The sulphur saga in TMC-1: Discovery of HCSCN and HCSCCH

Abstract

We report the detection, for the first time in space, of cyano thioformaldehyde (HCSCN) and propynethial (HCSCCH) towards the starless core TMC-1. Cyano thioformaldehyde presents a series of prominent a- and b-type lines, which are the strongest previously unassigned features in our Q-band line survey of TMC-1. Remarkably, HCSCN is four times more abundant than cyano formaldehyde (HCOCN). On the other hand, HCSCCH is five times less abundant than propynal (HCOCCH). Surprisingly, we find an abundance ratio HCSCCH/HCSCN of ∼ 0.25, in contrast with most other ethynyl-cyanide pairs of molecules for which the CCH-bearing species is more abundant than the CN-bearing one. We discuss the formation of these molecules in terms of neutral-neutral reactions of S atoms with CH₂CCH and CH₂CN radicals as well as of CCH and CN radicals with H₂CS. The calculated abundances for the sulphur-bearing species are, however, significantly below the observed values, which points to an underestimation of the abundance of atomic sulphur in the model or to missing formation reactions, such as ion-neutral reactions.

1. Introduction

The chemistry of sulphur-bearing molecules in cold interstellar clouds is an active area of research (Vidal et al. 2017; Cernicharo et al. 2021a,b). Recently, Cernicharo et al. (2021b) reported the detection of five new sulphur-bearing species in TMC-1, a source that presents an interesting carbon-rich chemistry that leads to the formation of long neutral carbon chain radicals and their anions, cyanopolyynes (see Cernicharo et al. 2020a,b and references therein), pure hydrocarbon cycles, and polycyclic aromatic hydrocarbons (Cernicharo et al. 2021c; Burkhardt et al., 2021; McCarthy et al. 2021). A large variety of S-bearing molecules are also present in this source, which leads to new questions regarding the formation routes to these species. It is clear that many ion-neutral and neutral-neutral reactions involving S-bearing species have to be studied to improve the reliability of chemical models (Petrie 1996; Bulut et al. 2021). Unveiling new sulphur-bearing species can also help to understand the chemistry of sulphur by comparing the observed abundances with those predicted by chemical models.

In this letter we report the discovery of two new sulphur-bearing molecules: HCSCN (cyano thioformaldehyde) and HC-SCCH (propynethial). Figure 1 shows the structure of these molecules. The derived abundances are compared with the oxygen analogues of these species, HCOCN (cyano formaldehyde) and HCOCCH (propynal). We discuss plausible reactions that could lead to the formation of these species with the aid of a chemical model.

Fig. 1. Structure of cyano thioformaldehyde, HCSCN, and ethynyl thioformaldehyde, HCSCCH, also known as propynethial.

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. A detailed description of the system is given by Tercero et al. (2021). The Q-band receiver consists of two high electron mobility transistor cold amplifiers covering the 31.0-50.3 GHz band with horizontal and vertical polarisations. The backends are 2×8×2.5 GHz fast Fourier transform (FT) spectrometers with a spectral resolution of 38.15 kHz that provide the whole coverage of the Q-band in both polarisations. The main beam efficiency varies from 0.6 at 32 GHz to 0.43 at 50 GHz.

The line survey of TMC-1 (α _J2000 = 4^h41^m41.9^s and δ _J2000 = +25°41′27.0″) in the Q-band was performed in several sessions. The first results (Cernicharo et al. 2020a,b) were based on two observing runs performed in November 2019 and February 2020. Two different frequency coverages were achieved, 31.0849.52 GHz and 31.98-50.42 GHz, in order to check that no spurious spectral ghosts are produced in the down-conversion chain. Additional data were taken in October 2020, December 2020, and January-April 2021. The observing procedure was frequency switching with a frequency throw of 10 MHz for the two first runs and of 8 MHz for all the others. Receiver temperatures in the runs achieved in 2020 vary from 22 K at 32 GHz to 42 K at 50 GHz. Some power adaptation in the down-conversion chains reduced the receiver temperatures in 2021 to 16 K at 32 GHz and 25 K at 50 GHz. The sensitivity of the survey varies along the Q-band between 0.25 and 1 mK, which is a considerable improvement over previous line surveys in the 31-50 GHz frequency range (Kaifu et al. 2004). The acronym we have assigned to this line survey is QUIJOTE (Q-band Ultrasensitive Inspection Journey to the Obscure Tmc-1 Environment).

The intensity scale used in this work, the antenna temperature $(T_A^{*}),$ was calibrated using two absorbers at different temperatures and the atmospheric transmission model ATM (Cernicharo 1985; Pardo et al. 2001). Calibration uncertainties of 10 % were adopted. All data were analysed using the GILDAS package².

3. Assignment of unidentified lines to HCSCN

Our millikelvin-sensitive line survey of TMC-1 has revealed hundreds of unknown features whose identification challenges our state-of-the-art understanding of the chemistry of this object (Cernicharo et al. 2021b,c). Line identification in this work was done using the catalogues MADEX (Cernicharo 2012), CDMS (Müller et al. 2005), and JPL (Pickett et al. 1998).

One surprising result in our data is the presence of a few strong unknown features with intensities around 10 mK. Three of them are in an excellent harmonic relation of 6:7:8. A fit to the standard Hamiltonian of a linear molecule provides a rotational constant, B, of 3082.8334±0.0029 MHz and a distortion constant, D, of 40.890±0.027 kHz. Such a large distortion constant points towards an asymmetric rotor. Hence, these three lines could correspond to the transitions K_a =0 of J_up =6, 7, and 8 of a molecule with (B + C)/2=3082.8 MHz. This suggests that the molecule could have a total of either four or five N, C, and/or O atoms. We explored around these features, looking for the K_a =1 lines. After some iterative work, we found six of these lines that can be fitted with a reduced number of rotational and distortion constants for a total of nine transitions. These lines are shown in Fig. 2. The assigned quantum numbers correspond to a-type transitions. The carrier is definitely an asymmetric rotor. Hence, we could expect to have b-type transitions if the molecule has a large μ_b. From the rotational and distortion constants derived from the a-type transitions, we predicted the spectrum for b-type transitions and found nine of them, as shown in Fig. 3. Moreover, these b-type transitions show a series of lines around the predicted frequencies, which is reminiscent of the hyperfine structure of an N-nucleus. Assuming this is the case, we fitted all these lines and obtained the rotational constants provided in Table 1. The derived values for χ_aa and χ_bb definitively correspond to a CN group. The hyperfine structure of the a-type transitions is not resolved with our spectral resolution. Hence, the three strongest components appear as a single feature. No c-type transitions were found at the level of sensitivity of our survey.

Fig. 2.

Selected a-type transitions of HCSCN. The abscissa corresponds to the rest frequency of the lines assuming a local standard of rest velocity of the source of 5.83 km s⁻¹. Frequencies and intensities for the observed lines are given in Table A.1. The ordinate is the antenna temperature, corrected for atmospheric and telescope losses, in millikelvin. The quantum numbers for each transition are indicated in the upper left corner of the corresponding panel. The red lines show the computed synthetic spectrum for this species for T_r =5 K and a column density of 1.3×10¹² cm⁻². Blanked channels correspond to negative features produced in the folding of the frequency switching data.

Fig. 3. Same as Fig. 2 but for the b-type transitions of HCSCN. The red lines show the computed synthetic spectrum for this species for T_r=5 K and a column density of 1.3×10¹² cm⁻².

Table 1. Spectroscopic parameters for HCSCN, HC³⁴SCN, and HCSCCH.

Constant	TMC-1a	Laboratory b	Merged fitc	Laboratory 34Sd	Bogey 1989 e	Zaleski 2012 f	HCSCCHg
A (MHz)	42910.051(11)	42910.0474(25)	42910.0478(21)	42604.3412h	43314.0052(34)	42909.9686(83)	42652.03185(40)
B (MHz)	3195.4016(17)	3195.40149(55)	3195.40171(26)	3116.7454(14)	3205.67686(24)	3195.3952(17)	3109.382386(28)
C (MHz)	2970.1303(17)	2970.13104(54)	2970.13067(28)	2900.6315(14)	2975.32470(23)	2970.1273(18)	2894.265480(28)
DJ (kHz)	1.2767(93)	1.2812(76)	1.2792(28)	1.228(30)	1.26557(17)	1.218(11)	1.1298295(90)
DJK (kHz)	-104.31(33)	-104.29(14)	-104.33(12)	-102.09(26)	-102.0177(24)	-105.78(38)	-104.85202(18)
d 1 (kHz)	-0.2288(64)	-0.2228(96)	-0.2277(25)	-0.247(41)	-0.228786(42)	0.2173(93)	-0.2110838(23)
χaa MHz	-3.46(16)	-3.5118(20)	-3.5117(20)	-3.4952(33)		-3.53(20)
χbb MHz	0.83(12)	0.8384(22)	0.8385(22)	0.828(11)		0.86(30)
Number of lines	32i	53	85	21	152	36	77
σ(kHz)	11.8	1.3	8.9	1.2	28.8	22.1	7.5
J max, Kamax	9,1	4,1	9,1	3,1	58,20		20,5
νmax (GHz)	50.204	25.112	50.204	18.376	279.98		51.3

^aValues between parentheses correspond to the uncertainties of the parameters in units of the last significant digits. Fit to the line frequencies measured in TMC-1.

^bFit to the laboratory line frequencies measured in this work.

^cMerged fit using our laboratory measurements and the frequencies measured in TMC-1. This is the recommended set of constants to predict the frequencies of HCSCN.

^dFit to the laboratory line frequencies of HC³⁴SCN measured in this work.

^eRotational constants for HCSCN from Bogey et al. (1989): They probably correspond to a vibrational state of HCSCN but not to the ground state of this species (see Appendix B). Additional distortion constants were obtained by these authors, but they are not relevant for the present work.

^fRotational constants for HCSCN from Zaleski et al. (2012). The d1 constant corresponds in this case to -δJ.

^gRotational constants for HCSCCH derived from a fit to the data observed for this species by Brown & Godfrey (1982), Cabtree et al. (2016), and Margulès et al. (2020). The full set of distortion constants are provided by Margulès et al. (2020).

^hFixed to the theoretical value.

ⁱThe a-type Transitions are unresolved and have been fitted to the centroid of the hyperfine components.

3.1. Identifying the carrier: Laboratory microwave spectroscopy experiments

We explored, using ab initio calculations, a significant number of the combinations of H_xC_yN_z and H_x C_yO_z without finding a good agreement with the constants of our new species.Molecules such as bent-HC₄N and c-C₃HCN have rotational constants close, within 10%, to ours. Several years ago we implemented a species in MADEX that matches our constants rather well. It is HCSCN (see Table 1), and it was observed in the laboratory by Bogey et al. (1989). Another species with very close constants to ours (see Table 1), HCSCCH, is also implemented in MADEX, and we found several of its lines in our data (see Sect. 4). Hence, HCSCN could be a good candidate for TMC-1. Unfortunately, with the rotational constants provided by Bogey et al. (1989), the species is not detected. We explored the literature for additional laboratory data for HCSCN and found a contribution to the Ohio State 67th International Symposium on Molecular Spectroscopy (Zaleski et al. 2012). In the presentation, the reported rotational constants for HCSCN agree perfectly with those derived in TMC-1 (see Table 1); Zaleski et al. (2012) also claim a difference between their data and those of Bogey et al. (1989). The detection of HCSCCH, and the good agreement between the TMC-1 constants and those of Zaleski et al. (2012), points to an uncorrect assignment of the HCSCN lines in the Bogey et al. (1989) work. A detailed search for a publication of Zaleski et al. data in a refereed journal has, regrettably, been unsuccessful.

In order to confirm that the carrier of our prominent unknown features in TMC-1 is indeed HCSCN, we performed laboratory measurements. The rotational spectrum of HCSCN was measured by a Balle-Flygare-type FT microwave spectrometer combined with a pulsed discharge nozzle (Endo et al. 1997; Cabezas et al. 2016). HCSCN was produced in a supersonic expansion by a pulsed electric discharge of a gas mixture of CH₃CN (0.2%) and CS2 (0.2%) diluted in Ar and by applying a voltage of 2.0kV through the throat of the nozzle source. We observed a total of nine a-type and two b-type Transitions (see Fig. B.1) at the predicted frequencies using the TMC-1 observed lines. To confirm the identity of the spectral carrier, we observed the HC³⁴SCN isotopic species in their natural abundance. Their rotational Transitions (see Fig. B.1) appear exactly at the expected frequencies considering the isotopic shift for ³⁴S. All the observed lines for HCSCN and HC³⁴SCN are shown in Tables B.1 and B.2.

We performed ab initio calculations at the MP2/6-311++G(d,p) level of theory to optimise the HCSCN molecular structure and calculate the electric dipole moment components. We obtained μ_a and μ_b values of 2.03 and 1.69 D. Details of these calculations are given in Appendix B. Our calculations also indicate that the observed rotational Transitions reported by Bogey et al. (1989) corresponds to the lowest excited vibrational state of HCSCN, ν ₉, which lies at 198.1 cm⁻¹ above the ground vibrational state.

The merged fit to the TMC-1 frequencies and the fit to our laboratory measurements provide a set of rotational constants (see Table 1) that can be used to predict the frequencies of HC-SCN up to 100 GHz with calculated uncertainties below 50 kHz for K_a=0 and 1. Within the present sensitivity of our survey, no lines of HC³⁴SCN, which are expected to be 20 times weaker than those of HCSCN, were found.

3.2. Column density of HCSCN

We used a model fitting technique (see Cernicharo et al. 2021a) in which the rotational temperature, the column density, and the μ_a/μ_b ratio are varied. A uniform brightness temperature source with a diameter of 80” was assumed (Fossé et al. 2001). The best fit provides a rotational temperature of 5.0±0.5 K, a dipole moment component ratio, μ_a/μ_b, of 0.70±0.05, and a column density of (6.2±0.5)/μ_a ²×10¹² cm⁻², where μ_a is the component of the dipole moment along the a-axis of the molecule. Adopting the value of the total dipole moment derived by our ab initio calculations and the μ_a/μ_b ratio of 0.7 obtained from our observations, we deriveμ_a=2.16D. Hence, the column density of HCSCN is (1.3±0.1)×10¹² cm⁻². Adopting a column density of H₂ of 10²² cm⁻² for TMC-1 (Cernicharo & Guélin 1987), the abundance of HCSCN relative to H₂ is 1.3 × 10⁻¹⁰.

The column density of cyano formaldehyde, HCOCN, in TMC-1 is (3.5±0.5)×10¹¹ cm⁻²(see Appendix C). Hence, the abundance ratio HCSCN/HCOCN is ∼4, which is surprising considering that the elemental gas-phase abundance of oxygen is expected to be much larger than that of sulphur. Moreover, the column density of formaldehyde is 5×10¹⁴ cm⁻² (Agúndez & Wakelam 2013), while that of H₂CS is 4.7×10¹³ cm⁻² (Cernicharo et al. 2021a). Hence, the abundance ratio HCSCN/HCOCN is a factor of 40 larger than that of H₂CS/H₂CO (∼0.1).

4. Detection of propynethial, HCSCCH

The detection of HCSCN suggests that other derivatives of thioformaldehyde could be present in TMC-1. The obvious candidate is ethynyl thioformaldehyde (HCSCCH), also known as propynethial. This molecule has been observed in the laboratory up to 630 GHz and J_max=107 (Brown & Godfrey 1982; Cabtree et al. 2016; Margulès et al. 2020). Its dipole moment components have been measured to be μ_α=1.763 D and μ_b=0.661 D (Brown & Godfrey 1982). Six lines of this molecule were found in our data, and they are shown in Fig. 4. The lines are considerably weaker than those of HCSCN. We searched for more Transitions of this species, but their expected intensity is below the present sensitivity of the survey. Nevertheless, the detection of all the strongest lines of HCSCCH in the Q-band is solid and robust. None of the detected lines is blended with any other feature from known species. This species was searched for towards different sources, including TMC-1, by Margulès et al. (2020) with negative results. However, the T_ransition they searched for towards TMC-1 has an upper energy level of 13.7K and is predicted to be very weak by our models. Assuming the same source parameters and rotational temperature as for HCSCN, we derive a column density for this species of (3.2±0.4)×10¹¹ cm⁻², which is eight times lower than the upper limit derived by Margulès et al. (2020). Therefore, we obtain an abundance ratio HCSCN/HCSCCH of 4±1. The column density of HCOCCH in this source is (1.5±0.2)×10¹² cm⁻²(see Appendix D). Hence, the abundance ratio HCOCCH/HCSCCH is 5±1, which is of the order of the H₂CO/H₂CS abundance ratio (∼10; see above).

Fig. 4. Same as Fig. 2 but for the observed Transitions of HCSCCH towards TMC-1. The red line shows the computed synthetic spectrum for propynethial assuming T_r = 5 K and N(HCCCHS) = 3.2 × 10¹¹ cm⁻².

HCSCCH is an isomer of H₂CCCS, a molecule recently found in TMC-1 by Cernicharo et al. (2021b). The column density of the latter has been derived to be (3.7±0.3)×10¹¹ cm⁻².

Hence, the abundance ratio between both isomers is ~1. The energy difference between them is 2 kcal mol⁻¹, the H₂CCCS isomer being the more stable one (Cabtree et al. 2016). Finally, we could expect HC₄CHS, which could result from the reaction of C₄H with H₂CS, to be present in TMC-1. Microwave laboratory data for this species are available (McCarthy et al. 2017). We searched for its strongest Transitions, but we only derive a 3σ upper limit to its column density of 10¹² cm⁻².

5. Discussion

The molecules HCSCN and HCSCCH are chemically related as they share an HCS subunit in which the carbon atom is bonded to a CN or a CCH group. Moreover, the oxygen analogues HCOCN and HCOCCH are also chemically related to the two aforementioned molecules since they result from the substitution of the sulphur atom with an oxygen atom. One can therefore conceive of similar chemical routes to these four molecules. A first formation pathway can be provided by

$$\text{O}+{\text{CH}}_2\text{CCH}\to \text{HCOCCH}+\text{H}$$ (1a)

$$\text{O}+{\text{CH}}_2\text{CN}\to \text{HCOCN}+\text{H}$$ (1b)

$$\text{S}+{\text{CH}}_2\text{CCH}\to \text{HCSCCH}+\text{H}$$ (1c)

$$\text{S}+{\text{CH}}_2\text{CN}\to \text{HCSCN}+\text{H}$$ (1d)

These reactions have not been studied, but one could reasonably expect them to be fast at low temperatures. Loison et al. (2016) assume that reaction (1a) is rapid and find that it is the main formation route to HCOCCH in cold interstellar clouds. A second formation route can be provided by

$${\text{H}}_2\text{CO}+\text{CCH}\to \text{HCOCCH}+\text{H}$$ (2a)

$${\text{H}}_2\text{CO}+\text{CN}\to \text{HCOCN}+\text{H}$$ (2b)

$${\text{H}}_2\text{CS}+\text{CCH}\to \text{HCSCCH}+\text{H}$$ (2c)

$${\text{H}}_2\text{CS}+\text{CN}\to \text{HCSCN}+\text{H}$$ (2d)

Quantum chemical calculations indicate that reaction (2a) could be rapid at low temperatures since the small calculated barrier is within the uncertainty of the calculations (Dong et al. 2005), while reaction (2b) is barrier-less (Tonolo et al. 2020). These authors also suggest that the reaction CN + CH₃CHO can plausibly form HCOCN, although it is probably less efficient than reaction (2b) since CH₃CHO is about 80 times less abundant than H₂CO in TMC-1 (Agúndez & Wakelam 2013).

We implemented the aforementioned neutral-neutral reactions with a rate coefficient of 10⁻¹⁰ cm³ s⁻¹ in a gas-phase chemical model, in which we adopted typical physical conditions of cold dark clouds (see e.g. Agúndez & Wakelam 2013) and used the chemical network UMIST RATE12 (McElroy et al. 2013), revised and expanded according to Loison et al. (2017) and Vidal et al. (2017). The chemical model reproduces the observed abundances of the O-bearing molecules HCOCCH and HCOCN reasonably well. Moreover, the observed abundance ratio HCOCCH/HCOCN of ~ 4 is of the same order as the observed abundance ratios CCH/CN (~ 10; Pratap et al. 1997) and CH₂CCH/CH₂CN (~ 6; Agúndez et al. 2021; Cabezas et al. 2021), which is consistent with the chemical routes proposed. In the case of the S-bearing molecules HCSCCH and HCSCN, the chemical model underestimates their abundances by two orders of magnitude, which may be related to an excessively low abundance of atomic sulphur calculated by the chemical model or to missing formation reactions, such as ion-neutral reactions.

6. Conclusions

We have reported the discovery of HCSCN and HCSCCH in TMC-1. While HCSCCH is five times less abundant than its oxygen analogue, propynal, we find that HCSCN is four times more abundant than HCOCN. Further studies on plausible formation reactions of HCSCN and HCSCCH are needed.

Fig. 5. Calculated fractional abundances as a function of time. Observed abundances in TMC-1 are indicated by dashed horizontal lines.