ALMA ACA study of the H 2 S/OCS ratio in low-mass protostars

Context. The identification of the main sulfur reservoir on its way from the diffuse interstellar medium to the cold dense star-forming cores and, ultimately, to protostars is a long-standing problem. Despite sulfur’s astrochemical relevance, the abundance of S-bearing molecules in dense cores and regions around protostars is still insufficiently constrained. Aims. The goal of this investigation is to derive the gas-phase H 2 S/OCS ratio for several low-mass protostars, which could provide crucial information about the physical and chemical conditions in the birth cloud of Sun-like stars. This may also shed new light onto the main sulfur reservoir in low-mass star-forming systems. Methods. Using Atacama Large Millimeter/submillimeter Array (ALMA) Atacama Compact Array (ACA) Band 6 observations, we searched for H 2 S, OCS, and their isotopologs in ten Class 0/I protostars with different source properties such as age, mass, and environmental conditions. The sample contains IRAS 16293-2422 A, IRAS 16293-2422 B, NGC 1333-IRAS 4A, RCrA IRS7B, Per-B1-c, BHR71-IRS1, Per-emb-25, NGC 1333-IRAS4B, Ser-SMM3, and TMC1. A local thermal equilibrium (LTE) model is used to fit synthetic spectra to the detected lines and to derive the column densities based solely on optically thin lines. Results. The H 2 S and OCS column densities span four orders of magnitude across the sample. The H 2 S/OCS ratio is found to be in the range from 0.2 to above 9.7. IRAS 16293-2422 A and Ser-SMM3 have the lowest ratio, while BHR71-IRS1 has the highest. Only the H 2 S/OCS ratio of BHR71-IRS1 is in agreement with the ratio in comet 67P/Churyumov–Gerasimenko within the uncertainties. Conclusions. The determined gas-phase H 2 S/OCS ratios can be below the upper limits on the solid-state ratios by as much as one order of magnitude. The H 2 S/OCS ratio depends in great measure on the environment of the birth cloud, such as UV-irradiation and heating received prior to the formation of a protostar. The highly isolated birth environment (a Bok globule) of BHR71-IRS1 is hypothesized as the reason for its high gaseous H 2 S/OCS ratio that is due to lower rates of photoreactions and more efficient hydrogenation reactions under such dark, cold conditions. The gaseous inventory of S-bearing molecules in BHR71-IRS1 appears to be the most similar to that of interstellar ices.


Introduction
Sulfur (S) is the tenth most abundant element in the Universe (S/H∼1.35×10−5 , Yamamoto 2017).It was first detected as carbon monosulfide (CS) in the interstellar medium (Penzias et al. 1971).S-bearing species have since been detected in different regions including molecular clouds (Navarro-Almaida et al. 2020;Spezzano et al. 2022), hot cores (Blake et al. 1987;Charnley 1997;Li et al. 2015;Codella et al. 2021;Drozdovskaya et al. 2018), comets (Smith et al. 1980;Bockelée-Morvan et al. 2000;Biver et al. 2021a,b), as well as starburst galaxies (NGC 253; Martín et al. 2005).The total abundance of an element in dust, ice, and gas is its cosmic abundance, also called its elemental abundance.The gas-phase abundance of atomic sulfur in diffuse clouds is comparable to the cosmic abundance of sulfur (∼10 −5 ; Savage & Sembach 1996;Howk et al. 2006).However, the observed abundance of S-bearing species in dense cores and protostellar environments is lower by a factor of ∼1000 (Snow et al. 1986;Tieftrunk et al. 1994;Goicoechea et al. 2006;Agúndez et al. 2018) in comparison to the total S-abundance in diffuse clouds.The forms and mechanisms behind this sulfur depletion on the environment being observed.Other possible reservoirs of sulfur have been proposed in the form of semi-refractory polymers up to S 8 (A'Hearn et al. 1983;Druard & Wakelam 2012;Calmonte et al. 2016;Shingledecker et al. 2020), hydrated sulfuric acid (Scappini et al. 2003), atomic sulfur (Anderson et al. 2013), and mineral sulfides, FeS (Keller et al. 2002;Köhler et al. 2014;Kama et al. 2019).On the other hand, chemical models of the evolution from cloud to dense core with updated chemical networks suggest that sulfur is merely partitioned over a diverse set of simple organo-sulfur ices and no additional form is required (Laas & Caselli 2019).Matching observed and modeled cloud abundances consistently for the full inventory of gaseous S-bearing molecules to better than a factor of 10 remains challenging (Navarro-Almaida et al. 2020).Laboratory experiments point to the importance of the photodissociation of H 2 S ice by UV photons leading to the production of OCS ice (Ferrante et al. 2008;Garozzo et al. 2010;Jiménez-Escobar & Muñoz Caro 2011;Chen et al. 2015) and S 2 in mixed ices (Grim & Greenberg 1987).Calmonte et al. (2016) claim to have recovered the full sulfur inventory in comets.
Sulfur-bearing species have been proposed to probe the physical and chemical properties of star-forming regions and to even act as chemical clocks (Charnley 1997;Hatchell et al. 1998;Viti et al. 2001;Li et al. 2015).However, it has since been shown that their abundance is sensitive to gas-phase chemistry and the availability of atomic oxygen, which puts their reliability as chemical clocks into question (Wakelam et al. 2004(Wakelam et al. , 2011)).Studying S-bearing molecules in young Class 0/I protostars is crucial for two reasons.Firstly, their inner hot regions thermally desorb all the volatile ices that are otherwise hidden from gasphase observations.Consequently, it is more likely to be able to probe the full volatile inventory of S-bearing molecules and investigate the "missing sulfur" reservoir.Secondly, these targets are a window onto the materials available for the assembly of the protoplanetary disk midplane and the cometesimals therein (Aikawa & Herbst 1999;Willacy 2007;Willacy & Woods 2009).This makes hot inner regions highly suitable targets for comparative studies with comets (Bockelée-Morvan et al. 2000;Drozdovskaya et al. 2019).
The main goal of this paper is to study the physical and chemical conditions in embedded protostars via the H 2 S/OCS ratio.A sample of 10 Class 0/I low-mass protostars with different physical properties (mass, age, environment) is considered.Such protostars are in their earliest phase of formation after collapse with large envelope masses.In this work, Atacama Large Millimeter/submillimeter Array (ALMA) Atacama Compact Array (ACA) Band 6 observations towards these 10 protostars are utilized.The H 2 S/OCS ratio is calculated from the column densities of H 2 S, OCS, and their isotopologs.The details of the observations, model, and model parameters used for synthetic spectral fitting are introduced in Section 2. The detected lines of major and minor isotopologs of H 2 S and OCS, their characteristics, and H 2 S/OCS line ratios are presented in Section 3. The discussion and conclusions are presented in Section 4 and 5, respectively.
The observed frequency ranges of the data are given in Table 1.Data cubes were processed through the standard pipeline calibration with CASA 5.4.0-68.For each source, the noise level has been calculated by taking the standard deviation of the flux in the frequency ranges where no emission lines were detected, i.e., regions with pure noise, in the spectral window containing the H 2 S, 2 2,0 -2 1,1 line.The noise level of the first data set is 21-32 mJy beam −1 channel −1 , and of the second data set is 7-13 mJy beam −1 channel −1 (Table 3).Both data sets have a flux uncertainty of 10%.The largest resolvable scale of the first and the second data sets are 26.2-29.2′′ and 24.6-29.0′′ , respectively.

Sources
The properties of the sources explored in this work are tabulated in  (Dzib et al. 2018).This source was studied thoroughly using ALMA under the Protostellar Interferometric Line Survey (PILS; Jørgensen et al. 2016) and many preceding observational campaigns (e.g., van Dishoeck et al. 1995;Caux et al. 2011).Both hot corinos around A and B are rich in a diverse set of complex organic molecules (Jørgensen et al. 2018;Manigand et al. 2020).The source IRAS 16293 A is itself a binary composed of sources A1 and A2 with a separation of 0.38 ′′ (54 au; Maureira et al. 2020).
IRAS4A is also a binary system, comprised of IRAS4A1 and IRAS4A2, separated by 1.8 ′′ (540 au; Sahu et al. 2019) in the Perseus molecular cloud, located at a distance of 299 pc (Zucker et al. 2018) in the south-eastern edge of the complex NGC 1333 (Looney et al. 2000).IRAS4A1 has a much higher dust density in its envelope than IRAS4A2, but both contain complex organic molecules (Sahu et al. 2019;De Simone et al. 2020).IRS7B is a low-mass source, with a separation of 14 ′′ (2 000 au) from IRS7A (Brown 1987), and ∼8 ′′ (1 000 au) from CXO 34 (Lindberg et al. 2014).It is situated in the Corona Australis dark cloud at a distance of 130 pc (Neuhäuser & Forbrich 2008).IRS7B has been shown to contain lower complex organic abundances as a result of being located in a strongly irradiated environment (Lindberg et al. 2015).
From the second set of sources, IRAS4B (sometimes labeled BI) has a binary component B ′ (or BII) that is 11 ′′ (3 300 au) away (Sakai et al. 2012;Anderl et al. 2016;Tobin et al. 2016).The separation between IRAS4B and IRAS4A is 31 ′′ (9 300 au; Coutens et al. 2013).IRAS4B displays emission from complex organic molecules (Belloche et al. 2020) and powers a highvelocity SiO jet (Podio et al. 2021).B1-c is an isolated deeply embedded protostar in the Barnard 1 clump in the western part of the Perseus molecular cloud at a distance of 301 pc (Zucker et al. 2018).B1-c contains emission from complex organic molecules and shows a high velocity outflow (Jørgensen et al. 2006;van Gelder et al. 2020).The next closest source, B1-a, is ∼100 ′′ (∼ 29 500 au) away (Jørgensen et al. 2006).BHR71 is a Bok globule in the Southern Coalsack dark nebulae at a distance of ∼200 pc (Seidensticker & Schmidt-Kaler 1989;Straizys et al. 1994).It hosts the wide binary system of IRS1 and IRS2 with a sepration of 16 ′′ (3 200 au; Bourke 2001; Parise et al. 2006;  Chen et al. 2008; Tobin et al. 2019).IRS1 displays pronounced emission from complex organic molecules (Yang et al. 2020).Emb-25 is a single source located in the Perseus molecular cloud (Enoch et al. 2009;Tobin et al. 2016).It does not show emission from complex organic molecules (Yang et al. 2021), but powers low-velocity CO outflows (Stephens et al. 2019).TMC1 is a Class I binary source, located in the Taurus molecular cloud (Chen et al. 1995;Brown & Chandler 1999) at a distance of 140 pc (Elias 1978;Torres et al. 2009).The separation between the two components, TMC1E and TMC1W, is ∼0.6 ′′ (∼85 au); and neither of the two display complex organic emission (van't Hoff et al. 2020).SMM3 is a single, embedded protostar located in the SE region in Serpens region; 436 pc away (Ortiz-León et al. 2018).The next closest-lying source is SMM6 at a separation Figure 1: ALMA pipeline-produced integrated intensity maps (color scale) with the line channels excluded, which are dominated by dust emission, for the studied sample of sources.On-source spectra are extracted by averaging the flux from the pixels within the circular area centered on 'X'.
Table 3: Location of the center position (in RA and Dec) and radius (in au and arcsecond) of each circular region, from which the spectra are extracted for the studied protostars.The pixel size (in arcsecond) for each source is also given.The noise levels in mJy beam −1 channel −1 and mJy beam −1 km s −1 of the protostars are deduced according to (flux in line-free channel i) 2 number of line-free channels and noise (mJy −1 beam −1 channel −1 ) × √ n × spectral resolution (km s −1 ), where n is the number of channels computed with of 20 ′′ (∼ 8 700 au; Davis et al. 1999;Kristensen et al. 2010;Mirocha et al. 2021).SMM3 launches a powerful jet (Tychoniec et al. 2021), but does not display complex organic molecule emission, which may be obscured by the enveloping dust (van Gelder et al. 2020).

Synthetic spectral fitting
For the spectral analysis, on-source spectra were extracted from the data cubes of the sources of the two data sets.Circular regions centered on source positions (in RA and Dec) with the radius of spectrum extraction corresponding to one-beam onsource are given in Table 3.The number of pixels in radius r of the circular region to be used was computed by dividing the radius of the circular region with the size of one pixel in arcsecond.The spectroscopy used for the targeted molecules and their isotopologs stems from the Cologne Database of Molecular Spectroscopy (CDMS; Müller et al. 2001Müller et al. , 2005;;Endres et al. 2016) 1 and the Jet Propulsion Laboratory (JPL) catalog (Pickett et al. 1998) 2 .Line blending in the detected lines was checked with the online database Splatalogue3 .Synthetic spectral fitting was performed with custom-made Python scripts based on the assumption of local thermal equilibrium (LTE).The input parameters include the full width halfmaximum (FWHM) of the line, column density (N), excitation temperature (T ex ), source size, beam size, and spectral resolution of the observations.Line profiles are assumed to be Gaussian.Further details are provided in section 2.3 of Drozdovskaya et al. (2022).For some sources, the number of free parameters (such as source size or T ex ) can be reduced based on information from other observing programs.These are detailed on a source-bysource basis in the corresponding Appendices.Typically, merely two free parameters were fitted at a time by means of a visual inspection and an exploration of a grid of possible values.The considered range for N was 10 13 -10 19 cm −2 in steps of 0.1 × N. Simultaneously, the FWHM of the synthetic fit was adjusted to match the FWHM of the detected line.For some sources (such as NGC 1333-IRAS4A and Ser-SMM3), excitation temperature could not be constrained.Hence, a grid of excitation temperatures was considered with a range between 50 and 300 K.
The line optical depth (τ) was calculated for the best-fitting combination of parameters to check for optical thickness.If a transition of a certain molecule was found to be optically thick, its column density was computed as the average of the column density of the main species derived from its minor isotopologs given by: where X is H 2 S or OCS and N(X) i is the column density of H 2 S or OCS derived from its minor isotopologs.Adopted isotopic ratios for the derivation of main isotopologs from minor isotopologs are given alongside Table 4.

Results
The spectral set up of the first data set allows the targeted sources to be probed for the emission of the main isotopologs of H 2 S and OCS, v=0, their minor isotopologs (HDS, HD 34 S H 33 2 S, H 34 2 S, 18 OCS, O 13 CS, OC 33 S, 18 OC 34 S), and also the vibrationally excited states of OCS (v 2 =1 ± ).Consequently, sources IRAS 16293 A, IRAS 16293 B, IRAS4A, and IRS7B were probed for all these species.The spectral set up of the second data set allows the other targeted sources (B1-c, BHR71-IRS1, Per-emb-25, IRAS4B, SMM3, and TMC1) to be probed for the main isotopologs of H 2 S and OCS, v=0, their minor isotopologs (HDS, HD 34 S, 18 OCS, OC 33 S, 18 OC 34 S, 18 O 13 CS), and vibrationally excited state of OCS (v 2 =1 − ).All the transitions of the detected molecules have E up in the range of 84 − 123 K and A i j values of 0.69 − 4.9 × 10 −5 s −1 .The details of the targeted molecular lines are presented in Appendix A. Note that the HDS lines probed in the two data sets are not the same -the first data set was probed for HDS, 14 2,12 -13 4,9 transition at rest frequency 214.325GHz, while the second data set was probed for HDS, 7 3,4 -7 3,5 and 12 5,7 -12 5,8 transitions at 234.046 and 234.528GHz, respectively.Nevertheless, the E up is high (> 400 K) for all three transitions of HDS; and it is not detected in any of these lines in any of the sources.The HD 34 S and OCS, v 2 =1 ± transitions also have high E up (> 400 K).HD 34 S was not detected in any of these lines in any of the sources, but OCS, v 2 =1 ± was detected in IRAS 16293 A, IRAS 16293 B, and IRAS4A (owing to high OCS column densities and higher sensitivity of the first data set).
All the main and minor isotopologs were detected in IRAS 16293 A, IRAS 16293 B, and IRAS4A except HDS, HD 34 S, 18 OC 34 S, and 18 O 13 CS.In IRS7B, only H 2 S was detected, the rest of the molecular lines were undetected including OCS, v=0.Whereas, only main S-bearing species, H 2 S and OCS, v=0, were detected in B1-c, BHR71-IRS1, and SMM3.IRAS4B showed the rotational transition of OC 33 S (J = 18 − 17) in addition to the H 2 S and OCS, v=0 lines.Emb-25 and TMC1 showed no detections of main S-bearing species and their minor isotopologs.Thus, 1-σ upper limits on the column densities of H 2 S and OCS, v=0 were derived for Emb-25 and TMC1; and an upper limit on the column density of OCS, v=0 was derived for IRS7B.Table A.1 provides the CDMS entry, transition quantum numbers, rest frequency, upper energy level, Einstein A coefficient, and the detection/non-detection of each line of all the targeted S-bearing molecules towards all of the sources in the sample.
In Figure 1, the pipeline-produced integrated intensity maps with the line channels excluded are shown for all the sources.These are dominated by the dust emission, but with some degree of contamination by line emission especially for some of the line-rich sources.The circular regions used to extract the spectra of each individual source are also shown.Pixel size of the integrated maps of the sources varies from 0.8 to 1.2 ′′ .To match the beam size of the observations, the radius of the circular regions was also varied from 3.0 to 4.3 ′′ .The spatial resolution of the presented ACA observations allowed the binary IRAS 16293 A and B (separated by 5.3 ′′ ) to be resolved as single sources; however, the resolution was not high enough to disentangle the binary components A1 and A2 of IRAS 16293 A. Similarly, the binary components of IRAS4A (with a separation of 1.8 ′′ ), and of TMC1 (with a separation of 0.6 ′′ ) could not be disentangled due to the spatial resolution not being high enough.All other sources are either single sources or binaries separated by large distances; hence, they are spatially resolved as individual sources.
The lower and upper uncertainties on the fitted column densities are derived assuming an error of ±20 K on the assumed excitation temperature and a 1-σ noise level.The analysis of the spectra extracted towards IRAS 16293-2422 B is presented in the following Section 3.   3) obtained with ALMA ACA at Band 6 frequencies (Table 1).A Doppler shift by v LSR = 2.7 km s −1 has been applied (Table 2).
windows towards IRAS 16293 B are shown in Figure 2.For the other sources, the analysis is presented in Appendices C through K.

IRAS 16293-2422 B
Towards IRAS 16293 B, the main S-bearing species (H 2 S and OCS, v=0) and all the targeted minor isotopologs are securely detected, except for HDS due to a very high E up value (1 277 K) and the double isotopologs of HD 34 S, 18 OC 34 S, 18 O 13 CS due to their low abundances (and high E up for the case of HD 34 S).
The detected transitions of H 2 S, 2 2,0 -2 1,1 and OCS, v=0, J = 19 − 18 are bright and optically thick (τ >>1).The H 34 2 S line is marginally optically thick (τ = 0.2), shown in Figure 3.The vibrationally excited OCS, v 2 =1 ± lines are detected.The lines of the detected molecules do not suffer from blending, except the H 33 2 S, 2 2,0,3 -2 1,1,3 transition at 215.512 GHz, which is contaminated by the CH 3 CHO, 11 2,9 -10 2,8 transition.HD 34 S, 7 3,4 -7 3,5 transition at 232.964 GHz is heavily blended with the CH 3 CN, v 8 = 1, J = 15 − 15, K = 7 − 5 transition.Most likely all the emission seen around the rest frequency of HD 34 S comes from CH 3 CN, because HD 34 S is a minor species ( 32 S/ 34 S=22, Wilson 1999, and D/H ∼0.04 incl.the statistical correction by a factor of 2 to account for the two indistinguishable D atom positions, Drozdovskaya et al. 2018) and the E up of this transition is high (416 K).The spectra of detected and undetected lines are in For the analysis of the targeted S-bearing molecules towards IRAS 16293 B, a T ex of 125 K is assumed.This value has been deduced to be the best-fitting on the basis of ALMA-PILS observations at higher spatial resolution obtained with the 12m array and a full inventory of S-bearing molecules (Drozdovskaya et al. 2018).A FWHM of 1 km s −1 is adopted, as it has been shown that this value consistently fits nearly all the molecules investigated towards the hot inner regions of IRAS 16293 B (e.g., Jørgensen et al. 2018).For the larger scales probed by the present ALMA ACA observations, a deviation by 2 km s −1 from this FWHM can be seen for optically thick lines.This broadening in FWHM is likely due to the opacity broadening effects, which are dominant in optically thick lines, but can be neglected in optically thin lines (Hacar et al. 2016).The synthetic spectral fitting has been carried out for two potential source sizes, 1 ′′ and 2 ′′ (Table 4).Column densities depend on the assumed source size and are lower for the larger source size.However, the N(H 2 S)/N(OCS) ratio is 1.3±0.27and 1.3±0.28 for source sizes of 1 ′′ and 2 ′′ , respectively.Thus, the ratio is independent of the assumed source size and is robustly determined with the ALMA ACA data.
For a source size of 2 ′′ , the column density of the vibrationally excited state of OCS, v 2 =1 ± derived for IRAS 16293 B (2.5×10 16 cm −2 ) is an order of magnitude lower than the OCS, v 2 = 1 column density (2.0×10 17 cm −2 ) derived in Drozdovskaya et al. (2018).For a source size of 1 ′′ , the here obtained value (8.5 × 10 16 cm −2 ) is in closer agreement with Droz-Table 4: Results from the modeling of synthetic spectra of the detected S-bearing species towards IRAS 16293 B for source sizes of 1 ′′ and 2 ′′ , an excitation temperature of 125 K, and a line width of 1 km s −1 .

Line Profiles
For the synthetic spectral modeling, Gaussian line profiles are assumed (Section 2.3).However, even for optically thin lines, a deviation from Gaussian line profiles is seen in some cases.
Two prominent examples are H 34 2 S and O 13 CS in IRAS 16293 A (Figure C.1), where the high spectral resolution of the data set clearly allows multiple peaks to be spectrally resolved in these lines.Likely, the reason for this is that this source is a compact binary (Maureira et al. 2020) with multiple components within the ACA beam of these observations.Another prominent example is the OCS, v = 0 line in Ser-SMM3 (Figure J.1), which has a double-peaked profile centered around the source velocity.Such a line profile is typical for a rotating structure around its protostar (which could be envelope or disk in nature).Detailed modeling of line profiles is out of scope of this paper, as additional observations would be necessary in order to achieve meaningful results.For the purpose of studying the H 2 S/OCS ratio, these effects are secondary and likely do not significantly affect the calculated ratio and the conclusions of this paper.For IRAS 16293 A, the column density of H 34 2 S is not used to get the column density of H 2 S, because it is computed to be partially optically thick.Meanwhile, the column density of OCS as obtained from O 13 CS is within a factor of 2 of what is obtained from OC 33 S and 18 OCS.For Ser-SMM3, the lack of constraints on the excitation temperature dominates the uncertainty in the H 2 S/OCS ratio.

H 2 S/OCS ratio determination
The column densities of H 2 S and OCS derived on the basis of ALMA ACA observations have been used to constrain the ra- tio of H 2 S to OCS (Table 5).It was possible to compute this ratio for five out of ten sources in the considered sample.Neither H 2 S nor OCS were detected in Emb-25 and TMC1, consequently the H 2 S/OCS ratio could not be constrained.The nondetection of OCS in IRS7B allowed to derive only a lower limit on the H 2 S/OCS ratio.Table 5 also contains the best-available estimates of the H 2 S/OCS ratio for the warm and cold components of B1-c, and for the cold component of BHR71-IRS1, although these numbers carry a higher level of uncertainty due to line opacity that could not be resolved on the basis of these observations.For the warm component of BHR71-IRS1, a lower limit on the H 2 S/OCS ratio could be computed.For further analysis, the sample has been divided into three sub-samples: compact binary, wide binary, and single, based on the separations between components of multiple sources or closest neighbours.

Discussion
Figure 4 displays the derived protostellar H 2 S/OCS ratios, as well as the cometary (67P/Churyumov-Gerasimenko, hereafter 67P/C-G) and interstellar ice (W33A and Mon R2 IRS2) H 2 S/OCS ratios.The derived protostellar H 2 S/OCS ratios span a range from 0.2 to above 9.7.The ratios show a variation of approximately one order of magnitude, being the lowest in IRAS 16293 A and SMM3, and the highest in BHR71-IRS1.
In the subsections 4.1 and 4.2, a comparison of the protostellar H 2 S/OCS ratios with this ratio in interstellar and cometary ices, respectively, is made.Comets are thought to preserve the chemical composition of the Sun's birth cloud (Mumma & Charnley 2011).By comparing the H 2 S/OCS ratio of comet 67P/C-G with the ratios in nascent solar-like protostellar systems, an assessment can be made whether such an inheritance is true in the case of S-bearing molecules.

Interstellar ices
Observations towards the cold, outer protostellar envelopes of high-mass protostars W33A and Mon R2 IRS2 are used to acquire the H 2 S/OCS ratio in interstellar ices.The ratio is computed based on the ice abundances of OCS detected as an absorption feature at 4.9 µm (Palumbo et al. 1995) (Smith 1991), respectively, yielding N solid (H 2 S)/N solid (H 2 O) <4.7×10 −3 .The H 2 S/OCS ratio in interstellar ices is poorly constrained due to the non-detection of solid H 2 S to date.The upper limits on the interstellar ices ratio are within the uncertainties of the cometary ices ratio.The derived protostellar ratios for all the sources are lower than the upper limit on the H 2 S/OCS ratio determined for interstellar ices except BHR71-IRS1 with H 2 S/OCS ratio exceeding the upper limit on the ratio for Mon R2 IRS2.

Comet 67P/Churyumov-Gerasimenko
Comets are thought to be the most unprocessed objects in the Solar System (Mumma & Charnley 2011).Cometary chemical composition has been shown to be similar to a degree to that of star-forming regions (Bockelée-Morvan et al. 2000;Drozdovskaya et al. 2019).Consequently, the cometary H 2 S/OCS ratio is thought to provide an independent measurement of this ratio in interstellar ices.The H 2 S and OCS abundances from the ESA Rosetta mission were used to compute the H 2 S/OCS ratio for the Jupiter-family comet 67P/C-G.The H 2 S and OCS abundances relative to H 2 O are 1.10±0.46%and 0.041 +0.082  −0.020 %, respectively (Rubin et al. 2019).These molecules are typical constituents of comets (Lis et al. 1997;Bockelée-Morvan et al. 2000;Boissier et al. 2007;Mumma & Charnley 2011).BHR71-IRS1 is the only protostar in the sample with a H 2 S/OCS ratio within the uncertainties of the cometary ices ratio.The H 2 S/OCS ratio for the other sources is at least an order of magnitude lower than for 67P/C-G, while even considering the high uncertainties on the cometary value.The availability of H 2 S relative to H 2 O in cometary ice (0.0064 − 0.0156) appears to be higher than in the interstellar ices towards Mon R2 IRS2 (< 0.0047).For W33A, the currently available upper limit (< 0.03) is less constraining and hence no conclusion can be drawn about how its ices compare to those of comet 67P/C-G.The relative ratio of H 2 S to OCS is only one window onto the inventory of S-bearing molecules in gas and ice at different stages of star and planet formation, meanwhile the overall availability relative to, for example, H 2 O is another window that requires dedicated exploration.

H 2 S/OCS ratio as an environmental (clustered/isolated) tracer
The measured gas-phase H 2 S/OCS ratios in the sample of young, low-mass protostars explored in this paper are predominantly lower (by as much as an order of magnitude) than the solidstate ratio measured through direct infrared observations of interstellar ices and indirectly via comets (Figure 4).There appears to be no correlation with binarity nor specific host cloud (Table 5).The dependence with evolutionary stage could not be properly explored, as the sample contains only one Class I source (TMC1), which did not result in detections of neither H 2 S nor OCS.
Close binary ( < 500 au) Wide binary (500-5000 au) Single source (within 5000 au) The highest ratio of ≥9.7 is found for the warm component (250 K) of BHR71-IRS1, which is a wide binary (∼ 3 200 au; Bourke 2001;Parise et al. 2006;Chen et al. 2008;Tobin et al. 2019) Class 0 protostar.The ratio in BHR71-IRS1 resides within the uncertainty of the cometary ratio, but is in-between the two upper limits derived for the interstellar ices.The overall envelope mass and bolometric luminosity of BHR71-IRS1 is comparable to those of other compact binary and wide binary systems.The similarity of its gas-phase H 2 S/OCS ratio to the ratio in ices may suggest that it is displaying the most recently thermally desorbed volatiles that have not been subjected to gas-phase processing for long.However, what makes BHR71-IRS1 stand out is that it is located in an isolated cloud, i.e., it is not associated with processes typical for clustered environments such as dynamical interactions, mechanical and chemical feedback from outflows, and enhanced irradiation.
Possibly, isolation resulted in lower irradiation of the ice grains during the prestellar phase in BHR71-IRS1, thus converting less H 2 S ices to OCS ices by photodissociation in the presence of CO ice.Consequently, leaving a higher H 2 S/OCS ratio in ices, which after evaporation resulted in a higher H 2 S/OCS ratio in the gas phase.Another reason could be more efficient hydrogenation chemistry in such a colder environment.On dust grains, hydrogenation is expected to be the most effective process leading to the formation of H 2 S (Wakelam et al. 2011;Esplugues et al. 2014).Hence, BHR71-IRS1 may have a higher H 2 S content and a lower OCS content, which results in a higher H 2 S/OCS ratio.Water deuteration is also higher by a factor of 2 − 4 in isolated protostars such as BHR71-IRS1 in comparison to those in clustered environments such as IRAS 16293 and IRAS4 (Jensen et al. 2019).
One alternative cause of lower H 2 S/OCS ratios towards clustered low-mass protostars could be local temperature differences in their birth clouds, e.g., due to enhanced irradiation from the neighbouring protostars.Laboratory experiments have proven that OCS forms readily in ices when interstellar ice-analogs are irradiated by high-energy photons (Ferrante et al. 2008;Garozzo et al. 2010;Jiménez-Escobar & Muñoz Caro 2011;Chen et al. 2015).This would lead to a lower H 2 S/OCS ratio.
Additionally, cosmic rays and other forms of radiation (UV and X-ray photons) are a ubiquitous source of ionization of the interstellar gas.It is a pivotal factor in the dynamical and chemical evolution of molecular clouds (Padovani et al. 2018(Padovani et al. , 2020)).Cosmic rays are not attenuated in the molecular clouds as strongly as UV photons (Ivlev et al. 2018;Padovani et al. 2018;Silsbee et al. 2018).Thus, dust grains in the interstellar medium can be heated by impinging cosmic rays, thereby heating up the icy grain mantles and resulting in calamitous explosions (Leger et al. 1985;Ivlev et al. 2015), thereby activating chemistry in solids (Shingledecker et al. 2017).Magnetohydrodynamic simulations have shown a higher cosmic ray production in protostars in a clustered environment (Kuffmeier et al. 2020), which would be consistent with the lower H 2 S/OCS ratios for such protostars found in this work.The results suggest that the H 2 S/OCS ratio traces the environment (isolated/clustered) of the protostellar systems.However, a follow-up study is needed as the sample consisted of only one isolated source.

Conclusions
This work probed a sample of ten low-mass protostars for the presence of H 2 S, OCS, and their isotopologs using ALMA ACA Band 6 observations.For 5 out of 10 protostars, the H 2 S/OCS ratio was firmly constrained and for an additional 3, best-possible estimates were obtained.This ratio is thought to be a potential chemical and physical clock of star-forming regions, which sheds light on the sulfur depletion that transpires from the diffuse medium to the dense core stage.The main conclusions are: -Main S-bearing species, H 2 S and OCS are detected in IRAS 16293-2422 A, IRAS 16293-2422 B, NGC 1333-IRAS4A, NGC 1333-IRAS4B, Per-B1-c, BHR71-IRS1, and Ser-SMM3.1-σ upper limits on the column densities of OCS are derived for RCrA IRS7B, TMC1, and Per-emb-25.1σ upper limits on the column densities H 2 S are derived for TMC1 and Per-emb-25.-The gas-phase H 2 S/OCS ratio ranges from 0.2 to above 9.7, and is typically at least one order of magnitude lower than that of ices.The lowest ratio is obtained for IRAS 16293 A and Ser-SMM3, while the highest for BHR71-IRS1.The environment of the natal cloud, prior to the onset of star formation, may have played a major role in the distribution of sulfur across various S-bearing molecules, which have resulted in an order of magnitude spread in the H 2 S/OCS ratio.-The upper limits derived for the interstellar ices (Mon R2 IRS2 and W33A) lie within the uncertainties of the cometary ices ratio, specifically that of comet 67P/Churyumov-Gerasimenko. The protostellar ratios are lower than the upper limits on the interstellar ices ratio and the cometary ices ratio by at least an order of magnitude for all sources except BHR71-IRS1.-The lower ratio in clustered protostellar regions could be due to elevated birth cloud temperatures or due to additional radiation from nearby protostars, thereby enhancing the photodissociation pathways from H 2 S to OCS. -The high H 2 S/OCS ratio in BHR71-IRS1 could be the result of less efficient photodissociation of H 2 S to OCS in the presence of CO ice in its isolated birth cloud or more efficient hydrogenation chemistry leading to more efficient H 2 S formation.
Figure 3: H 34 2 S line detected in IRAS 16293 B. The observed spectrum (in blue), rest frequency of the detected line (brown dashed line), spectroscopic uncertainty on the rest frequency of the detected line (yellow shaded region), and fitted synthetic spectrum (in pink) for source size: 2 ′′ , excitation temperature: 125 K, and FWHM: 1 km s −1 .

Figure 4 :
Figure4: N(H 2 S)/N(OCS) of the studied sources.Different symbols represent different types of sources, i.e., 'star' for close binary (< 500 au), 'square' for wide binary (500 − 5000 au), and 'circle' for single sources (within 5 000 au).The upper limits on the interstellar ice (W33A and Mon R2 IRS2) ratios are shown by downward arrows.The uncertainty on the H 2 S/OCS ratio in comet 67P/C-G is shown by the coral shaded region.The lower limit on the ratio in IRS7B is shown by an upwards arrow.The H 2 S/OCS ratios for the cold (cyan) and warm (orange) components of B1-c, and cold (cyan) component of BHR71-IRS1 are the best-available estimates pending opacity issues.These latter three data points do not have error bars associated to them in the figure to indicate that they are merely estimates.

Figure
Figure B.1: Observed spectra (in blue), rest frequency of the detected line (brown dashed line), spectroscopic uncertainty on the rest frequency of the detected line (yellow shaded region), blending species (green dash-dotted line and red dashed line), and fitted synthetic spectra (in pink) plotted for the sulfur-bearing species detected towards IRAS 16293-2422 B. N(H 2 S) and N(OCS) are used for the synthetic spectra of the optically thick lines.For the displayed fits, a source size of 2 ′′ is assumed.

Figure
Figure D.1: Observed spectra (in blue), rest frequency of the detected line (brown dashed line), spectroscopic uncertainty on the rest frequency of the detected line (yellow shaded region), blending species (green dash-dotted line and red dashed line), and fitted synthetic spectra (in pink) plotted for the sulfur-bearing species detected towards NGC 1333-IRAS4A for T ex = 150 K. N(H 2 S) and N(OCS) are used to fit the synthetic spectra of the optically thick lines.

Figure E. 2 :TFigure
Figure E.2: Observed spectra (in blue), rest frequency of the undetected line (brown dashed line), and spectroscopic uncertainty on the rest frequency of the undetected line (yellow shaded region) plotted for the sulfur-bearing species undetected towards RCrA IRS7B.Synthetic spectrum (in pink) fitted to the OCS, v=0 line with the 1-σ upper limit on its column density.

Table 2
. IRAS16293-2422 (hereafter, IRAS 16293) is a triple protostellar source, consisting of protostars A and B, separated by 5.3 ′′ (747 au; van der Wiel et al. 2019) and disklike structures around the two sources, located in the Rho Ophiuchi star-forming region at a distance of 141 pc 1 and Appendix 5. Full observed spectral Drozdovskaya et al. (2018)in a closer agreement with the column density of OCS, v = 0 (2.8×10 17 cm −2 ) derived inDrozdovskaya et al. (2018), also based on minor isotopologs, than for source size of 2 ′′ (7.0 × 10 16 cm −2 ).Drozdovskaya et al. (2018)used a smaller source size (0.5 ′′ ) to constrain the column densities of OCS and H 2 S.These comparisons suggest that the ALMA ACA observations in this work are subject to beam dilution, hence the column densities are likely somewhat underestimated.The olumn density of H 2 S could not be constrained to better than a factor of 10 inDrozdovskaya et al. (2018), namely 1.6 × 10 17 − 2.2 × 10 18 cm −2 . Ths due to the fact that only deuterated isotopologs of H 2 S were covered by the PILS observations and the D/H ratio of H 2 S is only constrained to within a factor of 10.Based on values of the H 2 S column densities for 1 ′′ and 2 ′′ source sizes obtained in this work, the lower estimate for the H 2 S column density inDrozdovskaya et al. (2018)seems to be more accurate.In turn, the H 2 S/OCS ratio obtained in this work (1.3) is closer to the lower end of the 0.7 − 7 range computed inDrozdovskaya et al. (2018).
(Wilson 1999)57(Wilson 1999).dovskaya et al(2018).Likewise, the OCS, v=0 column density determined from the minor isotopologs of OCS for a source size of 1 (Palumbo et al. 1997)elescope Facility (IRTF) and upper limits on the H 2 S abundance derived based on the non-detection of the 3.98 µm band.The column density of solid OCS with respect to solidH 2 O is N solid (OCS)/N solid (H 2 O) = 4×10 −4(Palumbo et al. 1995)and N solid (OCS)/N solid (H 2 O) = 5.5×10 −4(Palumbo et al. 1997)for W33A and Mon R2 IRS2, respectively.Based on the nondetection of solid H 2 S towards W33A in the Infrared Space Observatory (ISO) Spectra from the Short Wavelength Spectrometer (SWS), N solid (H 2 S)/N solid (H 2 O) solid<0.03(vander Tak et al.  2003).The upper limits on the solid H 2 S and solid H 2 O in Mon R2 IRS2 are < 0.2 × 10 17 and 42.7 × 10 17 cm −2

Table 5 :
H 2 S/OCS ratio for the studied sources, including their evolutionary class, binarity, environment, and the derived column densities of H 2 S and OCS for the stated excitation temperatures.The H 2 S/OCS ratios for the cold and warm components of B1-c, and the cold component of BHR71-IRS1 are the best-available estimates pending opacity issues.
Observed spectra (in blue), rest frequency of the detected line (brown dashed line), spectroscopic uncertainty on the rest frequency of the detected line (yellow shaded region), blending species (green dash-dotted line and red dashed line), and fitted synthetic spectra (in pink) plotted for the sulfur-bearing species detected towards IRAS 16293-2422 A. N(H 2 S) and N(OCS) are for the synthetic spectra of the optically thick lines.