betaFeedback

WorksPhysics and Astronomy

Solid-state formation of H2_2CS and CH3_3SH in molecular clouds: Carbon-initiated reactions with atomic and molecular hydrogen

Abstract

The solid-state formation pathways of thioformaldehyde (H2_2CS) and methanethiol (CH3_3SH) in dense molecular clouds remain poorly constrained. We investigate carbon-initiated reactions in hydrogen sulfide (H2_2S) ice in the presence of atomic or molecular hydrogen. We co-deposited H2_2S, ground-state atomic carbon, and H or H2_2 at 10 K under ultra-high vacuum. We monitored the reaction products using reflection absorption infrared spectroscopy (RAIRS) and characterized their thermal desorption using temperature-programmed desorption (TPD) with quadrupole mass spectrometry. Gas-phase ab initio calculations were used to assess pathways to CH3_3SH formation. Atomic carbon reacted with H2_2S at 10 K to form H2_2CS, detected as both monomers and H2_2CS-derived oligomers. With atomic H, successive H addition to H2_2CS leads to CH3_3SH. With molecular H2_2, CH3_3SH also forms efficiently. The results support an alternative pathway in which C reacts with H2_2 to form CH2_2, which abstracts H from H2_2S. The resulting CH3_3 and SH radicals can then recombine to form CH3_3SH. The infrared bands and TPD fragmentation patterns support the formation of H2_2CS and CH3_3SH in the ice at 10 K. These findings expand the set of viable solid-state pathways to sulfur-bearing molecules and constrain models of sulfur chemistry in prestellar environments.

Introduction

The detection of molecules in the interstellar medium provides fundamental constraints on the physical and chemical evolution of star- and planet-forming regions. Molecular inventories, abundance ratios, and spatial and thermal distributions encode the history of formation, processing, and transport of matter, making the origin of simple and complex species a central question in astrochemistry (e.g., [48, 113, 28]). Although most of the more than 350 molecules1 identified in interstellar and circumstellar environments have been detected in the gas phase through their rotational spectra, many abundant H-rich species are thought to form efficiently on icy grain surfaces such as H2OH_2O, CH4CH_4, NH3NH_3, and CH3OHCH_3OH ([32, 87, 37, 108, 41] (Oppolet al. 2008)).

Laboratory solid-state experiments under dense molecular cloud conditions have shown that complex organic molecules (COMs) can form via non-energetic atom-addition and abstraction chemistry, without requiring thermal diffusion of heavy radicals. Instead, reactive intermediates generated in close proximity can undergo prompt, non-diffusive radical-radical recombination reactions (e.g., [36, 25] (Chuang et al. 2016)). This mechanism has since been incorporated into modern astrochemical models ([57, 44, 12]). In the dense and cold interiors of prestellar cores, the majority of molecules are frozen out onto grain surfaces, so interstellar ices constitute the main chemical reservoir at this stage (e.g., [18, 7, 17]). Consistent with this picture, infrared observations, including recent JWST results, show that these ices are chemically rich, containing both simple molecules and signatures of O-bearing complex species ([11, 111, 78]).

In this context, solid-state sulfur chemistry remains a longstanding puzzle in its chemical evolution and potential carriers. The cosmic abundance of sulfur is derived from its ionized form, S+\mathrm{S}^{+}, in the diffuse interstellar medium, yielding S/H ∼1.7×10−5\sim1.7 \times10^{-5} ([35]), close to the solar value of S/H ∼1.3×10−5\sim1.3 \times10^{-5} ([6]). However, observations show that in dense clouds only a small fraction of sulfur is found in the gas phase, with depletion factors of two to three orders of magnitude relative to diffuse environments (e.g., [90, 55, 106, 68, 42]). This so-called missing sulfur problem implies the existence of major sulfur reservoirs that are not directly observable.

In the solid state, OCS is the only S-bearing molecule that has been securely identified in interstellar ices ([85, 78]), with abundances typically of 0.05–0.15% relative to H2O\mathrm{H}_{2}\mathrm{O} ice ([10, 78]), while SO2\mathrm{SO}_{2} has only been tentatively identified at inferred abundances below 0.1% relative to water ([78]), highlighting how little is known about solid-phase carriers of sulfur. Gaseous ionic or neutral sulfur atoms are expected to accrete and react with abundant and mobile H atoms on grain surfaces at low temperatures. Therefore, H2S\mathrm{H}_{2}\mathrm{S} is often assumed to be the main sulfur carrier in ices. However, observational constraints of solid H2S\mathrm{H}_{2}\mathrm{S} toward prestellar and protostellar sources yield only stringent upper limits, with H2S/H2O≲0.6\mathrm{H}_{2}\mathrm{S}/\mathrm{H}_{2}\mathrm{O} \lesssim0.6–0.7% (and as low as 0.13% in some lines of sight), implying that H2S\mathrm{H}_{2}\mathrm{S} accounts for at most a minor fraction of the solid-phase sulfur budget ([97, 56, 78] (Slavicinska et al. 2025)).

Moreover, laboratory and observational evidence indicates that a significant fraction of H2S\mathrm{H}_{2}\mathrm{S} does not remain intact, but is rapidly converted to other sulfur reservoirs. In particular, H2S\mathrm{H}_{2}\mathrm{S} readily undergoes acid–base reactions with NH3\mathrm{NH}_{3} at low temperatures to form ammonium hydrosulfide (NH4+SH−\mathrm{NH}_{4}^{+}\mathrm{SH}^{-}), a species that may account for up to ∼\sim17–18% of the sulfur budget in dense regions and could be a major carrier of the ubiquitous 6.85 μ\mum ice feature ([4, 107] (Slavicinska et al. 2025)). Additional sulfur may be locked into more refractory forms, including sulfur minerals and allotropes ([106, 61, 68, 94, 19, 16, 46]), but the detectability of these reservoirs remains highly uncertain.

In contrast to the limited number of sulfur carriers securely identified in interstellar ices, gas-phase observations reveal a diverse sulfur inventory. Sulfur-bearing molecules have been detected in a wide range of environments, from cold cores to hot cores, hot corinos, and Galactic Center molecular clouds, and include both simple species, increasingly complex organosulfur molecules and sulfur-bearing carbon chains ([1, 21]). Recent surveys have expanded this inventory with sulfur analogs of O-bearing species, including dimethyl sulfide (CH3SCH3\mathrm{CH}_{3}\mathrm{SCH}_{3}), the S analog of dimethyl ether ([92]), and larger cyclic S-bearing species such as 2,5-cyclohexadien-1-thione ([5]). These detections emphasize that sulfur can participate in rich organic chemistry even though its main solid-state and gas-phase reservoirs together remain insufficient to close the sulfur budget (e.g., [31, 89, 2]).

In the gas phase, sulfur-containing organic molecules such as thioformaldehyde (H2CS\mathrm{H}_{2}\mathrm{CS}) and methanethiol (CH3SH\mathrm{CH}_{3}\mathrm{SH}) are central to understanding how sulfur is incorporated into organic chemistry in the interstellar medium. As direct sulfur analogs of H2CO\mathrm{H}_{2}\mathrm{CO} and CH3OH\mathrm{CH}_{3}\mathrm{OH}, they have been detected in a wide range of environments, such as prestellar cores, massive hot cores and low-mass protostars (e.g., [74, 58, 31, 105]). Recent studies have further extended these detections to their deuterated counterparts, with HDCS observed in protostellar environments ([31]), and, more recently, the first interstellar detection of singly deuterated methanethiol (CH2DSH\mathrm{CH}_{2}\mathrm{DSH}) in IRAS 16293-2422 B ([14]).

The formation of H2CS\mathrm{H}_{2}\mathrm{CS} and CH3SH\mathrm{CH}_{3}\mathrm{SH} on grain surfaces has frequently been discussed within a successive hydrogenation framework, often by analogy with CO hydrogenation, in which sulfur-bearing precursors (e.g., CS, HCS, and H2CS\mathrm{H}_{2}\mathrm{CS}) are converted through stepwise H-addition reactions to more saturated products ([74, 80, 106, 69, 68, 73, 30]). Alternative atom/radical-addition routes, such as CH2+S→H2CS\mathrm{CH}_{2} + \mathrm{S} \rightarrow\mathrm{H}_{2}\mathrm{CS} and CH3+SH→CH3SH\mathrm{CH}_{3} + \mathrm{SH} \rightarrow\mathrm{CH}_{3}\mathrm{SH}, have also been included in astrochemical networks ([80]). However, direct laboratory evidence for these proposed solid-state pathways remains limited. Recent experiments have reported H2CS\mathrm{H}_{2}\mathrm{CS} and CH3SH\mathrm{CH}_{3}\mathrm{SH} formation upon hydrogenation of CS2\mathrm{CS}_{2} ([83]), and CH3SH\mathrm{CH}_{3}\mathrm{SH} and more complex S-bearing molecule formation upon electron irradiation of H2S\mathrm{H}_{2}\mathrm{S} and CH4\mathrm{CH}_{4} ice mixtures ([47]). Nevertheless, experiments directly probing hydrogenation networks starting from CS or H2CS\mathrm{H}_{2}\mathrm{CS} are still missing, in part because work with these reactive and/or hazardous sulfur-bearing precursors is technically challenging. As a result, the feasibility and product outcomes of these hydrogenation pathways under astrophysically relevant conditions remain experimentally poorly constrained.

Alternative formation routes involving the expected first-generation of S-bearing ice, like H2S\mathrm{H}_{2}\mathrm{S}, have also been proposed to contribute to H2CS\mathrm{H}_{2}\mathrm{CS} formation. Laas & Caselli [68] proposed that atomic carbon can react with H2S\mathrm{H}_{2}\mathrm{S} ice forming HCS which is efficiently converted into H2CS\mathrm{H}_{2}\mathrm{CS} through hydrogenation ([68]). This framework shifts the chemical entry point from prerequisite CS to carbon-initiated reactivity in ices involving H2S\mathrm{H}_{2}\mathrm{S}, which has also been experimentally studied in the gas-phase ([60, 29]).

This carbon-initiated H2S\mathrm{H}_{2}\mathrm{S} chemistry also motivates considering the role of the hydrogen reservoir itself in the ice reactivity ([26, 75]). While atomic H is well established as a key driver of grain-surface chemistry, particularly through successive hydrogenation and H-abstraction reactions (e.g., [108, 41, 87, 38, 28] (Chuang et al. 2016)), the potential contribution of molecular hydrogen (H2\mathrm{H}_{2}) to solid-state reaction networks remains much less constrained. Under appropriate conditions, H2\mathrm{H}_{2} can become chemically active when interacting with atomic carbon ([70, 40]). In particular, insertion of atomic carbon in H2\mathrm{H}_{2} has been proposed as an entry point to hydrocarbon chemistry in ices leading to CH4\mathrm{CH}_{4} formation ([70]), and recent astrochemical models suggest that C + H2\mathrm{H}_{2} reactions in interstellar ices can generate CH2\mathrm{CH}_{2} and CH3\mathrm{CH}_3 radicals, thus opening new pathways toward complex organic molecules [44, 12, 66]. Solid-state reactions following C+H2→CH2\mathrm{C} + \mathrm{H}_2 \rightarrow\mathrm{CH}_2 routes have been proposed as an efficient means to enrich the CHx\mathrm{CH}_x radicals in ices, yet they remain untested under laboratory ice-analog conditions. Their efficiency and role in driving subsequent ice chemistry therefore remain poorly constrained.

In this work, we experimentally investigate carbon-initiated solid-state routes linking H2S\mathrm{H}_2\mathrm{S}-containing ices to the formation of H2CS\mathrm{H}_2\mathrm{CS} and CH3SH\mathrm{CH}_3\mathrm{SH}. Using cryogenic ultra-high-vacuum co-deposition experiments, we (i) examine whether atomic carbon reacts directly with H2S\mathrm{H}_2\mathrm{S} to form H2CS\mathrm{H}_2\mathrm{CS} (C+H2S→H2CS\mathrm{C} + \mathrm{H}_2\mathrm{S} \rightarrow\mathrm{H}_2\mathrm{CS}), (ii) test successive H-atom addition to H2CS\mathrm{H}_2\mathrm{CS} as a route toward CH3SH\mathrm{CH}_3\mathrm{SH} (H2CS+2H→CH3SH\mathrm{H}_2\mathrm{CS} + 2\mathrm{H} \rightarrow\mathrm{CH}_3\mathrm{SH}), and (iii) probe an H2\mathrm{H}_2-enabled pathway in which C+H2\mathrm{C} + \mathrm{H}_2 chemistry generates the CH2\mathrm{CH}_2 intermediate which can subsequently react with H2S\mathrm{H}_2\mathrm{S} to form CH3SH\mathrm{CH}_3\mathrm{SH}. To support the interpretation of CH3SH\mathrm{CH}_3\mathrm{SH} formation in the experiments involving H2\mathrm{H}_2, we complement the laboratory measurements with gas-phase ab initio calculations. These calculations compare the singlet and triplet CH2+H2S\mathrm{CH}_2 + \mathrm{H}_2\mathrm{S} reaction channels and the alternative H2+H2CS\mathrm{H}_2 + \mathrm{H}_2\mathrm{CS} pathway, assessing their energetic feasibility and helping to constrain the mechanism underlying the experimental results. By probing these reaction channels under controlled interstellar ice analog conditions, this work assesses whether atomic carbon can drive the conversion of simple sulfur-bearing ices into more complex organosulfur species under dense-cloud conditions.

Experimental methods

Experiments were carried out using the SURFRESIDE3 ultra-high vacuum (UHV) setup. This setup has been optimized to experimentally study the solid-state atomic reactions on dust grains under dense cloud conditions and enables surface reactions in interstellar-ice analogs at cryogenic temperatures on a gold-plated substrate, while monitoring solid-state chemistry with Fourier-transform reflection absorption infrared spectroscopy (FT-RAIRS) and gas-phase species during temperature-programmed desorption using a quadrupole mass spectrometer (TPD-QMS). SURFRESIDE3 has been described in detail by Ioppolo et al. (2013) and Qasim et al. [88].

Interstellar ice analogs were grown on a gold-plated copper substrate mounted to the tip of a closed-cycle He cryostat at the center of the main chamber that is kept at a base pressure of ∼5×10−10\sim5 \times10^{-10} mbar. The substrate temperature can be controlled between 8 and 450 K by resistive heating and it is read by two Si-diode sensors that hold an absolute accuracy of ∼0.5\sim0.5 K. Co-deposition through independent inlet lines allowed us to combine ground-state C atoms (i.e., C(3P)\mathrm{C}(^3\mathrm{P}), H atoms, H2S\mathrm{H}_2\mathrm{S}, and H2\mathrm{H}_2. H2S\mathrm{H}_2\mathrm{S} was admitted through an all-metal leak valve. H atoms were supplied by a hydrogen atom beam source (HABS) using thermal cracking to produce the atom beam. The mixed beam (H + residual H2\mathrm{H}_2) was cooled by a nose-shaped quartz tube prior to the surface. For the H2\mathrm{H}_2 experiments, the same line and leak-valve setting were used with the dissociation filament off. The unchanged inlet setting maintains the nominal hydrogen supply to the source. For reporting purposes, Table 2 lists a nominal value of 0.5×10130.5 \times10^{13} molecules cm−2 s−1\mathrm{cm}^{-2}\ \mathrm{s}^{-1}, taken as half the calibrated H-atom flux under the idealized assumptions of complete dissociation in H-atom mode and equivalent beam transmission.

#ExperimentTT (K)H2S\mathrm{H_2S} (cm−2^{-2} s−1^{-1})C (cm−2^{-2} s−1^{-1})H (cm−2^{-2} s−1^{-1})H2\mathrm{H_2} (cm−2^{-2} s−1^{-1})H2S:C:H2/H\mathrm{H_2S:C:H_2/H}Time (s)
1H2S\mathrm{H_2S}10∼1×1013\sim1 \times10^{13}…\ldots…\ldots…\ldots20:−:−20:-:-1800
2H2S+C\mathrm{H_2S+C}10∼1×1013\sim1 \times10^{13}∼5×1011\sim5 \times10^{11}…\ldots…\ldots20:1:−20:1:-3600
3H2S+C+H\mathrm{H_2S+C+H}10∼1×1013\sim1 \times10^{13}∼5×1011\sim5 \times10^{11}∼1×1013\sim1 \times10^{13}…\ldots20:1:2020:1:203600
4H2S+C+H2\mathrm{H_2S+C+H_2}10∼1×1013\sim1 \times10^{13}∼5×1011\sim5 \times10^{11}…\ldots∼0.5×1013\sim0.5 \times10^{13}20:1:1020:1:103600
5CH3SH\mathrm{CH_3SH}10…\ldots…\ldots…\ldots…\ldots…\ldots3600

Table 2. Overview of the experiments performed in this work

A customized SUKO-A 40 C-atom source based on a commercial design (Dr. Eberl, MBE; Krasnokutski & Huisken [65]; Albar et al. [3]) produced a beam dominated by C(3P)\mathrm{C}(^3\mathrm{P}) atoms, with a cluster contamination below 1%1\%. The contribution of electronically excited C(1D)\mathrm{C}(^1\mathrm{D}) is expected to be negligible because its excitation energy is substantially higher than the operating temperature of the source [88, 70]. The beam was collimated onto the substrate by a set of apertures to limit the deposition of C atoms on the walls of the UHV chamber. CO and CO2\mathrm{CO}_2 are known contaminants in the C-beam and their deposition rates were estimated to be of the order of 1−3×1012 cm−2 s−11 - 3 \times10^{12}\ \mathrm{cm}^{-2}\ \mathrm{s}^{-1} and 2−5×109 cm−2 s−12 - 5 \times10^{9}\ \mathrm{cm}^{-2}\ \mathrm{s}^{-1}. The CO deposition rate was lower than that of H2S\mathrm{H}_2\mathrm{S} and H2/H2\mathrm{H}_2/\mathrm{H}_2, which is expected to favor encounters of incoming C atoms with these species over CO in a well-mixed ice. Comparable CO and CO2\mathrm{CO}_2 contamination levels were observed across the experiments, which were performed sequentially under similar conditions. The contribution of CO contamination is limited due to the non-detection of CO hydrogenation products, such as H2CO\mathrm{H}_2\mathrm{CO} and CH3OH\mathrm{CH}_3\mathrm{OH}. Each atom-beam line (HABS or C-source) was differentially pumped to a base pressure of ≤1×10−9\leq1 \times10^{-9} mbar and connected to the main UHV chamber through a shutter valve.

The mixed ices were deposited at 10 K and monitored in situ using FT-RAIRS with a 700-4000 cm−1\mathrm{cm}^{-1} range and 1 cm−1\mathrm{cm}^{-1} resolution. After deposition, the sample was heated at 5 K min−1\mathrm{min}^{-1} for TPD. The desorbing species were ionized at 70 eV and detected by a QMS, while the solid phase was tracked simultaneously by FT-RAIRS.

The gaseous samples used in this work were hydrogen sulfide (H2S\mathrm{H}_2\mathrm{S}, Linde Gas, ≥99.5%\geq99.5\%) and molecular hydrogen (H2\mathrm{H}_2, Linde Gas, ≥99.999%\geq99.999\%). Carbon filaments for the SUKO-A source were purchased from MBE Komponenten GmbH.

Experimental analysis

The absolute column density NN was derived using the Beer–Lambert relation:

NX=ln⁡10A′(X)∫Abs(ν) dν,(1)N_X = \frac{\ln10}{A'(X)} \int Abs(\nu)\,d\nu, \tag*{(1)}

where Abs(ν)Abs(\nu) is the integrated absorbance of IR features and A′A' is the apparent band strength of the transition. For H2S\mathrm{H}_2\mathrm{S}, we used a setup-specific band strength of A′=(4.7±0.1)×10−17 cm molecule−1A' = (4.7 \pm0.1) \times10^{-17}\ \mathrm{cm\ molecule}^{-1} for the applied RAIRS configuration, as determined by Santos et al. [91]. The H flux has been calibrated in situ for various H2\mathrm{H}_2 flow rates on the SURFRESIDE3 and determined to be around 1×10131 \times10^{13} atoms cm−2 s−1\mathrm{cm}^{-2}\ \mathrm{s}^{-1} under the applied experimental conditions. The flux of carbon atoms was determined prior co-deposition experiments when a new carbon filament was installed, following the method described by Qasim et al. [88]. Briefly, C atoms were co-deposited with a large excess of 18O2{}^{18}\mathrm{O}_2; the reaction between C and 18O2{}^{18}\mathrm{O}_2 yields 18C18O{}^{18}\mathrm{C}^{18}\mathrm{O} and 18O{}^{18}\mathrm{O}, with the latter reacting further with excess 18O2{}^{18}\mathrm{O}_2 to form 18O3{}^{18}\mathrm{O}_3. The 18O3{}^{18}\mathrm{O}_3 column density, derived from its RAIRS absorption band, was then used as an indirect measure of the C-atom flux reaching the substrate. All co-depositions were performed with comparable fluxes and deposition times, enabling qualitative comparison of relative band intensities across experiments.

The relative abundances of CH3SH\mathrm{CH}_3\mathrm{SH} and H2CS\mathrm{H}_2\mathrm{CS} were derived from the TPD-QMS data using the following method. For CH3SH\mathrm{CH}_3\mathrm{SH} and H2CS\mathrm{H}_2\mathrm{CS}, molecular ions were used at m/z=48m/z = 48 and 46, respectively, for quantification. The relative abundance ratio was obtained with the method described by Martín-Doménech et al. [77] using the ratio of the integrated TPD peak areas, A(m/z)A(m/z), corrected for the electron-impact ionization cross sections, σ+\sigma^+, the fragmentation factor, defined as the fraction of molecules that yield a fragment of mass mm in the mass spectrometer, FF(m)F_F(m), and the QMS sensitivity at the corresponding mass channels, SS (m/zm/z):

NXNY=A(m/z(X))A(m/z(Y))⋅σ+(Y)σ+(X)⋅FF(Y)FF(X)⋅S(m/z(Y))S(m/z(X))(2)\frac{N_X}{N_Y}=\frac{A(m/z(X))}{A(m/z(Y))}\cdot\frac{\sigma^+(Y)}{\sigma^+(X)}\cdot\frac{F_F(Y)}{F_F(X)}\cdot\frac{S(m/z(Y))}{S(m/z(X))} \tag*{(2)}

Following Martín-Doménech et al. [77], equal effective pumping speeds are assumed, so no additional correction for relative pumping speeds is applied in Eq. (2). Since experimental ionization cross sections at 70 eV were not available for these species, σ+\sigma^+ was estimated from the molecular polarizability α\alpha using the empirical relation σmax⁡+(X)=cα(X)\sigma^+_{\max}(X)=c\alpha(X), with c=1.48c=1.48 Å−1^{-1} (Hudson et al. [49]; Bull & Harland [13]). Adopting α=5.186\alpha=5.186 Å3^3 for CH3_3SH and α=4.218\alpha=4.218 Å3^3 for H2_2CS yields σ+\sigma^+(CH3_3SH) = 7.68 Å2^2 and σ+\sigma^+(H2_2CS) = 6.24 Å2^2. Fragmentation fractions were derived from available mass spectra databases, using NIST for CH3_3SH and SpectraBase for H2_2CS. The parameters adopted for all species are summarized in Table 1.

Speciesm/zm/zα\alpha (Å3^3)σ+\sigma^+ (Å2^2)FF(m)F_F(m)SS (m/zm/z)a^a
CH3_3SH485.1867.680.330.18
H2_2CS464.2186.240.390.17

Table 1. Parameters used to derive the relative abundances of CH3_3SH and H2_2CS from the TPD-QMS data.

An overview of the experiments performed is given in Table 2 including the deposition temperature, the atomic and molecular fluxes, and the deposition time.

Ab initio calculations

All electronic structure calculations employed the MOLPRO quantum chemistry package (Werner et al. [110]). The stationary structures on both singlet and triplet potential energy surfaces (PESs) of CH3_3SH were optimized at CCSD(T)-F12b/VTZ-F12 level of theory (Knizia et al. [64]; Feller & Peterson [39]). Harmonic vibrational analyses were subsequently done to confirm the nature of the optimized molecular structures as either minima (all real frequencies) or transition states (only one imaginary frequency); these calculations also yielded harmonic zero-point energies (ZPEs).

To understand the connections between the various transition states and associated minima, we performed one-dimensional PES cuts along suitable optimized reaction coordinate (ORC) paths (Varandas & Rocha [104]); this involves the selection of a single reactive coordinate responsible for the envisaged chemical process, with all other degrees of freedom being fully optimized at every selected discrete (fixed) value of such a coordinate (see, e.g., Varandas & Rocha [104]). Due to convergence difficulties of the Hartree-Fock (HF) and CCSD(T) wave functions at long-range regions of the PES, as well as the increasing multireference character of the wavefunction, ORC paths involving transition states connecting valence structures to dissociation channels were done at the complete active space self-consistent field (CASSCF) level of theory (Szalay et al. [101]) using the aug-cc-pV(T+d)Z (AVQZ) basis set (Dunning [33]; Kendall et al. [63]). The CASSCF active space consists of 14 electrons in 12 orbitals.

Final structures and energetics were all obtained at the CCSD(T)-F12b/VTZ-F12. Note that, for dissociation channels, asymptotic energies were computed as the sum of the energies of the isolated fragments at infinite separation, each evaluated at the CCSD(T)-F12b/VTZ-F12 level.

Results

Formation of H2_2CS and CH3_3SH

To directly address the reaction pathways introduced in Section 1, the RAIRS results are organized around three comparisons: the reaction of C atoms with H2_2S to form H2_2CS, the further hydrogenation promoted by additional H atoms toward CH3_3SH, and the formation of CH3_3SH through a CH2_2-initiated pathway enabled by H2_2 co-deposition. Figure 1 shows the IR spectra at 10 K for pure H2_2S, H2_2S + C (Exp. 2), and H2_2S + C + H (Exp. 3) experiments in the spectral range of 1500–690 cm−1^{-1}. For comparison, a reference spectrum of pure CH3_3SH at 10 K is shown at the top.

Figure 1. RAIRS spectra of the co-deposition experiments performed at 10 K for: (Exp. 1) H2S\mathrm{H_2S}, (Exp. 2) H2S\mathrm{H_2S} co-deposited with atomic C, (Exp. 3) H2S+C+H\mathrm{H_2S+C+H}, (Exp. 4) H2S+C+H2\mathrm{H_2S+C+H_2}, and (Exp. 5) CH3SH\mathrm{CH_3SH} reference, scaled by a factor of 0.1. Orange lines indicate the vibrational modes of CH3SH\mathrm{CH_3SH}. Dashed vertical lines highlight the new absorption features that appear in the H2S+C\mathrm{H_2S+C} experiments. Dotted lines mark characteristic H2S\mathrm{H_2S} and CH4\mathrm{CH_4} bands present in this spectral region.

(i) H2_2CS formation in the C + H2_2S experiment. In the spectral window of 1500–690 cm−1^{-1}, the IR spectrum of pure H2_2S (Exp. 1) is dominated by the ν2\nu_2 bending mode at 1170 cm−1^{-1} (Hudson & Gerakines [51]). Upon co-deposition of H2_2S with atomic carbon (Exp. 2), this feature remains present as a parent remnant, while several additional absorption bands appear, indicating the solid-state product formation at 10 K.

In Exp. 2, the feature at 995 cm−1^{-1} is assigned primarily to the CH2_2 wagging mode (ν4\nu_4) of monomeric H2_2CS, with a possible contribution from the nearly degenerate CH2_2 rocking mode (ν6\nu_6), in agreement with previous matrix-isolation and solid-state studies reporting this vibration at 987 cm−1^{-1} (Jacox & Milligan 1975; Suzuki et al. [100]) and near 990 cm−1^{-1} (Shao et al. [93]). The C=S stretching mode (ν3\nu_3) of H2_2CS is intrinsically much weaker; it is observed in this work at around 1046 cm−1^{-1}, compared to 1055.2 cm−1^{-1} reported by Suzuki et al. [100], a shift of 9 cm−1^{-1} attributable to matrix and ice-environment effects. This assignment is also consistent with recent high-level anharmonic calculations showing that the ν4\nu_4 mode near 1000 cm−1^{-1} is the strongest infrared feature of H2_2CS, whereas the ν3\nu_3 near 1060 cm−1^{-1} is significantly weaker (Guo et al. [45]).

Furthermore, the formation of H2_2CS is accompanied by the simultaneous detection of oligomeric forms of H2_2CS, (H2_2CS)x_x, at 740 cm−1^{-1}. This assignment is consistent with the matrix-isolation study of Jacox & Milligan (1975), who reported the strongest infrared absorption of the cyclic H2_2CS trimer, s-trithiane, (H2_2CS)3_3, at 735 cm−1^{-1}. However, because other H2_2CS oligomers may also absorb in this spectral region, the present feature cannot be attributed uniquely to s-trithiane. We therefore interpret the 740 cm−1^{-1} band as evidence for H2_2CS oligomerization, potentially including contributions from the cyclic trimer and other (H2_2CS)x_x species. Additional weaker bands are observed at 1023 and 1230 cm−1^{-1} in our experiments, although their exact origins remain uncertain. These features may tentatively arise from minor vibrational modes of H2_2CS-derived oligomers or clusters.

The detection of (H2_2CS)x_x indicates that newly formed H2_2CS monomers can undergo association or oligomerization under the present deposition conditions. This assignment is consistent with the propensity of H2_2CS to form oligomeric or clustered species, whereas isolated monomeric H2_2CS is typically generated in situ. (H2_2CS)3_3 is one of the stable oligomeric forms of H2_2CS and has been used as a commercially available precursor for H2_2CS generation via thermal decomposition at 700–800°C. Alternatively, monomeric H2_2CS can be generated by UV photolysis or thermal degradation of suitable precursors such as CH3_3SH, followed

by rapid trapping at low temperature (Jacox & Milligan 1975; Suzuki et al. 2007; Müller et al. 2019).

(ii) CH3SH\mathrm{CH_3SH} formation in the presence of atomic H. In the H2S+C+H\mathrm{H_2S+C+H} experiment (Exp. 3), several new absorption features appear exclusively when H atoms participate in the solid-state reactions, indicating the formation of hydrogenated products. For example, a band attributed to CH4\mathrm{CH_4} is observed near 1300 cm−1^{-1} resulting from the direct hydrogenation of C atoms [87].

In Exp. 3, the additional IR bands at 964, 1065, and 1438 cm−1^{-1} are assigned to the CH3\mathrm{CH_3} rocking mode (ν11\nu_{11}), the CH3\mathrm{CH_3} deformation/rocking mode (ν6\nu_6), and the CH3\mathrm{CH_3} antisymmetric deformation modes (ν4/ν10\nu_4/\nu_{10}) of CH3SH\mathrm{CH_3SH} (see the IR reference spectrum of Exp. 5). Other weaker IR transitions of CH3SH\mathrm{CH_3SH} at ∼\sim704 cm−1^{-1} (ν9\nu_9), ∼\sim800 cm−1^{-1} (ν7\nu_7), and ∼\sim1323 cm−1^{-1} (ν5\nu_5) bands, are barely detected due to the relatively high S/N ratio required to identify these weaker features. The observed band positions and relative prominence are consistent with the laboratory spectra reported of amorphous CH3SH\mathrm{CH_3SH} ice at 17 K [50].

(iii) CH3SH\mathrm{CH_3SH} formation in the presence of H2\mathrm{H_2}. In the IR spectrum of H2S+C+H2\mathrm{H_2S+C+H_2} (Exp. 4), the formation of H2CS\mathrm{H_2CS} is again confirmed by the characteristic absorption bands at 995 and 740 cm−1^{-1}, assigned to monomeric H2CS\mathrm{H_2CS} and oligomeric (H2CS\mathrm{H_2CS})x_x, respectively. The weaker features at 1046, 1024, and 1230 cm−1^{-1}, also observed in the C + H2S\mathrm{H_2S} experiment, are present as well. In addition, the CH4\mathrm{CH_4} band at 1300 cm−1^{-1} is detected, consistent with the additional reaction pathway initiated by C + H2\mathrm{H_2} →\rightarrow CH2\mathrm{CH_2}, followed by CH2\mathrm{CH_2} + H2\mathrm{H_2} →\rightarrow CH4\mathrm{CH_4}, as reported by Lamberts et al. [70]. Besides these products, CH3SH\mathrm{CH_3SH} is also clearly identified. The strongest diagnostic CH3SH\mathrm{CH_3SH} bands are observed at 964, 1065, and 1438 cm−1^{-1}, while the weaker features at 704 and 1323 cm−1^{-1} are less prominent, consistent with the relative band intensities observed in pure CH3SH\mathrm{CH_3SH} ice (Exp. 5). The positions of the bands and the overall spectral profile closely match those of the reference CH3SH\mathrm{CH_3SH} infrared spectrum.

Temperature-programmed desorption of products

After co-deposition, the ice samples were warmed up to 300 K with a temperature ramp of 5 K.min−1^{-1} to monitor the desorption behavior of the newly formed sulfur-bearing species. The top panels of Figure 2 show the TPD-QMS data obtained for the (a) H2_2S + C, (b) H2_2S + C + H2_2, and (c) H2_2S + C + H experiments (Exp. 2, 4 and 3). The bottom panels display the corresponding integrated intensities of the selected mass fragments, compared to the reference mass-fragmentation patterns of (d) CH3_3SH and (e) H2_2CS from the NIST6 and SpectraBase7 databases. The integrated intensities shown in panel (d) were obtained from Exp. 3 for the CH3_3SH desorption feature over 90–130 K, while those in panel (e) were obtained from Exp. 2 for the H2_2CS-derived oligomer desorption features over 130–170 K, 170–205 K, and 205–240 K, corresponding to the peaks at ∼150, 190, and 220 K, respectively.

Figure 2. Top: TPD mass spectra for ices co-deposited at 10 K in (a) H2_2S + C (Exp. 2), (b) H2_2S + C + H2_2 (Exp. 4), and (c) H2_2S + C + H (Exp. 3). Bottom: comparison of the integrated TPD peak intensities within four temperature intervals (112 K, 150 K, 190 K and 220 K) with the expected fragmentation patterns of CH3_3SH (d) and H2_2CS (e) from the literature. Error bars represent the uncertainties arising from the integration fits.

The first TPD-QMS feature, centered at ∼110 K, is attributed to CH3_3SH desorption. The CH3_3SH desorption temperature is in agreement with the value reported by Narayanan et al. [82] for thin multilayer 13^{13}CH3_3SH ices. Furthermore, the relative intensities of the selected mass signals (i.e., m/zm/z 44–48) are consistent with the standard CH3_3SH fragmentation pattern, as shown in Figure 2d.

In addition to the CH3_3SH peak, three TPD-QMS features are detected centered at 150, 190, and 220 K. All three peaks consist of highly similar mass-fragmentation patterns (Figure 2e), indicating that they originate from oligomeric species rather than from structurally distinct molecules. For lower mass channels (m/zm/z = 44–46), the observed fragmentation pattern closely resembles that of H2_2CS reported in the SpectraBase database and for the cyclic (H2_2CS)4_4 species in the NIST database. In mass spectral databases, H2_2CS (i.e., SpectraBase) and (H2_2CS)4_4 (i.e., NIST) exhibit nearly identical fragmentation patterns. Although these fragmentation patterns do not allow individual oligomers to be identified, their similarity supports the contribution of different H2_2CS-derived species, collectively denoted as (H2_2CS)x_x, to the higher-temperature desorption features.

These high-temperature desorption features occur well above the CH3_3SH desorption temperature under the same experimental conditions, suggesting that they correspond to species that are more strongly retained in the ice.

IR measurements during the TPD experiment provide complementary information on the temperature-dependent behavior of monomeric and oligomeric (H2_2CS)x_x. Figure 3 shows the integrated absorbance bands of the H2_2CS monomer band at 992 cm−1^{-1} and the oligomeric band at 740 cm−1^{-1} as a function of temperature during the TPD experiment.

Figure 3. Integrated absorbance of the H2_2CS monomer band at 992 cm−1^{-1} (circle) and the (H2_2CS)x_x band at 740 cm−1^{-1} (square) as a function of the temperature during TPD for ices deposited at 10 K, each normalized to its own initial integrated area. Error bars reflect band-integration uncertainties. Connecting lines are for visual guidance only.

The monomer band decreases rapidly between 30 and 80 K. In the same temperature range, weak m/zm/z = 45 and 46 signals are observed in the QMS data, indicating that a fraction of monomeric H2_2CS desorbs between ∼70 and 95 K (Fig. A.1 in Appendix § 7). This experimentally observed desorption range is broadly consistent with the thermal sublimation temperatures of ∼50–80 K predicted by Codella et al. [27] based on calculated H2_2CS binding energies of 3000–4600 K on amorphous water ice. The disappearance of the monomer band is not accompanied by a dominant TPD-QMS feature suggesting that desorption is not the dominant process responsible for the loss of monomeric H2_2CS from the ice. Instead, the temperature increase likely improves the mobility of the H2_2CS monomer leading to their oligomerization during warm-up. In addition, desorption of the remaining H2_2S matrix starting from 70 K may facilitate encounters between H2_2CS monomers, thereby promoting oligomerization. In contrast, the integrated IR band at 740 cm−1^{-1} increases in the same temperature range and persists up to 125 K. Its disappearance coincides with the desorption feature centered near 150 K in the TPD-QMS data, supporting a contribution from the corresponding H2_2CS-derived species to this high-temperature desorption feature. It is important to note that because the 740 cm−1^{-1} IR band may be shared by different (H2_2CS)x_x species, this correspondence does not permit an unambiguous assignment to a specific oligomer. The two weaker desorption features centered at approximately 190 and 220 K may originate from larger or more strongly bound H2_2CS-derived aggregates.

Production yields of H2_2CS and CH3_3SH

Figure 4 compares the abundance of CH3_3SH relative to H2_2CS for Exps. 2–4, calculated from the TPD-QMS data using Eq. (2). For each experiment, the H2_2CS abundance is normalized to unity, allowing the relative production of CH3_3SH under the different hydrogenation conditions to be compared.

Figure 4. Abundances of H2_2CS and CH3_3SH from TPD-QMS data under varying hydrogenation conditions, normalized to the H2_2CS abundance of each respective experiment. Conditions are ordered from low to high H/H2_2 availability. Error bars reflect uncertainties from peak integration and fitting.

In the H2_2S + C experiment (Exp. 2), CH3_3SH is detected at only approximately 10% of the H2_2CS abundance. The formation of a small amount of CH3_3SH in this experiment may result from residual H2_2 in the chamber, which is the main component of the background gas under UHV conditions. Its detection in the TPD-QMS data, despite the absence of corresponding bands in the IR spectrum, reflects the higher sensitivity of the QMS measurements.

Upon co-deposition of H2_2 in the H2_2S + C + H2_2 experiment (Exp. 4), the CH3_3SH/H2_2CS abundance ratio increases substantially to approximately 0.65. This shows that the increased availability of H2_2 promotes CH3_3SH formation, although H2_2CS remains the dominant sulfur-bearing product under these conditions. The highest relative CH3_3SH abundance is observed in the H2_2S + C + H experiment (Exp. 3), where CH3_3SH becomes as abundant as H2_2CS. Reactive H atoms can process carbon- and sulfur-bearing intermediates toward more saturated products. At the same time, H2_2 can also be present due to undissociated H2_2 from the H-atom beam and in-situ association of H atoms on the surface. The high CH3_3SH/H2_2CS ratio may therefore contain contributions from both H-atom-driven hydrogenation and H2_2-mediated chemistry, which cannot be disentangled from the present data.

Discussion

The obtained experimental results show the possible solid-state chemistry on dust grains involving astronomically relevant and abundant species, driving the transformation of H2_2S ice into organosulfur at 10 K. We demonstrate the efficient formation of H2_2CS and its oligomers, as well as the fully saturated species CH3_3SH, when C atoms accrete onto surfaces together with H2_2S and H or H2_2 at 10 K.

The formation of H2_2CS has previously been discussed in the literature via gas-phase and grain-surface chemistry [e.g. 60, 29, 69, 30]. In grain-surface networks, one proposed route involves successive H-atom addition to CS:

CS→+HHCS→+HH2CS(3)\mathrm{CS} \xrightarrow{+\mathrm{H}} \mathrm{HCS} \xrightarrow{+\mathrm{H}} \mathrm{H_2CS} \tag*{(3)}

In the theoretical network of Lamberts [69], which was largely treated without an explicit grain surface, H + CS →\rightarrow

HCS proceeds with a small gas-phase activation barrier of 3.6 kJ mol−1^{-1}. The subsequent H + HCS reaction is barrierless and can yield H2_2CS as well as cis- and trans-HCSH, while H-abstraction regenerating CS also constitutes a competing channel. Lamberts [69] additionally showed that the H + CS barrier changes only modestly when calculated on small water clusters. More recently, Di Genova et al. [30] investigated the hydrogenation network explicitly on water-ice clusters. The calculated

H + CS barrier changes only modestly upon adsorption, and inclusion of ZPE substantially lowers the effective surface barrier. Their kinetic calculations further indicate that tunneling enables efficient CS hydrogenation at 10–20 K.

In addition to sequential hydrogenation of CS, the present experiments demonstrate that H2_2CS can be formed from the interactions between atomic carbon and H2_2S in the solid state. The net reaction may be represented as

Equation not extracted
H2S+C→H2CS.(4)\mathrm{H_2S}+\mathrm{C}\rightarrow\mathrm{H_2CS}. \tag*{(4)}

While the experiments confirm that the reaction occurs, they do not constrain the underlying microscopic mechanism. Gas-phase studies suggest that C(3P)+H2S\mathrm{C}({}^{3}\mathrm{P})+\mathrm{H_2S} can form an HCSH intermediate, either through initial addition of C to sulfur followed by H-atom migration [60, 59], or through direct insertion of C into an S–H bond [43]. In the gas phase, HCSH mainly dissociates into HCS + H, whereas stabilization by the ice matrix may allow it to rearrange into H2_2CS [59]. The potential role of the ice matrix in facilitating this rearrangement is illustrated by the analog H2O+C→H2CO\mathrm{H_2O+C\rightarrow H_2CO} system studied by Molpeceres et al. [79], for which carbon insertion followed by ice-assisted isomerization was found to provide an efficient route to H2_2CO. However, the hydrogen-bond-mediated mechanism proposed for H2_2O-rich ice cannot be assumed to operate identically in pure H2_2S ice, and further theoretical calculations on the underlying H2_2S + C system in the solid state are required.

When reactive H atoms are available, such as in the case of the H2_2S + C + H experiments (Exp. 3), newly formed H2_2CS can be further processed toward more saturated sulfur-bearing species.

H2CS→+HCH2SH→+HCH3SH(5)\mathrm{H_2CS}\xrightarrow{+\mathrm{H}}\mathrm{CH_2SH}\xrightarrow{+\mathrm{H}}\mathrm{CH_3SH} \tag*{(5)}

The full H-addition and H-abstraction network linking H2_2CS to CH3_3SH was investigated in the gas-phase by Lamberts [69]. More recently, Di Genova et al. [30] extended this network to water-ice cluster models and showed that the successive H-addition reactions remain efficient at 10–20 K when tunneling is included. Lamberts [69] showed that H addition to H2_2CS preferentially forms CH2_2SH, with an activation barrier of 3.1 kJ mol−1^{-1}, whereas the competing H+H2CS→CH3S\mathrm{H+H_2CS\rightarrow CH_3S} channel has a higher barrier of 7.5 kJ mol−1^{-1}. More recently, Di Genova et al. [30] extended this network to water-ice cluster models and likewise found CH2_2SH to be the kinetically preferred product. Their kinetic calculations further show that H-atom tunneling allows these hydrogen-addition reactions to remain efficient at 10–20 K. Once CH2_2SH or CH3_3S is formed, further reaction with H can produce CH3_3SH.

In the H2_2S + C + H2_2 (Exp. 4) experiment, the successive hydrogenation via Eq. (5) cannot proceed due to the absence of H atoms. Alternatively, the active atomic carbon can react with partially trapped H2_2, resulting in the formation of triplet CH2_2 intermediates through the surface reaction:

C+H2→CH2.(6)\mathrm{C+H_2\rightarrow CH_2}. \tag*{(6)}

Laboratory and computational work by Krasnокutski et al. (2016) showed that Eq. (6) is barrierless in helium nanodroplets, and Simončič et al. [95] accordingly treated it as barrierless in their astrochemical models. However, Lamberts et al. [70] obtained activation barriers of 27–30 kJ mol−1^{-1} for a C–H2_2O complex and approximately 77 kJ mol−1^{-1} for a C–(H2_2O)3_3 complex. The difference arises from the distinct stabilization of the reactant and transition state by the local water environment, demonstrating that the barrier is sensitive to the C-atom binding site and ice structure. Using a larger (H2_2O)20_{20} cluster, Ferrero et al. [40] similarly found that C chemisorbs as 3^{3}C–OH2_2 and calculated a barrier of 70.3 kJ mol−1^{-1} for its reaction with H2_2. Together, these results indicate that the efficiency of Eq. (6) depends strongly on the adsorption state and local environment of the C atom.

The newly formed CH2_2 radicals have been suggested to be chemically reactive and to play an essential role in ice chemistry [57, 44, 95]. Hydrogen abstraction from H2_2 is expected to be barrier-mediated (49 kJ mol−1^{-1}) and thus inefficient under the experimental conditions studied [70], while the H-abstraction from H2_2S is expected to be more favorable, consistent with the efficient H-atom abstraction from H2_2S observed on icy surfaces, which proceeds through quantum tunneling across a barrier of approximately 12.7 kJ mol−1^{-1} [71, 84, 91]. CH2_2 may therefore trigger H-abstraction from H2_2S via

CH2+H2S→CH3+SH,(7)\mathrm{CH_2+H_2S\rightarrow CH_3+SH}, \tag*{(7)}

followed by radical recombination,

CH3+SH→CH3SH.(8)\mathrm{CH_3+SH\rightarrow CH_3SH}. \tag*{(8)}

Importantly, in this mechanism CH3_3SH formation does not require prior production and subsequent hydrogenation of H2_2CS, but instead depends on the availability of CH2_2 intermediates.

Computational results

To further assess the feasibility of this CH2_2-mediated pathway and to identify which reaction channel is the most favorable, quantum chemical calculations were performed for the relevant CH3_3SH formation routes.

Figure 5 shows the calculated stationary points for three reaction channels relevant to CH3_3SH formation: 1^{1}CH2_2 + H2_2S, 3^{3}CH2_2 + H2_2S, and H2_2 + H2_2CS. All energies are reported relative to the CH3_3SH minimum and include ZPE corrections. The H2_2 + H2_2CS channel is also included as an alternative pathway connecting H2_2CS to the CH3_3SH minimum; however, the high-lying transition states along this pathway indicate that the direct conversion of H2_2CS by H2_2 is unlikely to be efficient.

Figure 5. Stationary points on the ground-state singlet (gray) and triplet (green) potential energy surfaces of the CH3_3SH system and their conversion pathways. Minima and transition states are indicated with blue and red traces, respectively, while the dissociation channels are shown in black. Optimized structures and energies were obtained at CCSD(T)-F12b/VTZ-F12 level of theory; see text for details. Energies (in kcal mol−1^{-1}) are given relative to the global CH3_3SH minimum and include harmonic zero-point energy corrections. ISC stands for intersystem crossing.

For the CH2_2 + H2_2S reaction, the singlet and triplet entrance channels are located at 100.6 and 91.7 kcal mol−1^{-1}, respectively. The triplet pathway proceeds through a transition state at 93.2 kcal mol−1^{-1}, corresponding to a barrier of only 1.5 kcal mol−1^{-1} with respect to 3^{3}CH2_2 + H2_2S, and leads to the CH3_3 + SH radical pair at 73.1 kcal mol−1^{-1}.

At low temperatures of the ice surface, the formation of CH3_3SH from CH2_2 + H2_2S is likely dominated by the triplet abstraction pathway. The 3^{3}CH2_2 + H2_2S channel shows a very small activation barrier, allowing efficient formation of CH3_3 and SH radicals under cold conditions. These can then recombine and be stabilized within the ice matrix toward CH3_3SH formation. In contrast, an intersystem crossing (ISC) near the entrance channel provides possible access to the singlet PES but does not lead directly to CH3_3SH. Instead, the singlet pathway proceeds through the CH2_2SH2_2 intermediate, which must overcome a transition state at 89.7 kcal mol−1^{-1}, corresponding to a barrier of 27.5 kcal mol−1^{-1} relative to the intermediate. The competition between ISC and the rapid barrier crossing along the triplet pathway, together with the additional barrier separating CH2_2SH2_2 from CH3_3SH, makes the singlet route unlikely to compete with triplet H-abstraction followed by radical recombination.

In the context of the present experiments, the above gas-phase quantum chemistry calculations support a mechanism in which CH2_2, formed through the reaction of C atoms with H2_2, reacts with H2_2S mainly via triplet H-abstraction, yielding CH3_3 and SH radicals that subsequently recombine and are stabilized as CH3_3SH within the ice matrix.

Proposed pathways to H2_2CS and CH3_3SH

Figure 6 summarizes the reaction network inferred from the present experiments and highlights that sulfur chemistry in H2_2S-rich ices at 10 K can proceed via two complementary but mechanistically distinct regimes. This depends on whether reactive H atoms or only H2_2 molecules are available along with accreting C atoms.

Figure 6. Proposed reaction network for the formation of H2_2CS and CH3_3SH during co-deposition experiments involving C atoms, H atoms, and H2_2.

In the H-driven regime, atomic C reacts directly with H2_2S to form H2_2CS, most plausibly through carbon insertion into the S–H bond, i.e., Eq. (4). Once formed, H2_2CS becomes an intermediate that can undergo successive H-atom addition reactions to yield CH2_2SH/CH3_3S radicals and ultimately CH3_3SH. In this branch, H2_2CS is therefore both a primary product of the C + H2_2S chemistry and the precursor to more saturated sulfur-bearing species.

In the H2_2-driven regime, CH3_3SH formation cannot be explained by direct hydrogenation of H2_2CS, because reactive H atoms are absent. Instead, the key additional step is the formation of CH2_2 intermediates through the reaction of C with H2_2 (Eq. (6)). The next step involves H-abstraction from H2_2S to form CH3_3 and SH radicals (Eq. (7)), followed by radical recombination to produce CH3_3SH (Eq. (8)). In this branch, CH3_3SH formation proceeds independently of H2_2CS and instead depends on the availability of CH2_2 produced by C + H2_2 chemistry. The coexistence of H2_2CS, CH3_3SH and CH4_4 therefore reflects competition between insertion, hydrogenation, abstraction and radical–radical recombination processes on the ice surface.

The gas-phase calculations provide a reference for assessing the relative energies and activation barriers of the proposed reaction pathways. Previous calculations for H + CS → HCS on two water clusters [69] and H + H2_2S → H2_2 + HS in the presence of H2_2O or H2_2S clusters [71] found relatively small changes in activation energies compared with the corresponding gas-phase reactions, supporting the use of this approach as a first approximation. In the present experiments, the ice environment may additionally influence reactant orientations and facilitate stabilization of the reaction products through energy dissipation. Although the magnitude of these effects has not been evaluated explicitly for CH2_2 + H2_2S, the calculations support the proposed reaction pathway leading to the observed CH3_3SH formation.

Astrophysical implications

This laboratory work provides new experimental constraints on grain-surface reaction routes leading to the sulfur-bearing organic molecules H2_2CS and CH3_3SH. Although these species are observed in the gas phase, their grain-surface formation routes have only been partially explored theoretically and require systematic experimental studies. Therefore, the present experiments target chemical processes that may operate in environments where atomic carbon, molecular hydrogen, and sulfur-bearing species such as H2_2S can coexist on grain surfaces during the earliest stages of star formation.

Neutral C atoms are abundantly present in translucent clouds, typically associated with visual extinctions of AV∼1A_V \sim1–5 mag, with observed column-averaged abundances reaching C/H2_2 ∼2.2×10−5\sim2.2 \times10^{-5} at intermediate extinctions [103, 99]. Additionally, observational and theoretical studies have suggested that atomic carbon may remain spatially extended, possibly even to denser regions [72, 62, 112, 86, 15, 9]. In addition, atomic carbon can become available again in cold prestellar cores, where residual gas-phase CO can be destroyed by He+^+, producing C+^+ that is further neutralized to enrich neutral atomic carbon [90, 8]. Therefore, C-atom accretion and subsequent solid-state chemistry are highly expected during the onset of ice growth and at later prestellar stages.

Ultraviolet absorption studies indicate that sulfur abundance is only weakly depleted along diffuse and translucent sight lines [98, 55], indicating that little to no sulfur has been incorporated into icy mantles at this early, low-density stage. Observations of singly and doubly deuterated H2_2S toward Class 0 sources indicate that H2_2S formation on grains predates the onset of catastrophic CO freeze-out [20], which occurs only later, at the higher densities reached in prestellar cores. Taken together, these constraints suggest that H2_2S formation does not begin until the cloud has evolved beyond the translucent stage into the denser molecular gas that precedes prestellar core formation, at which point atomic carbon is still available for accretion onto H2_2S-bearing ices before it is locked into CO.

In dense molecular clouds, H atoms accrete onto grain surfaces and diffuse to react with available species, including H2_2S. When C atoms instead participate in reactions with H2_2S, they form H2_2CS, which can subsequently be hydrogenated by H atoms to yield CH3_3SH. However, H2_2 is several orders of magnitude more abundant than atomic H in these environments [99], and despite its comparatively low sticking coefficient, H2_2 is expected to also participate in solid-state reactions on grain surfaces. In particular, C + H2_2 → CH2_2 has been proposed to occur on grain surfaces, and the resulting CH2_2 can react with H2_2S to form CH3_3SH via a two-step process involving H-abstraction followed by radical recombination. These two routes – H-atom-driven hydrogenation of H2_2CS and H2_2-mediated formation via CH2_2 – can therefore operate simultaneously, with the relative importance of each mechanism depending on the local physical conditions. The present experiments show that this distinction matters, since CH3_3SH formation can occur through both pathways.

The distinction between H-atom-driven and H2_2-mediated chemistry may also have implications for isotopic fractionation. In prestellar cores, gas-phase deuterium fractionation enhances the H2_2D+^+/H3+_3^+ ratio well above the elemental D/H ratio [109], and this fractionation is subsequently transferred to reactive atomic and molecular species participating in grain-surface chemistry. By contrast, the bulk H2_2/HD reservoir remains close to the elemental D/H ratio, so pathways that use H2_2 directly as a reactant are, in principle, less likely to introduce strong deuterium enrichment than pathways relying on sequential H- and D-atom addition. Testing this expectation, however, requires a direct comparison of the reaction efficiencies of H versus D and of H2_2 versus D2_2 in these systems, which we identify as a valuable direction for future experimental work.

Recent astronomical observations also argue against a simple direct analogy between the hydrogenation of H2_2CS to CH3_3SH and that of H2_2CO to CH3_3OH [14]. In the prestellar core L1544, Bunn et al. report a lower limit of H2_2CS/CH3_3SH > 2760, which is already about an order of magnitude higher than the corresponding oxygen-bearing ratio H2_2CO/CH3_3OH =

282.3 ± 177.2. This large difference may reflect a combination of factors, such as efficient H2_2CS formation through the reaction of C atoms with H2_2S, limited availability of H atoms to drive its further hydrogenation, or efficient destruction of CH3_3SH once formed. The present experiments identify both H-atom-driven and H2_2-mediated routes to CH3_3SH in the laboratory, and further observational constraints on the H2_2CS/CH3_3SH ratio across

a range of sources will help clarify their relative contributions under interstellar conditions.

Together, these results indicate that CH3_3SH formation in interstellar ices need not rely solely on sequential H-atom addition, but can also proceed through H2_2-mediated pathways initiated by accreting C atoms.

The efficiency of the reaction routes identified here under interstellar ice conditions will depend on several factors that are not captured directly by the present experiments, including competition with other surface reactions, the binding and diffusion energies of the reactants, and the influence of local binding environments and molecular orientation on reaction probabilities (e.g., Enrique-Romero et al. [34]; Tsuge et al. [102]; Ferrero et al. [40]; Martínez-Bachs & Rimola [76]). Astrochemical modeling incorporating these factors will be needed to simulate the full network of surface chemistry induced by accreting C atoms in realistic astronomical environments and to assess the relative importance of the pathways identified in this work.

Conclusions

We present a systematic laboratory study of the formation of sulfur-bearing organic molecules induced by reactions of C atoms with H2_2S ice, in the absence and presence of H2_2 or H atoms on grain surfaces at 10 K. By combining RAIRS and TPD measurements, we securely identify the formation of H2_2CS and CH3_3SH and propose plausible reaction pathways for their production under dense molecular cloud conditions. Our main conclusions are summarized as follows: (i) The interaction of atomic carbon with H2_2S at 10 K efficiently forms H2_2CS. The IR spectra reveal H2_2CS in the forms of monomer and the H2_2CS-derived oligomers, (H2_2CS)x_x. The reaction may proceed through initial insertion of atomic carbon into an S–H bond, followed by rearrangement to H2_2CS.

(ii) CH3_3SH forms via hydrogenation of H2_2CS as demonstrated in the H2_2S + C + H experiments. The IR spectra at 10 K show characteristic CH3_3SH bands in the investigated spectral range, while the TPD-QMS data confirm a CH3_3SH desorption peak at ∼ 110 K by examining the corresponding fragmentation pattern with literature data.

(iii) An additional formation route to CH3_3SH is studied in the H2_2S + C + H2_2 experiment via the CH2_2-induced H-abstraction reaction of H2_2S, followed by CH3_3 + SH radical recombination. Theoretical calculations suggest that the CH2_2 + H2_2S → CH3_3 + SH reaction is possible, with a relatively low barrier of 1.5 kcal.mol−1^{-1} while the radical–molecule association pathway (i.e., CH2_2 + H2_2S → CH3_3SH) is not favored.

The experimental results suggest efficient reactions of H2_2S with C atoms and CH2_2 radicals on grain surfaces at 10 K, forming the organosulfurs H2_2CS and CH3_3SH. This surface chemistry enriches the interstellar sulfur reaction network and hints at an underexplored route to sulfur-bearing molecular complexity in ices. Given the large abundance of H2_2 relative to H atoms, newly formed CH2_2 radicals produced via the C + H2_2 reaction may open additional reaction channels that enhance the formation of complex molecules in interstellar ices. Future astrochemical modeling that accounts for the reaction network identified here will be essential to determine when carbon-driven and H2_2-mediated chemistry can efficiently contribute to sulfur-bearing organic formation in interstellar ices.

Acknowledgements. F.K. acknowledges support from the European Union’s Horizon Europe research and innovation program under the Marie Skłodowska-Curie Actions Postdoctoral Fellowship grant agreement No. # 101153804 (OR-CHID – Organic Reactions of Carbon with H2_2 in Interstellar Dust). F.K. thanks Paola Caselli, Silvia Spezzano, and Olli Sipilä for insightful discussions on the astronomical implications of this work. This work has been funded by NOVA and NWO-DANIII. J.C.S. was supported by the Heising-Simons Foundation through a 51 Pegasi b Fellowship. S.I. acknowledges support from the Danish National Research Foundation through the Centre of Excellence ‘InterCat’ (grant agreement No. DNRF150) …

Monomeric H2_2CS desorption signals

Figure A.1 shows an enlarged view of the TPD-QMS data. A weak, coincident rise in the m/z=45m/z = 45 and 46 signals is observed between approximately 70 and 95 K, consistent with the desorption of a small fraction of monomeric H2_2CS embedded in the ice.

Figure A.1. Enlarged view of the TPD-QMS data corresponding to panels (a)–(c) of Figure 2, showing only the m/z=45m/z = 45 ([HCS]+^{+}) and 46 ([H2_2CS]+^{+}) signals. Arrows mark the weak, coincident features between approximately 70 and 95 K, attributed to desorption of monomeric H2_2CS.

References

  1. [1]Agúndez, M., Cernicharo, J., & Güelin, M. 2014, A&A, 570, A45
  2. [2]Agúndez, M., Molpeceres, G., Cabezas, C., et al. 2025, A&A, 693, L20
  3. [3]Albar, J. D., Summerfield, A., Cheng, T. S., et al. 2017, Sci. Rep., 7, 6598
  4. [4]Altwegg, K., Combi, M., Fuselier, S. A., et al. 2022, MNRAS, 516, 3900
  5. [5]Araki, M., Sanz-Novo, M., Endres, C. P., et al. 2026, Nat. Astron., 10, 401
  6. [6]Asplund, M., Grevesse, N., Sauval, A. J., & Scott, P. 2009, ARA&A, 47, 481
  7. [7]Bergin, E. A. & Tafalla, M. 2007, ARA&A, 45, 339
  8. [8]Bialy, S. & Sternberg, A. 2015, MNRAS, 450, 4424
  9. [9]Bisbas, T. G., Schruba, A., & van Dishoeck, E. F. 2019, MNRAS, 485, 3097
  10. [10]Boogert, A. C. A., Brewer, K., Brittain, A., & Emerson, K. S. 2022, ApJ, 941, 32
  11. [11]Boogert, A. C. A., Gerakines, P. A., & Whitett, D. C. B. 2015, ARA&A, 53, 541
  12. [12]Borshcheva, K., Fedoseev, G., Punanova, A. F., et al. 2025, ApJ, 990, 163
  13. [13]Bull, J. N. & Harland, P. W. 2008, Int. J. Mass Spectrom., 273, 53
  14. [14]Bunn, H. A., Spezzano, S., Coudert, L. H., et al. 2025, ApJL, 980, L13
  15. [15]Burton, M. G., Ashley, M. C. B., Braiding, C., et al. 2015, ApJ, 811, 13
  16. [16]Carrascosa, H., Muñoz Caro, G. M., Martín-Doménech, R., et al. 2024, MNRAS, 533, 967
  17. [17]Caselli, P., Pineda, J. E., Sipilä, O., et al. 2022, ApJ, 929, 13
  18. [18]Caselli, P., Walmsley, C. M., Tafalla, M., Dore, L., & Myers, P. C. 1999, ApJL, 523, L165
  19. [19]Cazaux, S., Carrascosa, H., Muñoz Caro, G. M., et al. 2022, A&A, 657, A100
  20. [20]Ceccarelli, C., Caselli, P., Bockelée-Morvan, D., et al. 2014, in Protostars and Planets VI, ed. H. Beuther, R. S. Klessen, C. P. Dullemond, & T. Henning, 859–882
  21. [21]Cernicharo, J., Cabezas, C., Agúndez, M., et al. 2021, A&A, 648, L3
  22. [22]Chuang, K.-J. 2018, PhD thesis, Leiden University, Leiden, The Netherlands
  23. [25]Chuang, K.-J., Fedoseev, G., Qasim, D., et al. 2020, A&A, 635, A199
  24. [26]Chuang, K.-J., Fedoseev, G., Qasim, D., et al. 2018, A&A, 617, A87
  25. [27]Codella, C., Podio, L., Garufi, A., et al. 2020, A&A, 644, A120
  26. [28]Cuppen, H. M., Linnartz, H., & Ioppolo, S. 2024, ARA&A, 62, 243
  27. [29]Deeyamulla, M. P. & Husain, D. 2006, J. Photochem. Photobiol. A: Chem., 184, 347
  28. [30]Di Genova, G., Pareras, G., Dominguez-Dalmases, C., et al. 2026, A&A, 711, A269
  29. [31]Drozdovskaya, M. N., van Dishoeck, E. F., Jørgensen, J. K., et al. 2018, MNRAS, 476, 4949
  30. [32]Dulieu, F., Amiaud, L., Congiu, E., et al. 2010, A&A, 512, A30
  31. [33]Dunning, T. H. 1989, J. Chem. Phys., 90, 1007
  32. [34]Enrique-Romero, J., Ceccarelli, C., Rimola, A., et al. 2021, A&A, 655, A9
  33. [35]Esteban, C., Peimbert, M., García-Rojas, J., et al. 2004, MNRAS, 355, 229
  34. [36]Fedoseev, G., Cuppen, H. M., Ioppolo, S., Lamberts, T., & Linnartz, H. 2015a, MNRAS, 448, 1288
  35. [37]Fedoseev, G., Ioppolo, S., Zhao, D., Lamberts, T., & Linnartz, H. 2015b, MNRAS, 446, 439
  36. [38]Fedoseev, G., Qasim, D., Chuang, K.-J., et al. 2022, ApJ, 924, 110
  37. [39]Feller, D. & Peterson, K. A. 2013, J. Chem. Phys., 139, 084110
  38. [40]Ferrero, S., Ceccarelli, C., Ugliengo, P., Sodupe, M., & Rimola, A. 2024, ApJ, 960, 22
  39. [41]Fuchs, G. W., Cuppen, H. M., Ioppolo, S., et al. 2009, A&A, 505, 629
  40. [42]Fuente, A., Riviere-Marichalar, P., Beitia-Antero, L., et al. 2023, A&A, 670, A114
  41. [43]Galland, N., Caralp, F., Rayez, M.-T., et al. 2001, J. Phys. Chem. A, 105, 9893
  42. [44]Garrod, R. T., Jin, M., Matis, K. A., et al. 2022, ApJS, 259, 1
  43. [45]Guo, T., Xu, J., Li, L., Cao, Z., & Zhang, Y. 2023, Russ. J. Phys. Chem. A, 97, 1953
  44. [46]Herath, A., McAnally, M., Turner, A. M., et al. 2025, Nat. Commun., 16, 5571
  45. [47]Herath, A., Turner, A. M., McAnally, M., Wang, J., & Kaiser, R. I. 2026, Phys. Chem. Chem. Phys., 28, 5094
  46. [48]Herbst, E. & van Dishoeck, E. F. 2009, ARA&A, 47, 427
  47. [49]Hudson, J. E., Hamilton, M. L., Vallance, C., & Harland, P. W. 2003, Phys. Chem. Chem. Phys., 5, 3162
  48. [50]Hudson, R. L. 2016, Phys. Chem. Chem. Phys., 18, 25756
  49. [51]Hudson, R. L. & Gerakines, P. A. 2018, ApJ, 867, 138
  50. [55]Jenkins, E. B. 2009, ApJ, 700, 1299
  51. [56]Jiménez-Escobar, A. & Muñoz Caro, G. M. 2011, A&A, 536, A91
  52. [57]Jin, M. & Garrod, R. T. 2020, ApJS, 249, 26
  53. [58]Jørgensen, J. K., van der Wiel, M. H. D., Coutens, A., et al. 2016, A&A, 595, A117
  54. [59]Kaiser, R. I., Ochsendel, C., Head-Gordon, M., & Lee, Y. T. 1999, J. Chem. Phys., 110, 2391
  55. [60]Kaiser, R. I., Sun, W., & Suits, A. G. 1997, J. Chem. Phys., 106, 5288
  56. [61]Kama, M., Shorttle, O., Jermyn, A. S., et al. 2019, ApJ, 885, 114
  57. [62]Keene, J., Blake, G. A., Phillips, T. G., Huggins, P. J., & Beichman, C. A. 1985, ApJ, 299, 967
  58. [63]Kendall, R. A., Dunning, T. H., & Harrison, R. J. 1992, J. Chem. Phys., 96, 6796
  59. [64]Knizia, G., Adler, T. B., & Werner, H.-J. 2009, J. Chem. Phys., 130, 054104
  60. [65]Krasnokutski, S. A. & Huisken, F. 2014, Appl. Phys. Lett., 105, 113506
  61. [66]Krasnokutski, S. A., Jäger, C., & Henning, T. 2020, ApJ, 889, 67
  62. [68]Laas, J. C. & Caselli, P. 2019, A&A, 624, A108
  63. [69]Lamberts, T. 2018, A&A, 615, L2
  64. [70]Lamberts, T., Fedoseev, G., van Hemert, M. C., et al. 2022, ApJ, 928, 48
  65. [71]Lamberts, T. & Kästner, J. 2017, J. Phys. Chem. A, 121, 9736
  66. [72]Langer, W. 1976, ApJ, 206, 699
  67. [73]Le Gal, R., Öberg, K. I., Loomis, R. A., Pegues, J., & Bergner, J. B. 2019, ApJ, 876, 72
  68. [74]Majumdar, L., Gratier, P., Vidal, T., et al. 2016, MNRAS, 458, 1859
  69. [75]Martín-Doménech, R., Delfranco, A., Öberg, K. I., & Rajappan, M. 2024, ApJ, 962, 107
  70. [76]Martínez-Bachs, B. & Rimola, A. 2026, ApJL, 1000, L58
  71. [77]Martín-Doménech, R., Manzano-Santamaría, J., Muñoz Caro, G. M., et al. 2015, A&A, 584, A14
  72. [78]McClure, M. K., Rocha, W. R. M., Pontoppidan, K. M., et al. 2023, Nat. Astron., 7, 431
  73. [79]Molpeceres, G., Kästner, J., Fedoseev, G., et al. 2021, J. Phys. Chem. Lett., 12, 10854, published online November 2 2021
  74. [80]Müller, H. S. P., Belloche, A., Xu, L.-H., et al. 2016, A&A, 587, A92
  75. [82]Narayanan, S., Piacentino, E. L., Öberg, K. I., & Rajappan, M. 2025, ApJ, 986, 10
  76. [83]Nguyen, T., Molpeceres, G., Oba, Y., et al. 2026, MNRAS, 547, stag396
  77. [84]Oba, Y., Tomaru, T., Lamberts, T., Kouchi, A., & Watanabe, N. 2018, Nat. Astron., 2, 228
  78. [85]Palumbo, M. E., Tielens, A. G. G. M., & Tokunaga, A. T. 1995, ApJ, 449, 674
  79. [86]Papadopoulos, P. P., Thi, W.-F., & Viti, S. 2004, MNRAS, 351, 147
  80. [87]Qasim, D., Fedoseev, G., Chuang, K.-J., et al. 2020a, Nat. Astron., 4, 781
  81. [88]Qasim, D., Witlox, M. J. A., Fedoseev, G., et al. 2020b, Rev. Sci. Instrum., 91, 054501
  82. [89]Rodríguez-Almeida, L. F., Rivilla, V. M., Jiménez-Serra, I., et al. 2021, ApJL, 912, L11
  83. [90]Ruffle, D. P., Hartquist, T. W., Caselli, P., & Williams, D. A. 1999, MNRAS, 306, 691
  84. [91]Santos, J. C., Linnartz, H., & Chuang, K.-J. 2023, A&A, 678, A112
  85. [92]Sanz-Novo, M., Rivilla, V. M., Endres, C. P., et al. 2025, ApJL, 980, L37
  86. [93]Shao, X., Wu, Z., Wang, L., Rauhut, G., & Zeng, X. 2023, J. Phys. Chem. Lett., 14, 10450
  87. [94]Shingledecker, C. N., Lamberts, T., Laas, J. C., et al. 2020, ApJ, 888, 52
  88. [95]Simončič, M., Semenov, D., Krasnokutski, S., Henning, T., & Jäger, C. 2020, A&A, 637, A72
  89. [97]Smith, R. G. 1991, MNRAS, 249, 172
  90. [98]Snow, T. P., Black, J. H., van Dishoeck, E. F., et al. 1996, ApJ, 465, 245
  91. [99]Snow, T. P. & McCall, B. J. 2006, ARA&A, 44, 367
  92. [100]Suzuki, E., Yamazaki, M., & Shimizu, K. 2007, Vib. Spectrosc., 43, 269
  93. [101]Szalay, P. G., Müller, T., Gidofalvi, G., Lischka, H., & Shepard, R. 2012, Chem. Rev., 112, 108
  94. [102]Tsuge, M., Molpeceres, G., Aikawa, Y., & Watanabe, N. 2023, Nat. Astron., 7, 1351
  95. [103]van Dishoeck, E. F. & Black, J. H. 1989, ApJ, 340, 273
  96. [104]Varandas, A. J. C. & Rocha, C. M. R. 2018, Phil. Trans. R. Soc. A, 376, 20170145
  97. [105]Vastel, C., Quénard, D., Le Gal, R., et al. 2018, MNRAS, 478, 5514
  98. [106]Vidal, T. H. G., Loison, J.-C., Jaziri, A. Y., et al. 2017, MNRAS, 469, 435
  99. [107]Vitorino, J., Loison, J.-C., Wakelam, V., Congiu, E., & Dulieu, F. 2024, MNRAS, 533, 52
  100. [108]Watanabe, N. & Kouchi, A. 2002, ApJL, 571, L173
  101. [109]Watson, W. D. 1974, ApJ, 188, 35
  102. [110]Werner, H.-J., Knowles, P. J., Knizia, G., et al. 2020, MOLPRO, version 2020.1, A Package of Ab Initio Programs, see: http://www.molpro.net
  103. [111]Yang, Y.-L., Green, J. D., Pontoppidan, K. M., et al. 2022, ApJL, 941, L13
  104. [112]Zmuidzinas, J., Betz, A. L., Boreiko, R. T., & Goldhaber, D. M. 1988, ApJ, 335, 774
  105. [113]Öberg, K. I. & Bergin, E. A. 2021, Phys. Rep., 893, 1

Paper details

vlnc:W_rpuxu3rb
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents