. 2008; Rolffs et al. 2011). Detections of several COMs, including O-bearing molecules CH3OH, 13CH3OH, CH3CCH, CH3CHO, CH3OCH3, HCOOCH3, CH3CHO, CH3COCH3, trans-HCOOH, C2H5CN, C2H3CN, and NH2CHO were reported in these surveys. Sulfur chemistry, probed by H2CS, SO, and SO2 emissions, was studied by Hatchell et al. (1998a). Suzuki et al. (2016) detected CH2NH in this source as well.

In our survey, we report the rigorous detection of 13 COMs, in addition to 29 smaller molecules. For O-bearing molecules, we detected CH3OH with a cold and a warm component, as well as 13CH3OH, CH3OCH3, H2CCO, HCOOCH3, trans-HCOOH, C2H5OH, CH3CHO, and CH3COCH3. For N-bearing molecules, we detected CH3CN (both ground state and ${v}_{8}=1$), 13CH3CN, C2H3CN, C2H5CN, CH2NH, HCCCN, and NH2CHO. Because of heavy line blending, the FWHMs of C2H5OH, CH3CHO, and CH3COCH3 were fixed to 10 km s−1 for the fit to converge. Our result of detected COMs agrees well with previous line surveys. The cold CH3OH component is at 21 K, while the warm component is at 258 K. All CH3OH emission features could be modeled well with LTE. HCOOCH3 is slightly colder than the warm CH3OH at 233 K. Other O-bearing COMs are slightly colder than the warm CH3OH, with 13CH3OH at 179 K, and CH3OCH3, H2CCO, and trans-HCOOH all between 130 and 150 K, agreeing with the result for CH3OCH3 from Ikeda et al. (2001). The temperatures of N-bearing COMs span a wide range. Methyl cyanide is one of the hottest: the main species is at 668 K, the ${v}_{8}=1$ state is at 733 K, and the 13C isotopologue is at 298 K. N-bearing COMs also suffer from a large temperature uncertainty due to heavy line blending. The methyl cyanide temperature is found to be much higher than the literature value of 148 K (Olmi et al. 1996b). In our survey, we observed emission features from methyl cyanide lines up to K = 12 in the $J=12\to 11,J=13\to 12$, and $J=14\to 13$ ladders. The higher K lines observed as compared to Olmi et al. (1996b) cause the higher rotational temperature of methyl cyanide in our analysis. CH2NH is at 654 K, but with a large temperature uncertainty because of line blending. C2H3CN is much hotter than C2H5CN, the former being at 507 K, and the latter being at 177 K. NH2CHO is at 248 K. For S-bearing molecules, we detected H2CS, OCS, SO, and SO2. They are significantly colder than the N-bearing species, with the hottest SO2 at 162 K. CH3CCH is also present in large abundance at a moderate temperature of 59 K.

The FWHMs of these molecular lines are relatively wide, spanning 7 to 12 km s−1. The line shape is symmetric, agreeing with Olmi et al. (1993, 1996b). The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with two Gaussian components.

4.24. Orion-KL

The molecular lines and chemistry in Orion-KL have been extensively studied by numerous groups (Sutton et al. 1985, 1995; Menten et al. 1986; Blake et al. 1987; Turner 1989, 1991; Mangum et al. 1990; Minh et al. 1993; Groesbeck 1995; Dickens et al. 1997; Schilke et al. 1997, 2001; Slysh et al. 1999; Ikeda et al. 2001; Lee et al. 2001; Liu et al. 2001, 2002; Alakoz et al. 2002; Minier & Booth 2002; White et al. 2003; Beuther et al. 2004, 2006; Charnley 2004; Comito et al. 2005; Friedel et al. 2005; Persson et al. 2007; Friedel & Snyder 2008; Remijan et al. 2008; Goddi et al. 2009; Crockett et al. 2010; Gupta et al. 2010; Margulès et al. 2010; Favre et al. 2011a, 2011b; Zapata et al. 2011; Friedel & Widicus Weaver 2012; Brouillet et al. 2013; Crockett et al. 2014; Favre et al. 2014; Frayer et al. 2015; Morris et al. 2016; Suzuki et al. 2016; Tahani et al. 2016). We observed this source with a pointing centered on the Orion Hot Core as described by Blake et al. (1986). Due to the extreme complexity of the kinematics of Orion-KL, the line shapes are asymmetric and non-Gaussian. Our global fit for Orion-KL required more manual configuration for the fitting parameter space, and the overall fit residual was not as clean as the other sources we analyzed. For several molecules such as HCCCN, HCCCN (${v}_{7}=1$), and CH2NH, Gaussian components with fixed FWHM and $\delta v$ were necessary to constrain the overall line shape of the transitions. There were also numerous lines that could not be modeled or were only partially modeled. Nonetheless, as a benchmark source, our detection of many COMs, including CH3OH, 13CH3OH, CH3OCH3, HCOOCH3, CH3COCH3, trans-HCOOH, C2H3CN, C2H5CN, and CH3CN, agrees with the previous literature. The asymmetric H2CO ${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$ line was modeled with three Gaussian components.

4.25. Sgr B2(N-LMH)

Sgr B2(N-LMH) is by far the most chemically complex source observed. A large portion of reported COMs were first detected in Sgr B2(N), which is known to harbor the largest number of COMs of any sightline investigated thus far. There are numerous literature studies of both comprehensive broadband line surveys and detailed studies for individual molecules for Sgr B2(N): to list a few, for example, Goldsmith & Linke (1981), Miao & Snyder (1997), Mehringer et al. (1997), Nummelin et al. (1998a, 2000), Cernicharo et al. (1999), Hollis et al. (2000, 2006), Liu et al. (2001), Snyder et al. (2002), Remijan et al. (2002), Halfen et al. (2006, 2011, 2013), Apponi et al. (2006), Requena-Torres et al. (2006), Belloche et al. (2007, 2008, 2009, 2013), Leigh et al. (2007), Halfen & Ziurys (2008), Remijan et al. (2008), Martín-Pintado et al. (2001), Friedel et al. (2004), Favre et al. (2014), and Neill et al. (2014). Because of the extreme spectral complexity and numerous previous studies that have targeted this source, we only provide the deconvolved spectrum of Sgr B2 (N-LMH) without performing a global spectral analysis.

5. Discussion

Given that this set of sources was targeted because they are known to be rich in COMs, it does not come as a surprise that many of these sources have extremely rich chemistry. This chemical richness means that trends that offer insight into chemical pathways that have long been assumed for star-forming regions can be discerned. An overview of the results in graphical form is given in Figure 3. Here we plot the column density for a given molecule as a circle. The area of each circle is scaled to the magnitude of the column density. The color of each circle corresponds to the temperature, the scale for which is shown on the right. Sources are sorted on the horizontal axis by type, and within each type the chemical complexity of the sources is arranged to increase from left to right. The molecules are sorted on the vertical axis, grouped into the O-bearing, N-bearing, and S-bearing species. Within each group, molecules on the top are more frequently detected in the sources. As such, trends are easily discernible in the data set. For molecules with multiple temperature and velocity components, we plot concentric circles with each ring representing one component. The column density of molecules in Orion-KL was divided by 10 before plotting.

Figure 3.

Figure 3. Visualization of Table 3: column density and LTE temperature. Sources are on the horizontal axis, and molecules are on the vertical axis. The areas of the circles are proportional to the column density of the molecules. The circles for Orion-KL are decreased by a factor of 10 in area to fit the overall scale of the plot. For molecules with multiple temperature components, concentric circles are drawn, where each ring represents one component. The color of the circles indicates the LTE temperature of the molecules, with the scale given to the right. The full-scale plot is presented in the top panel, and the zoomed-in region, highlighted by the orange frame, is presented in the bottom panel.

Standard image High-resolution image

Here we will not attempt a detailed discussion of each molecule detected in each source and the corresponding implications, but rather will save that level of detail for subsequent papers that examine a particular chemical or physical subset from this data set. Nonetheless, we can immediately find general trends in these results that offer glimpses into the power of such comprehensive observational studies. The most striking features of these results as gleaned from Figure 3 are as follows:

  • 1.  
    Although most sources are chemically complex, this is not universally true. Among different source types, hot cores present the most complex chemistry, indicated by the large number of COMs observed in these sources. Protostars present some level of complexity. Hot corinos are quiet in this survey, likely due to their small source sizes and therefore severe beam dilution effect. Some observed complexity is expected, given that these sources were chosen because COMs had already been detected in previous observations. But the level of complexity in some of these sources is comparable to that of Orion-KL. The most striking case is that of GAL 10.47+00.03, which displays a remarkable array of COMs.
  • 2.  
    Methanol, CH3CCH, H2CS, OCS, SO, and SO2 are nearly ubiquitous in this sample set. Methanol, SO, and SO2 are known to be indicators of grain mantle disruption (Charnley 1997; Hatchell et al. 1998a; Buckle & Fuller 2003), hence it is possible that these other ubiquitous molecules are also indicators of this process. It is particularly noteworthy that all of the major sulfur-bearing species are on this list.
  • 3.  
    CH3OH is ubiquitous, as is expected. For most hot cores, two CH3OH temperature components can usually be found: a cold one of 10–30 K and a warm one ranging from 100 to 300 K. The temperature of the warm CH3OH appears to be a variable dependent on the size and evolutionary stage of the central star. The difference between the temperatures for cold and warm methanol is significant. Since methanol is abundant, its line opacity may introduce systematic error in the rotational temperature values determined under the optically thin assumption. Nevertheless, the line shapes for the methanol lines we observed are Gaussian, indicating that the lines are not completely opaque. Based on our beam-averaged results, the cold methanol component usually has a maximum optical depth (τ) value of 0.1–0.5, while the warm component has a τ on the order of 0.01. If beam dilution was considered, the τ value would be higher, which could make the strongest cold methanol lines marginally optically thick. But overall this will not change the qualitative result that two methanol temperatures coexist in our sources. The observation of two methanol temperature components is also supported by results in the literature, i.e., in Orion-KL (Blake et al. 1986) and G31.41 (Isokoski et al. 2013). 13CH3OH was only detected in massive star-forming regions with a high CH3OH abundance.
  • 4.  
    CH3CCH is nearly as ubiquitous and abundant as CH3OH, and it is always cold compared to other COMs that were detected. This may indicate that the emission of CH3CCH arises from the extended envelope around the hot cores. CH3CCH is highly abundant, on the same order of magnitude as CH3OH. CH3CCH, as a small carbon chain molecule, is thought to be a tracer for the cold chemistry in dense clouds (Kuiper et al. 1984). A "warm carbon chain chemistry" (WCCC; Sakai et al. 2008) was proposed to explain the detection of carbon chains, including CH3CCH, in several envelopes of low-mass star-forming regions with a moderate temperature around 30 K (Sakai et al. 2008, 2009a, 2009b). The prevalence of CH3CCH in all of the hot cores may be support for this WCCC, where regeneration mechanisms of the carbon chain molecules during the warm-up phase may be important (Hassel et al. 2011). Further studies of the formation mechanisms of CH3CCH around hot cores are warranted.
  • 5.  
    The S-bearing molecules, H2CS, OCS, SO, and SO2 are nearly ubiquitous in this sample set, which is to be expected since they are structurally simpler than COMs, and therefore have stronger spectral lines. S-bearing molecules are proposed to be indicators of grain mantle disruption and material injection into the gas phase during the star formation process, and therefore can be used as a chemical clock to trace the evolutionary stage of the source (Charnley 1997; Hatchell et al. 1998a; Buckle & Fuller 2003). However, Wakelam et al. (2004) argued that this chemical clock approach is not straightforward.In our surveys, SO is always present. SO and OCS are definitively detected in Orion-KL, but not included in Figure 3 because the narrower spectral coverage for Orion-KL provides insufficient information to perform LTE analysis on these molecules. H2CS correlates well with SO across our sources, while OCS and SO2 do not always coexist. In NGC 1333 IRAS 4A, NGC 6334–38, NGC 6334–43, and G24.33+00.11 MM1, we have only detected OCS but not SO2. Since SO2 is generally less abundant than OCS and SO in other sources, the non-detection of SO2 in NGC 1333 IRAS 2A may simply be caused by its low line intensity, which is below our detection limit. NGC 6334–38, NGC 6334–43, and G24.33+00.11 MM1, however, display significant OCS abundances. In HH 80–81 and NGC 7538, we only detect SO2 and not OCS. For sources where all four S-bearing molecules are detected, the abundance ratios vary dramatically from source to source. For example, GAL 10.47+00.03 and W51, compared to other hot cores, have significantly more OCS than the rest of the S-bearing molecules. Orion-KL is the anomaly on the other extreme: it has an extensive amount of SO2, which is as abundant as CH3OH. In general, we conclude that the abundance of S-bearing small molecules cannot be explained by a simple correlation with the evolutionary stage of the star-forming region. Our observations agree with the argument addressed by Wakelam et al. (2004).
  • 6.  
    Nearly all molecules detected that have a CN group are found at elevated temperatures relative to O-bearing COMs. Despite the prevalence of these species in these sources, they are present at much lower abundances than their O-bearing counterparts. The variation of the CH3CN temperature across sources is also large. Unlike CH3CCH, which is always cold, or CH3OCH3 and HCOOCH3, which are always warm but only present in chemically evolved sources, CH3CN is detected in nearly every source, and its temperature correlates with the source type. C2H3CN and C2H5CN, two "interstellar weeds" thought to be ubiquitous in the ISM, are only detected in a handful of massive star-forming regions with the richest chemistry and most elevated temperatures (Crockett et al. 2014). It could be that the cyanides larger than CH3CN are both warm and compact, and therefore beam diluted beyond the point of detection here. The abundance of C2H5CN is low relative to CH3CN, despite their nearly equal abundance in Orion-KL.
  • 7.  
    There may be a chemical link between dimethyl ether (CH3OCH3) and methyl formate (HCOOCH3) through grain-surface chemistry (see Garrod et al. 2008 for a detailed discussion). However, these two molecules do not display obvious trends in abundance or temperature. In general, CH3OCH3 is more abundant and colder than HCOOCH3, but the ratio between the CH3OCH3 and HCOOCH3 column densities is not consistent throughout our sample set. This result is unexpected, given that they are both presumed to form from grain-surface processes involving the CH3O radical. The model by Garrod & Widicus Weaver (2013) showed that the abundances of CH3OCH3 and HCOOCH3 peak at different temperatures during the warm-up phase, and their peak abundance is largely affected by the warm-up speed. In a fast warm-up period, HCOOCH3 has a higher peak abundance than that of CH3OCH3. In contrast, in a medium or slow warm-up period, CH3OCH3 has a higher peak abundance than that of HCOOCH3. In addition, beam dilution can contribute to this inconsistency if CH3OCH3 and HCOOCH3 have different spatial distributions. It is unclear if scaling by the spatial scale of the emission would result in column densities that are correlated; additional interferometric observations are warranted. Nevertheless, if the model (Garrod & Widicus Weaver 2013) is correct, the abundance ratio between CH3OCH3 and HCOOCH3 may then probe the warm-up speed of hot cores.
  • 8.  
    Previous interferometric studies of HCOOCH3 in Orion indicated that it is formed on grain surfaces and detected in the colder, extended regions rather than in the hot core (Friedel & Widicus Weaver 2012; Favre et al. 2014). In our results, however, HCOOCH3 is shown to be present at warmer temperatures than CH3OCH3 and most of the other O-bearing COMs that are detected. This result is consistent with the findings in other massive star-forming regions such as GAL 24.78, GAL 75.78, W3(H2O), and GAL 12.91 (Bisschop et al. 2007; Isokoski et al. 2013). In these surveys, dimethyl ether is consistently found at temperatures lower than 150 K, while the temperature of methyl formate is consistently higher than that of dimethyl ether. In Sgr B2(N), however, Belloche et al. (2013) detected dimethyl ether at 130 K and methyl formate at 80 K. Our results agree with Bisschop et al. (2007) and Isokoski et al. (2013), but the reason for this trend and the reason for Sgr B2(N) being an outlier are unclear. It is possible that we could sufficiently describe the HCOOCH3 spectral intensity with one temperature component due to its weaker overall line intensities, and this temperature is in fact an average of two different temperatures. If a warm component can be found in multiple sources, it is not impossible that the gas-phase formation mechanisms for HCOOCH3 (Horn et al. 2004) are still relevant, although they do not account for all of the observed HCOOCH3 abundance. Further interferometric investigation of the HCOOCH3 spatial distribution as a function of temperature in sources other than Orion-KL is warranted.
  • 9.  
    Complex O-bearing species such as C2H5OH, CH3CHO, trans-HCOOH, and NH2CHO are not highly abundant in these sources. They are only detected in a few sources. Indeed, complex molecules are more difficult to detect, because of their weaker line intensities and their lower abundances. But our result has implications on the possibility of more complex chemistry, as many of these species are thought to be chemical precursors to more complex molecules. If these molecules are even less abundant than what was anticipated before, the formation routes for more complex molecules are uncertain.
  • 10.  
    Interstellar acetone has previously only been detected in Sgr B2 (N-LMH) (Snyder et al. 2002), Orion-KL (Friedel et al. 2005), GAL 10.47+00.03 (Rolffs et al. 2011), and GAL 31.41+0.31 (Isokoski et al. 2013). We have confirmed the detections in Orion, GAL 10.47+00.03, and GAL 31.41+0.31 with a full-band analysis, and also detected acetone in GAL 034.3+00.2. An additional paper offering the details of the acetone analysis in these surveys is in preparation (L. Zou & S. L. Widicus Weaver 2017, in preparation).
  • 11.  
    The complexity of chemistry may not be directly correlated with elevated temperature. NGC 6334–29 and NGC 6334-I(N) have similar chemistry, but NGC 6334-I(N) is significantly colder than NGC 6334–29. In NGC 6334-I(N), CH3OCH3 and CH3CHO are present at 35 K and 20 K, respectively. In NGC 6334–29, CH3OCH3 and CH3CHO are present at 135 K and 182 K, respectively. The cold CH3OCH3 and CH3CHO indicate the possible grain-surface chemistry origin of these O-bearing molecules, but what unique process causes their sublimation into the gas phase in NGC 6334-I(N) in such a cold environment is not clear. Likewise, GAL 034.3+0.02 presents similar chemical complexity to GAL 31.41+0.31, but its average COM temperature is lower.
  • 12.  
    The non-LTE CH3OH emissions, ${8}_{-\mathrm{1,8}}\to {7}_{\mathrm{0,7}}\,({E}_{\mathrm{low}}\,=78\ {\rm{K}})$ at 229.76 GHz and ${11}_{\mathrm{0,11}}\to {10}_{\mathrm{1,10}}\,({E}_{\mathrm{low}}=141\ {\rm{K}})$ at 250.51 GHz, are found primarily in shocks and H ii regions. The ${8}_{-\mathrm{1,8}}\to {7}_{\mathrm{0,7}}$ line can be modeled by our LTE fit only in W3(H2O), GAL 034.3+0.02, GAL 31.41 +0.31, GAL 10.47+00.03, and Orion-KL. These sources seem to be the richest sources in chemistry. The elevated methanol temperature in these sources may contribute to the flux of this warm methanol line. The ${11}_{\mathrm{0,11}}\to {10}_{\mathrm{1,10}}$ lines can be modeled by our LTE fit only in GAL 75.78 +0.34 and GAL 10.47+00.03, and the emission does not appear in B1-b, L1448 MM-1, or NGC 1333 IRAS 2A; in G24.33+00.11 MM1. In all other sources, they are observed in fluxes much larger than the CH3OH simulation with the LTE model.Several postulations may explain these unexpected emissions, but more evidence is needed to support any of these postulations. The two methanol lines may overlap with the lines from another unknown molecular carrier, but searches in the molecular line databases did not return likely molecular candidates; these lines are not included because of interference with spectral artifacts. It is not likely that these two lines arise from an extended source in the line of sight of our target source, as they would usually have distinctive velocities that can be resolved in our surveys. The lower state energies of these two lines are also not extremely cold, which means that they are not likely to be emitted by some unknown methanol source under LTE, as one would expect to see more methanol lines in our spectral coverage with detectable intensities under this LTE temperature.

6. Conclusion

We conducted broadband CSO spectral line surveys of 30 star-forming regions in the 1.3 mm band (220–267 GHz). The deconcolved spectra are presented. For each spectrum, we performed a global LTE analysis for molecules with a sufficient number of spectral lines and intensities, using the GOBASIC program. Some molecular tracers, such as CH3OH, CH3CCH, and SO, are ubiquitously detected in these sources. The difference in chemical complexity between hot cores, hot corinos, and Class 0 protostars is verified. Within each type of star-forming region, however, unexpected variations in the molecular complexity, abundance, and temperatures, are present. No straightforward correlations between different COMs can be drawn, as it seems that each source demonstrates a unique chemistry. Because of the large size of the single-dish beam, line intensities are diluted, and spatial distributions of different molecules cannot be resolved. To reveal the details of the astrochemistry in these star-forming regions, more information from interferometric studies on these sources is required so that source size can be properly included in the analysis.

Support for this work was provided in part by the National Science Foundation CAREER Award CHE-1150492. Support for the development of GOBASIC was provided by NASA Herschel OT1 Analysis Program RSA No. 1428755, the Emory Chemistry Early Career Research Grant, Emory SURE and SIRE programs (funding from Howard Hughes Medical Institute grant No. 52005873), and the Emory College Undergraduate Research Matching Grant. S.L.W.W., J.C.L., L.Z., B.M.H., N.W., and B.A.M. were supported for this work by S.L.W.W.'s startup funds, provided by Emory University. J.L.S., M.L.R., T.N.C., and J.A.K. were also supported by S.L.W.W.'s Emory startup funds, and by the Emory SURE and SIRE programs (funding from Howard Hughes Medical Institute grant No. 52005873). B.A.M. is a Jansky Fellow of the National Radio Astronomy Observatory, which is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. D.C.L. acknowledges support by NASA (Herschel OT funding) through an award issued by JPL/Caltech. We thank Frank Rice, Jonas Zmuidzinas, and Geoffrey Blake for technical support with the prototype receiver and assistance in data collection for the Orion-KL spectrum. We thank Martin Emprechtinger, Peter Schilke, and Claudia Comito for helpful discussions regarding the analysis of line surveys and deconvolution of DSB spectra, Thomas Anderson for assistance during part of the observations, Gordon Stacey for collegiality with time-sharing the telescope during rapid weather transitions on Maunakea, and Thomas Phillips for his general guidance and support. We gratefully acknowledge the support and services offered by the CSO technical and support staff, in particular Simon Radford, Richard Chamberlin, Ed Bufil, Brian Force, Hiroshige Yoshida, and Diana Bisel. We also thank the support staff from Caltech and Emory. This material is based upon work performed at the Caltech Submillimeter Observatory, which was operated by the California Institute of Technology under cooperative agreement with the National Science Foundation (AST-0838261).

Software: GOBASIC (version 5.0), Rad et al. (2016), CLASS (versions jun13a and jul14a) , (GILDAS 2017).

Footnotes

  • 10 

    The CSO was operated by the California Institute of Technology under contract from the National Science Foundation.

Please wait… references are loading.
10.3847/1538-4365/aa8098