Searches for Interstellar HCCSH and H₂CCS

Despite being one of the most abundant elements in the interstellar medium (ISM), the sulfur content in cold molecular clouds can only account for ∼0.1% of that seen in warm, diffuse clouds (Tieftrunk et al. 1994; Ruffle et al. 1999), which have values comparable to the solar abundance (Bilalbegović & Baranović 2015). To account for this "missing" sulfur, many hypotheses have been put forward regarding hidden sinks and reservoirs, including ices, dusty grains, and unknown molecular species (Martín-Doménech et al. 2016).

The search for condensed-phase sulfur has largely been unsuccessful to date. Simple molecules (such as OCS) sequestered in interstellar ice grains, for example, can only account (Boogert et al. 1997; Martín-Doménech et al. 2016) for a small amount of the total cosmic abundance (∼4%). In cometary ices, the principle sulfur-bearing species appears to be ${{\rm{H}}}_{2}{\rm{S}}$, with an abundance of ∼1.5% relative to water (Bockelée-Morvan et al. 2000), whereas FeS appears to be a major reservoir in the cometary grains themselves (Dai & Bradley 2001). FeS grains have also been detected in protoplanetary disks, suggesting this reservoir may be widespread, and not unique to solar system objects (Keller et al. 2002). Several recent observational and modeling studies hypothesize ${{\rm{H}}}_{2}{\rm{S}}$ may be a substantial sulfur reservoir in interstellar ices as well (Holdship et al. 2016; Vidal et al. 2017). While ${{\rm{H}}}_{2}{\rm{S}}$ is a well-known interstellar species in the gas phase (Thaddeus et al. 1972), there has been no definitive condensed-phase observation of it outside the solar system (Boogert et al. 2015). Recent modeling work by Laas & Caselli (2019) has shown that the sulfur depletion in cold clouds can be reproduced without the need for a significant buildup of ${{\rm{H}}}_{2}{\rm{S}}$ in ices, suggesting this species may play a less substantial role than previously assumed. Further investigations of condensed-phase sulfur species will undoubtedly be a strong science driver for the forthcoming James Webb Space Telescope.

New gas-phase sulfur-bearing molecules remain one of the most promising explanations for the missing sulfur content, and one that is addressable by extant radio facilities, assuming the molecules possess a permanent dipole moment and the necessary laboratory work is both available and complete (Cazzoli et al. 2016). Of the more than 200 known interstellar and circumstellar molecules, only 23 contain at least one sulfur atom (Table 1). Perhaps more striking, of the 94 molecules with five or more atoms, 3 contain sulfur, whereas 30 contain at least one oxygen atom (McGuire 2018). In total, sulfur-containing species comprise ∼10% of the known molecular inventory of any size, whereas one-third of all molecules are oxygen-containing. Despite the lower interstellar abundance of sulfur relative to oxygen (Cameron 1973), the large disparity between the number of S- and O-bearing species is striking, and may indicate detections of large, sulfur-bearing species in space are simply limited by the lack of precise laboratory rest frequencies.

Table 1.  Known Interstellar and Circumstellar Molecules Containing at Least One Sulfur Atom and References to the First Detection of Those Species

Molecule	References
CS	Penzias et al. (1971)
SO	Gottlieb & Ball (1973)
SiS	Morris et al. (1975)
NS	Gottlieb et al. (1975), Kuiper et al. (1975)
${\mathrm{SO}}^{+}$	Turner (1992)
${\mathrm{SH}}^{+}$	Benz et al. (2010)
SH	Neufeld et al. (2012)
${\mathrm{NS}}^{+}$	Cernicharo et al. (2018)
OCS	Jefferts et al. (1971)
${{\rm{H}}}_{2}{\rm{S}}$	Thaddeus et al. (1972)
${\mathrm{SO}}_{2}$	Snyder et al. (1975)
${\mathrm{HCS}}^{+}$	Thaddeus et al. (1981)
${{\rm{C}}}_{2}{\rm{S}}$	Saito et al. (1987)
${{\rm{S}}}_{2}{\rm{H}}$	Fuente et al. (2017)
HCS	Agúndez et al. (2018)
HSC	Agúndez et al. (2018)
${{\rm{H}}}_{2}\mathrm{CS}$	Sinclair et al. (1973)
HNCS	Frerking et al. (1979)
${{\rm{C}}}_{3}{\rm{S}}$	Yamamoto et al. (1987)
HSCN	Halfen et al. (2009)
${\mathrm{CH}}_{3}\mathrm{SH}$	Linke et al. (1979)
${{\rm{C}}}_{5}{\rm{S}}$	Bell et al. (1993)
${\mathrm{CH}}_{3}{\mathrm{CH}}_{2}\mathrm{SH}$	Kolesniková et al. (2014)

Download table as:  ASCIITypeset image

Large sulfur-bearing species (by interstellar standards) are perhaps improbable as substantial reservoirs of gas-phase sulfur, given the generally lower abundance of large molecules relative to simpler species (see, e.g., Belloche et al. 2013). Detailed studies of their formation chemistry, however, can provide insights into the underlying abundances (and potential reservoirs) of as-yet-undetected simpler precursor molecules. As an example, one of the most promising ways to probe formation pathways is through the study of isomeric families. Comparisons of the [${{\rm{H}}}_{4}$, ${{\rm{C}}}_{2}$, ${{\rm{O}}}_{2}$] isomers glycolaldehyde, methyl formate, and acetic acid, for instance, have been used in attempts to infer branching fractions in the UV photodissociation of methanol (${\mathrm{CH}}_{3}\mathrm{OH}$), from which most of the key precursor species for the [${{\rm{H}}}_{4}$, ${{\rm{C}}}_{2}$, ${{\rm{O}}}_{2}$] family are expected to form (Laas et al. 2011).

Here, we present an observational counterpart to the recent laboratory investigation of the [${{\rm{H}}}_{2}$, ${{\rm{C}}}_{2}$, S] isomeric family of molecules. The three lowest-energy isomers in this family are ${{\rm{H}}}_{2}\mathrm{CCS}$ (thioketene), HCCSH (ethynethiol), and c-${{\rm{H}}}_{2}{{\rm{C}}}_{2}{\rm{S}}$ (thiirene). The microwave and millimeter-wave spectra of ${{\rm{H}}}_{2}\mathrm{CCS}$ have been previously reported by Winnewisser & Schäfer (1980), and several of the coauthors of the present work recently reported the laboratory spectrum of HCCSH from the microwave to submillimeter wavelengths (Lee et al. 2018). Efforts to measure the spectrum of c-${{\rm{H}}}_{2}{{\rm{C}}}_{2}{\rm{S}}$ are currently underway. Here, we summarize the astronomical search for ${{\rm{H}}}_{2}\mathrm{CCS}$ and HCCSH in a number of interstellar sources spanning the evolutionary spectrum from dark clouds to high-mass star-forming regions (HMSFRs).

HCCSH would appear to be a promising candidate for astronomical detection and potential sink for sulfur for two reasons: (1) this isomer might be formed directly and efficiently via a recombination reaction involving two well-known astronomical radicals, SH and CCH (Yamada et al. 2002), and (2) HCCSH does not possess a simple, readily identifiable rotational spectrum, since the dipole moment along its a-inertial axis is nearly zero (0.13 D), while that along the b-axis is significantly larger (0.80 D). As a consequence, unlike ${{\rm{H}}}_{2}\mathrm{CCS}$ whose rotational spectrum is characterized by a series of lines separated in frequency by ratios of integers at low-frequency, the same lines of HCCSH are extremely faint; instead, its b-type lines in the millimeter and submillimeter bands are much more intense [i.e., (μ_b/μ_a)² ∼ 35].

Electronic structure calculations (Lee et al. 2018) predict that the most stable isomeric arrangement of [${{\rm{H}}}_{2}$, ${{\rm{C}}}_{2}$, S] is ${{\rm{H}}}_{2}\mathrm{CCS}$, followed by HCCSH roughly 56 kJ mol⁻¹ (∼6770 K) higher in energy, with the three-membered heterocycle c-${{\rm{H}}}_{2}{{\rm{C}}}_{2}{\rm{S}}$ lying 132 kJ mol⁻¹ (∼15,900 K) above ground. Nevertheless, HCCSH can be formed by an exothermic and barrierless reaction involving SH and CCH, via reaction (R1) (Lee et al. 2018):

$$\begin{eqnarray}&&\mathrm{SH}\ +\mathrm{CCH}\ \longrightarrow \mathrm{HCCSH}.\end{eqnarray} \tag{ R1 }$$

This may result in preferential production in astronomical sources in which the two radicals are prominent, although in the gas phase it is possible that the exothermicity may result in the dissociation of a nontrivial fraction of the product, necessitating grain-surface production. The closely related [${{\rm{H}}}_{2}$, ${{\rm{C}}}_{3}$, O] isomers provide a dramatic illustration of this effect. Although l-propadienone (${{\rm{H}}}_{2}\mathrm{CCCO}$) has not been observed in space despite considerable efforts (Loomis et al. 2015), isoenergetic propynal (HCCC(O)H) is prominent (Irvine et al. 1988a). Further, the considerably less stable cyclopropenone (c-${{\rm{H}}}_{2}{{\rm{C}}}_{3}{\rm{O}}$) has also been detected in space (Hollis et al. 2006). Taken together, these findings highlight the importance of kinetic factors in molecule formation and destruction.

Thioketene (${{\rm{H}}}_{2}\mathrm{CCS}$) has a linear heavy-atom backbone with C_2v symmetry and ortho-para spin statistics. For this reason, it only has a-type rotational lines, but its dipole moment is sizable, 1.01(3)D (Winnewisser & Schäfer 1980). At low temperature, its rotational spectrum consists of relatively closely spaced triplets that are harmonically related spaced by B + C, or about 11.2 GHz. Laboratory measurements provide rest frequencies up to 230 GHz, which, given the rigidity of the molecule, can be extrapolated with reasonable confidence to better than 0.5 km s⁻¹ up to ∼450 GHz.

Although HCCSH has the same heavy-atom backbone as ${{\rm{H}}}_{2}\mathrm{CCS}$, owing to a different bond order along the chain, it is an asymmetric top with a bent backbone, but one still close to the prolate limit (κ = −0.999), as defined by κ, Ray's asymmetry parameter (Ray 1932). A purely prolate molecule (cigar shaped, and the most common in the ISM; McGuire 2018) has κ = −1, whereas a purely oblate molecule (disk shaped) has κ = 1. In contrast to ${{\rm{H}}}_{2}\mathrm{CCS}$, HCCSH has a relatively small calculated dipole moment along its a-axis (0.13 D), but that along the b-axis (0.80 D) is significant (Lee et al. 2018). Because the A rotational constant is large, 291.4 GHz, the most intense rotational lines lie at millimeter wavelengths. The detailed spectroscopic analysis of its a- and b-type transitions up to 660 GHz was presented in Lee et al. (2018).

Column densities were calculated assuming a single-excitation temperature model following the formalisms of Hollis et al. (2004a) using Equation (1), with corrections due to optical depth as described in Turner (1991):

$$\begin{eqnarray}&&{N}_{T}=\displaystyle \frac{1}{2}\displaystyle \frac{3k}{8{\pi }^{3}}\sqrt{\displaystyle \frac{\pi }{{\rm{l}}{\rm{n}}2}}\displaystyle \frac{{Qe}^{{E}_{{\rm{u}}}/{T}_{{\rm{e}}{\rm{x}}}}{\rm{\Delta }}{T}_{b}{\rm{\Delta }}V}{B\nu {S}_{ij}{\mu }^{2}}\displaystyle \frac{1}{1-{\textstyle {\textstyle {\textstyle \tfrac{{{\rm{e}}}^{h\nu /({kT}_{{\rm{e}}{\rm{x}}})}-1}{{{\rm{e}}}^{h\nu /({kT}_{{\rm{b}}{\rm{g}}})}-1}}}}}.\end{eqnarray} \tag{ 1 }$$

Here, N_T is the total column density (cm⁻²), Q is the partition function, T_ex is the excitation temperature (K), E_u is the upper state energy (K), ΔT_b is the peak intensity (K), ΔV is the full-width at half-maximum of the line (km s⁻¹), B is the beam filling factor, ν is the frequency (Hz), S_ij is the intrinsic quantum mechanical line strength, μ is the permanent dipole moment (Debye¹² ), and T_bg is the background continuum temperature (K). For the upper limits presented here, a simulated spectrum was generated using the parameters described for each observation below. The strongest predicted line in the observational spectrum was then used to calculate the 1σ upper limit, taking the rms noise value at that point as the value of the brightness temperature ΔT_b.

For HCSSH and ${{\rm{H}}}_{2}\mathrm{CCS}$, we consider both the rotational partition function and the contribution from low-lying vibrational states. The total partition function is calculated according to Equation (2):

$$\begin{eqnarray}&&Q={Q}_{\mathrm{rot}}\times {Q}_{\mathrm{vib}}\times {Q}_{\mathrm{elec}},\end{eqnarray} \tag{ 2 }$$

where Q_rot, Q_vib, and Q_elec are the rotational, vibrational, and electronic components, respectively. Under interstellar conditions, we assume Q_elec = 1. The rotational partition function was calculated via a direct summation of states as described by Gordy & Cook (1984) and Equation (3):

$$\begin{eqnarray}&&{Q}_{r}=\displaystyle \sum _{i=0}^{{\rm{\infty }}}(2J+1){{\rm{e}}}^{-{E}_{i}/{kT}_{{\rm{e}}{\rm{x}}}},\end{eqnarray} \tag{ 3 }$$

where E_i is the energy of the ith rotational state. The vibrational correction was calculated according to Equation (4):

$$\begin{eqnarray}&&{Q}_{{\rm{v}}{\rm{i}}{\rm{b}}}=\displaystyle \prod _{i=1}^{3N-6}\displaystyle \frac{1}{1-{{\rm{e}}}^{-{E}_{v,i}/{kT}_{{\rm{e}}{\rm{x}}}}},\end{eqnarray} \tag{ 4 }$$

where E_v,i is the energy of the ith excited vibrational state. Here, we consider only the lowest five vibrational states as those higher in energy make a negligible contribution to Q_vib.

The vibrational (harmonic) energies were calculated at the CCSD(T)/cc-pVQZ level of theory. For HCCSH, the lowest five vibrational states lie at 356, 416, 727, 848, and 942 cm⁻¹ above ground, while for ${{\rm{H}}}_{2}\mathrm{CCS}$, they fall at 295, 352, 578, 696, and 718 cm⁻¹. The rotational, vibrational, and total partition functions used for the column density calculations at each temperature are listed in Tables 2 and 3 along with other molecular parameters required for the calculations.

Table 2.  Upper Limits to HCCSH and the Line Parameters Used to Calculate Them in Each of the Sets of Observations

Source	Frequencya	Transition	Eu	${S}_{{ij}}{\mu }^{2}$	Q (Qrot, Qvib)b	NT	N(${{\rm{H}}}_{2}$)	${X}_{{{\rm{H}}}_{2}}$	References
	(MHz)	(${J}_{{K}_{a},{K}_{c}}^{{\prime} }\mbox{--}{J}_{{K}_{a},{K}_{c}}^{{\prime\prime} }$)	(K)	(Debye2)		(cm−2)	(cm−2)		N(${{\rm{H}}}_{2}$)
NGC 6334I	293352.8	161,15–160,16	85.8	10.4	3894 (2824, 1.38)	≤1.4 × 1016	⋯	⋯	⋯
Sgr B2(N)	618565.8	331,33–320,32	307.9	11.4	35808 (8441, 4.24)	≤3.9 × 1017	1 × 1024	≤4 × 10−7	1
Sgr B2(N)	844982.2c	212,19–211,20	176.7	7.1	35808 (8441, 4.24)	≤1.4 × 1017	1 × 1024	≤1 × 10−7	1
Orion-KL	618565.8	331,33–320,32	307.9	11.4	12664 (5442, 2.32)	≤3.1 × 1015	3.9 × 1023	≤8 × 10−9	2
Orion-KL	850042.2c	162,14–161,15	126.6	5.3	12664 (5442, 2.32)	≤3.3 × 1015	3.9 × 1023	≤8 × 10−9	2
Barnard 1	241629.5	31,3–40,4	16.9	1.0	57 (57, 1.00)	≤2.3 × 1012	1.5 × 1023	≤2 × 10−11	3
IRAS 4A	230425.1	41,4–50,5	19.0	1.3	173 (173, 1.00)	≤1.3 × 1013	3.7 × 1023	≤4 × 10−11	3
L1157B1	288898.1	101,9–100,10	42.9	6.7	855 (837, 1.02)	≤5.3 × 1012	1 × 1021	≤5 × 10−9	3
L1157mm	207853.8	61,6–70,7	24.7	1.9	855 (837, 1.02)	≤7.7 × 1012	6 × 1021	≤1 × 10−9	3
L1448R2	207853.8	61,6–70,7	24.7	1.9	855 (837, 1.02)	≤2.1 × 1013	3.5 × 1023	≤6 × 10−11	4
L1527	162066.9	101,10–110,11	42.6	3.2	75 (75, 1.00)	≤3.1 × 1012	2.8 × 1022	≤1 × 10−10	4
L1544	87882.4	80,8–70,7	19.0	0.1	57 (57, 1.00)	≤1.6 × 1013	5 × 1021	≤3 × 10−9	5
SVS13A	207853.8	61,6–70,7	24.7	1.9	149 (149, 1.00)	≤9.0 × 1015	3 × 1024	≤3 × 10−9	6
TMC1	150487.4	111,11–120,12	48.3	3.6	34 (34, 1.00)	≤2.9 × 1013	1 × 1022	≤3 × 10−9	3

Notes.

^aTypical experimental accuracy of the millimeter-wave measurements was ∼50 kHz. ^bCalculated at the excitation temperature assumed for the source. See Table 4 in Appendix A. ^cExtrapolated beyond the upper range (660 GHz) of the laboratory measurements.

References. [1] Lis & Goldsmith (1990), [2] Crockett et al. (2014), [3] Cernicharo et al. (2018), [4] Jørgensen et al. (2002), [5] Vastel et al. (2014), [6] Chen et al. (2009).

Download table as:  ASCIITypeset image

Table 3.  Upper Limits to ${{\rm{H}}}_{2}\mathrm{CCS}$ and the Line Parameters Used to Calculate Them in Each of the Sets of Observations

Source	Frequencya	Transition	Eu	${S}_{{ij}}{\mu }^{2}$	Q (Qrot, Qvib)b	NT	N(${{\rm{H}}}_{2}$)	${X}_{{{\rm{H}}}_{2}}$	References
	(MHz)	(${J}_{{K}_{a},{K}_{c}}^{{\prime} }-{J}_{{K}_{a},{K}_{c}}^{{\prime\prime} }$)	(K)	(Debye2)		(cm−2)	(cm−2)		N(${{\rm{H}}}_{2}$)
NGC 6334I	292685.4c	261,25–251,24	203.2	78.2	9051 (5456, 1.66)	≤4.7 × 1015	⋯	⋯	⋯
Sgr B2(N)d	22407.9	20,2–10,1	1.6	2.0	31 (31, 1.00)	≤6.4 × 1012	1 × 1024	≤6 × 10−12	1
Sgr B2(N)e	100828.5	90,9–80,8	24.2	9.0	12106 (6395, 1.89)	≤2.9 × 1016	1 × 1024	≤3 × 10−8	1
Sgr B2(N)f	494982.4c	441,43–431,42	548.3	132.5	121982 (16448, 7.42)	≤2.7 × 1017	1 × 1024	≤3 × 10−7	1
Orion-KL	494982.4c	441,43–431,42	548.3	132.5	4410 (3466, 1.27)	≤3.2 × 1015	3.9 × 1023	≤8 × 10−9	2
Barnard 1	89626.7	80,8–70,7	19.4	8.2	104 (104, 1.00)	≤5.0 × 1011	1.5 × 1023	≤3 × 10−12	3
IRAS 4A	89626.7	80,8–70,7	19.4	8.2	334 (334, 1.00)	≤2.5 × 1012	3.7 × 1023	≤7 × 10−12	3
L1157B1	213921.0	191,18–181,17	116.2	58.0	1725 (1648, 1.05)	≤3.1 × 1012	1 × 1021	≤3 × 10−9	3
L1157mm	211738.3	191,19–181,18	115.1	58.0	1725 (1648, 1.05)	≤2.4 × 1012	6 × 1021	≤4 × 10−10	3
L1448R2	100311.4	91,9–81,8	37.6	27.2	1725 (1648, 1.05)	≤4.2 × 1012	3.5 × 1023	≤1 × 10−11	4
L1527	78825.9	71,6–61,5	28.6	21.0	139 (139, 1.00)	≤3.1 × 1011	2.8 × 1022	≤1 × 10−11	4
L1544	89626.7	80,8–70,7	19.4	8.2	104 (104, 1.00)	≤2.2 × 1011	5 × 1021	≤4 × 10−11	5
SVS13A	100828.5	90,9–80,8	24.2	9.2	286 (286, 1.00)	≤9.1 × 1015	3 × 1024	≤3 × 10−9	6
TMC1	134430.3	120,12–110,11	41.9	12.2	57 (57, 1.00)	≤5.5 × 1012	1 × 1022	≤6 × 10−10	3

Notes.

^aWithin the range of the measurements (60–230 GHz), Winnewisser & Schäfer (1980) claim a typical accuracy of ∼16.5 kHz. ^bCalculated at the excitation temperature assumed for the source. See Table 5 in Appendix A. ^cExtrapolated beyond the upper range (230 GHz) of the laboratory measurements. ^dGBT (PRIMOS) observations. ^eIRAM 30 m observations. ^fHerschel HIFI observations.

References. [1] Lis & Goldsmith (1990), [2] Crockett et al. (2014), [3] Cernicharo et al. (2018), [4] Jørgensen et al. (2002), [5] Vastel et al. (2014), [6] Chen et al. (2009).

Download table as:  ASCIITypeset image

The simulated spectrum of HCCSH, with arbitrary abundance, at three different temperatures appropriate to interstellar environments (10, 80, and 200 K) is shown in Figure 1, along with the associated observing bands of the most sensitive telescope facilities in those frequency ranges: the Robert C. Byrd 100 m Green Bank Telescope (GBT), the Atacama Large Millimeter/submillimeter Array (ALMA), and the GREAT instrument on board the Stratospheric Observatory for Infrared Astronomy (SOFIA). Due to its geometry, the most intense transitions fall in ALMA Bands 7 and 10 and the 1.4 THz window of the SOFIA GREAT receiver, as well as archival coverage of the Herschel HIFI instrument. Here, we examined observations from ALMA, Herschel, and the GBT targeting the HMSFRs NGC 6334I, Sgr B2(N), and Orion-KL. We also searched the line survey data toward sources from the publicly available Astrochemical Surveys at IRAM (ASAI) Large Project conducted with the IRAM 30 m telescope.¹³ The exact physical parameters assumed for each of the sources examined here are given in Tables 4 and 5 in Appendix A.

Zoom In Zoom Out Reset image size

4.1. NGC 6334I

4.2. Orion-KL

4.3. Sgr B2(N)

4.4. ASAI Sources

In addition to the formation of HCCSH via reaction (R1), as proposed by Lee et al. (2018), several additional reaction pathways would appear at least qualitatively plausible. For example, both quantum chemical (Ochsenfeld et al. 1999) and laboratory work (Galland et al. 2001) have shown that HCS/HSC can be readily produced by a reaction involving atomic carbon:

$$\begin{eqnarray}&&{{\rm{H}}}_{2}{\rm{S}}\ +{\rm{C}}\ \longrightarrow \mathrm{HCS}/\mathrm{HSC}\ +{\rm{H}}.\end{eqnarray} \tag{ R2 }$$

It is not unreasonable, therefore, to speculate that HCS/HSC could further react with atomic carbon, followed by hydrogenation, to yield both HCCSH and ${{\rm{H}}}_{2}\mathrm{CCS}$ on grain surfaces. Indeed, successive hydrogenation reactions beginning from CS or ${{\rm{C}}}_{2}{\rm{S}}$ could also yield HCS/HSC intermediates, or perhaps even HCCSH and ${{\rm{H}}}_{2}\mathrm{CCS}$ directly (see, e.g., Lamberts 2018). Further quantum chemical and laboratory work exploring the efficiency, branching ratios, and rates of these reactions would certainly help to shed light on the viability of these pathways.

It is also important to consider the potential destruction pathways of these species as well, and in this context it may be enlightening to compare the [${{\rm{H}}}_{2}$, ${{\rm{C}}}_{2}$, S] species with a similar family of isomers having the formula [${{\rm{H}}}_{2}$, ${{\rm{C}}}_{3}$, O]. As mentioned earlier, one rather longstanding astronomical mystery has been why the most stable of these isomers, propadienone (${{\rm{H}}}_{2}\mathrm{CCCO}$) has thus far remained undetected in space (Loomis et al. 2015; Loison et al. 2016), despite detections of two higher-energy forms, propynal (HCCCHO) and cyclopropenone (c-${{\rm{H}}}_{2}{{\rm{C}}}_{3}{\rm{O}}$; Irvine et al. 1988a; Hollis et al. 2006) there. Recently, an ab initio study involving reactions between atomic hydrogen and propadienone and propynal revealed unexpectedly that only the addition to propadienone, i.e.,

$$\begin{eqnarray}&&{\rm{H}}\ +{{\rm{H}}}_{2}\mathrm{CCCO}\ \longrightarrow {{\rm{H}}}_{2}\mathrm{CCHCO},\end{eqnarray} \tag{ R3 }$$

was barrierless and exothermic (Shingledecker et al. 2019). Moreover, it was found that the radical formed via (R3) could again subsequently react with H to form propenal (${\mathrm{CH}}_{2}\mathrm{CHCHO}$), a species which in fact has been observed by Hollis et al. (2004b) in Sgr B2(N), where the other isomers of [${{\rm{H}}}_{2}$, ${{\rm{C}}}_{3}$, O] have been detected. Thus, reaction (R3) likely keeps the abundance of propadienone low, both on grains—where it serves as a precursor to propenal—and in the gas, where the association product likely dissociates.

Similarly, given the high mobility of atomic hydrogen on grain surfaces—particularly in warm environments—it may be the case that ${{\rm{H}}}_{2}\mathrm{CCS}$ and/or HCCSH are efficiently destroyed by H. Intriguingly, the detection of the related saturated species ${\mathrm{CH}}_{3}{\mathrm{CH}}_{2}\mathrm{SH}$ in Orion-KL by Kolesniková et al. (2014) hints at the successive hydrogenation of a CCS backbone, with either of the linear [${{\rm{H}}}_{2}$, ${{\rm{C}}}_{2}$, S] isomers potentially serving as precursors. Detailed calculations of these types of reactions could therefore reveal whether such kinetic effects might play a role in explaining the observational results described here.

Given that the structurally analogous ${{\rm{H}}}_{2}\mathrm{CCC}$, ${{\rm{H}}}_{2}\mathrm{CCN}$, and ${{\rm{H}}}_{2}\mathrm{CCO}$ are all known interstellar species (Turner 1977; Irvine et al. 1988b; Cernicharo et al. 1991), the nondetections of both HCCSH and ${{\rm{H}}}_{2}\mathrm{CCS}$ is somewhat surprising. In a few cases, there are hints of emission at appropriate frequencies; nevertheless, no signal was seen that could be even tentatively assigned with any confidence to either HCCSH or ${{\rm{H}}}_{2}\mathrm{CCS}$. Upper limits to the column densities were established for each molecule in each source and are summarized in Tables 2 and 3, along with the pertinent line parameters. The observational spectra around each transition used to calculate the upper limits, as well as a simulation of the molecular spectra using those upper limits, are shown in Figures 2 and 3 in Appendix B.

Given the large number of nondetections presented here, we conclude that neither ${{\rm{H}}}_{2}\mathrm{CCS}$ nor HCCSH are substantial interstellar sulfur reservoirs. These species, along with c-${{\rm{H}}}_{2}{{\rm{C}}}_{2}{\rm{S}}$, however, remain reasonable candidates for interstellar detection in sensitive observations, but probably in a source very rich in sulfur-bearing species. We note that the upper limits reported here could be substantially improved if additional laboratory spectroscopy is performed. As shown in Figure 1, the laboratory data for both HCCSH and ${{\rm{H}}}_{2}\mathrm{CCS}$ fail to cover the strongest transitions of these species in warm environments. In sources where line confusion is not an issue, the errors introduced by extrapolation would likely not preclude detection, but would instead require additional observing time owing to the need for wider frequency coverage. In line-confused sources, however, such as NGC 6334I in ALMA Band 10, frequency extrapolation cannot be made with confidence, making any purported detection highly tenuous. Similarly, while the strong b-type branch of HCCSH near 1.4 THz is likely to be identifiable with SOFIA even if the frequencies are slightly uncertain, the search space would be substantially narrowed by laboratory measurements in this band. Finally, a search for the higher-energy cyclic isomer, c-${{\rm{H}}}_{2}{{\rm{C}}}_{2}{\rm{S}}$, is currently impossible due to the lack of any laboratory spectra.

A search for HCCSH and ${{\rm{H}}}_{2}\mathrm{CCS}$ in a number of astronomical line surveys based on newly reported laboratory rest frequencies obtained with a combination of microwave and submillimeter spectroscopy is presented. Nondetections are reported in all sources, suggesting that these molecules are not substantial reservoirs of sulfur, nor can they be readily used to infer sulfur chemistry. A possible explanation for the absence of HCCSH is an analogous destruction pathway to l-propadienone by barrierless, exothermic reaction with atomic hydrogen in the solid state. The detection of the strongest transitions of HCCSH, which arise at THz frequencies, remains a possibility, and would be substantially aided by additional enabling laboratory work.

This paper makes use of the following ALMA data: ADS/JAO.ALMA#2015.A.00022.T and #2017.1.00717.S. ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada) and NSC and ASIAA (Taiwan) and KASI (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO and NAOJ. The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. The Green Bank Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. Support for B.A.M. was provided by NASA through Hubble Fellowship grant #HST-HF2-51396 awarded by the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., for NASA, under contract NAS5-26555. C.N.S. thanks the Alexander von Humboldt Stiftung/Foundation for their generous support. M.C.M. acknowledges support from NASA grants NNX13AE59G and 80NSSC18K0396, and NSF grant AST–1615847. M.-A. M.-D. is thankful to the Programme National "Physique et Chimie du Milieu Interstellaire" (PCMI) of CNRS/INSU with INC/INP co-funded by CEA and CNES for support.

The physical parameters assumed for each source examined here are provided in Tables 4 and 5.

Figures 2 and 3 show the transition used to calculate the upper limit in each source, simulated using the upper limit column density and parameters given in Tables 2–5.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size