Deep K-band Observations of TMC-1 with the Green Bank Telescope: Detection of HC₇O, Nondetection of HC₁₁N, and a Search for New Organic Molecules

Radio spectroscopy is a powerful and rigorous technique for the detection of new molecules in the dense interstellar medium. Organic molecules are commonly observed in many different astronomical sources (Herbst & van Dishoeck 2009) and the fact that many of the known interstellar organics are also present in protoplanetary disks and in comets (e.g., Biver et al. 2015; Cochran et al. 2015; Öberg et al. 2015; Altwegg et al. 2016; Walsh et al. 2016) supports the view that interstellar clouds could plausibly be the first formation sites for the prebiotic molecules that may have been delivered to the early Earth by comets (e.g., Mumma & Charnley 2011).

The dark interstellar cloud TMC-1 has proven to be the archetype for studies of cold cloud organic chemistry (Kaifu et al. 2004; McElroy et al. 2013; Sakai et al. 2013). Observations show that the inventory of TMC-1 is dominated by long carbon-chain molecules: the cyanopolyynes (${\mathrm{HC}}_{2n+1}{\rm{N}}$, $n=1,2,3,4;$ CH₃C₃N, CH₃C₅N), carbon-chain radicals (C₂H, C₃H, C₄H, C₅H, C₆H, C₈H, C₂N, C₃N, C₅N) and related anions, linear carbenes (H₂CCC, H₂CCCC, H₂CCCCCC), polyacetylenic molecules (CH₃C₂H, CH₃C₄H), and smaller ring molecules such as c-C₃H and c-C₃H₂ (Gratier et al. 2016; Ohishi 2016).

Cyanodecapentayne (HC₁₁N) has long been considered to be the largest gas-phase molecule identified in dense interstellar clouds. The first reported measurement of HC₁₁N in TMC-1 was by Bell et al. (1997), who detected two apparent transitions ($J=39-38$ and $J=38-37$) in the Ku band using the NRAO 140-foot telescope, from which they derived a column density of $2.8\times {10}^{11}\,{\mathrm{cm}}^{-2}$ (assuming a rotational temperature of 10 K). However, there have since been no reported confirmations of HC₁₁N in the literature, either in TMC-1 or any other astronomical object. Loomis et al. (2016) searched for and failed to detect six transitions of HC₁₁N in the Ku band in TMC-1 using the Green Bank Telescope (GBT), including the $J=39-38$ and $J=38-37$ transitions claimed by Bell et al. (1997). They obtained a ($2\sigma$) column density upper limit of $9.4\times {10}^{10}\,{\mathrm{cm}}^{-2}$, and concluded that the relative lack of HC₁₁N compared with the smaller cyanopolyynes may be due to the propensity of carbon chains longer than 10 C-atoms to form closed ring structures.

TMC-1 also contains shorter carbon chains with O and S terminations: C₂O, C₃O, C₂S, and C₃S (Ohishi et al. 1991; Kaifu et al. 2004). Theoretical models of interstellar anion chemistry have predicted that longer carbon-chain oxides could be present in TMC-1. Cordiner & Charnley (2012) predicted detectable column densities for C₆O, HC₆O, C₇O, and HC₇O in dense interstellar clouds, where they were theorized to form as a result of associative electron detachment (AED) reactions between carbon-chain anions (H)C_n⁻ and atomic oxygen. Such reactions have been observed in the laboratory by Eichelberger et al. (2007), and although these reactions were found to be rapid, the product branching ratios are currently unknown. McGuire et al. (2017a) considered the production of HC_nO radicals ($n=3\mbox{--}7$) based on a chemistry initiated by radiative association reactions between hydrocarbon cations and CO. The importance of these reaction mechanisms can be assessed by searching for, and measuring the abundances of, the predicted carbon-chain oxide molecules.

Aromatic compounds play a fundamental role in prebiotic chemistry (e.g., Ehrenfreund et al. 2006). However, despite strong observational evidence for their presence in high-excitation astronomical environments via infrared emission (Tielens 2008), there have been no firm spectroscopic identifications of specific polycyclic aromatic hydrocarbon (PAH) molecules in the interstellar medium so far. Aromatic nitrogen heterocycles are of particular prebiotic importance and, in particular, pyrimidine could be a possible precursor of DNA nucleobases. Experiments have shown that energetic processing of interstellar ice analogs containing pyrimidine can form the nucleobases uracil, thymine, and cytosine (Materese et al. 2013; Nuevo et al. 2014), as well as the N-heterocycles quinoline and isoquinoline (Materese et al. 2015). As a permanent electric dipole moment results from the presence of a heteroatom in the aromatic structure, rotational transitions are possible; however, such heterocycles have not yet been detected, despite dedicated searches in star-forming regions and circumstellar envelopes (Kuan et al. 2003; Charnley et al. 2005; Brünken et al. 2006).

Experimental and theoretical studies indicate that gas-phase formation of pyrimidine and benzene could be viable in dark clouds (Bera et al. 2009; Jones et al. 2011). The recent discovery of benzene and naphthalene in the volatile ices of comet 67P (K. Altwegg 2016, private communication) provides impetus to renew searches for aromatic interstellar compounds.

During the preparation of this article, McGuire et al. (2017a) published a detection of the carbon-chain oxide HC₅O in TMC-1 using the GBT K-band receiver, as well as a tentative detection of HC₇O. The first detection of benzonitrile (C₆H₅CN) has also been submitted for publication (McGuire et al. 2017b).

In this paper, we report detections and column densities for HC₇O and two isotopologues of HC₇N. We also report nondetections of HC₁₁N, HC₆O, C₆O, C₇O, malononitrile, quinoline, isoquinoline, and pyrimidine, for which column density upper limits are presented.

Spectra were obtained toward the TMC-1 cyanopolyyne peak (at J2000 R.A. 04:41:42, decl. +25:41:27) using the Robert C. Byrd Green Bank Telescope (GBT) in October–November 2011. Observations were conducted using the central beam of the K-band focal plane array (KFPA) receiver, with a half-power beam width of $\approx 37^{\prime\prime}$ on the sky. The GBT spectrometer was configured with a channel width of 12.2 kHz and $4\times 50\,\mathrm{MHz}$ spectral windows. As a result of the GBT's dynamic scheduling system, observations were performed under consistently good weather conditions, with system temperatures (T_sys) in the range of 35–50 K and zenith opacities of 0.02–0.05. Pointing was checked every 1–2 hr, and the mean pointing error over the 64 hr allocated to our program was $5.2^{\prime\prime}$. Spectra in the range of 18.1–19.8 GHz were obtained using frequency switching, and the range of 20.2–20.5 GHz was observed by position switching, using an uncontaminated reference position offset by $17^{\prime}$ NE from the cyanopolyyne peak.

Data reduction was performed using the GBTIDL software, which included opacity correction, frequency-alignment of individual scans in the LSR frame, and T_sys-weighted averaging. Reduced spectra were subsequently corrected for the main beam efficiency of 0.92. On-source integration times ∼10 hr per spectral setting resulted in rms noise levels of 1–2 mK per channel. Targeted spectral line frequencies are shown in Table 1. We used line lists from the Cologne Database for Molecular Spectroscopy (Müller et al. 2001) and Jet Propulsion Laboratory Molecular Spectroscopy web pages,⁷ augmented, where necessary, by our own extrapolations outside the laboratory data sets. The energy level quantum numbering schemes adopted here differ for the various molecules and are detailed in the primary spectroscopic references. For HC₇O, we obtained additional GBT KFPA data from the recent study of McGuire et al. (2017a).

Table 1.  Targeted Line Frequencies, Transitions, and Upper-state Energies

Rest Freq.	Species	Transition	Eu
(MHz)			(K)
18106.283	HC7O	$17-16e$	7.5
18106.312	HC7O	$16-15e$	7.5
18107.873	HC7O	$17-16f$	7.5
18107.900	HC7O	$16-15f$	7.5
18334.033	C7O	16 − 15	7.5
18596.764	HC11N	55 − 54	25.0
18638.616	HC5N	7 − 6	3.6
18828.319	HC6O	$12-11e$	5.5
18828.336	HC6O	$11-10e$	5.5
18829.311	HC6O	$12-11f$	5.5
18829.327	HC6O	$11-10f$	5.5
18992.942	C9H7N	${10}_{\mathrm{0,10}}-{9}_{\mathrm{0,9}}$, $F=10-9$	5.2
18992.990	C9H7N	${10}_{\mathrm{0,10}}-{9}_{\mathrm{0,9}}$, $F=9-8$	5.2
18993.002	C9H7N	${10}_{\mathrm{0,10}}-{9}_{\mathrm{0,9}}$, $F=11-10$	5.2
19071.386	C6O	${11}_{12}-{10}_{11}$	5.9
19203.671	HC7O	$18-17e$	8.4
19203.699	HC7O	$17-16e$	8.4
19205.267	HC7O	$18-17f$	8.4
19205.294	HC7O	$17-16f$	8.4
19479.903	C7O	17 − 16	8.4
19785.41	CH2(CN)2	${4}_{\mathrm{1,3}}-{4}_{\mathrm{0,4}}$	3.6
20278.94	C4H4N2	${12}_{\mathrm{9,3}}-{12}_{\mathrm{8,4}}$	44.8
20287.345	HC11N	60 − 59	29.7
20292.487	HC313CC3N	18 − 17	9.3
20294.271	HC413CC2N	18 − 17	9.3
20303.946	HC7N	18 − 17	9.3
20465.490	HC6O	$13-12e$	6.5
20465.504	HC6O	$12-11e$	6.5
20466.662	HC6O	$13-12f$	6.5
20466.676	HC6O	$12-11f$	6.5
20529.076	i-C9H7N	${11}_{\mathrm{1,11}}-{10}_{\mathrm{1,10}}$, $F=11-10$	6.1
20529.097	i-C9H7N	${11}_{\mathrm{1,11}}-{10}_{\mathrm{1,10}}$, $F=10-9$	6.1
20529.113	i-C9H7N	${11}_{\mathrm{1,11}}-{10}_{\mathrm{1,10}}$, $F=12-11$	6.1

Notes. Primary spectroscopic sources for molecular line frequencies: HC₇O and HC₆O—Mohamed et al. (2005), C₇O—Ogata et al. (1995), HC₁₁N—Travers et al. (1996), C₉H₇N and i-C₉H₇N—Kisiel et al. (2003), C₆O—Ohshima et al. (1995), CH₂(CN)₂—Cox et al. (1985), C₄H₄N₂—Kisiel et al. (1999), HC₅N—Bizzocchi et al. (2004), HC₇N and its isotopic species—McCarthy et al. (2000). The CH₂(CN)₂ and C₄H₄N₂ lines are blends of many hyperfine components due to the presence of two nitrogen nuclei.

Download table as:  ASCIITypeset image

Column densities were obtained using spectral line models constructed for each species, based on Gaussian (thermally broadened) opacity profiles (e.g., Nummelin et al. 2000). The cloud radial velocity (5.81 $\mathrm{km}\ {{\rm{s}}}^{-1}$) and Doppler FWHM (0.37 $\mathrm{km}\ {{\rm{s}}}^{-1}$) were obtained from a least-squares fit to the high signal-to-noise HC₇N $J=18-17$ line (shown in Figure 4). The column density for each species was varied until the best fit to the observed spectral data was obtained; $1\sigma$ uncertainties were derived using Monte Carlo noise replication and refitting with 500 replications (as described by Cordiner et al. 2013). For species with no detectable emission, we adopted an approach similar to Loomis et al. (2016), by constructing a spectral model (with fixed radial velocity and line FWHM), and increasing the column density from zero until a ${\chi }^{2}$ value of $2\sigma$ was reached. For species with multiple lines in different spectral windows, the fit was generated for all lines simultaneously, weighted by the rms noise level in each window.

Rotational temperatures were taken from the literature where available: 5 K was adopted for HC₅N and 6 K for HC₇N based on Bell et al. (1998), and 10 K for HC₁₁N based on Bell et al. (1997) and Loomis et al. (2016). Although the kinetic temperature of TMC-1 CP has been well established at 10 K (e.g., Pratap et al. 1997), species with large dipole moments (such as the cyanopolyynes and related asymmetric species with 3–9 C atoms) undergo rapid rotational cooling that can cause their rotational excitation temperatures (T_rot) to fall away from thermodynamic equilibrium. As discussed by Bell et al. (1998), the magnitude of this effect for the cyanopolyynes is primarily dependent on the rotational constant (inverse of the molecule length). Thus, for species containing a chain of six or seven C-atoms, we adopted ${T}_{\mathrm{rot}}=6$ K. For the remaining molecules, we adopted ${T}_{\mathrm{rot}}=10$ K. Our calculated column densities, abundances, and upper limits are given in Table 2. The TMC-1 H₂ column density of ${N}_{{{\rm{H}}}_{2}}={10}^{22}$ cm⁻¹ was taken from Cernicharo & Guelin (1987).

Table 2.  TMC-1 (CP) Molecular Column Density/Abundance Measurements and Upper Limits

Species	Name	Trot (K)	N (cm−2)	$N/{N}_{{{\rm{H}}}_{2}}$
HC11N	Cyanodecapentayne	10	<7.4 × 1010	<7.4 × 10−12
HC7N	Cyanohexatriyne	6	1.36 × 1013	1.36 × 10−9
HC313CC3N	13C-Cyanohexatriyne	6	(1.2 ± 0.2) × 1011	(1.2 ± 0.2) × 10−11
HC413CC2N	13C-Cyanohexatriyne	6	(1.4 ± 0.2) × 1011	(1.4 ± 0.2) × 10−11
HC5N	Cyanobutadiyne	5	5.3 × 1013	5.3 × 10−9
C6O	Hexacarbon monoxide	6	<5.2 × 1010	<5.2 × 10−12
C7O	Heptacarbon monoxide	6	<2.6 × 1010	<2.6 × 10−12
HC6O	1-oxo-hexa-1,3,5-triynyl	6	<1.5 × 1011	<1.5 × 10−11
HC7O	1-oxo-hepta-2,4,6-triynyl	6	(7.8 ± 0.9) × 1011	(7.8 ± 0.9) × 10−11
CH2(CN)2	Malononitrile	10	<7.3 × 1010	<7.3 × 10−12
C4H4N2	Pyrimidine	10	<4.2 × 1013	<4.2 × 10−9
C9H7N	Quinoline	10	<6.2 × 1011	<6.2 × 10−11
i-C9H7N	Isoquinoline	10	<9.4 × 1011	<9.4 × 10−11

Download table as:  ASCIITypeset image

3.1. Carbon Chain Oxides

To obtain improved sensitivity to the cumulenone radical species HC₇O, our observational data (covering the $F=17-16$ and 16 − 15 e and f transitions) were combined with additional GBT data (covering the $F=18-17$ and 17 − 16 e and f transitions) from the study of McGuire et al. (2017a) for a total of eight HC₇O transitions, in the form of two pairs of hyperfine doublets. The data were rebinned to a common frequency resolution and transformed to the (LSR) velocity scale. At the resolution of our observed spectra, the hyperfine doublet structure of each line was not well resolved, so the mean rest frequency was used for each doublet. The average spectrum for the four doublets is shown in Figure 1, with a well-defined, statistically significant peak at the TMC-1 systemic velocity of 5.8 $\mathrm{km}\ {{\rm{s}}}^{-1}$ (marked with a vertical dotted line). The integrated line intensity is 3.8 ± 0.39 mK $\mathrm{km}\ {{\rm{s}}}^{-1}$, which corresponds to a $9.5\sigma$ detection.

Zoom In Zoom Out Reset image size

A spectral line model was constructed to fit all eight HC₇O hyperfine components, shown in Figure 2. Allowing both the HC₇O column density (N) and radial velocity (v) to vary, the best-fitting model results were $N({\mathrm{HC}}_{7}{\rm{O}})=(7.8\pm 0.9)\,\times {10}^{11}$ cm⁻², and $v=5.83\pm 0.05$ $\mathrm{km}\ {{\rm{s}}}^{-1}$.

Zoom In Zoom Out Reset image size

Three possible formation mechanisms have been proposed for these carbon-chain oxides. The observations reported here were originally motivated by the theoretical models of Cordiner & Charnley (2012) who considered the formation of C₆O, C₇O, HC₆O, and HC₇O as a result of reactions between O atoms and carbon-chain anions, as measured in the laboratory by Eichelberger et al. (2007). For reactions of oxygen with the polyyne anions C_nH⁻ $n=2\mbox{--}7)$, the most likely product channels are

$$\begin{eqnarray}&&{{\rm{C}}}_{n}{{\rm{H}}}^{-}+{\rm{O}}\longrightarrow {\mathrm{HC}}_{n}{\rm{O}}+{e}^{-}\end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray}&&{{\rm{C}}}_{n}{{\rm{H}}}^{-}+{\rm{O}}\longrightarrow {{\rm{C}}}_{n}{{\rm{O}}}^{-}+{\rm{H}}\end{eqnarray} \tag{ 2 }$$

$$\begin{eqnarray}&&{{\rm{C}}}_{n}{{\rm{H}}}^{-}+{\rm{O}}\longrightarrow {{\rm{C}}}_{n-1}{{\rm{H}}}^{-}+\mathrm{CO}.\end{eqnarray} \tag{ 3 }$$

The chemical model for TMC-1 presented by Cordiner & Charnley (2012) predicted a maximum HC₇O column density of $\approx 1.5\times {10}^{13}$ cm⁻² at a chemical age of $t\approx 1\times {10}^{5}$ years, as well as potentially detectable abundances of the other three molecules. This calculated maximum HC₇O abundance is about a factor of 19 higher than the observed value reported here, and the modeled abundances of C₆O, C₇O, and HC₆O are also significantly higher than our derived upper limits. However, comparison of the observed HC₇O abundance ($\approx 8\times {10}^{-11}$) with the TMC-1 model (Figure 4 of Cordiner & Charnley 2012) indicates that it can be reproduced at $t\approx 3\times {10}^{5}$ years, at which point the C₆O and C₇O abundances are also within the upper limits reported here. Although at $t\approx 3\times {10}^{5}$ years the HC₆O abundance is comparable to that of HC₇O, in conflict with our observed abundance limits, it should be noted that the Cordiner & Charnley (2012) model also overpredicted the C₆H⁻ abundance (by a factor of 6.7, compared with the observed value from Brünken et al. 2007). If the modeled C₆H⁻ abundance was less by a similar factor, then the predicted HC₆O abundance would drop to just below our observed upper limit as a result of the close chemical relationship between these species.

McGuire et al. (2017a) considered the chemistry of the HC_nO radicals ($n=3\mbox{--}7$) based on an extension of the reactions suggested by Adams et al. (1989). In this case, the cumulenone radicals are produced in radiative association reactions between hydrocarbon cations and CO:

$$\begin{eqnarray}&&{{\rm{C}}}_{n-1}{{{\rm{H}}}_{2}}^{+}+\mathrm{CO}\longrightarrow {{{\rm{H}}}_{2}{\rm{C}}}_{n}{{\rm{O}}}^{+}+\nu \end{eqnarray} \tag{ 4 }$$

$$\begin{eqnarray}&&{{\rm{C}}}_{n-1}{{{\rm{H}}}_{3}}^{+}+\mathrm{CO}\longrightarrow {{{\rm{H}}}_{3}{\rm{C}}}_{n}{{\rm{O}}}^{+}+\nu \end{eqnarray} \tag{ 5 }$$

followed by electron dissociative recombination of the ions. In a model for TMC-1, McGuire et al. (2017a) found that their observed HC₅O column density and their HC₇O upper limit were best reproduced at $t\approx 2\times {10}^{5}$ years. However, at this time, HC₄O was calculated to be about 37 times more abundant than observed, and HC₆O was a factor of 3 × 10⁻³ below the observed upper limit.

It has been suggested that the presence of CCO and CCCO in TMC-1, as well as perhaps other carbon-chain oxides, could originate in the injection of surface-formed cumulenone molecules from dust grains (Markwick et al. 2000). On cold dust grains, carbon atom additions, starting from HCO, could possibly lead to long cumulenone HC_nO radicals that would form cumulenone molecules, H₂C_nO, following an H atom addition (see Charnley 1997). Following desorption, the cumulenone molecules would be protonated and the product ions subject to electron dissociative recombination reactions, leading to C_nO and HC_nO radicals, as well as H₂C_nO molecules, all being present in the gas. Thus, in this case, the HC_nO radicals would have both a grain-surface and a gas-phase origin. However, the nondetection of the propynonyl radical (HC₃O) in TMC-1 (McGuire et al. 2017a), and consistently negative results from searches for propadienone (H₂C₃O) in molecular clouds, including TMC-1 (Irvine et al. 1988; Brown et al. 1992; Loison et al. 2016), tend to strongly disfavor this scenario.

Theoretical models for carbon-chain oxide chemistry will require laboratory measurements of the rates and branching ratios of key formation reactions. Generally, the product distributions of electron dissociative reactions involving ${{{\rm{H}}}_{m}{\rm{C}}}_{n}{{\rm{O}}}^{+}$ ions are important and, for anion chemistry, an assessment of the efficiency of the associative electron detachment (AED) channel (Reaction 1) would be very informative. Future deep searches in TMC-1 and elsewhere for these molecules will permit the elucidation of their exact formation pathways; for example, a detection of HC₆O would help distinguish between the anionic and cationic formation mechanisms.

3.2. Cyanopolyynes

3.2.1. HC₁₁N Confirmed Nondetection

Our GBT K-band spectra support the results of Loomis et al. (2016): we searched at the frequencies of the higher-energy $J=55-54$ and $J=60-59$ transitions but found no evidence for any HC₁₁N emission. Our spectra are shown in Figure 3 with a model based on the HC₇N line FWHM and radial velocity, with the temperature and column density of Bell et al. (1997) overlaid. Our spectral line modeling implies a column density upper limit of $7.4\times {10}^{10}\,{\mathrm{cm}}^{-2}$ (at 95% confidence), which is somewhat lower than the upper limit obtained by Loomis et al. (2016), and almost four times less than the value of $2.8\times {10}^{11}\,{\mathrm{cm}}^{-2}$ claimed by Bell et al. (1997).

Zoom In Zoom Out Reset image size

We performed our search for HC₁₁N at the same coordinates as Loomis et al. (2016), so our combined results are consistent with a lack of HC₁₁N at the TMC-1 cyanopolyyne peak. It should be noted, however, that the detection claimed by Bell et al. (1997) was based on observations using a significantly larger ($2\buildrel{\,\prime}\over{.} 4$) telescope beam, offset from the nominal cyanopolyyne peak by about 1 s in R.A. and $9^{\prime\prime}$ in decl. It is therefore conceivable that their larger, slightly offset telescope beam was sensitive to HC₁₁N emission that was missed by the present study and that of Loomis et al. (2016). The previous detection of HC₁₁N in TMC-1 therefore cannot be completely refuted until additional observations are performed, ideally covering the same area on the sky as those made by Bell et al. (1997) using the NRAO 140-foot telescope.

3.2.2. New HC₇N isotopologues

The spectra shown in Figure 4 reveal the presence of the $J=18-17$ lines of HC₃¹³CC₃N and HC₄¹³CC₂N at the 6–7σ level. This represents the first reported detection of individual isotopologues of HC₇N in any environment we are aware of. The only prior observation of ¹³C-substituted HC₇N was in TMC-1 by Langston & Turner (2007) using the GBT, who stacked the spectra of various HC₇N isotopologues (with ¹³C at different positions in the carbon chain), to obtain a mean ¹²C/¹³C ratio of ${87}_{-19}^{+37}$ for this molecule. We derive abundance ratios for the specific isotopologues HC₇N/HC₃¹³CC₃N = 110 ± 16 and HC₇N/HC₄¹³CC₂N = 96 ± 11. These values are notably high, and indicate depletion of ¹³C relative to the local interstellar elemental ¹²C/¹³C ratio of 60–70 (Lucas & Liszt 1998; Milam et al. 2005). Depleted ¹²C/¹³C ratios have previously been observed in TMC-1 for the closely related molecules HC₃N and HC₅N, as well as CCH and CCS (Sakai et al. 2013). Modeling efforts are in progress to understand this isotopic depletion, which seems to be a common characteristic of carbon chains in cold interstellar gas (Yoshida et al. 2015; Araki et al. 2016).

Zoom In Zoom Out Reset image size

Based on our new results, the depletion of ¹³C in HC₇N (resulting in a higher ¹²C/¹³C ratio), seems to be stronger than for the shorter cyanopolyynes. For HC₇N, our mean ¹²C/¹³C ratio is 103 ± 6, compared with the mean value of 94 ± 6 for HC₅N (Taniguchi et al. 2016) and 70 ± 5 for HC₃N (Takano et al. 1998). However, it should be noted that previous studies have found the ratio to be significantly smaller for the carbon atom in the CN end-group than for the other C atoms along the chain (Taniguchi et al. 2016). Excluding the CN end-group, the mean ratios are 97 ± 7 for HC₅N and 77 ± 7 for HC₃N. Additional observations will be required to determine if this is also the case for HC₇N.

Zoom In Zoom Out Reset image size

The current understanding regarding the origin of ¹³C depletion in carbon-chain-bearing species is that it likely occurs due to gas-phase isotope exchange reactions, whereby ¹³C is preferentially incorporated into other molecules. For instance, it is known that the ¹²C/¹³C ratio in (atomic) ${{\rm{C}}}^{+}$ becomes depleted due to the incorporation of the heavier isotope into CO, through the exchange of ¹²C and ¹³C in the reaction between ¹³C⁺ and ¹²CO (Langer et al. 1984). It has thus been hypothesized that as gas-phase ${{\rm{C}}}^{+}$ becomes depleted in ¹³C, any molecules whose chemistry is closely linked with ${{\rm{C}}}^{+}$ should also show isotopic depletion (Sakai et al. 2013).

It has been suggested that the analogous exchange reaction of ¹³C⁺ with ¹²CN could produce the observed (relative) ¹³C enrichment of the CN group in cyanopolyynes (Takano et al. 1998; Araki et al. 2016), but the viability of this mechanism has yet to be conclusively demonstrated in dense molecular clouds. Based on the relative similarities of the ¹²C/¹³C ratios in the different C atoms of HC₅N in TMC-1, Taniguchi et al. (2016) deduced that the main route to synthesizing the observed HC₅N is via hydrocarbon ion (plus N-atom) chemistry. The trend for lower ¹³C fractions in progressively larger cyanopolyynes is therefore indicative of a diminishing ¹³C content in hydrocarbon ions of increasing size.

3.3. Nitrogen Heterocycles

We have performed deep searches for large organic molecules in TMC-1 with the GBT. Our data, when combined with that of McGuire et al. (2017a), allows us to confirm the previous tentative detection of HC₇O. The measured HC₇O column density and the derived upper limits for C₆O, HC₆O, and C₇O place constraints on chemical models in which these molecules form through reactions involving negatively and/or positively charged hydrocarbon ions. We also confirm the previously reported nondetection of HC₁₁N at the cyanopolyyne peak by Loomis et al. (2016). We find that the detected ¹³C isotopologues of HC₇N are depleted in ¹³C. Reproduction of these ¹²C/¹³C ratios, as well as those measured in shorter carbon-chain molecules (CCH, CCS, HC₃N, and HC₅N), present a challenge for models of interstellar isotopic fractionation. Finally, our nondetections of nitrogen heterocycles in a cold molecular cloud complement previous nondetections of aromatic molecules in star-forming regions and in the envelopes of evolved stars. This raises the question: why have no aromatic compounds yet been detected by astronomical millimeter/submillimeter spectroscopy? Most previous searches have involved targeting molecules in which the heteroatom, specifically N, is incorporated in an aromatic ring. While our derived abundance of pyrimidine is not particularly restrictive with regard to its possible presence in TMC-1, the recent detection of benzonitrile in TMC-1 by McGuire et al. (2017b) and of toluene in comet 67P/Churyumov–Gerasimenko (Altwegg et al. 2017) suggests that interstellar chemistry may favor the presence of heteroatoms as functional side-groups, rather than within the ring structure. Future deep searches in TMC-1 with the GBT should provide further insights into the inventory of the heaviest interstellar molecules.

This work was supported by the Goddard Center for Astrobiology and NASA's Origins of Solar Systems and Exobiology programs.