Introduction

Since the first discovery in 19891, discharges of liquid and hydrate-coated CO2 have been observed in various submarine hydrothermal systems2,3,4,5,6,7 (Supplementary Movie 1). Isotopic and compositional analyses have revealed that the venting CO2 derives from marine carbonate and mantle carbon, with a minor contribution from sedimentary organic matter2,8. In the deep reaction zones of submarine hydrothermal systems, elevated temperature, and pressure allow even molar-level dissolution of CO2 into hydrothermal fluids, whereas their upward flow and mixing with seawater reduce CO2 solubility, facilitating the phase separation of CO23,9. The phase separation of CO2 in subseafloor upwelling regions has also been suggested from fluid inclusions in ~3.5-Gyr-old hydrothermal precipitates from the Pilbara craton, Australia10,11. The Archean oceanic crust records extensive carbonatization under CO2-rich seawater12,13. Considering theoretical and laboratory simulations suggesting comparably, or even further, CO2-enriched surface environments on the Hadean Earth14,15,16, widespread occurrence of CO2 vents and pools is a likely feature of the Hadean Ocean floor9.

Our group recently hypothesized that such generation and transportation of CO2 fluids were a crucial circumstance for the chemical evolution of life9. Although it has long been suggested that life originated in deep-sea hydrothermal vent environments17,18,19, the apparently water-filled condition has posed many serious difficulties in experimental verification of prebiotic organic synthesis20. Indeed, the expansion of the narrow chemical space accessible in water has been a major challenge in the broad field of organic chemistry21,22. Although low water activities within hydrothermal chimney pores may have promoted dehydration/polymerization of building blocks of life (e.g., amino acids)23,24, this possibility remains to be tested.

Supercritical CO2 (hereafter referred to as scCO2) offers an alternative measure against the water problem. Beyond warm and moderately pressurized conditions (>31.0 °C and >7.38 MPa)25, CO2 is present as a supercritical fluid, exhibiting a versatile solvent function owing to its intermediate property between liquid and gas. The unique characteristics, coupled with the ease of preparation, have led to the use of scCO2 in a variety of chemical and industrial processes that are difficult to be operated in a simple aqueous solution26. It has been suggested that the formation of amphiphilic vesicles at the scCO2–water interface in a tectonic fault zone, and the subsequent selection and accumulation of life’s building blocks within the vesicles facilitated the origin of life27,28,29,30. Although there is no experimental support for the proposed supply of primordial ingredients, this idea gives an important insight into the role of scCO2 in the emergence of biological functions (e.g., compartmentalization). Thus, liquid and scCO2 formed through subseafloor hydrothermal processes possibly diversified prebiotic chemistry and resultant materials/systems available for the origin of life.

Here, we demonstrate methanethiol synthesis under a simulated scCO2 condition in a primordial subseafloor. Methanethiol is a necessary ingredient for the synthesis of methyl thioacetate, a presumptive prebiotic precursor of acetyl-CoA31. Harboring a “high-energy” bond in the thioester structure, acetyl-CoA occupies the central position in cellular metabolism: it is a key component of six out of seven representative carbon fixation pathways, taking part in diverse biosynthetic processes involving C–C bond formation, esterification, and amide bond formation32,33,34. Recent works led by Moran and Martin suggested a one-pot pyruvate synthesis from gaseous CO2 and hydrogen (H2) on pure metallic iron (Fe0), pure metallic nickel (Ni0), and their alloys (e.g., Ni3Fe) and oxides (e.g., Fe3O4) as a protometabolic acetyl-CoA pathway35,36,37,38,39. The proposed reaction, however, provides no explanation for the origin and antiquity of thioester-dependent metabolism. Thioesters may have played central roles in protometabolism not only as carbon sources but also as energy currencies in a manner analogous to adenosine triphosphate (ATP)40,41,42, serving as an entry point of phosphate into metabolism43,44.

However, no geochemical evidence of abiogenic methanethiol has been observed in hydrothermal fluids from mid-ocean ridge systems45. Although hydrothermal experiments by Heinen and Lauwers46,47 synthesized methanethiol from CO2 and hydrogen sulfide (H2S) on iron sulfide or Fe0, the yield was much less than 0.1% based on H2S. Specifically, the total yield of thiols including methanethiol was 0.060% when a serum bottle (60 mL) containing 2.6 mmol of CO2, 120 μmol of H2S, 114 μmol of FeS, and 10 mL H2O was incubated at 90 °C (see Table VII in ref. 46). Assuming the distributions of CO2 and H2S between the gas- and aqueous-phases calculated thermodynamically48,49, their initial concentrations are predicted to be 17 and 2.3 mM, respectively. Loison et al.50 showed gas chromatography–mass spectroscopy (GC–MS) evidence of methanethiol formed from carbon monoxide (CO) and H2S in the presence of nickel sulfide (NiS), but no yield was reported. The reductions of potassium cyanide (KCN), potassium cyanate (KOCN), potassium thiocyanate (KSCN), and carbonyl sulfide (CS2) by mackinawite (FeS) or oxalic acid are alternative means of methanethiol production51,52, but availabilities of these stating materials in primordial undersea environments are unclear. To the best of our knowledge, there is no aqueous-phase mechanism demonstrated so far that supports an ample and sustained supply of methanethiol as envisioned in the protometabolism scenario53 and the relevant thermodynamic predictions54,55.

In contrast, methanethiol was recently detected in liquid CO2-containing fluid inclusions in ~3.5-Gyr-old hydrothermal precipitates with the carbon isotopic composition suggesting an abiotic origin56. This fact, along with the geological availability and chemical versatility of scCO2 described above, led us to the idea that methanethiol was sourced from subseafloor scCO2. Indeed, a gas-phase reaction of CO, H2S, and H2 on a molybdenum sulfide (MoS2) catalyst is a well-known artificial means of methanethiol production57. CO is generated from CO2 and H2 through the water–gas shift reaction (CO2 + H2 ⇌ CO + H2O) even in a water-filled condition58. Mo is a trace but ubiquitous component of natural basalt and komatiite59,60,61,62. Although the rock-containing Mo is predominantly in the +6 valence state (i.e., presents as molybdate (MoO42–))63,64, molybdate is readily converted to MoS2 with H2S as the reductant and source of sulfur (see the “Methods” section). Mo is also a common trace element in modern and ancient hydrothermal sulfide deposits65,66. Up to 6791 ppm of Mo has been found in the pyrite deposits66. Moreover, Mo is a vital element for modern biochemistry functioning in molybdenum cofactor and nitrogenase67,68. It is widely recognized that the presence of water often severely suppresses catalytic reactions21,22. However, we show below that, with MoS2 deposited in a carbonate-NaCl solution, methanethiol formation still effectively proceeds from H2S, H2, and scCO2 as the sole carbon source. Our result is the first experimental evidence supporting the significance of subseafloor scCO2 in abiotic organic synthesis. Based on the obtained result, with the known prebiotic processes in hydrothermal vent environments, we propose a new scenario involving liquid and scCO2 for the emergence of protometabolism on the primordial ocean floor.

Results

MoS2 was prepared by the reduction and sulfidation of molybdate under scCO2 (see Methods). The obtained MoS2 was a poorly crystallized molybdenite forming submicro- to several micrometer-sized aggregates with occasional rose-like structures (Supplementary Fig. 1). The MoS2 (50 mg) was deposited in 3.43 mL of an aqueous solution of sodium sulfide (Na2S) and hydrogen chloride (HCl) (400 mmol L–1 for each) in a glass vial. The vial was encased in a titanium reaction vessel (Supplementary Fig. 2) with the headspace (34.3 mL) filled with CO2 and H2 (6.0 and 0.3 MPa at 25 °C, respectively), and heated at 200 °C for up to 14 days. At the reaction temperature, the dissolved concentration and partial pressure of CO2 were estimated to be 0.8 mol kg–1 and 14 MPa, respectively25,69, which met the condition for CO2 to be supercritical (>31.0 °C and >7.38 MPa)25. The introduced H2 set the initial dissolved H2 concentration at 200 °C to be 5 mM70, which was within a typical range of H2 concentration in fluids from submarine hydrothermal systems71,72. Considering the pKa of H2S and CO2 at the reaction condition48,73, the neutralization of HS– was expected to lead the solution pH to 6.2, a slightly higher value than the neutral pH at 200 °C (5.6). After a certain period of heating, the vessel was cooled under air, depressurized, and opened to collect the aqueous solution. The headspace gas was sampled into a sealed glass vial just before the depressurization.

The GC–MS analysis of the gas sample detected methanethiol as a major organosulfur product (Fig. 1a). Throughout the 14-day reaction, the amount of methanethiol increased steadily up to 109 μmol (Fig. 1b), which corresponded to 7.9% of the initial amount of H2S. Other identified gas-phase products include CO, methane (CH4), dimethyl sulfide (CH3SCH3), and trace amounts of carbonyl sulfide (OCS) and CS2 (Fig. 1a and Supplementary Figs. 4 and 5). Dimethyl disulfide (CH3SSCH3) was also seen, but it was likely formed through methanethiol oxidation by atmospheric oxygen (O2) at the timing of gas sampling and/or sample injection into the GC. In the aqueous solution, formate (HCOO–) was observed as the dominant dissolved product (Supplementary Figs. 6 and 7). Formate exhibited the highest concentration at the shortest reaction duration (86 mmol L–1 after the 3-day reaction) and subsequently reduced the value (Fig. 1c). Considering the gradual depletion of H2 during the experiment (Fig. 1c), formate was expected to rapidly attain thermodynamic equilibrium with H2 and CO2 for its formation and decomposition (CO2 + H2 ⇌ HCOO– + H+) under the examined condition. Besides formate and methanethiol, the proton nuclear magnetic resonance (1H NMR) spectra of sample solutions indicated the occurrence of several dissolved species at low concentrations (Supplementary Fig. 6), but biologically relevant organic compounds such as pyruvate and glyoxylate were not identified.

Fig. 1: Product characteristics of the scCO2 experiments.
figure 1

a GC–MS chromatogram of the sulfur-containing compounds measured after the 7-day reaction of H2, H2S, and CO2 in the presence of MoS2 at 200 °C. See Supplementary Fig. 4 for the mass spectra of identified species. The yields of products are shown in b together with the results after the 3- and 14-day experiments. c The amounts of H2, CO, CH4, and HCOO– after the three reaction durations. The yield on the left y-axis (0–350 μmol) refers to CO, CH4, and HCOO–, while the corresponding dissolved concentration of HCOO– is indicated on the right. d The yields of methanethiol, CO, CH4, and HCOO– under various reaction conditions. w and w/o denote with and without, respectively. d represents day. All runs were conducted under scCO2 at 200 °C. The error bars in a–d were determined based on multiple independent runs under the same reaction conditions (see the “Methods” section).

Full size image

When either H2S, H2, or MoS2 was absent, very little or no detectable methanethiol was formed (<0.1 μmol; Fig. 1d). The yield of methanethiol was negligibly low (<0.1 μmol) when MoS2 was replaced with mackinawite (FeS) or a synthetic Mo-free komatiite. Regarding the reported gas-phase production of methanethiol from CO, H2S, and H257, the proposed catalytic process on MoS2 involves OCS formation from CO and H2S and the hydrogenation of OCS with H atoms provided by H2 dissociation. The necessity of H2 with H2S and MoS2 in our experiment (Fig. 1d), together with the detection of CO and OCS (Fig. 1a and Supplementary Figs. 4 and 5), suggests that a similar mechanism occurred in our binary system of CO2 and H2O. Additionally, because the run without H2S produced CH4 in the greatest amount (Fig. 1d), a surface-bound methyl formed through CO adsorption and hydrogenation—a likely intermediate in the CH4 formation— possibly served as a component of methanethiol. The steeper increase in the yield of dimethyl sulfide (CH3SCH3) at the later stage of reaction duration (Fig. 1b) corroborates the formation and gradual accumulation of the surface-bound methyl on MoS2. Interestingly, the coexistence of H2S, H2, and MoS2 also favored formate production (Fig. 1d). Formate is likely a by-product, not an intermediate on the route to methanethiol: when 13C-labeled formate (100 mmol L–1) was used as a starting material with H2S and MoS2 (H2 was not introduced in this run), the 7-day reaction at 200 °C led to the formation of a small amount of methanethiol (2.9 μmol); however, its mass spectrum was close to that of non-labeled standard (Supplementary Fig. 8). A probable interpretation is that formate served as a source of H2 (HCOO– + H+ → CO2 + H2). Given its standard Gibbs energy of reaction (ΔrG° = 7.8 kJ mol–1 at 200 °C, the total pressure of ~15 MPa, and pH 6.2)48,74, the formate oxidation is predicted to lead to a dissolved concentration and partial pressure of H2 of 0.4 mmol kg–1 and 0.04 MPa, respectively, at equilibrium. The corresponding amount of H2 in the vapor headspace is 0.3 mmol. Because this thermodynamic calculation does not take into account the reactions that consume H2 (e.g., methanethiol synthesis), the actual amount of H2 gas quantified after the 7-day reaction (0.21 mmol) was lower than the calculation. Nevertheless, the by-product 13CO2 must be significantly less than the amount of CO2 initially introduced (0.343 mmol at the maximum vs. ~150 mmol). Thus, its impact on the mass spectrum of methanethiol is negligible. Elucidation of the catalytic mechanism on MoS2 is beyond the scope of the present study, but further isotopic and kinetic assessment of the products, in combination with the spectroscopic/computational analyses of the surface process, should be an important future work.

After the scCO2 experiments, no significant morphological and crystallographic changes of MoS2 was discerned (Supplementary Fig. 9). In contrast, a complete conversion of mackinawite to pyrrhotite was observed after the 7-day reaction at 200 °C (Supplementary Fig. 10). This oxidative alteration of mackinawite was accompanied by H2 generation: the run without H2 gas as the starting material under otherwise identical condition resulted in the appearance of 114 μmol of H2 from 500 mg of mackinawite. The pyrrhotite formation and concomitant H2 production also occurred on komatiite through its 7-day exposure to H2S under scCO2 (Supplementary Fig. 11; again, H2 was excluded from the starting material). The resultant H2 from 500 mg of komatiite was 22 μmol, whereas the amount of H2 generated in the absence of H2S was only 2.5 μmol. The substantial carbonatization of komatiite to siderite (FeCO3) is another interesting outcome in the presence of H2S (Supplementary Fig. 11).

It is also noteworthy that methanethiol was formed from gaseous CO2: when a low pressure of CO2 (0.34 MPa at 25 °C) was used in place of 6 MPa of CO2, the 7-day reaction at 200 °C under otherwise identical conditions resulted in the formation of methanethiol in a yield of 1.7% based on H2S (or 23 μmol). 3.2 μmol of CO and 1.1 μmol of CH4 were also generated in this run.

Discussion

Mechanism of MoS2-catalyzed methanethiol synthesis under scCO2

Our experiment demonstrated that MoS2 effectively catalyzes methanethiol synthesis from CO2, H2S, and H2 under scCO2, even with the MoS2 deposited in a carbonate-NaCl solution. This result is in striking contrast to the report by Mul et al.75, in which the presence of water vapor strongly suppressed methanethiol synthesis from gaseous CO and H2S on a vanadium oxide (V2O5) catalyst due to hydrolysis of the reaction intermediate, OCS (OCS + H2O → CO2 + H2S). Unlike V2O5, MoS2 is capable of serving as a S source as well as a catalyst, generating a surface-bound OCS through CO adsorption76,77,78,79,80. Its subsequent hydrogenation with H atoms provided by H2 dissociation forms a surface-bound methylthiolate (CH3S–). Once methanethiol is desorbed, the resultant S vacancy is immediately repaired by H2S78, thereby a repeated catalytic cycle is realized with a limited hydrolytic loss of intermediate species. The OCS formation from CO and H2S is also catalyzed by FeS81. Thus, the observed superiority of MoS2 over FeS for methanethiol synthesis (Fig. 1d) is likely due to the MoS2’s unique capability of facilitating the formation and hydrogenation of surface-bound OCS.

Because methanethiol was formed under gaseous CO2 as well as under scCO2, the phase of CO2 source is not a crucial factor for the surface catalytic reaction. Nevertheless, the high density of scCO2 is expected to be advantageous to methanethiol synthesis owing to the resultant concentration of CO via the CO2/CO equilibrium (CO2 + H2 \(\rightleftharpoons\) CO + H2O). The greater yield of methanethiol obtained from scCO2 than from gaseous CO2 (61 vs. 23 μmol after the 7-day reaction at 200 °C) supports this consideration.

Another synthetic route from CO to methanethiol proposed so far is CS2 formation through OCS disproportionation and its subsequent hydrogenation57:

$$2{{\rm{OCS}}} \, \rightleftharpoons \, {{\rm{C}}}{{{\rm{S}}}}_{2}+{{\rm{C}}}{{{\rm{O}}}}_{2}$$
(1)
$${{\rm{C}}}{{{\rm{S}}}}_{2}+3{{{\rm{H}}}}_{2}\to {{\rm{C}}}{{{\rm{H}}}}_{3}{{\rm{SH}}}+{{{\rm{H}}}}_{2}{{\rm{S}}}$$
(2)

This process is unlikely to proceed, to a great extent, under scCO2 because of the low equilibrium constant for Eq. (1) (0.24 at 200 °C and 15 MPa; the gas-phase thermodynamic parameters were used because of the lack of the corresponding parameters for aqueous species82). In our experiment, the generated CS2 was trace (<0.5 μmol) regardless of the use of H2 as the starting material.

In addition to the mechanistic perspective discussed above, the significance of H2 is acknowledged from the ΔrG° for methanethiol synthesis from CO2, H2S, and H2 under the experimental condition (200 °C and the total pressure of ~15 MPa) that is significantly lower than the value without H248,49,54:

$${{\rm{C}}}{{{\rm{O}}}}_{2}\left({{\rm{aq}}}\right)+3{{{\rm{H}}}}_{2}\left({{\rm{aq}}}\right)+{{{\rm{H}}}}_{2}{{\rm{S}}}\left({{\rm{aq}}}\right)\to {{\rm{C}}}{{{\rm{H}}}}_{3}{{\rm{SH}}}\left({{\rm{aq}}}\right)+2{{{\rm{H}}}}_{2}{{\rm{O}}}\left({{\rm{l}}}\right),\\ {\Delta }_{r}{{\rm{G}}}^{\circ}={{\mbox{-}}}68.0\,{{\rm{kJ}}}\; {{\rm{mo}}}{{{\rm{l}}}}^{-1}$$
(3)
$${{\rm{C}}}{{{\rm{O}}}}_{2}\left({{\rm{aq}}}\right)+4{{{\rm{H}}}}_{2}{{\rm{S}}}\left({{\rm{aq}}}\right)\to {{\rm{C}}}{{{\rm{H}}}}_{3}{{\rm{SH}}}\left({{\rm{aq}}}\right)+3{{{\rm{S}}}}^{0}\left({{\rm{s}}}\right)+2{{{\rm{H}}}}_{2}{{\rm{O}}}\left({{\rm{l}}}\right),\,\\ {\Delta }_{r}{{\rm{G}}}^{\circ}=67.9\,{{\rm{kJ}}}\; {{\rm{mo}}}{{{\rm{l}}}}^{-1}$$
(4)

where (aq), (l) and (s) denote the dissolved, liquid, and solid states, respectively. The strongly negative ΔrG° for Eq. (3) supports the feasibility of abundant formation of methanethiol in subseafloor scCO2 environments as long as the necessary materials (H2, H2S, and MoS2) are available.

Availabilities of H2S, H2 and MoS2 in subseafloor scCO2 environments

Liquid and scCO2 flowing out of the seafloor occasionally contain high concentrations of H2S at the mole fractions of up to 3% due to degassing of the underlying magma1,2. H2S as well as CO2 are the dominant non-aqueous components in fluid inclusions in ~3.5-Gyr-old hydrothermal precipitates from the Pilbara Craton, Australia56. The molar ratio of H2S/CO2 set in our experiment (0.9%) is consistent with these observations of modern and ancient hydrothermal systems rich in CO2 and H2S.

H2 is a common reductive component in fluids from submarine hydrothermal systems. A well-known geochemical mechanism of H2 production is serpentinization, a subseafloor water–rock interaction causing H2O reduction by coupling with Fe2+ oxidation. Although a high concentration of CO2 suppresses H2 generation through low-temperature serpentinization (<~300 °C) owing to the precipitation of Fe2+-bearing carbonates, the high-temperature serpentinization (>~300 °C) generates extraordinarily high concentrations of H2 even in CO2-saturated seawater83,84.

Alternative source of H2 is the reaction between H2S and iron sulfide minerals. In our scCO2 experiment without H2 gas as the starting material, 500 mg of mackinawite generated 114 μmol of H2 through its conversion to pyrrhotite. The amount of H2 (114 μmol) corresponded to 2.0 mol% of the mackinawite used. It is expected that mackinawite initially reacted with H2S to form pyrite and H285,86:

$${{\rm{F}}}{{\rm{eS}}}\left({{\rm{s}}}\right)+{{{\rm{H}}}}_{2}{{\rm{S}}}\left({{\rm{aq}}}\right)\to {{\rm{Fe}}}{{{\rm{S}}}}_{2}\left({{\rm{s}}}\right)+{{{\rm{H}}}}_{2}\left({{\rm{aq}}}\right)$$
(5)

H2 then served as a reductant for the pyrite-to-pyrrhotite conversion87:

$${{\rm{F}}}{{\rm{e}}}{{{\rm{S}}}}_{2}\left({{\rm{s}}}\right)+\left(1{{{-}}}{x}\right){{{\rm{H}}}}_{2}\left({{\rm{aq}}}\right)\to {{\rm{Fe}}}{{{\rm{S}}}}_{1+{x}}\left({{\rm{s}}}\right);+\left(1{{{-}}}{x}\right){{{\rm{H}}}}_{2}{{\rm{S}}}\left({{\rm{aq}}}\right) \, \left(0\, < \, x < \, 0.125\right)$$
(6)

The residual H2 accounts for our observation of mackinawite-derived H2 generation:

$${{\rm{F}}}{{\rm{eS}}}\left({{\rm{s}}}\right)+{x}{{{\rm{H}}}}_{2}{{\rm{S}}}\left({{\rm{aq}}}\right)\to {{\rm{Fe}}}{{{\rm{S}}}}_{1+{x}}\left({{\rm{s}}}\right)+{x}{{{\rm{H}}}}_{2}\left({{\rm{aq}}}\right) \, \left(0 \, < \,{x} \, < \,0.125\right)$$
(7)

The pyrrhotite formation and concomitant H2 production also occurred on komatiite through its 7-day exposure to H2S under scCO2. Thus, komatiite and mackinawite in H2S-rich environments can serve as the source of H2 for methanethiol synthesis (Supplementary Fig. 12). Given the thermodynamic equilibrium between mackinawite and siderite (FeCO3(s) + H2S(aq) ⇌ FeS(s) + CO2(aq) + H2O(l)), mackinawite is more stable than siderite when the activity ratio of H2S/CO2 exceeds 0.0064 under our experimental condition (Supplementary Fig. 13). Thus, unless the dissolved CO2 is extremely concentrated, several to dozens of mmol kg–1 of H2S, which are typical H2S concentrations in fluids from submarine hydrothermal systems71, are sufficient for mackinawite precipitation and the subsequent H2 generation.

Natural basalt and komatiite contain Mo as a trace but ubiquitous component at the reported concentrations up to 18.7 ppm59. Importantly, komatiite was substantially carbonatized through our scCO2 experiment in the presence of H2S (Supplementary Fig. 11). Because Mo does not form a stable carbonate mineral, hydrothermal processes of komatiite in subseafloor scCO2 environments likely extract the rock-containing Mo into the fluids. Once released Mo is sulfurized to form MoS2, and the surface can initiate the synthetic steps to methanethiol, as demonstrated in the present study (Fig. 1).

Significance of scCO2 in the initiation of protometabolism

It has been suggested that methanethiol played at least three key roles in the origin and early evolution of life: as (1) an ingredient of methyl thioacetate (see the “Introduction” section), (2) a metal chelator for delivering hydrothermally derived metals to the deep ocean88, and (3) a nutrient for microbial chemoautotrophic activities89,90. The first possibility can be realized in H2-rich alkaline hydrothermal vent environments with the geoelectrochemical mechanism91,92, in which the strongly negative electric potentials generated by H2 oxidation (H2 → 2H+ + 2e–) at the fluid-mineral interface facilitate CO2-to-CO reduction and methanethiol carbonylation at the mineral-seawater interface with NiS partially electroreduced to metal (NiS_PERM) working as the catalyst93. This methanethiol conversion efficiently proceeds even at room temperature and neutral pH, facilitating the accumulation of methyl thioacetate while suppressing the decomposition.

It may be worth referring to other reported experiments for prebiotic thioester synthesis to clarify the significance of the above-mentioned mechanism. Chandru et al.94 examined the hydrolysis of methyl thioacetate and thioacetic acid under simulated hydrothermal vent conditions and confirmed that these chemicals are unstable at elevated temperatures and extreme pHs. Their conclusion, “it is unlikely that these species could have accumulated abiotically to any significant extant” is consistent with the experiment by Huber and Wächtershäuser31 for a CO–methanethiol reaction at 100 °C, in which no thioester synthesis was observed in most of the examined conditions although its hydrolysis product—acetic acid—was obtained in the yield as high as 40% based on the initial amount of methanethiol.

In contrast, the geoelectrochemical mechanism demonstrated by Kitadai et al.93 facilitates both the CO2-to-CO reduction and the thioester synthesis through CO–methanethiol reaction even at room temperature and neutral pH. For example, through a 7-day electroreduction in a CO2-saturated NaCl solution, a coprecipitate of NiS and CoS was found to generate and accumulate surface-bound CO, from which methyl thioacetate was produced in a yield of up to 56% via the subsequent 7-day incubation with methanethiol (100 mM) at room temperature and pH 793. To the best of our knowledge, there is no experimental demonstration of thioester synthesis from CO and thiol under mild aqueous conditions except for this work. Other prebiotic mechanisms of thioester synthesis proposed so far include the reactions of thiols with glyceraldehyde95 or α-keto acids (e.g., pyruvate)96. The latter reaction requires an oxidant (S2O82–) or UV light. These processes are unlikely to have occurred sustainably in primordial deep-sea environments.

In the present ocean hydrothermal systems, H2-rich alkaline fluids largely derive from low-temperature (<~200 °C) water–rock interaction between ultramafic rocks (e.g., peridotite) and CO2-depleted seawater97,98,99,100. Beneath the primordial ocean floor with hotter mantle convection (Fig. 2a), an alternative mechanism of fluid basification was a high-temperature (>300 °C) hydrothermal alteration of basalt and komatiite under CO2-rich conditions84,101. A few hundred mmol kg–1 of CO2 suffices to generate alkaline pH (>9) through anorthite decalcification and calcite precipitation101. Additionally, hydrothermal fluids are expected to be basified in association with the phase separation between H2O and CO2 owing to the following reactions:

$${{\rm{H}}}{{\rm{C}}}{{{\rm{O}}}}_{3}^{-}\left({{\rm{aq}}}\right)+{{{\rm{H}}}}^{+}+\left({{\rm{aq}}}\right) \, \rightleftharpoons \, {{\rm{C}}}{{{\rm{O}}}}_{2}\left({{\rm{aq}}}\right)+{{{\rm{H}}}}_{2}{{\rm{O}}}\left({{\rm{l}}}\right)$$
(8)
$${{\rm{C}}}{{{\rm{O}}}}_{2}\left({{\rm{aq}}}\right) \, \rightleftharpoons \, {{\rm{scC}}}{{{\rm{O}}}}_{2}$$
(9)
Fig. 2: Proposed geological mechanism for the generation, transportation, and prebiotic reaction of methanethiol in a primordial ocean hydrothermal system.
figure 2

a The hot and active mantle convection induced a massive upwelling of komatiite/basaltic melt beneath the ocean floor130,131. Infiltration of CO2-rich seawater into the solidified (ultra)mafic rocks led to the precipitation of carbonate minerals while generating H2 via serpentinization and sulfidation. When the infiltrated fluids reached the deep reaction zone, they incorporated a substantial amount of CO2 degassed from underlying magma. The highly carbonatized oceanic crust formed through the oceanic ridge volcanism11,12,13 also supplied CO2 via carbonate dissolution and decomposition. b In an upwelling region, the decrease in hydrostatic pressure caused the phase separation of CO29, setting the stage for methanethiol synthesis from scCO2, H2, and H2S on MoS2 precipitates. The concomitantly generated alkaline hydrothermal fluids discharged into the metal-rich ocean to form a hydrothermal chimney with a powerful geoelectrochemical energy91,92. c At the chimney–seawater interface, methanethiol diffused from liquid/scCO2 underwent (electro)catalytic reactions with CO2 to form methyl thioester93, a potential ingredient and energy currency for protometabolism.

Full size image

When a subseafloor upwelling region at 300 °C is considered, CO2 is soluble up to 3.0 mol kg–1 under the hydrostatic pressure of 50 MPa, whereas the solubility reduces to 0.6 mol kg–1 as the pressure decreases to 15 MPa102. As the pKa of CO2 significantly exceeds the neutral pH at the high-temperature condition (8.5 vs. 5.7 at 300 °C and 15 MPa), the phase separation of CO2 from upwelling fluids would enhance and maintain alkaline pH, unless the fluids are significantly cooled by the mixing with seawater. If the generated alkaline fluids are rich in H2 and H2S, their discharges into a metal-rich ocean must result in the formation of sulfidic hydrothermal chimneys with powerful geoelectrochemical energy. On the present ocean floor, discharges of liquid and scCO2 are occasionally observed in close vicinity to hydrothermal vents, with some portions being trapped in the chimney cavities and flange structures103 (Supplementary Movie 1). Such spatial association between CO2 and hydrothermal fluids is expected to facilitate the involvement of scCO2-derived materials in the (electro)chemical processes at the chimney–seawater interface.

Taken together, the following four consecutive steps are envisioned as a possible entry route of methanethiol into protometabolism in primordial ocean hydrothermal vent environments (Fig. 2b and c): (1) the generation and transportation of methanethiol through the upward flux of CO2, (2) the discharges of liquid and scCO2 in the vicinity of hydrothermal vents and chimneys, (3) the dissolution and diffusion of methanethiol from the discharged CO2 into the surrounding seawater, and (4) the (electro)catalytic reaction of methanethiol with CO2 to form methyl thioacetate, a potential ingredient and energy currency for protometabolism. Thus, our experimental results, coupled with the relevant geological evidence, suggest that protometabolism was not a mere aqueous-phase CO2 reduction on a mineral catalyst; it was a consequence of the active hydrothermal circulation beneath the primordial ocean floor (Fig. 2a) causing multiple geochemical processes including the generation of electric current and the flow of scCO2. It is noteworthy that the density of liquid CO2 exceeds that of seawater below ~3000 m104. Thus, the delivery of methanethiol is feasible on seafloor environments shallower than the threshold. Given that the mean ocean depth in the Hadean era exceeded the present level105, hydrothermal systems on oceanic plateaus may be more accessible than the ocean ridge systems in this regard.

In addition to methanethiol, the pools and bubbles of liquid/scCO2 are expected to provide a high concentration of CO2 to the surrounding seawater and maintain the pH at neutral to slightly acidic. Such a situation is favorable for the electroreduction of CO2 to CO106,107 and the conversion of metal sulfides to the corresponding zero-valent metals93,108,109. The generated CO, as well as the CO, delivered through the scCO2 flux (Fig. 1c), is useful for various abiotic reactions110. Another by-product, OCS (Fig. 1a), is capable of promoting chemical bonding to form peptides, pyrophosphate, and aminoacyl phosphates111,112,113,114 and facilitates the formation of urea from CO and ammonia115. Moreover, methanethiol in scCO2 possibly facilitates the leaching of (rare) metals from subseafloor rocks owing to the strong chelation ability88. The resultant supply of (rare) metals with methanethiol, CO, and OCS must be beneficial to abiotic organic synthesis, and also to the chemolithoautotrophic microbial activities89,90. Furthermore, if amphiphilic molecules were available, the selection and accumulation of primordial ingredients at the scCO2–seawater interface is a conceivable process beneficial to the origin of life27,28,29,30.

Until the publication of a recent study9, liquid and scCO2 were not considered in scenarios of the origin of life in submarine hydrothermal settings. Nonetheless, experimental simulations of prebiotic hydrothermal processes have often been conducted in reaction vessels with the vapor headspace; the adequacies of the experimental conditions have been questioned because the products inevitably derive from the gas as well as aqueous-phase reactions20,116,117. Their reassessment in terms of the reproducibility in subseafloor scCO2 should be a meaningful approach to ascertain their realistic contributions to the origin of life. The Fischer–Tropsch-type hydrocarbon synthesis118,119, for example, would be a promising reaction attainable in scCO2.

Last but not least, the formation of methanethiol from gaseous CO2 observed in our experiment suggests that abiotic formation of methanethiol is feasible even in shallow aquatic environments as long as H2, H2S, and MoS2 are available. This reaction may partially account for the sedimentary organosulfur compounds detected on Mars120,121.

Methods

All chemicals were purchased from FUJIFILM Wako Pure Chemical Corporation as reagent grade except for 13C-labeled sodium formate (H13COONa) and a standard gas of OCS (10 vol% OCS in Helium), which were sourced from the Taiyo Nippon Sanso Corporation. Deaerated Milli-Q water (18.2 MΩ cm) was used as the solvent. A gas cylinder of CO2 (47 L, 6 MPa, >99.995% purity) was obtained from Taiyo Nippon Sanso Corporation. H2 gas (>99.99996%) was generated with NM Plus 100 Hydrogen Generator (LNI Schmidlin SA). To prevent oxidation by atmospheric O2, all aqueous solutions were prepared in an anaerobic chamber filled with nitrogen (N2) and H2 gases (the volumetric ratio = 96:4).

Preparation of MoS2

To simulate the formation of MoS2 in a subseafloor scCO2 environment, MoS2 was prepared by the reduction and sulfidation of molybdate under scCO2. The reaction in the presence of H2 is expressed as follows:

$${{\rm{M}}}{{\rm{o}}}{{{\rm{O}}}}_{4}^{2-}+2{{{\rm{H}}}}_{2}{{\rm{S}}}+{{{\rm{H}}}}_{2}+2{{{\rm{H}}}}^{+}\to {{\rm{Mo}}}{{{\rm{S}}}}_{2}+4{{{\rm{H}}}}_{2}{{\rm{O}}}$$
(10)

We conducted 7-day heating of the reactants at 200 °C, as described below.

A glass bottle containing 8 mL of an aqueous solution of sodium molybdate (400 mM), Na2S (1600 mM), and HCl (1600 mM) was encased in a titanium reaction vessel (Supplementary Fig. 2). After tightly sealing the vessel with a polytetrafluoroethylene (PTFE) O-ring compressed between the vessel and cap, the headspace (29.7 mL) was evacuated using a rotary pump (GCD-136X, ULVAC). This was followed by the addition of 0.3 MPa of H2 and the subsequent introduction of CO2 from a gas cylinder (47 L) filled with 6 MPa of CO2. The mass of introduced CO2 was 5.1 ± 0.5 g. Thereafter, the vessel was heated at 200 °C for 7 days in a laboratory oven (DSN-111, ISUZU).

H2 is expected to be completely miscible with scCO2 under the reaction condition122. According to the solubility data of CO2 compiled by dos Santos et al.123, CO2 solubility increases with pressure without a gap at the critical pressure (73.8 bar). The dissolved CO2 has been characterized with Raman spectroscopy by Truche et al.124. Except for the intensity, the spectral profile of CO2 exhibited no significant dependence on the pressure in the range from 17 to 102 bar at 200 °C, whereas the band position was placed at a clearly different wavenumber from that of gaseous CO26. Thus, it is unlikely that the CO2 dissolved under scCO2 behaves as a gas: the phase of the CO2 source (gas or supercritical) would have no direct impact on the behavior of dissolved CO2.

After the 7-day reaction, the vessel was cooled under air, depressurized, and opened in an anaerobic chamber. Black precipitate of MoS2 was found in a clear, colorless solution. Note that the molybdate-to-MoS2 conversion does not necessarily require H2 because H2S works as the reductant as well as the source of sulfur:

$${{\rm{M}}}{{\rm{o}}}{{{\rm{O}}}}_{4}^{2-}+3{{{\rm{H}}}}_{2}{{\rm{S}}}+2{{{\rm{H}}}}^{+}\to {{\rm{Mo}}}{{{\rm{S}}}}_{2}+4{{{\rm{H}}}}_{2}{{\rm{O}}}+{{{\rm{S}}}}^{0}$$
(11)

Indeed, the 7-day reaction without H2 under otherwise identical conditions resulted in the precipitation of MoS2 in a yellow-colored aqueous solution. The color was indicative of polysulfides, dissolved sulfur chains (Sn2–; n = 2–8) formed as a result of the reaction of H2S with elemental sulfur (S0)114,115. Additionally, the formation of S0 was confirmed by the X-ray diffraction (XRD) measurement of MoS2 formed in the absence of H2 (Supplementary Fig. 14). H2 reduces S0 to H2S under hydrothermal conditions125,126:

$${{{\rm{S}}}}^{0}+{{{\rm{H}}}}_{2}\to {{{\rm{H}}}}_{2}{{\rm{S}}}$$
(12)

The summation of Eqs. (11) and (12) give Eq. (10). Thus, the presence of H2 suppressed S0 formation on MoS2, as confirmed by the XRD measurement (Supplementary Fig. 1).

To completely eliminate S0, the MoS2 prepared in the presence of H2 was separated from the supernatant solution, dispersed in 50 mL of 1 M Na2S in a sealed glass vial (100 mL), sonicated for 1 h, and rinsed with deaerated Milli-Q water on a PTFE membrane filter (pore size = 0.2 μm) with the eluate evacuated under vacuum. The rinsed MoS2 was dried under vacuum, stored inside an anaerobic chamber, and used for the methanethiol synthesis experiment.

Preparation of komatiite

Komatiite was prepared in accordance with the method reported by Shibuya et al.127. An aluminum-depleted composition (Supplementary Table 1) was chosen because it was a representative komatiite composition in the Archean Ocean crust127. Silicon dioxide (SiO2), titanium dioxide (TiO2), aluminum oxide (Al2O3), ferric oxide (Fe2O3), manganese dioxide (MnO2), magnesium oxide (MgO), calcium carbonate (CaCO3), sodium carbonate (Na2CO3), and potassium carbonate (K2CO3) were used as the source of metals. These powders were mixed in a ratio consistent with the composition presented in Supplementary Table 1; placed in a platinum–rhodium (Pt–Rh) crucible, and heated at 1000 °C for 1 h for decarboxylation. Thereafter, the temperature was raised to 1600 °C at the rate of 200 °C h–1, maintained at that level for 1 h, decreased to 1450 °C at the rate of 100 °C h–1, and maintained at that level for 1 h. During the heating, a gas mixture of H2 and CO2 flowed to maintain the O2 fugacity at the quartz–fayalite–magnetite (QFM) buffer condition. The sample was then quenched to room temperature, crushed in a quartz mortar, and sieved to obtain the fraction with a particle size smaller than 90 μm.

Preparation of mackinawite

Mackinawite was prepared by adding 100 mmol L–1 Na2S dropwise into 100 mmol L–1 FeCl2 under vigorous stirring to a final volume ratio of 1:1. The obtained precipitate was separated from the supernatant solution by centrifugation (8000 rpm, 10 min), dried under vacuum, and stored inside an anaerobic chamber.

Solid characterization

XRD patterns of solid samples were measured using an X-ray diffractometer with Cu Kα radiation (MiniFlex II, Rigaku). All runs were conducted with 2θ ranging from 10° to 90° using 0.02° 2θ step with a scan rate of 0.1 min−1. To prevent oxidation by atmospheric O2 during the measurement, the samples were shielded in an air-sensitive sample holder (Rigaku). Peak identifications were made based on the reference patterns stored in the software (PDXL-2, Rigaku). The reference patterns are presented in Supplementary Figs. 1, 9, 10, and 11 with the measured XRD data.

The scanning electron microscopy (SEM) imaging and energy dispersive X-ray spectroscopy (EDS) analysis were performed on a Helios G4 UX (Thermo Fisher Scientific) equipped with an Octane Elite Super (C5) EDS detector (AMETEK) and a cryogenic stage with a preparation chamber (PP3010T, Quorum). The carbon electrode with the solid sample deposited was trimmed to ~3 × 3 mm and mounted on an aluminum stub by double-sided carbon tape in an anaerobic chamber. Thereafter, the stub was carried in a liquid N2-filled plastic bag to the SEM instrument, mounted on a transfer shuttle, and inserted into the SEM stage after vacuum evacuation. To acquire surface-sensitive high-resolution images and elemental maps, an acceleration voltage of 1 kV was applied for morphological imaging, and an acceleration voltage of 5 or 20 kV for EDS analysis under a reduced pressure of <1 × 10−4 Pa at room temperature (23 ± 2 °C). The acceleration voltage of 20 kV was applied to average the spatial elemental distribution of solid samples, as the characteristic X-ray is generated several μm from the point irradiated at this voltage128.

Methanethiol synthesis under scCO2

In a typical experiment, 50 mg of MoS2 was deposited in 3.43 mL of an aqueous solution of Na2S and HCl (400 mmol L–1 for each) in a glass vial. The vial was encased in a titanium reaction vessel (Supplementary Fig. 2) with the headspace (34.3 mL) filled with CO2 and H2 gases (6.0 MPa and 0.3 MPa at 25 °C, respectively), and heated at 200 °C for 7 days (see the section “Preparation of MoS2” for more details). The mass of introduced CO2 was 6.5 ± 0.2 g. At the reaction temperature, the inner pressure of vessel was measured to be around 14.3 MPa (Supplementary Fig. 15), well above the critical pressure of CO2 (7.38 MPa)25.

Thereafter, the vessel was cooled under air. The headspace gas was taken into a gas-tight syringe (50 mL, Hamilton), and injected in a pre-evacuated glass vial (13 mL) for storage. Excess pressure was released through a stainless needle during gas injection. The vessel was then depressurized, and transferred into an anaerobic chamber filled with N2 and H2 gases (the volumetric ratio = 96:4), in which MoS2 was separated from the sample solution, rinsed with deaerated Milli-Q water, dried under vacuum, and stored. To ensure reproducibility, this typical experiment was conducted three times. Differences in the yield of methanethiol among the three independent runs were <15%.

To confirm that CO2 and H2 immediately form a homogeneous mixture under the reaction temperature (200 °C), we also performed a short-term experiment (3 h) and sampled the headspace gas into several glass vials after cooling the vessel under air. The molar ratios of H2/CO2 quantified by the GC-BID analysis (0.030 ± 0.003) suggested a complete mixing of CO2 and H2 in the reaction vessel.

Methanethiol synthesis from 13C-labeled formate

The typical experiment described above was repeated using 13C-labeled formate as a starting material without the addition of H2 gas. Specifically,13C-labeled formate (3.43 mL, 100 mmol L–1) was encased in the titanium reaction vessel with Na2S (3.43 mL, 400 mmol L–1), HCl (3.43 mL, 400 mmol L–1), MoS2 (50 mg), and CO2 (34.3 mL, 6.0 MPa at 25 °C), and heated at 200 °C for 7 days. After cooling the vessel under air, a portion of the headspace gas was sampled into a sealed glass vial, and analyzed by GC–MS (see below) to obtain the mass spectrum of methanethiol produced (Supplementary Fig. 8).

H2 production via hydrothermal alteration of mackinawite and komatiite

The typical experiment described above was repeated with 500 mg of mackinawite or komatiite in place of MoS2 without the addition of H2 gas. Specifically, 500 mg of mackinawite or komatiite was encased in the titanium reaction vessel with Na2S (3.43 mL, 400 mmol L–1), HCl (3.43 mL, 400 mmol L–1), and CO2 (34.3 mL, 6.0 MPa at 25 °C), and heated at 200 °C for 7 days. After cooling the vessel under air, a portion of the headspace gas was sampled into a sealed glass vial. The vessel was then depressurized, and transferred into an anaerobic chamber filled with N2 and H2 gases (the volumetric ratio = 96:4), in which the mineral sample was separated from the sample solution, rinsed with deaerated Milli-Q water, dried under vacuum, and stored. XRD and EDS measurements were performed to characterize the alteration of mineral samples (Supplementary Figs. 10 and 11).

Sample analysis

The gas-phase sulfur compounds were analyzed using a GC (7890B, Agilent) equipped with a quadrupole mass spectrometer (JMSQ1500GC, JEOL). In a typical run, 50 μL of gas sample was injected at a split ratio of 1:10, and flowed into a column (DB-Sulfur SCD column, Agilent) with He (purity > 99.99995%) at a constant rate of 2.8 mL min−1. The column temperature was initially maintained at 35 °C for 3.5 min, and thereafter, raised to 245 °C at a rate of 20 °C min−1. The positive-mode electrospray ionization (ESI) was applied, and the scan range was set to m/z 20–100 Da. Compounds were identified by comparing the observed retention times and mass spectra with those of respective standards prepared from commercial reagents/gases, and were quantified based on the calibration curves (Supplementary Fig. 3). The obtained data were used to calculate the mole fractions of target compounds in the sample gas, which were then multiplied by the amount of gas-phase CO2 plus H2—the predominant gas species—in the reaction vessel to determine the yield of each target (Fig. 1b and d). The amount of CO2 was calculated based on the mass of CO2 in the reaction vessel considering the dissolution of CO2 in the sample solution25,69, whereas that of H2 was quantified via the GC measurement described below.

H2, CO, and CH4 were quantified using the Shimadzu GC system equipped with a barrier discharge ionization detector (BID) (Nexis GC-2030). In a typical run, 50 μL of gas sample was injected at a split ratio of 1:1, and flowed into a column (MICROPACKED-ST, Shinwa) with He (purity > 99.99995%) at a constant rate of 7 mL min−1. The column temperature was initially maintained at 35 °C for 2.5 min, and thereafter, raised to 270 °C at the rate of 20 °C min−1, and maintained for 1.75 min (16 min in total). A chromatogram for a standard gas sample is shown in Supplementary Fig. 5 with obtained calibration curves.

The liquid-phase products were characterized with proton NMR (1H NMR) spectroscopy using the Bruker AVANCE NEO 400 spectrometer (400 MHz) at the sample temperature of 298.0 K. 540 μL of sample solution was mixed with 60 μL of D2O containing 5 mM of 3-(trimethylsilyl)−1-propanesulfonic acid-d6 sodium (DSS-d6), and placed in an NMR tube (5 mm outside diameter; Wilmad-LabGlass). DSS-d6 was used for the calibration of the 0-ppm position and to quantify the product concentrations as an internal standard. The one dimensional 1H nuclear Overhauser effect spectroscopy (1D-NOESY)129 was applied to suppress the water signal.

The concentration of formate was quantified with a high-performance liquid chromatography (HPLC) system equipped with a photodiode array detector (SPD-M40, Shimadzu). An Aminex HPX-87H column (300 × 7.8 mm, BIO-RAD) was used at 30 °C. 5 mM of sulfuric acid was flowed as the eluent at the rate of 0.6 mL min–1. A chromatogram of formate is presented in Supplementary Fig. 7 with the obtained calibration curve.