Thiols in the Interstellar Medium: First Detection of HC(O)SH and Confirmation of C₂H₅SH

The fitting procedure was performed considering the total emission of any other identified molecule in the spectral survey. This evaluation was carried out introducing all the compounds already detected in the ISM and in G+0.693 (see Requena-Torres et al. 2006, 2008; Zeng et al. 2018; Rivilla et al. 2019; Jiménez-Serra et al. 2020). For the analysis, it was also assumed a cosmic microwave background temperature (T_bg) of 2.73 K and no background continuum source (Zeng et al. 2020).

The global fit of all rotational lines to the observed data is shown in blue lines in Figure 1, while in red we show the fit of the individual lines of t-HC(O)SH, and the observational data in black. As shown in both Figure 1 and Table 1, we have detected a total of nine a-type transitions in the 7 mm and 3 mm bands, each one of them detected above the 5σ level in integrated intensity. Note that the SLIM synthetic spectrum shows small deviations with respect to the observations. These deviations are larger that the ones obtained for CH₃SH and g-C₂H₅SH (see Figures 3 and 4 in Appendix A). This is due to the fact that the t-HC(O)SH lines are significantly weaker than those of CH₃SH and g-C₂H₅SH (3–7 mK versus 5–12 mK and >40 mK, respectively). However, the fit of all clean t-HC(O)SH lines is consistent with the noise. We also note that for the three slightly blended lines of t-HC(O)SH, the contribution of the other molecules is less than 10% of the total intensity. Despite that the remaining two appear blended, the predicted LTE intensities are consistent with the observed lines. Note that the rest of the t-HC(O)SH lines covered within the observed frequency range are not shown due to strong blending issues.

Zoom In Zoom Out Reset image size

Table 1. Lines of t-HC(O)SH Detected toward G+0.693 with Their Corresponding Quantum Numbers (QNs), Logarithm of the Einstein Coefficients ($\mathrm{log}\ {A}_{\mathrm{ul}}$) , Degeneracy (${g}_{{\rm{u}}}$), and Energy (${E}_{{\rm{u}}}$) of the Upper State

Rest Frequency	QNs a	gu	Eu	$\mathrm{log}\ {{\rm{A}}}_{{ul}}$	rms	$\int {T}_{A}^{* }d\nu$ b	S/N b	Comments
(MHz)			(K)	(s−1)	(mK)	($\mathrm{mK}\,\mathrm{km}\,{{\rm{s}}}^{-1}$)
34248.82	${3}_{\mathrm{1,3}}\to {2}_{\mathrm{1,2}}$	7	4.3	−6.4784	1.4	56	7.3	clean transition
35915.48	${3}_{\mathrm{1,2}}\to {2}_{\mathrm{1,1}}$	7	4.4	−6.4164	1.4	58	7.6	blended with HOCH2CH2OH
45659.99	${4}_{\mathrm{1,4}}\to {3}_{\mathrm{1,3}}$	7	6.0	−6.0647	2.4	88	13.4	clean transition
46737.73	${4}_{\mathrm{0,4}}\to {3}_{\mathrm{0,3}}$	7	3.4	−6.0064	2.6	122	8.6	clean transition
71800.18	${6}_{\mathrm{1,5}}\to {5}_{\mathrm{1,4}}$	13	11.3	−5.4428	3.5	121	6.3	clean transition
81630.08	${7}_{\mathrm{0,7}}\to {6}_{\mathrm{0,6}}$	15	11.8	−5.2587	3.4	153	8.2	slightly blended with CH3CH2CHO
83749.29	${7}_{\mathrm{1,6}}\to {6}_{\mathrm{1,5}}$	15	14.8	−5.2342	3.4	116	11.8	slightly blended with CH3COOH
91251.84	${8}_{\mathrm{1,8}}\to {7}_{\mathrm{1,7}}$	17	18.1	−5.1166	1.7	104	11.2	slightly blended with (CH2OH)2 and unidentified species
93505.09	${8}_{\mathrm{2,7}}\to {7}_{\mathrm{2,6}}$	17	26.5	−5.1061	1.4	46	6.0	blended with CH3 18OH

Notes. The integrated signal ( $\int {T}_{A}^{\ast }d\nu$) and rms noise level are also provided and used to calculate the signal-to-noise ratio (S/N) of the detections.

^a ${J}_{{K}_{a}^{{\prime\prime} }{K}_{c}^{{\prime\prime} }}^{{\prime\prime} }\to {J}_{{K}_{a}^{{\prime} }{K}_{c}^{{\prime} }}^{{\prime} }$ (where the double prime indicates the upper state). ^b S/N is calculated from the integrated signal and noise level $\sigma =\mathrm{rms}\,\sqrt{\delta v\,{FWHM}}$, where $\delta v$ is the velocity resolution of the spectra.

Download table as:  ASCIITypeset image

The physical parameters obtained from the fit are listed in Table 2. The values obtained for T_ex and FWHM are consistent with those obtained previously for other molecular species toward this cloud (${T}_{\mathrm{ex}}\,\sim \,$5–20 K and line widths around 20 km s⁻¹; Zeng et al. 2018; Rivilla et al. 2020). The low ${T}_{\mathrm{ex}}$ indicates that the emission of the molecules in G+0.693 is subthermally excited as a result of the low H₂ densities of this source (${T}_{\mathrm{kin}}\,\sim \,$150 K; Requena-Torres et al. 2006; Zeng et al. 2018).

Table 2. Physical Parameters of the Species Derived by LTE Analysis in madcuba

Molecular Formula	N	Tex	vLSR	FWHM	Abundance
	(×1013 cm−2)	(K)	$(\mathrm{km}\,{{\rm{s}}}^{-1})$	$(\mathrm{km}\,{{\rm{s}}}^{-1})$	($\tfrac{N({\rm{X}})}{N({{\rm{H}}}_{2})}\times {10}^{-10}$)
${\bf{trans}}-{\bf{HC}}({\bf{O}}){\bf{SH}}$	1.6 ± 0.1	10 ± 1	69.0 a	21.0 a	1.2 ± 0.2
$\mathrm{cis}-\mathrm{HC}({\rm{O}})\mathrm{SH}$	≤0.3	10.0	69.0 a	21.0 a	≤0.2
${\rm{g}}-{{\rm{C}}}_{2}{{\rm{H}}}_{5}\mathrm{SH}$	4 ± 2	10 ± 5	69.0 a	20.0 a	3 ± 1
${\rm{a}}-{{\rm{C}}}_{2}{{\rm{H}}}_{5}\mathrm{SH}$	≤2.3	9.91	69.0 a	20.0 a	≤1.7
${\mathrm{CH}}_{3}\mathrm{SH},\ \ {{\rm{K}}}_{{\rm{a}}}=0$	46.8 ± 0.5	8.5 ± 0.1	68.0 ± 0.1	21.2 ± 0.3	⋯
${\mathrm{CH}}_{3}\mathrm{SH},\ \ {{\rm{K}}}_{{\rm{a}}}=1$	18.6 ± 0.7	14.9 ± 0.5	68.8 ± 0.3	22.0 ± 0.7	⋯
${\mathrm{CH}}_{3}\mathrm{SH}$	65 ± 2	⋯	⋯	⋯	48 ± 5
${\rm{t}}-\mathrm{HCOOH}$	20 ± 4	10 ± 2	68 ± 2	22 ± 5	15 ± 4

Note.

^a Value fixed in the fit.

Download table as:  ASCIITypeset image

For t-HC(O)SH, the fitted column density gives (1.6 ± 0.1)×10¹³ cm⁻². In addition, we derived a 3σ upper limit of ≤3 × 10¹² cm⁻² for c-HC(O)SH (Table 2), as no clear transition has been detected within our data set. For the calculation, we have assumed the same ${T}_{\mathrm{ex}}$, ${v}_{\mathrm{LSR}}$, and FWHM as for t-HC(O)SH. A ratio c-HC(O)SH/t-HC(O)SH ≤ 0.2 was obtained.

This molecule was tentatively detected in Orion KL (Kolesniková et al. 2014), based on a few isolated transitions assigned to this isomer. We report here an unambiguous detection of this isomer that confirms it presence in the ISM. We have listed in Table 4 (Appendix A) the transitions measured toward G+0.693. Note that eight of the targeted lines are totally clean (Figure 3, Appendix A) and above the 5σ level in integrated intensity. We derived a total column density of (4 ± 2) × 10¹³ cm⁻² (Table 2).

In the case of CH₃SH, the the fit was carried out separating the targeted lines into its K_a  = 0 and K_a  = 1 levels ¹⁴ . The brightest transitions fall into the 3 mm and 2 mm bands as presented in Figure 4 and Table 5 (Appendix B). Note that the agreement between the predicted and observed spectra is excellent for all clean transitions. The column density using the K_a  = 0 and K_a  = 1 levels is (6.5 ± 0.2) × 10¹⁴ cm⁻² (Table 2). This gives a ratio of CH₃SH/g-C₂H₅SH = 16 ± 7. Note that the column density of K_a  = 2 and remaining transitions in CH₃SH contribute less than a 10% in the total determined.

Finally, we note that both ¹³C and ³⁴S isotopologues of CH₃SH have also been detected with a few clean lines and will be presented in a forthcoming paper (L. Colzi et al. 2021, in preparation).

To derive the fractional abundances of these species relative to H₂, we have assumed a H₂ column density of ${N}_{{{\rm{H}}}_{2}}=1.35\,\times {10}^{23}\,$ cm⁻² (Martín et al. 2008). This gives ∼ 3 × 10⁻¹⁰, ∼1 × 10⁻¹⁰, and ∼5 × 10⁻⁹ for g-C₂H₅SH, t-HC(O)SH, and CH₃SH, respectively (Table 2).

In Figure 2 we have plotted these values and compared them with the abundances obtained for their OH molecular analogs, namely, C₂H₅OH (ethanol), HC(O)OH (formic acid), and CH₃OH (methanol). We obtained a ratio CH₃OH/CH₃SH = 23, C₂H₅OH/C₂H₅SH = 15, and HC(O)OH/HC(O)SH = 13. Although these OH analogs are more abundant by a factor ≥10, the resulting trend is strikingly similar. Note, however, that this trend is lost when we compare molecules such as carbon monosulfide (CS) and thioformaldehyde (H₂CS) with their O-bearing analogs, resulting in H₂CO/H₂CS = 3.5 and CO/CS = 3.5 × 10³ (Figure 2).

Zoom In Zoom Out Reset image size

Since CH₃OH, C₂H₅OH, H₂CO, CO, and CS are optically thick toward G+0.693, the column density of these molecules was inferred from CH₃ ¹⁸OH, ¹³CH₃CH₂OH, H₂C¹⁸O, C¹⁸O, and ¹³C³⁴S. It was assumed ¹⁶O/¹⁸O = 250 and ¹²C/¹³C = 21 (Armijos-Abendaño et al. 2015) and ³²S/³⁴S = 22 (Wilson 1999). For HC(O)OH, the main isotopologue of HC(O)OH was employed as the ratio H¹²C(O)OH/H¹³C(O)OH, which is consistent with the ¹²C/¹³C ratio. For the sulfur analogs, except CS, we do not expect opacity issues due to their much lower abundances. The fit and transitions selected for these molecules are listed in Appendix B (Figures 5 and 6 and Tables 6, 7, 8, 9, and 10).

We have compared different column density ratios obtained between the S-bearing compounds measured toward G+0.693 and two other sources, namely, Orion KL (Kolesniková et al. 2014) and Sgr B2(N2) (Müller et al. 2016; see Table 3).

From this table, we find that all abundance ratios measured in G+0.693 are strikingly similar. For instance, when we compare SH-bearing molecules with their OH analogs (CH₃OH/CH₃SH and C₂H₅OH/C₂H₅SH) we recover the trend already found in Figure 2, which might indicate a similar chemistry between OH- and SH-bearing species. The CH₃OH/CH₃SH ratio measured in G+0.693 is a factor of 5 lower than those found in Sgr B2(N2) and Orion KL, which indicates that G+0.693 is richer in sulfur-bearing species than the two massive hot cores. This may be related to the fact that the chemistry of this cloud is affected by large-scale shocks (Requena-Torres et al. 2006; Zeng et al. 2018). Since sulfur is heavily depleted on grains (possibly in the form of S₈ and other sulfur allotropes; Shingledecker et al. 2020), the sputtering of dust grains induced by shocks could liberate a significant fraction of the locked sulfur.

The C₂H₅OH/C₂H₅SH ratio is comparable to the value determined toward Orion KL while there is a difference by a factor of 10 between G+0.693 and the upper limit toward Sgr B2(N2). All these ratios suggest that sulfur is less incorporated into molecules than oxygen, which again might be due to the fact that a significant fraction of sulfur is locked up in grains.

In order to understand the chemistry of S-bearing compounds, different models have been proposed based on the simple molecules detected in the gas phase (Müller et al. 2016; Gorai et al. 2017; Lamberts 2018; Laas & Caselli 2019). CH₃SH is thought to be formed on grain surfaces by sequential hydrogenations starting from CS:

$$\begin{eqnarray*}&&\mathrm{CS}\ +\ {\rm{H}}\ \to \ \mathrm{HCS}\ \end{eqnarray*}$$

$$\begin{eqnarray*}&&\mathrm{HCS}\ +\ {\rm{H}}\ \to \ {{\rm{H}}}_{2}\mathrm{CS}\end{eqnarray*}$$

$$\begin{eqnarray*}&&{{\rm{H}}}_{2}\mathrm{CS}\ +\ {\rm{H}}\ \to \ {\mathrm{CH}}_{3}{\rm{S}}\end{eqnarray*}$$

$$\begin{eqnarray*}&&{\mathrm{CH}}_{3}{\rm{S}}\ +\ {\rm{H}}\ \to \ {\mathrm{CH}}_{3}\mathrm{SH}.\end{eqnarray*}$$

The last step could also yield other products, but theoretical calculations show a branching ratio (br) of ∼75% for CH₃SH (Lamberts 2018). This is a viable mechanism since approximately half of the CS present in the ices is available to undergo hydrogenation, while the other half is converted to OCS (Palumbo et al. 1997). Once formed, CH₃SH could remain stored in the ices until released by grain sputtering in the large-scale shocks present in G+0.693 (Requena-Torres et al. 2006; Zeng et al. 2020).

Likewise, ethyl mercaptan is proposed to be formed by radical–radical reactions (Müller et al. 2016; Gorai et al. 2017) such as

$$\begin{eqnarray*}&&{{\rm{C}}}_{2}{{\rm{H}}}_{5}\ +\ \mathrm{SH}\ \to \ {{\rm{C}}}_{2}{{\rm{H}}}_{5}\mathrm{SH}\end{eqnarray*}$$

$$\begin{eqnarray*}&&{\mathrm{CH}}_{3}\ +\ {\mathrm{CH}}_{2}\mathrm{SH}\ \to \ {{\rm{C}}}_{2}{{\rm{H}}}_{5}\mathrm{SH}.\end{eqnarray*}$$

Observations of the reactants could give us a hint of the dominant reaction based on the measured column densities. However, chemical modeling is needed to understand the efficiency of these formation routes. Note that CH₂SH is the main product of the hydrogenation of H₂CS (Lamberts 2018).

To our knowledge, there is no information available in the literature about the chemistry of HC(O)SH. However, we can make a guess and assume similar astrochemical pathways for the SH-based species as the ones for their OH analogs. A possible pathway to HC(O)SH could mimic the formation of HC(O)OH. Ioppolo et al. (2011) showed that the formation of HC(O)OH starts from CO and the OH radical in the ice. The thiol-equivalent reaction would be

$$\begin{eqnarray*}&&\mathrm{CO}\ +\ \mathrm{SH}\ \to \ \mathrm{HSCO}\mathop{\to }\limits^{{\rm{H}}}\mathrm{HC}({\rm{O}})\mathrm{SH}.\end{eqnarray*}$$

The first step has been found to be efficient by Adriaens et al. (2010), but further experimental and/or theoretical work is needed to investigate whether HSCO could be hydrogenated further.

Another possibility could be

$$\begin{eqnarray*}&&\mathrm{HCO}\ +\ \mathrm{SH}\ \to \ \mathrm{HCOSH}\end{eqnarray*}$$

or

$$\begin{eqnarray*}&&\mathrm{OCS}\ \mathop{\to }\limits^{2{\rm{H}}}\mathrm{HC}({\rm{O}})\mathrm{SH}.\end{eqnarray*}$$

The first reaction was initially proposed for HC(O)OH by Garrod & Herbst (2006) and could be a viable mechanism since SH can be formed on surfaces from S+H or H+H₂S via tunneling (Vidal et al. 2017). The second involves the sequential hydrogenation of OCS, which is a molecule detected in ices (Palumbo et al. 1997) and it is moderately abundant in the gas phase in G+0.693 (N > 10¹⁵ cm⁻²; Armijos-Abendaño et al. 2015). Theoretical studies about these chemical networks are currently under work and will be presented in a forthcoming paper (G. Molpeceres et al. 2021, in preparation).

As it was previously stated, thioacids and thioesters have been proposed as key agents in the polymerization of amino acids into peptides and proteins (Foden et al. 2020; Muchowska & Moran 2020). In some of these works, it is stressed the importance of cystein, HSCH₂CH(NH₂)COOH, as the primary organic source of sulfide in biology and a key catalyst in the abiotic polymerization of peptides. Smaller thiols such as CH₃SH and C₂H₅SH are also believed to play a key role in prebiotic chemistry and in theories about the origin of life, since they are precursors for the synthesis of the amino acids methionine and ethionine (Parker et al. 2011). In turn, HC(O)SH could be an important ingredient in the phosphorylation of nucleosides as demonstrated by Lohrmann & Orgel (1968).

The idea of an extraterrestrial delivery of SH compounds onto Earth, is supported by the detection of CH₃SH in Murchinson carbonaceous chondrite (Tingle et al. 1991) and in the coma of the 67P/Churyumov–Gerasimenko comet (Calmonte et al. 2016). In the latter, C₂H₆S was also detected (either in the form of C₂H₅SH or CH₃SCH₃, dimethyl sulfide) with an abundance ratio CH₃SH/C₂H₆S ∼ 10 that is consistent with the value obtained in G+0.693 (Table 3). This resemblance could be possibly indicate a presolar origin of these compounds. Still, further studies are required within other regions in the ISM and planetary bodies to make a proper connection.

In summary, not only have our observations confirmed the presence of sulfur-bearing complex organics such as CH₃SH and C₂H₅SH, but they also have revealed the existence of the simplest thioacid known, HC(O)SH, in the ISM.