TMC-1, the starless core sulfur factory: Discovery of NCS, HCCS, H₂CCS, H₂CCCS, and C₄S and detection of C₅S

Abstract

We report the detection of the sulfur-bearing species NCS, HCCS, H₂CCS, H₂CCCS, and C₄S for the first time in space. These molecules were found towards TMC-1 through the observation of several lines for each species. We also report the detection of C₅S for the first time in a cold cloud through the observation of five lines in the 31-50 GHz range. The derived column densities are N(NCS) = (7.8±0.6)×10¹¹ cm⁻², N(HCCS) = (6.8±0.6)×10¹¹ cm⁻², N(H₂CCS) = (7.8±0.8)×10¹¹ cm⁻², N(H₂CCCS) = (3.7±0.4)×10¹¹ cm⁻², N(C₄S) = (3.8±0.4)×10¹⁰ cm⁻², and N(C₅S) = (5.0±1.0)×10¹⁰ cm⁻². The observed abundance ratio between C₃S and C₄S is 340, that is to say a factor of approximately one hundred larger than the corresponding value for CCS and C₃S. The observational results are compared with a state-of-the-art chemical model, which is only partially successful in reproducing the observed abundances. These detections underline the need to improve chemical networks dealing with S-bearing species.

1. Introduction

The cold dark core TMC-1 presents an interesting carbon-rich chemistry that leads to the formation of long neutral carbon-chain radicals and their anions, as well as cyanopolyynes (see Cernicharo et al. 2020a,b and references therein). The carbon chains CCS and CCCS are particularly abundant in this cloud (Saito et al. 1987; Yamamoto et al. 1987) and also exist in the envelopes of carbon-rich circumstellar envelopes (Cernicharo et al. 1987). TMC-1 is also peculiar due to the presence of protonated species of abundant large carbon chains such as HC₃O⁺ (Cernicharo et al. 2020c), HC₅NH⁺ (Marcelino et al. 2020), HC₃S⁺ (Cernicharo et al. 2021a), and CH₃CO⁺ (Cernicharo et al. 2021b). The number of sulfur-bearing species detected to date in TMC-1 is small compared to oxygen- and nitrogen-bearing species (see, e.g. McGuire et al. 2019). In fact, the chemistry of sulfur-bearing molecules is strongly dependent on the depletion of sulfur (Vidal et al. 2017). Many reactions involving S⁺ with neutrals as well as radicals with S and CS have to be studied to achieve a better chemical modelling of sulfur-bearing species (Petrie 1996; Bulut et al. 2021). Nevertheless, the main input to understand these chemical processes is to unveil new sulfur-bearing species in the interstellar medium as well as to understand their formation paths and their role in the chemistry of sulfur.

In this letter we report the discovery, for the first time in space, of the following five new sulfur-bearing species: NCS, HCCS, H₂CCS, H₂CCCS, and C₄S. The detection of C₅S in a cold dark cloud is also reported for the first time. A detailed observational study of the most relevant S-bearing species in this cloud is accomplished. We discuss these results in the context of state-of-the-art chemical models.

2. Observations

New receivers, built within the Nanocosmos project¹, and installed at the Yebes 40m radio telescope were used for the observations of TMC-1. The receivers and the spectrometers have been described by Tercero et al. (2021). The observations needed to complete the Q-band line survey towards TMC-1 (α _J2000 = 4^h41^m41.9^s and δ _J2000 = +25°41’27.0”) were performed in several sessions during December 2019 and January 2021. All data were analysed using the GILDAS package². The observing procedure has been previously described (see, e.g. Cernicharo et al. 2021a,b). The IRAM 30m data come from a line survey performed towards TMC-1 and B1 and the observations have been described by Marcelino et al. (2007) and Cernicharo et al. (2013).

The intensity scale, antenna temperature ($T_A^{*}$) was calibrated using two absorbers at different temperatures and an atmospheric transmission model (ATM; Cernicharo 1985; Pardo et al. 2001). Calibration uncertainties have been adopted to be 10 %. The nominal spectral resolution of 38.15 kHz was used for the final spectra.

3. Results and discussion

Line identification in our TMC-1 survey has been performed using the MADEX catalogue (Cernicharo 2012), the Cologne Database of Molecular Spectroscopy catalogue (Müller et al. 2005), and the JPL catalogue (Pickett 1998). A description of the methods used to fit the data and to derive column densities is provided in Appendix A.

3.1. HCCS and NCS

Among the unidentified spectral features, we have found a couple separated by 1 MHz that perfectly match the frequencies of the strongest hyperfine components of the J=7/2-5/2 transition of HCCS (²Π₃/₂). This species was observed in the laboratory by Kim et al. (2002) and the prediction of its rotational spectrum is available in the CDMS (Müller et al. 2005) and MADEX (Cernicharo 2012) catalogues. Taking into account the perfect match in frequencies, the narrow linewidth of the emission features in this source, and the perfect match in the relative intensities of the two hyperfine components, the possibility of a fortuitous coincidence is very low. Hence, we conclude that these two lines arise from HCCS. The observed lines are shown in Fig. 1 and the derived line parameters are given in Table D.1. The synthetic spectrum on Fig. 1 corresponds to T_r=5 K and N(HCCS) = 6.8×10¹¹ cm⁻². We searched for HCCCS but only a 3σ upper limit to its column density of 2.4×10¹¹ cm ² can be given (see Table 1).

Fig. 1.

Observed lines of HCCS (panel a) and NCS (panels b and c) towards TMC-1. The abscissa corresponds to the rest frequency assuming a local standard of rest velocity of the source of 5.83 km s⁻¹ (see text). Blanked channels correspond to negative features produced in the frequency switching data folding. The ordinate is the antenna temperature corrected for atmospheric and telescope losses in milliKelvin. Spectral resolution is 38.15 kHz. The red lines show the synthetic spectrum of HCCS and NCS for a rotational temperature of 5 K, a linewidth of 0.6 km s⁻¹, and a column density of 6.8×10¹¹ cm⁻² and 7.8×10¹¹ cm⁻², respectively.

Table 1. Column densities of sulfur-bearing species in TMC-1.

Molecule	Trot (K)	N obs (cm−2)	N calc (cm −2) e
CSa *	10.0	(3.50±0.40)×1014	3.1 × 1014
C34S*	10.0	(1.45±0.10)×1013
13C34S*	10.0	(1.45±0.20)×1011
HCS	5.0b	(5.50±0.50)×1012	5.1 × 1010
HSC	5.0b	(1.30±0.20)×1011
NCS	5.0b	(7.80±0.60)×1011	2.5 × 1010
HCCS	5.0b	(6.80±0.60)×1011	3.1 × 1010
H2CSc	10.0	(4.70±0.40)×1013	9.5 × 1012
CC34Sc	10.0	(1.80±0.18)×1012
o-H2CCS	7.0±1.0	(6.00±0.60)×1011	}1.8 × 1012
p-H2CCS	7.0±1.0	(1.80±0.20)×1011
0-H2CCCS	10.0±1.0	(3.00±0.30)×1011	}7.8 × 1010
p-H2CCCS	10.0±1.0	(6.50±0.60)×1010
CCS*	5.1±0.2	(5.50±0.65)×1013	3.8 × 1011
CC34S*	3.6±0.4	(5.00±0.50)×1012
CCCS*	5.8±0.2	(1.30±0.13)×1013	1.1 × 1013
CCC34S*	6.7±0.2	(5.30±0.50)×1011
C4S	7.0±1.0	(3.80±0.50)×1010	1.6 × 1010
C5S	7.0±1.0	(5.00±1.00)×1010	1.9 × 1011
HCS+ *	10.0	(1.00±0.10)×1013	1.7 × 1011
HC34s+ *	10.0	(7.10±0.70)×1011
HC3s+ *	10.0±2.0	(4.00±1.50)×1011	2.8 × 1011
HSCN	5.0	(5.80±0.60)×1011	3.7 × 109
HNCS	5.0	(3.80±0.40)×1011	1.2 × 1010
HC3Sd	5.0	≤2.40×1011	1.5 × 1011
HCNSd	10.0	≤6.00×1010	3.9 × 106
HSNCd	10.0	≤2.00×1010
HSCS+d	5.0	≤4.00×1012
HCCSHd	7.0	≤3.00×1012
HCCCSHd	7.0	≤2.40×1011

Notes.Entries in bold face correspond to the molecular species detected for the first time in space, or in TMC-1 (C₅S).

^(*)Data from Cernicharo et al. (2021a).

^(a)Derived from C³⁴S and the C₃S/C₃ ³⁴S abundance ratio.

^(b)Rotational temperature assumed identical to that of CCS.

^(c)Column density refers to the ortho or para species. Assuming an ortho/para ratio of 3, the total column density can be estimated by multiplying the para value by a factor of 4.

^(d)Upper limits correspond to 3<values derived assuming the indicated rotational temperature and a linewidth of 0.6 km s⁻¹.

^(e)Maximum fractional abundance calculated in the 10⁵-10⁶ yr range was converted to column density using N(H₂) = 10²² cm⁻² (Cernicharo & Guélin 1987).

Thiocyanogen, NCS, has been observed in the laboratory (Amamo & Amano 1991; McCarthy et al. 2003; Maeda et al. 2007), but never detected in space. Only one rotational transition lies within the Q-band, the J=5/2-3/2 transition of NCS in its² Π₃/₂ ladder. Figure 1 shows the three hyperfine components of this transition observed in TMC-1. The match between observations and the synthetic spectrum, corresponding to T_r=5 K and N(NCS) = 7.8×10¹¹ cm⁻², is remarkably good, ensuring the discovery of this sulfur compound in TMC-1. The oxygen analogue of thiocyanogen, the isocyanate radical (NCO), was detected in cold core L483 by Marcelino et al. (2018). Unfortunately, NCO does not have lines in the 31-50 GHz range.

3.2. H₂CCS and H₂CCCS

Prompted by the detection of HCCS, we searched for other sulfur-bearing species in our survey. McGuire et al. (2019) searched for H₂CCS, thioketene, in TMC-1. They obtained an upper limit to its column density of 5.5×10¹² cm⁻². Unfortunately they used a transition with an upper level energy around 40 K, which is not the best for the physical conditions of TMC- 1. In this work we report the discovery of thioketene in space through observations of transitions with smaller upper level energies.

Spectroscopic laboratory data of thioketene (Georgiou et al. 1979; Winnewiser & Schäfer 1980; McNaughton et al. 1996) and H₂CCCS (Brown et al. 1988) were used to predict the frequencies of these species and implement them in the MADEX code. The dipole moments adopted for H₂CCS and H₂CCCS are 1.02 D (Georgiou et al. 1979) and 2.06 D (Brown et al. 1988), respectively. We detected the four ortho and two para transitions of H₂CCS expected in our data, as shown in Fig. 2. For H₂CCCS (propadienthione), six ortho and three para transitions are detected (see seven of them in Fig. 3). The derived line parameters for both species are given in Table D.1.

Fig. 2.

Observed transitions of H₂CCS towards TMC-1. The abscissa corresponds to the velocity in km s⁻¹ of the source with respect to the local standard of rest. Line parameters are given in Table D.1. The ordinate is the antenna temperature corrected for atmospheric and telescope losses in milliKelvin. Blanked channels correspond to negative features produced by the folding of the frequency switching observations. The vertical blue dashed line indicates the v_LSR of the cloud (5.83 kms⁻¹). The red line spectra correspond to the synthetic model spectrum for each line adopting T_r=7 K, N(o-H₂CCS) = 6.0×10¹¹ cm⁻², and N(p- H₂CCS) = 1.8×10¹¹ cm⁻² (see text).

Fig. 3.

Selected transitions of H₂CCCS towards TMC-1. The abscissa corresponds to the velocity in km s⁻¹ of the source with respect to the local standard of rest. Line parameters are given in Table D.1. The ordinate is the antenna temperature corrected for atmospheric and telescope losses in milliKelvin. Blanked channels correspond to negative features produced by the folding of the frequency switching observations. The vertical blue dashed line indicates the v_LSR of the cloud (5.83 kms⁻¹). The red line spectra correspond to the synthetic model of the emission obtained for T_r =10 K, N(o-H₂CCS) = 3.0×10¹¹ cm⁻², and N(p-H₂CCS) = 6.5×10¹⁰ cm⁻².

An analysis of the data through a standard rotation diagram provides rotational temperatures of 7±1 K and 10±1 K for H₂CCS and H₂CCCS, respectively. For H₂CCS, we derived N(ortho) = (6.0±0.6)×10¹¹ cm⁻² and N(para) = (1.8±0.2)×10¹¹ cm⁻². Hence, the ortho/para ratio for this species is 3.3±0.7. For H₂CCCS, we derived N(ortho) = (3.0±0.3)×10¹¹ cm⁻² and N(para) = (6.5±0.5)×10¹⁰ cm⁻², respectively. The ortho/para abundance ratio derived for this species is 4.6±0.8. These ortho/para ratios are compatible with the 3/1 value expected from the statitiscal spin degeneracies. They imply that non-significant enrichment of the para species is produced through reactions of H₂CCS and H₂CCCS with ${\text{H}}_3^{+}$, or other protonated molecular cations. The synthetic spectrum computed with these parameters for both species is shown in red in the different panels of Figures 2 and 3. The derived abundance ratio between H₂CCS and H₂CCCS is ≃2, which is very different from the H₂CCO/H₂CCCO abundance ratio of > 130 derived in the same source by Cernicharo et al. (2020c), but very similar to the C₂S/C₃S abundance ratio of ~3 derived here (see Table 1).

3.3. C₄S and C₅S

Taking the large column density derived for CCS and C₃S into account (Cernicharo et al. 2021a), the next member of this family, C₄S, is a potential candidate to be present in TMC-1. The spectroscopic laboratory data used to predict the spectrum of C₄S are from Hirahara et al. (1993) and Gordon et al. (2001). A dipole moment of 4.03 D was computed through ab initio calculations by Pascoli & Lavendy (1998) and Lee (1997). Nineteen transitions of this species have frequencies within the range of our line survey. The line by line search through our data provides a clear detection for the N_u=10 up to 13, J_u=N_u+1 transitions (see Fig. B.1). A rotational temperature of 7±1 K was derived from these lines. The model fitting method (see Appendix A) was used to derive a column density of (3.8±0.4)×10¹⁰ cm⁻². The stacked spectrum obtained for these lines and for the weaker J_u=N and J_u=N-1 transitions are shown in panels (b) and (c) of Fig. B.1, respectively.

The next member of the CnS family, C₅S, was tentatively detected towards the carbon-rich star CW Leo by Bell et al. (1993) and confirmed by Agúndez et al. (2014). Laboratory spectroscopy from Gordon et al. (2001) and a dipole moment of 4.65 D (Pascoli & Lavendy 1998) have been adopted. Five lines from J_u=17 up to 21 have been detected (see Fig. B.2). The model fitting provides a rotational temperature of 7±2K and a column density of (5.0±1.0)×10¹⁰ cm⁻². The abundance ratio between C₄S and C₅S is of the order of unity, and the abundance ratio C2S/C₃S/C₄S/C₅S is 5500/1300/3.8/5.0. The change in abundance for C₄S and C₅S relative to C₂S is of three orders of magnitude. This result is very different than the one obtained by Agúndez et al. (2014) for the carbon-rich star IRC+10216. In this object, the derived C₃S/C₅S ratio is ~ 1-10 (depending on the assumed rotational temperature), versus ~260 in TMC-1. However, the C₂S/C₃S abundance ratio is the same for both sources, that is to say ~3. The radical C₄S has not been detected yet in IRC+10216 (Agúndez et al. 2014). S-bearing carbon chains do not follow the smooth decrease in abundance observed in cold dark clouds and circumstellar envelopes for other carbon chains such as cyanopolyynes (HC_2n+1N; a factor 3-5 between members of this molecular species).

3.4. Chemical models

The chemistry of sulfur-bearing molecules in cold dark clouds was recently discussed by Vidal et al. (2017), Vastel et al. (2018), and Laas & Caselli (2019), based on new chemical network developments. These studies revealed that the chemistry of sulfur strongly depends on the poorly constrained degree of depletion of this element in cold dense clouds. Chemical networks are relatively incomplete when dealing with S-bearing species. For example, from the six species detected in this work, only C₄S is included in the chemical networks RATE12 (UMIST; McElroy et al. 2013) and kida.uva.2014 (KIDA; Wakelam et al. 2015). Vidal et al. (2017) made an effort to expand the number of reactions involving S-bearing species significantly by including several of the molecules reported here. These authors, however, discussed only a small number of sulfur compounds and did not provide calculated abundances for any of the six species reported in this work. We therefore carried out chemical modelling calculations to describe the chemistry of the new sulfur-bearing molecules detected. We used the gas-phase chemical network RATE12 from the UMIST database (McElroy et al. 2013), expanded with the set of gas-phase reactions involving S-bearing species constructed by Vidal et al. (2017). We included additional reactions to describe the chemistry of NCS and C₅S, which was not treated by Vidal et al. (2017), assuming a similar chemical kinetics behaviour to NCO and C₃S, respectively (see Table E.1). Our main purpose is to see whether state-of-the-art gas-phase chemical networks can explain the abundances of the S-bearing species discovered. We adopted typical parameters of cold dark clouds: a gas kinetic temperature of 10 K, a volume density of H₂ of 2 × 10⁴ cm⁻³, a visual extinction of 30 mag, a cosmic-ray ionisation rate of H₂ of 1.3 × 10⁻¹⁷ s⁻¹, and ‘low-metal’ elemental abundances (see, e.g. Agúndez & Wakelam 2013). We therefore adopted a relatively low gas-phase abundance of sulfur, 8 × 10⁻⁸ relative to H.

Calculated column densities for S-bearing species are compared with observed values in Table 1. In Fig. 4 we compare the abundances calculated for the six S-bearing species detected in this study with the values derived from the observations. The peak abundances calculated are within one order of magnitude of the observed abundance for H₂CCS, H₂CCCS, C₄S, and C₅S. The two species for which the chemical model severely disagrees with observations are NCS and HCCS, in which case calculated abundances are lower than observed by more than one order of magnitude. These two species have low calculated abundances because they are assumed to react quickly with O atoms, with rate coefficients ≳ 10⁻¹⁰ cm³ s⁻¹. This same behaviour has been previously found for other radicals detected in cold dark clouds such as HCCO (Agúndez et al. 2015) and C₂S (Cernicharo et al. 2021a; see also Table 1). In these cases, the fast destruction with neutral atoms, including O, resulted in calculated abundances well below the observed values. These facts suggest that either the abundance of O atoms calculated by gas-phase chemical models of cold clouds is too high or O atoms are not as reactive with certain radicals as currently thought. While the low-temperature reactivity of O atoms with closed-shell S-bearing species such as CS has been shown to be low (Bulut et al. 2021), its reactivity with radicals is very poorly known.

Fig. 4.

Calculated fractional abundances of the six S-bearing species reported in this work as a function of time. Horizontal dotted lines correspond to observed values in TMC-1 adopting a column density of H₂ of 10²² cm⁻² (Cernicharo & Guélin 1987).

The radical NCS is formed in the chemical model through the reactions CN + SO and N + HCS. The kinetics and product distribution of these reactions is, however, largely unconstrained. This uncertainty also affects the chemical network involving the family of CHNS isomers, for which the chemical model underestimates the column densities (see Table 1). In the chemical models of dark clouds performed by Adande et al. (2010) and Vidal et al. (2017), the main formation pathways to these molecules involve grain surface reactions, which are not included in our chemical network.

The formation of the other five molecules reported in this work, which are S-bearing carbon chains and thus can be represented by the formula H_mC_nS, occurs through two types of chemical routes. The first one involves neutral-neutral reactions. Along this pathway, HCCS is formed by the reactions C + H₂CS and OH + C₃S, the reaction S + C₂H₃ yields H₂CCS, while H₂CCCS is mostly formed by the reaction S + CH₂CCH. On the other hand, the carbon chain C₄S is formed through the reactions S + C₄H and C + HC₃S, while C₅S is produced in the reactions C₄H + CS and S + C₅H. It must be noted that the kinetics and product distribution of these reactions is very poorly known. Most of these reactions are assumed to proceed with capture rate coefficients by Vidal et al. (2017).

The second pathway consists of reactions involving cations, which ultimately lead to the ion H_pC_n S⁺, which dissociatively recombines with electrons to yield H_mC_nS, where typically p = m + 1. In this route, the ions H₂CCS⁺, H₃CCS⁺, H₃CCCS⁺, HC₄S⁺, and HC₅S⁺ are the precursors of HCCS, H₂CCS, H₂CCCS, C₄S, and C₅S, respectively. The abovementioned ions are in turn formed through reactions between atomic S with hydrocarbon ions and S⁺ with neutral hydrocarbons. In the same line of the reactions discussed above, there are large uncertainties regarding these reactions, for which rate coefficients and product distributions are taken from Vidal et al. (2017). The chemical network is probably incomplete in that it misses important reactions involving S and S⁺. Moreover, reactions on grain surfaces, which are considered by Vidal et al. (2017) but are not taken into account here, could play an important role.

4. Conclusions

In this work, we present the detection of five new sulfur-bearing molecules in TMC-1: NCS, HCCS, H₂CCS, H₂CCCS, and C₄S. In addition, the species C₅S previously found only towards carbon-rich circumstellar envelopes is also detected in a cold dark cloud for the first time. Chemical models fail to reproduce the observed column densities which implies that the chemical networks are incomplete and that laboratory and theoretical work has to be performed in order to understand the chemistry of sulfur in cold prestellar cores.