Long Carbon Chains in the Warm Carbon-chain-chemistry Source L1527: First Detection of C₇H in Molecular Clouds

Carbon-chain molecules account for ∼40% of the interstellar molecular species detected to date. They are abundant in starless dark clouds and circumstellar envelopes of carbon-rich late-type stars. The starless dark cloud Taurus Molecular Cloud-1 Cyanopolyyne Peak (TMC-1 CP) is the richest known source of carbon-chain molecules and is where most of the carbon-chain molecules detected in interstellar clouds exist (e.g., Kaifu et al. 2004). Therefore, this source has been used as a testbed for carbon-chain chemistry. Various carbon-chain molecules have also been detected in the similar starless dark cloud Lupus-1A (Sakai et al. 2010b). The low temperature of ∼10 K and the young evolutionary stages of both clouds provide favorable conditions for the production of carbon-chain molecules. Recently, carbon-chain molecules have also been detected in dense and warm regions around low-mass protostars. The carbon-chain molecules are thought to be produced from CH₄ evaporated from dust grains via a mechanism called warm carbon-chain chemistry (WCCC), for which representative sources are L1527 and IRAS 15398-3359 (e.g., Sakai et al. 2008).

Hassel et al. (2008) conducted chemical model calculations and confirmed that the evaporation of CH₄ in a lukewarm region near a protostar triggers enhancement of carbon-chain molecules. This model better reproduces the molecular abundances observed in L1527 than the model of the remnants of a cold starless core. Aikawa et al. (2008) also revealed the enhancement of various carbon-chain molecules due to the evaporation of CH₄ near a protostar with the chemical model of a dynamically evolving cloud. Even though the basic chemistry in L1527 has been characterized by these observational and modeling efforts, the remnant species formed in the starless core phase may also contribute to the chemical composition in L1527 to some extent. Such a contribution has recently been suggested by the ¹³C isotope observations of HC₃N (Araki et al. 2016). To fully understand the chemical composition of the prototypical WCCC source L1527, a more comprehensive study of the chemical composition of L1527 and a comparison with TMC-1 CP are needed.

Here, we focus on longer carbon-chain molecules. In general, the column density ratios (N_L1527/N_TMC-1) between the WCCC source L1527 and TMC-1 CP for observed molecules with carbon chains tend to decrease with the carbon-chain length. This trend can be seen in cyanopolyynes H–(C≡C)_n–C≡N (n = 1, 2, 3, 4), cumulene carbenes C_nH₂ (n = 3, 4), radical carbon chains C_nH (n = 4, 5, 6) (Sakai et al. 2007, 2008), and carbon-chain anions C_nH⁻ (n = 4, 6) (McCarthy et al. 2006; Sakai et al. 2007; Cordiner et al. 2013). However, cumulene carbene C₆H₂ (n = 6) appears to follow an opposite trend, i.e., the column density of C₆H₂ in L1527 is comparable to that in TMC-1 CP (Sakai et al. 2007). These results suggest that the WCCC regions also harbor abundant populations of "long" carbon chains as in the case of cold starless clouds. Therefore, a more thorough investigation of longer carbon-chain molecules will provide clues to our understanding of the contribution of the remnant species. With this in mind, we conducted sensitive observations of long carbon-chain molecules toward L1527. In these observations, we detected several long carbon-chain molecules, C₇H, C₆H, CH₃C₄H, and C₆H₂, as well as the cyclic species C₃H and C₃H₂O.

The rotational spectral lines in the 42.0–44.8 GHz region listed in Table 1 were observed simultaneously using the Robert C. Byrd 100 m Green Bank Telescope (GBT) of the National Radio Astronomy Observatory⁷ from 2015 March to October. We observed the IRAS 04368+2557 position in L1527: (α2000.0, δ2000.0) = (04^h 39^m 5389, 26°03'11.0'') (Sakai et al. 2008). A dual-polarization Q-band receiver with an IF bandwidth of 4 GHz was used. The system temperature during the observations was in the range of 70 to ∼100 K depending on the elevation angle of the telescope and the weather conditions. The half-power beam width of the telescope is 169 at 44.8 GHz, and the main-beam efficiencies are 0.83 and 0.81 below and above 44.0 GHz, respectively (GBT Support Staff 2014). The pointing of the telescope was checked by observing the continuum sources J0336+3218 and J0319+4130 every 1.5 hr, and the pointing accuracy was kept within 5''. The VErsatile GBT Astronomical Spectrometer (VEGAS) was used as a backend at the 23.44 MHz bandwidth and 1.4 kHz resolution mode. The resolution corresponds to a velocity resolution of 0.15–0.16 km s⁻¹ after 16-channel smoothing, which is sufficient to resolve a line width for this cloud (∼0.5 km s⁻¹). A frequency-switching mode with a frequency offset of ±2.5 MHz was employed for the observations. The intensity scale was calibrated using noise injection, and the spectra of the right- and left-handed circular polarizations were averaged to obtain the final spectrum.

Table 1.  Molecular Lines Observed in L1527

Species	Transition	Frequencya,b	TMB	Δvc	Wd	rms	VLSR
		(GHz)	(K)	(km s−1)	(K km s−1)	(mK)	(km s−1)
CH3C4H	J = 11–10, K = 0	44.785974	0.0441(4)	0.574(6)	0.0270(5)	2.0	5.9
	K = 1	44.785537	0.0391(4)	0.525(6)	0.0219(5)	2.0	5.9
	K = 2	44.784226	0.0083(3)	0.77(4)	0.0068(6)	2.0	5.8
C7H 2Π1/2	J = 24.5–23.5e,f	42.848418e,f	...	...	...	2.1	...
		42.850080e,f	...	...	...	1.9	...
	J = 25.5–24.5e,f	44.597314e,f	...	...	...	1.6	...
		44.598980e,f	...	...	...	2.1	...
	Average	...	0.00459(18)	0.75(3)	0.0037(3)	0.9	5.6
C6H2	${J}_{{K}_{{\rm{a}}},{K}_{{\rm{c}}}}={16}_{\mathrm{1,16}}\mbox{--}{15}_{\mathrm{1,15}}$	42.976147	0.0211(5)	0.492(13)	0.0111(5)	1.8	6.1
	160,16–150,15	43.030509	0.0095(3)	0.900(29)	0.0091(5)	2.0	6.0
	161,15–151,14	43.083943	0.0172(4)	0.567(15)	0.0104(5)	2.1	6.0
c-C3H	${N}_{{K}_{{\rm{a}}},{K}_{{\rm{c}}}}={2}_{\mathrm{1,1}}\mbox{--}{2}_{\mathrm{1,2}}$
	J = 2.5–2.5, F = 3–3	44.610393	0.0806(8)	0.516(6)	0.0443(5)	4.3	5.6
	J = 2.5–2.5, F = 2–2	44.610873	0.0494(9)	0.439(9)	0.0231(5)	4.3	5.6
c-C3H2O	${J}_{{K}_{{\rm{a}}},{K}_{{\rm{c}}}}={3}_{\mathrm{0,3}}\mbox{--}{2}_{\mathrm{0,2}}$	42.031940	0.0217(5)	0.343(10)	0.0079(2)	2.0	5.9
	${J}_{{K}_{{\rm{a}}},{K}_{{\rm{c}}}}={3}_{\mathrm{1,2}}\mbox{--}{2}_{\mathrm{1,1}}$	44.587398	0.0521(9)	0.451(9)	0.0250(5)	4.5	6.0
HCS+	J = 1–0	42.674195	0.0729(6)	0.3358(29)	0.0261(4)	1.7	6.0
C6H	2Π3/2, J = 15.5–14.5f	42.970432	0.0894(4)	0.854(4)	0.0813(4)	1.8	5.9
	2Π3/2, J = 15.5–14.5e	42.977089	0.0899(3)	0.812(3)	0.0777(3)	1.8	5.9
	2Π1/2, J = 15.5–14.5e	43.294790	0.0128(3)	0.850(24)	0.0116(3)	1.9	5.8

Notes. The numbers in parentheses are the 1σ errors in units of the last significant digit.

^aRest frequency. ^bThe Cologne Database for Molecular Spectroscopy (Müller et al. 2001, 2005). ^cFWHM obtained by Gaussian fit. ^d $W=\int {T}_{\mathrm{MB}}{dv}$. ^eMcCarthy et al. (1997). ^fAverage of two hyperfine components. The line width of 0.75 km s⁻¹ includes the hyperfine splitting of 0.2 km s⁻¹.

Download table as:  ASCIITypeset image

Because the C₄H and CCH emission lines are distributed over a 30''–40'' region around the center of L1527, as reported by Sakai et al. (2008, 2010a), the emitting region of the carbon-chain molecules is expected to be larger than the beam size of the GBT in the present frequency region. Therefore, the beam dilution effect was not considered in the analysis of the observed lines. The local thermodynamic equilibrium (LTE) condition was assumed to estimate the column densities of the observed molecules. The method used for the column-density calculation is the same as that reported by Turner (1991).

Because we observed only one rotational transition or a few rotational transitions with similar upper-state energies for each molecular species, we need to assume the excitation temperature in the calculation of the column density. For C₆H, the excitation temperature is reported to be 12 K on the basis of the observations of six lines (Sakai et al. 2007), while for HC₃N it is reported to be 12.1 K (Araki et al. 2016). These excitation temperatures are lower than the gas kinetic temperature expected in this region as described later, indicating that the lines are subthermally excited. Under this condition, the excitation temperature is determined by the collisional excitation and the radiative deexcitation (i.e., the Einstein A coefficient), and therefore it depends on the H₂ density and the dipole moment. The C₆H and HC₅N molecules have relatively large dipole moments, 5.536 D and 3.73 D, respectively, which are roughly comparable to those of C₇H (5.945 D; Woon 1995), C₆H₂ (6.2 D; Maluendes & McLean 1992), and c-C₃H₂O (4.39 D; Benson et al. 1973). Therefore, we simply assumed the excitation temperature of C₆H (12 K) for these molecules. Conversely, the dipole moments of CH₃C₄H and c-C₃H are 1.2071 D and 2.4 D, respectively, which are smaller than that of C₆H. Therefore, the excitation temperature could be higher for these two species. Nevertheless, we assumed an excitation temperature of 12 K for them and present the deviation of the column density due to a variation of the excitation temperature by 1 K.

3.1. C₇H

So far, C₇H has been detected only in the circumstellar envelope of IRC+10216 (Guélin et al. 1997). Despite sensitive observations, C₇H has not yet been detected in the starless dark cloud TMC-1 CP (Bell et al. 1999; Dickens et al. 2001). In this study, we observed the four lines of C₇H in L1527, marginally detecting them with a signal-to-noise ratio of approximately two, as shown in Figure 1. By stacking the four spectra in the velocity axis, the line clearly appears with a signal-to-noise ratio (S/N) of five, therefore securing the detection of C₇H. This marks the first detection of C₇H in molecular clouds.

Zoom In Zoom Out Reset image size

C₇H has been searched for toward TMC-1 CP. Bell et al. (1999) observed the J = 10.5–9.5 line (18.4 GHz) with the NRAO 43 m telescope at a root-mean-square (rms) noise sensitivity of 33 mK (T_B), and Dickens et al. (2001) observed the J = 12.5–11.5 line (21.9 GHz) with the DSN 70 m telescope at an rms noise sensitivity of 7 mK (T_B). However, C₇H was not detected in these observations. Conversely, our sensitive observations with the GBT in combination with the stacking analysis using the four lines permitted us to achieve an rms noise sensitivity of 1 mK. This allows us to detect the faint C₇H line in this region.

We derive a column density for C₇H of 6(1) × 10¹⁰ cm⁻² by assuming the excitation temperature of 12 K (Sakai et al. 2007) and using the dipole moment μ = 5.945 D (Woon 1995), where the error represents 1σ derived from the error of the integrated intensity of the averaged line. When the assumed excitation temperature is changed by 1 K, the column density varies by 13%. The fractional abundance relative to H (2N (H₂)) is as low as 1 × 10⁻¹², where the H₂ column density (2.8 × 10²² cm⁻²) is taken from Jørgensen et al. (2002). In the case of TMC-1 CP, the upper limit of the column density has been reported to be 1.5 × 10¹¹ cm⁻² (Bell et al. 1999; Dickens et al. 2001). Therefore, the column density ratio N_L1527/N_TMC-1 is higher than 0.4, as shown in Figure 2.

Zoom In Zoom Out Reset image size

3.2. C₆H

The two rotational lines of the ²Π_3/2 electronic ground state and the one rotational line of the ²Π_1/2 electronic excited state for C₆H were detected in L1527, as shown in Figure 3. This is the first detection of C₆H in the ²Π_1/2 state in molecular clouds, and the lines in the ²Π_1/2 state have currently only been detected in IRC+10216 (Saito et al. 1987). The derived column densities of the ²Π_3/2 and ²Π_1/2 states of C₆H are 1.08(2) × 10¹² cm⁻² and 1.58(4) × 10¹¹ cm⁻², respectively, assuming the excitation temperature of 12 K (Sakai et al. 2007) and the dipole moment μ of 5.536 D (Woon 1995), where the errors are 1σ derived from the errors of the integrated intensities of the lines. Therefore, the derived temperature between the ²Π_3/2 and ²Π_1/2 states is 11.3(3) K based on the energy separation of these two electronic states of 15.04 cm⁻¹ (21.6 K) reported by Linnartz et al. (1999). Because the gas kinetic temperature is 20(2) K, as described later, the ²Π_1/2 state is subthermally populated.

Zoom In Zoom Out Reset image size

3.3. CH₃C₄H

The rotational transitions J = 11–10 (K = 0, 1, 2) of CH₃C₄H were detected, as shown in Figure 4. The K = 1 and 2 levels (para) are readily thermalized to the gas kinetic temperature via collisions with H₂ molecules because the radiative interconversion between the two levels is extremely slow owing to the symmetry of the molecule. Therefore, the gas kinetic temperature is evaluated to be 20(2) K using the relative intensities of the K = 1 and 2 lines. Note that the relative intensity between the K = 0 (ortho) and 1 (para) lines is not adequate to determine the gas kinetic temperature because the relative intensity depends on both the gas kinetic temperature and the ortho-to-para ratio. The gas kinetic temperature is reported to be 13.9 K using the lines of CH₃C₂H (Sakai et al. 2008). Using the dipole moment μ_a = 1.2071 D (Bester et al. 1984) and the assumption of the excitation temperature within each K ladder of 12 K (Sakai et al. 2007), the column densities of the K = 0, 1, and 2 ladders are derived to be 2.71(5) × 10¹² cm⁻², 2.21(5) × 10¹² cm⁻², and 6.9(6) × 10¹¹ cm⁻², respectively, with the errors representing 1σ derived from the errors of the integrated intensities of the lines. The contributions of the K > 2 levels to the column density are negligible. For example, the contribution of the K = 3 level is expected to be only 6% of the K = 0 level. Under the assumption of no population in the K > 2 levels, the derived ortho-to-para ratio is 0.94(4), which is roughly consistent with a statistical column-density ratio of one. Therefore, the total column density of CH₃C₄H is determined to be 5.6(2) × 10¹² cm⁻², where a deviation of 1 K for the assumed excitation temperature causes a change in the total column density of 1.8%.

Zoom In Zoom Out Reset image size

3.4. Cumulene Carbene C₆H₂

3.5. Cyclic-C₃H

The cyclic-C₃H radical (hereafter c-C₃H) was first detected in TMC-1 CP by Yamamoto et al. (1987) and has been observed toward several galactic molecular clouds (Mangum & Wootten 1990), in the photodissociation region of the Horsehead Nebula (Pety et al. 2012), in diffuse clouds (Liszt et al. 2014), and in the circumstellar envelope of IRC+10216 (Cernicharo et al. 2000; Ford et al. 2004). In L1527, several emission lines have already been detected (K. Yoshida et al. 2017, in preparation). We observed the two emission lines of the ${N}_{{K}_{{\rm{a}}},{K}_{{\rm{c}}}}={2}_{\mathrm{1,1}}\mbox{--}{2}_{\mathrm{1,2}}$ transition (J = 2.5–2.5, F = 3–3 and 2–2) in L1527, as shown in Figure 5. The column density of c-C₃H is evaluated to be 5.65(9) × 10¹² cm⁻² under the assumption of an excitation temperature of 12 K (Sakai et al. 2007) and the dipole moment μ_a of 2.4 D (Yamamoto et al. 1987). When the assumed excitation temperature is changed by 1 K, the column density varies by 4%. This column density is comparable to that (6 × 10¹² cm⁻²) in TMC-1 CP (Yamamoto et al. 1987).

Zoom In Zoom Out Reset image size

3.6. Cyclic-C₃H₂O

The column densities of the various carbon-chain molecules in L1527 and TMC-1 CP are summarized in Table 2, and the column density ratios N_L1527/N_TMC-1 are plotted in Figure 2. While the column densities of the small species CH₃C₂H and c-C₃H₂ in L1527 are comparable to those in TMC-1, their ratios tend to decrease as the carbon-chain length increases, i.e., the longer carbon-chain molecules are less abundant. For example, the N_L1527/N_TMC-1 of CH₃C₄H is 0.3 (Remijan et al. 2006) and that of CH₃C₂H is approximately 1 (Irvine et al. 1981; Sakai et al. 2008). This trend is also seen in the cases of C_nH (n = 4, 6), cyanopolyynes H–(C≡C)_n–C≡N (n = 2, 3, 4), as reported by Sakai et al. (2008), and carbon-chain anions C_nH⁻ (n = 4, 6; McCarthy et al. 2006; Sakai et al. 2007; Cordiner et al. 2013).

However, the N_L1527/N_TMC-1 of C₇H is ≥0.40, i.e., C₇H is relatively abundant in L1527 despite its relatively long carbon-chain length. This means that this molecule is relatively abundant in L1527. According to the UMIST Database for Astrochemistry 2012 (McElroy et al. 2013) and the KInetic Database for Astrochemistry (KIDA; Wakelam et al. 2012), we find the following reaction schemes producing C₇H via shorter carbon chains (Smith et al. 2004; Walsh et al. 2009):

$$\begin{eqnarray}&&{{\rm{C}}}_{7}{{\rm{H}}}_{2}^{+}+{{\rm{e}}}^{-}\to {{\rm{C}}}_{7}{\rm{H}}+{\rm{H}},\end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray}&&{\rm{C}}+{\mathrm{HC}}_{6}{\rm{H}}\to {{\rm{C}}}_{7}{\rm{H}}+{\rm{H}},\end{eqnarray} \tag{ 2a }$$

$$\begin{eqnarray}&&{\rm{C}}+{{\rm{C}}}_{6}{{\rm{H}}}_{2}\to {{\rm{C}}}_{7}{\rm{H}}+{\rm{H}},\end{eqnarray} \tag{ 2b }$$

and

$$\begin{eqnarray}&&{\rm{C}}+{{\rm{C}}}_{6}{{\rm{H}}}^{-}\to {{\rm{C}}}_{7}{\rm{H}}+{{\rm{e}}}^{-},\end{eqnarray} \tag{ 3 }$$

where HC₆H is triacetylene. In reaction (1), ${{\rm{C}}}_{7}{{\rm{H}}}_{2}^{+}$ is produced by various condensation reactions of small hydrocarbon ions/neutrals. The rate coefficients for reactions (2a) and (3) are 3 × 10⁻¹⁰ cm³ s⁻¹ (Loison et al. 2014) and 1 × 10⁻⁹ cm³ s⁻¹ (Harada & Herbst 2008), respectively. Note that chemical model calculations do not generally distinguish the two isomers (HC₆H and C₆H₂).

The column density of C₆H⁻ in L1527 is reported to be 5.8 × 10¹⁰ cm⁻² (Sakai et al. 2007), which is comparable to that of 1 × 10¹¹ cm⁻² in TMC-1 CP (McCarthy et al. 2006). Here, the abundance of C₆H⁻ can be less than that of HC₆H because an anion species "M⁻" is expected to be less abundant than a neutral species "HM," e.g., the abundance of C₃N⁻ is estimated to be two orders of magnitude less than that of HC₃N according to the model by Harada & Herbst (2008). Therefore, the contribution of reaction (3) would be minor in comparison to those of reactions (1), (2a), and (2b). HC₆H would be abundant in L1527 because it is a stable carbon-chain molecule, and the overabundance of HC₆H would originate from the abundant C₇H in L1527. In fact, C₆H₂, the isomer of HC₆H, is also relatively abundant in the cumulene carbene series C_nH₂ despite its long carbon-chain length, as shown in Figure 2. C₆H₂ can be produced by the protonation of HC₆H by "HX⁺" and subsequent electron recombination (Langer et al. 1997). Hassel et al. (2008) estimated the C₆H₂/HC₆H ratio to be 0.02 by employing the abundance ratio of the carbon isomer cumulene carbene C₃H₂ to the total abundance of the C₃H₂ isomers. Even though it is uncertain whether this ratio can actually be used for the C₆H₂ case, C₆H₂ is likely less abundant than HC₆H. If so, reaction (2a) would be more important than reaction (2b). In any case, these reactions ((2a) and (2b)) may contribute to the production of C₇H, even though reaction (1) always operates to some extent. Note that HC₆H cannot be observed in the radio wavelengths owing to its lack of a permanent dipole moment.

The fractional abundances of C_nH (n ≥ 4) are compared with the results of the model calculation of L1527 reported by Hassel et al. (2008), who calculated the molecular abundances using the warm-up model. The observed fractional abundances of the shorter molecules C₂H and C₄H agree with the calculated ones in the case of the warmer temperature T_max = 30 K, which is the maximum dust temperature in the model simulation. Conversely, the observed values of the longer molecules C₅H, C₆H, and C₇H, including a substantial abundance gap of an order of magnitude between C₆H and C₇H, are better reproduced in the lower temperature case with T = 10 K, as shown in Figure 6. If the model accurately represents the findings, the spatial distributions of the longer carbon-chain molecules C_nH (n ≥ 5) could differ from those of the shorter species in L1527. Short carbon chains C_nH (n ≤ 4) are generally produced by the WCCC mechanism, while the larger ones, including C₇H, could be a remnant component from the cold starless core phase.

Zoom In Zoom Out Reset image size

The observed fractional abundances of the molecules CH₃C₂H, CH₃C₄H, C₃H₂, and C₄H₂ roughly agree with the calculated abundances in the case of the warmer temperature T_max = 30 K (Hassel et al. 2008). These molecules can be attributed to WCCC products, while HC₃N could be a remnant of the cold starless core phase (Araki et al. 2016). The remnants are observed together with the WCCC products toward the protostar. Sakai et al. (2009) reported that the WCCC source was formed by the faster collapse of the parent core, and Hassel et al. (2008) suggested that a faster collapse allows molecules to survive as remnants. The present results provide us with a further verification of the fast collapse scenario of L1527.

If the variance in the behavior of the column-density ratios between TMC-1 CP and L1527 in Figure 2 originates from the difference in the production schemes, the spatial distributions of the remnant molecules would differ from those of the WCCC products. Conversely, if the distributions of the longer carbon-chain molecules are the same as those of the shorter species, reactions producing some longer carbon-chain molecules (e.g., C₇H and C₆H₂) would need to be involved in the WCCC model to reproduce the observed abundances of the carbon-chain molecules in the warm region. In any case, our understanding of the chemistry of long carbon-chain molecules is far from complete not only in the WCCC source but also in cold molecular clouds. Further observational and modeling studies are awaited.

We thank the staff at the National Radio Astronomy Observatory for help with the observations. We are grateful to the anonymous reviewer for valuable comments. We thank Dr. David Frayer for his help in the observations with the GBT. MA thanks Grant-in-Aid for Scientific Research on Innovative Areas (grant No. 25108002), Grant-in-Aid for Scientific Research (C) (grant No. 15K05395), and the support from the grant from Institute for Quantum Chemical Exploration.