15.2.1 Bulk Rock Investigations

The first mention of the presence of CM in mafic and ultramafic rocks came from the vast Russian literature published since the 1950s. At that time, the so-called Russian–Ukrainian School had spent significant effort in exploring hydrocarbons in mafic and ultramafic rocks to support a deep mantle origin hypothesis for hydrocarbons deposits. We refer to Sephton and HazenReference Sephton, Hazen, Hazen, Jones and Baross² for a historical perspective and an inventory of achievements in this field. Although a large part of these studies remains untranslated, a review of the occurrences of putatively abiotic condensed naphtides showed a large diversity of compounds associated with volcanic and igneous rocks in a wide range of geodynamic contexts.Reference Zubkov³⁷ These compounds include aromatic and aliphatic hydrocarbons and their N-, O-, and S-bearing derivatives. Notably, they mainly occur in mafic and ultramafic rocks. The first study targeting oceanic rocks has focused on the Rainbow (36°14ʹN) and Logatchev (14°45ʹN) hydrothermal fields (Mid-Atlantic Ridge (MAR)).Reference Pikovskii, Chernova, Alekseeva and Verkhovskaya³⁸ In addition to the presence of low-molecular-weight alkanes and isoalkanes, this study reported the presence of CM in serpentinized rocks, in metalliferous sediments, and in secondary Fe–Ni sulfides. The Fe–Ni sulfides presented the highest concentrations and the broadest diversity of viscous and solid organic compounds depicted as resinous bitumen. Notably, a compositional relationship was established between these resinous bitumen and hydrothermal crystalline hydrocarbons identified as karpatite (C₂₄H₁₂) and idrialite (C₂₂H₁₄) found in the same rock samples and thought to derive from serpentinization reactions.Reference Pikovskii, Chernova, Alekseeva and Verkhovskaya³⁸

Karpatite (Figure 15.1) and idrialite correspond to polycyclic aromatic hydrocarbon (PAH) minerals. PAH minerals are the most typical molecular organic minerals with well-defined chemical compositions and crystallographic properties.Reference Echigo and Kimata³⁹ These molecular crystals can form thanks to the high stability of PAHs in hydrothermal fluids,Reference Stein⁴⁰ allowing transport without thermal degradation and subsequent concentration and crystallization at the hydrothermal discharge zone.Reference Murdoch^41, Reference Echigo, Kimata and Maruoka⁴² Coronene and phenanthrene, the precursors of karpatite and of a number of natural molecular organic minerals (Table 15.1), were shown to derive from the hydrothermal alteration of organic matter present in sediments,Reference Simoneit and Lonsdale⁴³ but can also be produced during basalt cooling (Section 15.3.4).Reference Zolotov and Shock^44, Reference Zolotov and Shock⁴⁵ Although the mechanisms of PAH concentration prior to crystallization is still unknown, occurrences of molecular organic minerals may attest to localized high concentrations in (poly)aromatic compounds in the oceanic lithosphere.

Figure 15.1 (a) A vein of karpatite (yellow crystals) surrounded by quartz (white crystals) and cinnabar (red spots). Scanning electron micrographs of (b) the broken surface of native karpatite and (c) its layering at the end of the layered structure.

Reproduced with permission of Springer Nature, from Potticary et al. (2017), Sci Rep, 7, 9867, figures 1b, 2a,b.Reference Potticary, Jensen and Hall⁴⁶

More recently, three studies characterized by bulk molecular approaches (1) serpentinized peridotites and gabbroic rocks recovered at the Atlantis Massif (30°N MAR) at both the Lost City hydrothermal field and Hole U1309D (Expeditions 304/305, Integrated Ocean Drilling Program (IODP)),Reference Delacour, Früh-Green, Bernasconi, Schaeffer and Kelley²⁶ (2) peridotites from the Ashadze (12°58ʹN) and Logatchev hydrothermal sites along the MAR,Reference Bassez, Takano and Ohkouchi⁵³ and (3) the fossil ocean–continent transition recorded in the Swiss Alps.Reference Mateeva⁵⁴ All of these studies used gas chromatography and/or gas chromatography–mass spectrometry. Based on the presence of biomarkers (e.g. pristane, phytane, squalane, hopanes, and steranes) or higher relative abundances of n-C₁₆ to n-C₂₀ alkanes,Reference Delacour, Früh-Green, Bernasconi, Schaeffer and Kelley^26, Reference Mateeva⁵⁴ amino acids, and long-chain n-alkanesReference Bassez, Takano and Ohkouchi⁵³ identified in the solvent-extracted fraction, they all concluded that the organic carbon was of biological origin, as supported by TOC isotopic analysis of the oceanic rocks.Reference Alt^25, Reference Schwarzenbach, Früh-Green, Bernasconi, Alt and Plas²⁸ However, biological contamination may have overprinted any possible abiotic geochemical signatures,Reference Delacour, Früh-Green, Bernasconi, Schaeffer and Kelley²⁶ since these oceanic basement rocks have been affected by long-lived hydrothermal alteration and present-day microbial ecosystems.Reference Klein^29–Reference Pasini³¹

15.2.2 In Situ Investigations at the Microscale

Development of in situ techniques currently allows imaging of rock-hosted CM along with their relationship with mineral parageneses. The scarce data for oceanic rocks report that the carbon phases exhibit a structural organization varying from amorphous to well-organized graphite-like material (Figure 15.2a and b). Organic compounds have also been detected as thin films or gel-like materials embedding minerals and filling microfractures (Figure 15.2c). As is detailed below, some have an origin that is still debated, depending on the geological setting and textural criteria.

Figure 15.2 Examples of occurrences of organic carbon in serpentinized oceanic rocks. (a) Scanning electron micrograph of O-bearing condensed carbonaceous matter (CCM) abiotically formed jointly with hematite (Hem) and saponite (Sap) during the low-T alteration (T < 150°C) of oceanic serpentinites of the Ligurian Tethyan ophiolites. (b) Associated elemental distributions of carbon (red) and iron (green) within the square area in (a). Reproduced with permission of Springer Nature, from Sforna et al. (2018), Nat Commun, 9, 5049, figure 2c and d.Reference Sforna⁵⁵ (c) Transmission electron micrograph showing polyhedral serpentine (pol-spt) sections wet by a jelly film of organic carbon interfacing between the pol-spt and an andraditic hydrogarnet (H-adr) in serpentinites from the MAR (4-6°N). In these rocks, organic carbon was shown to mediate the nucleation and growth of polyhedral and polygonal serpentine from the hydrogarnet.

Reproduced with permission of Elsevier, from Ménez et al. (2018), Lithos, 323, 262–276.Reference Ménez⁵⁶

The Hyblean basaltic diatreme (Sicily, southern Italy), recognized as a paleo-oceanic serpentinite-hosted hydrothermal system comparable to those found along slow-spreading ridge segments,Reference Scribano, Sapienza, Braga and Morten⁵⁷ is one of the most studied places where the presence of large amounts of putatively abiotic carbonaceous compounds has been related to hydrothermal activity. A series of papers reported the occurrence of heavy hydrocarbons in deep-seated xenoliths including metasomatic gabbroic xenolithsReference Ciliberto⁴⁷ and highly serpentinized peridotite xenoliths.Reference Scirè⁵⁸ In the gabbroic xenoliths, the presence of saturated aliphatic and aliphatic–aromatic hydrocarbons was highlighted by bulk rock analysis using electron impact-direct pyrolysis mass spectra. These observations were supported by Fourier-transform infrared (FTIR) spectroscopy showing carbon-cored clayey vesicles and dull-black, tiny pellets associated with the clayey mat pervasively intruding into fractures of the host rocks.Reference Ciliberto⁴⁷ Organic crystalline phases were also detected using X-ray diffraction (Table 15.1), and thermal decrepitation allowed the analysis of hydrocarbons trapped in fluid inclusions. Similarly, in the extensively serpentinized and carbonated ultramafic xenoliths, microscopic accumulations of sulfur-bearing organic matter commonly occur between hydrothermal minerals (i.e. secondary calcites and fibrous phyllosilicates), forming occasionally coarse bituminous patches.Reference Scirè⁵⁸ The organic matter was estimated to represent 3–6% of the whole rock. Micro-FTIR spectroscopy highlighted the presence of condensed aromatic rings with aliphatic tails consisting of a few C atoms, which is suggestive of asphaltene-like structures. Such structures were confirmed by solubility tests in toluene and n-hexane, thermo-gravimetric analyses and differential thermal analysis. X-ray photoelectron spectroscopy (XPS) also indicated minor S and O functional groups.

In these studies, it has been proposed that asphaltenes and high-molecular-weight hydrocarbons, respectively, derived from in situ aromatization or progressive polymerization and polycondensation reactions of light aliphatic hydrocarbons formed by FTT reactions.Reference Ciliberto^47, Reference Scirè⁵⁸ The hydrocarbon-bearing xenoliths possibly represent any level of the hydrothermal system, including its deepest and hottest parts at ~400°C.Reference Ciliberto⁴⁷ Nevertheless, both biotic and abiotic origins could have been possible. It is challenging to infer the possible processes and formation conditions giving rise to heavy hydrocarbon accumulations in such a complex geologic setting. Mantle xenoliths experienced transport and decompression during the formation of the diatreme, with possible consequences for CM that are unrelated to hydrothermal circulation.

Indeed, CM commonly occurs in basalts and mantle xenoliths, which were thoroughly investigated during the 1980s and 1990s.Reference Mathez^59–Reference Tingle, Mathez and Michael⁶² At that time, the aim was to assess the mantle carbon content by using a suite of techniques including electron microscopy, chemical imaging, XPS, and carbon isotopes,Reference Mathez^59, Reference Mathez, Dietrich and Irving^61–Reference Muenow⁶³ as well as thermal desorption surface analysis by laser ionization and low-energy electron diffraction.Reference Tingle, Mathez and Michael⁶² In these rocks, CM was observed as discrete platy lumps of up to 20–200 µm in size or as thin, amorphous films of a few nanometers located on quench-produced crack surfaces, grain boundaries, and the walls of fluid inclusions. They consisted dominantly of graphite-intercalated compounds along with ill-defined complex mixtures of graphite-like compounds and organic material composed of C, H, and possibly N. All of these studies concluded that CM had an abiotic origin based on: (1) its preferential concentration on sulfide spherules attached to vesicle walls,Reference Mathez and Delaney⁶⁰ where the sulfides may have played a catalytic role in organic compound formation and concentration as also observed for hydrothermal sulfides;Reference Pikovskii, Chernova, Alekseeva and Verkhovskaya³⁸ or (2) the close association of carbon with Si, Al, alkalis, halogens, and/or transition metals, which are elements that were likely present in the volcanic gas at the origin of these carbon accumulations.Reference Tingle, Mathez and Michael⁶² Again, to account for the production of CM in basalts and mantle xenoliths, these studies invoked: (1) FTT reactions involving volcanic gases degassed of host lava and reacting with fresh and chemically active crack surfaces formed by thermal stresses during eruption, decompression, and cooling; and (2) subsequent evolution of the condensate during cooling. Although heterogeneous catalysis at the mineral surface was the favored hypothesis, organics may have been alternatively assimilated into the volcanic gases prior to eruption and deposited on cracks formed during eruption and cooling.Reference Tingle, Mathez and Michael⁶²

The possible role of mineral surfaces in the abiotic formation of CM in the oceanic lithosphere was pointed out in a series of recent studies. Various associations between minerals and condensed CM were documented using scanning electron microscopy (SEM) and Raman spectroscopy within magma-impregnated, mantle-derived serpentinites of the Ligurian Tethyan ophiolites, in a context of common occurrences in the lower oceanic crust (Figure 15.2a and b).Reference Sforna⁵⁵ Three distinct types of CM in paragenetic equilibrium with low-T mineralogical assemblages have been sequentially formed at decreasing T during the hydrothermal alteration of the rock assemblage. The first type corresponds to thin films of aliphatic chains, coating hydroandraditic garnets in bastitized pyroxenes. The second type forms micrometric aggregates associated with the alteration rims of spinel and plagioclase. The third and most massive type appears as large aggregates (up to 200 μm in size) bearing highly aromatic carbon and short aliphatic chains associated with hematite and Fe-saponite assemblages replacing the pseudomorphoses after plagioclase (Figure 15.2a and b). The systematic association of a given type of CM with a specific mineral paragenesis indicates that condensed CM precipitated simultaneously to the growth of the host mineralogical assemblage, overall in favor of an abiotic endogenesis. The mineral formation would have been accompanied by the production of H₂ able to reduce carbon species, as is the case when ferric hydrogarnets and ferric serpentines formReference Andreani, Munoz, Marcaillou and Delacour^64–Reference Plümper, Beinlich, Bach, Janots and Austrheim⁶⁶ or when spinel oxidizes to Cr-magnetite forming ferritchromite rims.Reference Mayhew, Ellison, McCollom, Trainor and Templeton⁶⁷ This possible abiotic pathway was not considered in former studies performed on hydrogarnet-hosted carbonaceous compounds that suggested, based on their biological Raman and FTIR spectroscopy signatures, that this disordered CM may have resulted from the hydrothermal alteration of cryptoendolithic microbial ecosystems.Reference Ménez, Pasini and Brunelli^30, Reference Pasini³¹ Overall, while the involved mechanisms and reactants are still unknown, this recent study emphasizes a key role of local parageneses forming unique microenvironments prone to the synthesis of abiotic hydrocarbon.Reference Sforna⁵⁵ The productivity could have been also controlled by the local presence of catalysts such as Cr³⁺ in hydrogarnets or Fe-saponite in addition to specific redox conditions among the large redox potential gradients existing in serpentinizing systems down to the microscale (Figure 15.3).Reference Andreani, Munoz, Marcaillou and Delacour⁶⁴

Figure 15.3 (a) Evolution of H₂ concentration in hydrothermal fluid as a function of serpentinization degree estimated from mass balance calculation on Fe²⁺ and Fe³⁺ in abyssal peridotite. (b) Differences in H₂ concentrations, generated by the spatial heterogeneity of reaction, create a redox gradient between partly and fully serpentinized areas down to the micrometric scale.

Modified with permission of Elsevier, from Andreani et al. (2013), Lithos, 178, 70–83, figure 9b.Reference Andreani, Munoz, Marcaillou and Delacour⁶⁴

The catalytic role of phyllosilicates in abiotic organic synthesis has been suggested by the syngenetic link between phyllosilicates and organic compounds in several hydrothermal systems. The abundance of long-chain aliphatic and aromatic hydrocarbons (identified using FTIR spectroscopy) has been reported in a diapir of saponite-dominated clays intruded in a diatremic tuff–breccia deposit.Reference Manuella, Carbone and Barreca⁶⁸ The clays formed during the high-T (350–400°C) hydrothermal alteration of mafic and ultramafic lithologies in the Hyblean crustal basement. Based on the close association of hydrocarbons with hydrothermal clays, the lack of fossils, and the resemblance with the previously described xenolith-hosted organic compounds,Reference Ciliberto^47, Reference Scirè⁵⁸ a possible abiogenic origin was suggested for the hydrocarbons via FTT reactions mediated by the organoclay. A close association of organic compounds with clays minerals such as saponite has been previously described in interplanetary dustReference Pizzarello, Cooper, Flynn, Lauretta and McSween⁶⁹ and in carbonaceous meteorites that have undergone significant aqueous alteration processes.Reference Le Guillou and Brearley^70–Reference Zega⁷² A similar association was recently highlighted at a micrometric scale using FTIR spectroscopy in oceanic serpentinites collected at ~170 m depth below seafloor during the IODP Expeditions 304/305 targeting the Atlantic Massif.Reference Pisapia, Jamme, Duponchel and Ménez⁷³

Overall, all of these studies highlight that hydrothermally derived organic carbon trapped within the upper lithosphere as heavy and aromatic compounds can be chemically and structurally diverse, although formation mechanisms are not yet well-known. Whether aromatization occurs in situ from CO₂/CO and H₂ during serpentinization or derives from progressive polymerization and polycondensation reactions of light aliphatic ± aromatic hydrocarbons inherited from higher‑T FTT reactions still needs to be addressed. This latter scenario implies that reactants for the low-T (<400°C) hydrothermal synthesis of abiotic organic compounds that may occur during serpentinization might be more diverse than H₂ and CO₂/CO and account for aliphatic and aromatic compounds, as supported by the presence of molecular organic minerals in serpentinized rocks (Sections 15.2.1 and 15.2.2).Reference Pikovskii, Chernova, Alekseeva and Verkhovskaya^38, Reference Ciliberto⁴⁷

15.2.3 Carbon in Fluid Inclusions Trapped in the Oceanic Lithosphere

Fluid inclusions are used to document the composition of high-T fluids (either of magmatic or seawater origin) circulating in the lower crustal component of hydrothermal systems. Hence, they can inform on the intermediate abiotic processes that may happen at higher T and then feed with diverse reactants the low-T hydrothermal reactions, including organic synthesis. Fluid inclusions are made of liquid and/or gas with or without tiny daughter minerals that are trapped within a crystal structure during either primary crystallization or secondary healing of fluid-filled cracks. They occur throughout the gabbroic and peridotitic plutonic crustReference Kelley and Früh-Green⁷⁴ and have been vastly studied by Raman spectroscopy and microthermometry. In ophiolites, fluid inclusions formed in sub-seafloor rocks may be preserved during tectonic uplift and maintain their original chemical signatures. This has been shown for pure methane fluid inclusions occurring in olivine from partially serpentinized harzburgites and dunites from the Nidar ophiolite complex (eastern Ladakh, India)Reference Sachan, Mukherjee and Bodnar⁷⁵. In oceanic rocks, all studies have pointed to the presence of CH₄ ± H₂ ± CO₂ within fluid inclusions,Reference Kelley^9, Reference Kelley and Früh-Green^10, Reference Kelley and Früh-Green^74, Reference Vanko, Stakes, Herzen, Fox, Palmer and Robinson^76, Reference Kelley, Karson, Cannat, Miller and Elthon⁷⁷ although propane and ethane were also sometimes reported in the gaseous phases.Reference Bortnikov and Piestrzyhski⁷⁸ Analyses of fluid inclusions in primitive olivine gabbros, oxide gabbros, and evolved granitic rocks recovered from the slow-spreading Southwest Indian Ridge at Ocean Drilling Program Hole 735B (Atlantis II fracture zone) recorded CH₄ concentrations 15–40 times those of hydrothermal vent fluids from sediment-poor environments and of basalt-derived volcanic gases.Reference Kelley and Früh-Green^10, Reference Kelley and Früh-Green⁷⁴ These studies also frequently reported the presence of disordered graphite, carbonaceous compounds, and putative graphite coating inclusion walls.Reference Kelley^9, Reference Kelley and Früh-Green^74, Reference Vanko, Stakes, Herzen, Fox, Palmer and Robinson⁷⁶

While several processes were proposed to explain the presence of organic compounds in high-T hydrothermal fluids,Reference Elthon^79, Reference Nakamura⁸⁰ fluid inclusions are generally believed to represent evolved magmatic fluids dominated by CO₂ or CO.Reference Kelley^9, Reference Kelley and Früh-Green^10, Reference Kelley and Früh-Green⁷⁴ Indeed, subsequent to devolatilization and re-equilibration, graphite precipitation is promoted by re-speciation of magmatic CO₂, and the formation of CH₄-enriched fluids is promoted by cooling from 800°C to 500°C of these H₂-rich fluids. Alternatively, re-speciation and reduction of entrapped CO₂-bearing magmatic fluids by diffusion of external H₂ into the inclusions during cooling was also considered.Reference Kelley^9, Reference Kelley and Früh-Green⁷⁴ Once trapped within crystals as fluid inclusions, the fluids can still evolve following reactions between the fluid and the host mineral. Direct evidence for in situ H₂ production and strongly reducing conditions within serpentinizing olivine fluid inclusions was provided for the Mineoka ophiolite complex considered as an analog to serpentinite-hosted hydrothermal vent systems.Reference Katayama, Kurosaki and Hirauchi⁸¹ In the latter study, mineral-phase equilibria indicated that CH₄–H₂-bearing fluids were trapped under equilibrium conditions at T below 300°C, and the absence of CO₂ was suggestive of extensive reduction of CO₂ to CH₄ within the inclusions.Reference Katayama, Kurosaki and Hirauchi⁸¹ Whether CH₄ was also produced at higher T remains an open question. The presence of H₂ and CH₄ together with secondary mineral microinclusions was similarly reported in olivine and orthopyroxene crystals from a harzburgite from the northern Oman ophiolite considered to be similar to abyssal peridotite.Reference Miura, Arai and Mizukami⁸² In olivine, the mineral inclusions mainly consisted of lizardite and brucite with small amounts of magnetite, while in orthopyroxene they were made of talc and chromian spinel. The differential in situ production of reduced gas and secondary phases was related to the presence of either magnetite or a magnetite component in chromium spinels.Reference Miura, Arai and Mizukami⁸² Alternatively, both H₂ and CH₄ inherited from high-T magmatic processes could have hydrothermally evolved differentially depending on the reactivity of their mineral host (e.g. its capacity to produce in situ H₂ or to form mineral byproducts able to catalyze organic synthesis, as discussed in Section 15.2.2).

The presence of reduced carbon species in fluid inclusions within oceanic gabbros and mantle peridotites could represent a potentially important source of organic compounds in hydrothermal fluids.Reference Wang, Reeves, McDermott, Seewald and Ono^7, Reference McDermott, Seewald, German and Sylva^8, Reference Kelley and Früh-Green¹⁰ In particular, CH₄-rich aqueous fluids trapped in the oceanic lithosphere within fluid inclusions recently redrew attention as they could constitute the main source of methane venting at unsedimented mid-ocean ridge hydrothermal fields.Reference Kelley^9, Reference Kelley and Früh-Green^10, Reference Kelley and Früh-Green^74, Reference Kelley, Karson, Cannat, Miller and Elthon^77, Reference Kelley, Gillis and Thompson⁸³ Similar processes were invoked to explain elevated concentrations of both CH₄ and H₂ in fluids from the Menez Gwen (37°50ʹN, MAR), Lucky Strike (37°17ʹN, MAR), and Piccard (Mid-Cayman Ridge) vent fields usually referred to as basalt-hosted vents supposedly less rich in reduced gases compared to ultramafic rock-hosted vents,Reference Konn, Charlou, Holm and Mousis²⁰ despite the possible production of non-negligible amounts of H₂ by diking-eruptive events.Reference Holloway and O’Day⁸⁴

Overall, in agreement with Section 15.2.2, studies on fluid inclusions emphasize the need to consider a larger range of reactants in addition to CO₂/CO and H₂ for low-T hydrothermal synthesis of abiotic organic compounds. This includes methane and light hydrocarbons, but also potentially reactive graphitic phases. Due to their tiny size and their entrapment in minerals, the latter are extremely challenging to analyze, and most of the previous studies seldom went further in characterizing the chemical diversity and degree of structural order of this CM. The precipitation of highly crystalline hydrothermal graphite from aqueous fluids containing CO₂ and CH₄ was reported at T as low as 500°C during the propylitic hydrothermal alteration of volcanic host rocks.Reference Luque⁸⁵ Other studies, targeting different geodynamical contexts, have shown that precipitation of CM in fluid inclusions, notably on their walls, can lead to graphitic material with varying degrees of crystallinity and disorder.Reference Pasteris^86–Reference Wopenka and Pasteris⁸⁹ Disordering in graphite is principally caused by in-plane defects and/or heteroatoms (e.g. O, N, S).Reference Beny-Bassez and Rouzaud⁹⁰ While crystalline graphite was shown to be highly refractory and chemically inert, the chemical reactivity of graphitic carbon increases with structural disorder and the abundance of heteroatoms and unsaturations.Reference Beyssac and Rumble⁹¹ Hence, this possibly impacts the diversification of abiotic organic compounds observed in the altering oceanic lithosphere.

15.3.1 Experimental Approach

Most of the experimental work dedicated to hydrothermal organic synthesis has focused on fluids, especially on the production of volatile organic compounds, using either monophasic (pure liquid) or biphasic (liquid + gas) systems. Solid mineral phases have not been systematically introduced in the experimental devices, and potential carbonaceous phases were rarely characterized after experiments. When present, minerals were either used as reactants, such as olivine to produce H₂, or as redox buffers such as the hematite–magnetite (HM), hematite–magnetite–pyrite (HMP), pyrite–pyrrhotite–magnetite (PPM), or quartz–fayalite–magnetite (QFM) assemblages. Minerals have also been introduced as potential catalysts of organic reactions.Reference Shipp, Gould, Shock, Williams and Hartnett^92–Reference Yang, Gould, Williams, Hartnett and Shock⁹⁴ Unfortunately, they were never characterized after experiments, precluding the finding of CM occurrences.

Hence, the role of minerals on the formation of carbonaceous compounds under hydrothermal conditions, as emphasized from natural observations, has remained largely unexplored. This is partly explained by the analytical difficulty of detecting such a small organic fraction within the solid products (concentration close to or below detection limits) and of locating and characterizing it with high-resolution methods. In addition, contamination issues (from the experimental setup or deriving from organic compounds preexisting in the minerals used; e.g. trapped in fluid inclusions) remain central when looking for low levels of organic products whose nature is unknown and difficult to address.

The role of the experimental container (vessel or capsules) on CO₂ reduction or H₂ production reactions under hydrothermal conditions is potentially non-negligible while not clearly established. Stainless steel reactors have been shown to accelerate FTT reactionsReference Voglesonger, Holloway, Dunn, Dalla-Betta and O’Day^95, Reference McCollom and Seewald⁹⁶ compared to quartz, glass, or Au and TiO₂ reactors. However, stainless steel reactor walls may be passivated after several hours of experimental runs.Reference Voglesonger, Holloway, Dunn, Dalla-Betta and O’Day⁹⁵ That is also the case for titanium, whose oxidation can be a source of H₂ if a preoxidation is not done prior to experiments. Gold, in the form of nanorods or particles, is well known to catalyze aqueous CO₂ reduction to CO during electrochemical experiments,Reference Chen, Li and Kanan^97, Reference Liu⁹⁸ but it is not usually considered for hydrothermal experiments, probably because CO was either not measured or not abundant in products, and also because the gold liner is seen as a smooth surface. Hence, vessels or capsules made of gold or oxidized titanium are usually thought to be the most inert materials and are often preferred at high pressure (P) and T to stainless steel, platinum, alumina, or hastelloy (Ni–Fe-rich alloy). The effect of hastelloy may depend on the reaction of interest. It was infrequently used for FTT reactions, probably because of the potential catalytic properties of the Ni-rich metal alloy, but it was used for H₂ production.Reference Marcaillou, Munoz, Vidal, Parra and Harfouche⁹⁹ Its catalytic effect has not yet been demonstrated for such reactions. The use of Teflon™ especially may be a direct source of carbon contamination with increasing T. Concerning reactor permeability to gas, H₂ has a high diffusivity in most metals, but H₂ loss is expected to be relatively low at T < 400°C.Reference Chou¹⁰⁰

Hence, there is no homogeneity in the type of reactor used in the literature, which often precludes comparison between results, especially when the mineral effect has to be unraveled. In addition, the possible precipitation of a carbonaceous phase on the reactor wall has never been investigated.

15.3.2 Carbon-Bearing Reactants in Experiments

Oceanic hydrothermal systems developing near ridge axes have two main sources of inorganic carbon: (1) mantle-derived carbon delivered by magmatic activity in which CO₂ dominates in the C–O–H system, with very minor CO and CH₄;Reference McCollom, Hazen, Jones and Baross¹ and (2) dissolved inorganic species that result from the equilibration of atmospheric CO₂ and seawater. Hence, aqueous CO₂ (CO_2(aq)) and bicarbonate and carbonate ions (ΣCO₂) are usually the most abundant species in solutions and are the preferential sources of carbon used in experiments and models designed to test abiotic organic synthesis under hydrothermal conditions mimicking natural systems.

Nevertheless, ΣCΟ₂ are not the only single-carbon compounds possibly available at equilibrium in hydrothermal fluids, especially if H₂ is available. Experiments have shown that the speciation of aqueous single-carbon compounds in the C–O–H system for T = 150–300°C and P = 35 MPa is controlled by reactions between ΣCO₂, CO, ΣHCOOH (HCOOH, formic acid + HCOO^–, formate), CH₂O (formaldehyde), CH₃OH (methanol), and CH₄ (Figure 15.4).Reference Seewald, Zolotov and McCollom¹⁰¹ Indeed, the water–gas shift reaction (15.1) under aqueous hydrothermal conditions leads to the formation of ΣHCOOH as an intermediate product (Figure 15.4):

Figure 15.4 Oxidation states of carbon in some single-carbon organic compounds. The water–gas shift reaction (15.1) is represented along with the successive reversible redox reactions that control the speciation of single-carbon compounds under hydrothermal conditions.Reference Zolotov and Shock^44, Reference Seewald, Zolotov and McCollom¹⁰¹

ΣHCOOH reaches a redox-dependent equilibrium with methanol, formaldehyde, and CH₄ within few days at T > 150°C, but may need years at T < 100°C.Reference Seewald, Zolotov and McCollom¹⁰¹ Formaldehyde and CH₄ were close to below detection limits in the experiments, suggesting kinetic inhibition at least for the P–T range tested.Reference Seewald, Zolotov and McCollom¹⁰¹ In the absence of CH₄, the relative concentrations of single-carbon compounds in fluids strongly depended on T, H₂ fugacity (fH₂), and pH. Under neutral and acidic conditions, CO₂ largely dominated, with minor amounts of CO and ΣHCOOH (several orders of magnitude < CO₂) at 350°C. At 150°C and similar pH, methanol is predicted to be only one order of magnitude below CO₂ with minor CO and ΣHCOOH, but it could exceed CO₂ if the H₂ concentration increased to values similar to those measured at ultramafic rock-hosted hydrothermal vents (i.e. ~10 mMReference Fouquet, Rona, Devey, Dyment and Murton¹⁰²). Under alkaline conditions, the formate concentration can be equivalent to bicarbonate with decreasing T or increasing H₂ concentrations.

Direct formation of methanol (~1 mM max) from a H₂–CO₂-rich vapor phase can also occur, with a limited conversion rate (10^–4%) at 300–350°C (18 MPa), provided magnetite surfaces are available.Reference Voglesonger, Holloway, Dunn, Dalla-Betta and O’Day⁹⁵ This could be realistic at mid-ocean ridges (e.g. following dike emplacement with volatile exsolution and migration in adjacent oxide gabbro or in serpentinized peridotites where magnetite is abundant). Magnetite shows decreasing surface reactivity with time and can be regenerated with increasing T, suggesting that it may serve as a catalyst during successive diking events, for instance. The authors do not preclude possible intermediates such as CO and ΣHCOOH in the redox reaction, but these compounds were not detected or measured, respectively, and no graphitic phase or alkanes were observed.Reference Voglesonger, Holloway, Dunn, Dalla-Betta and O’Day⁹⁵

These works show that a large variety of single-carbon species can be available in fluids in addition to ΣCO₂ (mainly methanol and formate), especially in natural serpentinizing systems with highly variable pH and local H₂ levels.Reference Andreani, Munoz, Marcaillou and Delacour⁶⁴ This is particularly true in low-T serpentinizing environments that may be dominated by methanol or formate depending on pH, provided equilibrium is reached. This diversity reinforces the conclusions of Section 15.2 about the need to consider a wider range of carbon-bearing reactants for organic synthesis in natural systems and consequently in experiments. Natural systems also pointed to the possible availability of PAHs, CM, and deep magmatic CH₄ for low-T hydrothermal reactions. Except for formic acid, whose decomposition in aqueous fluid (Figure 15.4) is frequently exploited to experimentally produce H₂ and CO₂, these compounds have rarely (e.g. methanolReference Williams, Canfield, Voglesonger and Holloway¹⁰³) or never been used as reactants in geologically relevant experiments.

CH₄, which is abundant in hydrothermal fluids or cold seepages associated with the serpentinization of mantle rocks,Reference Etiope and Sherwood Lollar^3, Reference Fouquet, Rona, Devey, Dyment and Murton¹⁰² was usually considered as the reduction product of inorganic carbon sources rather than as a possible reactant. An important experimental effort was deployed to reproduce methane synthesis by FTT reactions under moderate to low-T conditions (≤500°C), but systematically failed to produce abundant CH₄. The natural H₂/CH₄ ratio measured in serpentinization-related oceanic hydrothermal environments (<30)Reference Fouquet, Rona, Devey, Dyment and Murton¹⁰² remains much lower than experimental ones (>500; except values at 42Reference Horita and Berndt¹⁰⁴ and 17;Reference Ji, Zhou and Yang¹⁰⁵ e.g. see McCollomReference McCollom¹⁰⁶ for a review). This difference was proposed as an indication of the abiotic versus biotic origin of CH₄ in natural systems, where intense H₂ consumption and associated CH₄ production can be attributed to biological activity.Reference Oze, Jones, Goldsmith and Rosenbauer¹⁰⁷ This simplistic criterion was refuted because it cannot account for the complex processes occurring in the environment,Reference Lang¹⁰⁸ and it also disregarded some other parameters. The H₂/CH₄ ratio may vary with T, decreasing as T increases from 200°C to 500°C at 300 MPa.Reference Huang¹⁰⁹ The effect of pressure has been seldom investigated, but a few studies have shown that increased pressure can account for a significant increase in the CH₄ yield between 100 and 350 MPa.Reference Lazar, Cody and Davis¹¹⁰ CH₄ formation can also compete with CO₂ carbonation depending on fH₂, CO₂ partial pressure, and T. At 200°C, CH₄ formation by CO₂ reduction is limited by H₂ production during olivine alteration if the system is supersaturated with respect to carbonate because the kinetics of carbonate precipitation are faster. Finally, the micromolar levels of CH₄ often found in experiments might not even be produced by in situ reactions, but instead may represent contamination. This is often difficult to assess since blank experiments are rarely provided and ¹³C-labeled experiments are scarce. Contamination issues are critical in very-low-T experiments (<100°C) where product levels are much lower and display contrasting values of H₂ and CH₄ despite similar protocols.Reference Miller^13, Reference Mayhew, Ellison, McCollom, Trainor and Templeton^67, Reference Hellevang, Huang and Thorseth^111–Reference McCollom and Donaldson¹¹⁵ The most complete investigation conducted so far at T ≤ 100°C led to the formation of low-molecular-weight organic acids (mainly formate and acetate); no CH₄ was observed during H₂ production by serpentinization.Reference Miller¹³ This suggests that at such low-T and moderately alkaline conditions (pH ~8), the inorganic carbon (CO₂, CO, or bicarbonate) tends to equilibrate with organic acids instead of CH₄ in a metastable assemblage as previously described for higher T (150–350°C).Reference McCollom and Seewald^14, Reference McCollom and Seewald^15, Reference Seewald, Zolotov and McCollom¹⁰¹

Hence, the high H₂/CH₄ ratio observed in most experiments (e.g. run at ~300°C, 30–50 MPa) highlights the kinetic inhibition of CH₄ formation from the reduction of dissolved CO₂ species or formate under aqueous hydrothermal conditions.Reference McCollom and Seewald^14, Reference McCollom and Seewald¹⁵ For this reason, the role of realistic catalysts (magnetite, hematite, chromite, Fe–Ni alloys, or sulfides) has been tested to circumvent this limitation. Most of them allowed very limited conversion of inorganic carbon to CH₄ (<1% max) and even less for longer-chain alkanes (≪0.1% max) under laboratory timescales (several months), with a final CO₂/CH₄ ratio ranging from 100Reference Berndt, Allen and Seyfried¹¹⁶ to 1000.Reference McCollom and Seewald^14, Reference McCollom and Seewald^15, Reference Foustoukos and Seyfried^117–Reference Lazar, McCollom and Manning¹¹⁹ Lower values of CO₂/CH₄, as small as ~0.1–1.0 and 10, were obtained in studies using Fe–Ni alloys (porous awaruite)Reference Horita and Berndt¹⁰⁴ or Ni-sulfidesReference Lazar, McCollom and Manning¹¹⁹ and Co-bearing magnetite,Reference Ji, Zhou and Yang¹⁰⁵ respectively, pointing to some very specific catalysts possibly available in serpentinized peridotites. Conversely to CO₂, hematite, magnetite, and Fe–Ni alloys did not affect the stability of formate and formic acid, at least for T of 170–260°C.Reference McCollom and Seewald¹⁵

The presence of a gaseous phase in the system, which creates conditions closer to the industrial Fisher–Tropsch process, also seems to favor a higher yield of CH₄ and the formation of more complex hydrocarbons,Reference McCollom and Seewald^15, Reference McCollom and Seewald^96, Reference McCollom¹⁰⁶ even in the absence of a catalystReference Lazar, Cody and Davis¹¹⁰ (if olivine and gold reactor walls are excluded). At very-low-T conditions,Reference Etiope and Ionescu¹²⁰ Ru-bearing chromitite exhibits efficient catalytic properties for CH₄ production under gaseous conditions. The slow kinetics of the first step (CO₂ to CO reduction) in a gas phase is already well known in the industry, and chemists have strived to improve catalyst efficiency, notably by increasing the local concentration of CO₂ on the catalyst surface by altering the catalyst’s nature, shape, and size.Reference Liu^98, Reference Lu¹²¹

Overall, these experiments seldom report on the potential presence of heavier organic compounds or on the organic content of the liquidReference Rushdi and Simoneit¹²² and solid phases, if any are present (see Section 15.3.1), leaving the possibility for other type of products and reaction mechanisms to account for low CH₄ contents. Mineral reactants also need to be systematically introduced in future works for a more realistic approach that fits better with the conditions of natural systems (Section 15.2).

15.3.3 Experimental Occurrences of Carbonaceous Material

Very few experiments have investigated CM phases and their relationship with minerals using bulk or in situ methods such as SEM or transmission electron miscroscopy (TEM) with or without XPS, Raman, and FTIR spectroscopy. As for the characterization of natural occurrences, these methods are seldom used together, limiting comparisons, and the exact nature of CM is usually unidentified.

As described in Section 15.2.2, in the oceanic lithosphere, CM can be first deposited at high T (>400°C) on mineral surfaces and cracks during magmatic degassing at depth and subsequent migration and cooling of mantle-derived fluids. Experiments designed to reproduce this process simulated the sudden cooling of a C–O–H magmatic gas over freshly cracked olivines.Reference Tingle and Hochella¹²³ Evidence for direct carbon precipitation on newly formed cracks has been reported despite thermodynamically unfavorable experimental conditions for graphite precipitation.Reference Tingle and Hochella¹²³ In this study, it was proposed that freshly cracked surfaces of olivine, and possibly of other silicates, offer chemically active areas facilitating the heterogeneous nucleation of CM during interaction with C–O–H gases, at least for initial T between 400°C and 800°C. The deposited carbonaceous film displayed a more complex structure than graphite. Bonding was dominated by C–C and C–H species under the more oxidizing conditions tested, whereas the more reducing ones showed similar amounts of C–C, C–H, C–O, and metal–C species such as carbides (SiC or MgC).Reference Tingle and Hochella¹²³ Carbides, mainly observed at 400°C, could have acted as reaction intermediates for FTT reactions, but this has not been proven since the analyses of companion gases could not be performed in this study.

The thermal decomposition of siderite under water vapor conditions (300°C) and saturated vapor pressure (P_sat) also resulted in the formation of CM.Reference McCollom¹²⁴ Siderite alone provides both H₂ from water reduction by Fe²⁺ oxidation and inorganic carbon transformed into a reduced carbon phase (15.2):

Although siderite may not be abundant in the oceanic lithosphere, other carbonates are present and may locally decompose (e.g. during magmatic injections) and react if H₂ is available. After solvent extraction, the organic products were identified as dominantly alkylated and hydroxylated aromatic compounds,Reference McCollom¹²⁴ which considerably differs from the FTT products (aliphatic chains) that would have been catalyzed by the abundantly formed magnetite. In addition, discrepancies between the relative H/C ratios of reactants and products suggested the presence of an unidentified product with a low H/C ratio, which could be an insoluble C-rich phase remaining in the solid phase.Reference McCollom¹²⁴

Under aqueous conditions more representative of low-T hydrothermal processes, two types of experiments report the formation of a poorly crystallized CM: (1) alteration of ferromagnesian silicates by a CO₂-enriched fluid; and (2) carbonate dissolution experiments.

Experiments involving the carbonation of olivine (400–500°C, 100 MPa, static capsulesReference Dufaud, Martinez and Shilobreeva¹²⁵) and of a sandstone made of Fe²⁺-rich volcanic clasts (100°C, 10 MPa, flow-through reactorReference Luquot, Andreani, Gouze and Camps¹²⁶) reported the precipitation of a poorly crystallized graphitic phase (Figure 15.5a and b). In both experiments, the graphitic phase accounts for a non-negligible part of the reaction products in addition to phyllosilicates and carbonates. As in natural systems (Section 15.2.2), it is embedded in phyllosilicates, corresponding to serpentine (Figure 15.5a) or Fe-rich chlorite (chamosite; Figure 15.5b) here, closely associated with magnetite when present.Reference Luquot, Andreani, Gouze and Camps¹²⁶ CM formation was attributed to the reduction of CO₂ (or CO at high P–TReference Dufaud, Martinez and Shilobreeva¹²⁵) by H₂ initially present in solution (Eqs. (15.3) and (15.4))Reference Dufaud, Martinez and Shilobreeva¹²⁵ or by Fe²⁺-bearing minerals.Reference Luquot, Andreani, Gouze and Camps¹²⁶

Figure 15.5 Examples of CM that precipitated in hydrothermal experiments. (a) Poorly crystallized graphitic phase (round particles) formed during high-T (400–500°C) carbonation of olivine. (b) Amorphous carbon precipitation during low-T (100°C) carbonation of a sandstone made of Fe²⁺-rich volcanic clasts. (c, d) Two different types of CM precipitated during low-T (200–300°C) siderite dissolution: a poorly structured hydrated or hydrogenated carbonaceous phase (c) and a more ordered graphitic phase (d).

Reproduced with permission of Elsevier, from (a) Dufaud et al. (2009), Chem Geol, 265, 79–87, figure 6,Reference Dufaud, Martinez and Shilobreeva¹²⁵ (b) Luquot et al. (2012), Chem Geol, 294–295, 75–88, figure 10,Reference Luquot, Andreani, Gouze and Camps¹²⁶ and (c, d) Milesi et al. (2015), Geochim Cosmochim Acta, 154, 201–211, figures 4 and 5a.Reference Milesi¹²⁷

The two protocols described above resulted in different oxygen fugacities (fO₂). It was estimated to be close to the CCO buffer (C as graphite–CO) in the high-P–T runs,Reference Dufaud, Martinez and Shilobreeva¹²⁵ which prevented the formation of magnetite, in opposition to the low-P–T runs.Reference Luquot, Andreani, Gouze and Camps¹²⁶ This indicates that magnetite is not mandatory for the formation of graphitic material, whose nature remains poorly constrained, however. Organic volatiles have not been investigated in these studies.

Siderite dissolution experiments at 200°C and 300°C and 50 MPa described two types of carbonaceous products (Figure 15.5c and d)Reference Milesi¹²⁷. A poorly structured hydrated or hydrogenated CM occurred without spatial relationship with minerals in all runs (Figure 15.5c). At 200°C, a more ordered graphitic carbon formed on the surface of the neo-formed magnetite grains or near iron oxides at the siderite surface (Figure 15.5d). CO₂ was the only gas detected at 200°C, while additional small amounts of H₂ and CH₄ were detected at 300°C. A blank experiment showed that trace amounts (<mM) of dissolved organic compounds were present as contaminants.Reference Milesi¹²⁷ Although they were much less abundant than CO₂, there remain open questions regarding whether they contributed to the formation of low levels of CH₄ and the presence of carbonaceous phases. Whatever the case may be, the results of this work fit well with kinetic inhibition of CH₄ formation from CO₂ and H₂ at low T (<400°C; Section 15.3.2) and suggest a preferential precipitation of CM with decreasing T through reactions (15.3) and (15.5).

The inefficiency of magnetite as a FTT reaction catalyst is also pointed out here (see Section 15.3.2). Poisoning of the magnetite surface by carbon coatings (Figure 15.5d) at low T has been questioned; alternatively, the carbon coating may only form when CH₄ cannot (i.e. at low T). A syngenetic relationship between CM and CH₄ has also been proposed but not demonstrated. Equations (15.3) and (15.6) on the one hand and (15.7) and (15.8) on the other show two reversal genetic links. The second one (Eqs. (15.7) and (15.8)) appears less likely under the experimental conditions tested (≤300°C) according, again, to the kinetic inhibition hampering CH₄ formation (15.7).

Equation (15.6), which suggests that carbon coating may serve as an intermediate step to CH₄ formation at least at 300°C where its absence would result from its full consumption, may be a more relevant hypothesis, as already mentioned in Section 15.2.3. Similarly, it was previously suggested that background carbon, possibly more reduced than graphite on the magnetite surface, facilitates CH₄ formation.Reference Fu, Sherwood Lollar, Horita, Lacrampe-Couloume and Seyfried¹¹⁸

At lower T compatible with life, experimental data are lacking on the formation of CM, while this is the condition under which it is the most difficult to discriminate the origin of CM in natural systems (Section 15.2). In serpentinization experiments run at ≤100°C (see Section 15.3.2), no CH₄ was produced, and the strong decrease of CO₂ and H₂ after 3 months was not fully explained.Reference Miller¹³

The limited data available to describe the association of or interplay between organic volatiles and CM limit the conclusions so far on their genetic relationship.

If all occurrences of CM in experimental solid products are due to in situ abiotic reactions from an inorganic source (not contamination), they suggest that at least small amounts of CM easily form under a wide range of hydrothermal conditions during short runs of a few days to a few months (100–500°C, 10–110 MPa, fO₂ between CCO and QFM buffers). The rare available images report different aspects (spherules, film-like surficial layers, or porous accumulations) similar to some of those observed in natural systems (Figure 15.2). They seem to condense from gas or precipitate from fluid preferentially on the mineral surfaces where they possibly evolve. Minerals used in experiments (mainly magnetite) were not as diverse as in nature (Section 15.2), and other mineral surfaces such as hematite, sulfides, hydrogarnets, phyllosilicates, epidote, and quartz, not yet introduced in experiments, may allow the formation of carbonaceous compounds, possibly when the reducing conditions are not optimal.Reference Pikovskii, Chernova, Alekseeva and Verkhovskaya^38, Reference Sforna^55, Reference Mathez and Delaney^60, Reference Luque^85, Reference Tingle and Hochella¹²³

In agreement with natural observations (Section 15.2.2), one experimental study has demonstrated that smectite clays, and particularly montmorillonites, can promote and preserve organic compounds formed under seafloor hydrothermal conditions (300°C, 100 MPa).Reference Williams, Canfield, Voglesonger and Holloway¹⁰³ While ¹³CH₄ was the dominant organic product in the gas-phase, aromatic compounds, including PAHs (up to C₂₀), were extracted from the solid products that reacted for 6 weeks with labeled ¹³C-methanol (10.4 M). The absence of organic material within the illite sample compared to the smectite one showed that the clay interlayer properties exert a control on organic synthesis, notably the negatively charged layer of smectite. The exact reaction path was not constrained; the role of alkanes as intermediate reactants or products and the presence of methanol as a major reactant are worth noting. This study also highlighted the possible protection of organic compounds by clays from thermal alteration at 300°C, making them available for further reactions under lower-T conditions.

15.3.4 Thermodynamic Predictions

The nature of hydrothermal CM observed in natural and experimental works related to the alteration of mantle-derived rocks is not clarified yet. It ranges from graphite to very different forms in terms of structural order, nature and quantity of heteroatoms, unsaturations, and hence reactivity. There has been a considerable amount of work to quantify the thermodynamics of C–C bond formation in graphite and other phases.Reference Amend, LaRowe, McCollom and Shock³² We selected here the main results related to CM formation under hydrothermal conditions for comparison with previous sections.

The formation of graphite is predicted from C–O–H fluids under a wide range of hydrothermal conditions, provided there is no kinetic inhibition. Graphite can readily precipitate from the fluid; this is favored by a decrease in temperature or an increase in pressure along with a decrease in water content.Reference Rumble^36, Reference Zolotov and Shock^45, Reference Ferry, Baumgartner, Carmichael and Eugster^128, Reference Holloway¹²⁹ Both T decrease (deep fluid ascent) and water decrease (e.g. during hydration reactions) correspond well to natural oceanic scenarios in the deeper part of the lithosphere. Graphite formation results from redox reactions unless the reactant carbon is at the same oxidation state (i.e. zero). This does not necessarily require very strong reducing conditions, and graphite can be produced through either oxidation or reduction reactions (Figure 15.4).Reference Zolotov and Shock^44, Reference Frost¹³⁰ In addition to graphite, CO₂ or CH₄ should be the dominant carbon species in hydrothermal fluids depending on temperature and oxygen fugacity. For a given oxygen fugacity, CO₂ should predominate at high-T conditions, while CH₄ is favored at the lower TReference Shock^35, Reference McCollom¹³¹ according to the equilibrium described in (15.7).

When methane and graphite formation is inhibited (i.e. when the kinetic barriers hamper access to a stable equilibrium,Reference Shock³⁵ as in most natural and experimental systems; Sections 15.2, 15.3.2, and 15.3.3), there is considerable potential for the hydrothermal synthesis of a wide range of metastable organic molecules provided that fH₂ is high enough.Reference Milesi, McCollom and Guyot^33, Reference Zolotov and Shock^44, Reference Zolotov and Shock⁴⁵ The reversible character of several organic reactions in hydrothermal fluids contributes to this variety (e.g. hydration/dehydration or oxidation/reduction reactions).Reference Shock¹³² Calculations have shown that organic synthesis of metastable phases is even possible in environments where the amount of H₂ is usually below those observed during the alteration of mantle-dominated environments;Reference Zolotov and Shock^44, Reference Zolotov and Shock⁴⁵ as an example, the rapid cooling of volcanic gases containing CO, CO₂, H₂O, and H₂ below ~250°C provides a thermodynamic drive for the abiotic synthesis of metastable hydrocarbons such as condensed n-alkane and PAHs, depending on the H/C ratio of the gas. Such products may relate to the complex CM and molecular organic minerals observed in natural samples and in CH₄-poor experiments. In more reducing environments, the energetic drive of metastable compound formation is increased and reaction temperatures can be shifted to higher values.Reference Zolotov and Shock^44, Reference Zolotov and Shock⁴⁵ Decreasing temperature allows for metastable mixtures of different condensed phases, in agreement with the variety of compounds found in natural rock textures (Section 15.2.2).

The potential abiotic formation of CM during serpentinization that is central to this chapter has been investigated by comparing experimental fluid compositions with natural hydrothermal compositions across a wide range of CO₂ concentrations.Reference Milesi, McCollom and Guyot³³ Predicted mineral assemblages consisted of serpentine and brucite with or without magnetite and carbonates depending on the relative activities of H₂ and CO₂. CO_2(aq) was considered at equilibrium with graphitic compounds that share the thermodynamic properties of graphite and a hydrogenated aromatic carbon compound (i.e. anthracene, a PAH made of three benzene rings) in order to mimic the experimental products (Figure 15.5c and d).Reference Milesi, McCollom and Guyot^33, Reference Milesi¹²⁷ Results show that serpentinization fluids can equilibrate with both the graphitic and the hydrogenated carbon when the formation of alkanes is prohibited (kinetic inhibition). As the precipitation of a hydrogenated carbon compound requires a higher hydrogen activity, its formation predicted after the serpentinization reaction significantly progresses (i.e. >70% complete), similarly to the classical serpentinization degree attained in oceanic peridotites (e.g. Andreani et al.Reference Andreani, Mével, Boullier and Escartín¹³³).