The Production of Methane, Hydrogen, and Organic Compounds in Ultramafic-Hosted Hydrothermal Vents of the Mid-Atlantic Ridge

Abstract Both hydrogen and methane are consistently discharged in large quantities in hydrothermal fluids issued from ultramafic-hosted hydrothermal fields discovered along the Mid-Atlantic Ridge. Considering the vast number of these fields discovered or inferred, hydrothermal fluxes represent a significant input of H2 and CH4 to the ocean. Although there are lines of evidence of their abiogenic formation from stable C and H isotope results, laboratory experiments, and thermodynamic data, neither their origin nor the reaction pathways generating these gases have been fully constrained yet. Organic compounds detected in the fluids may also be derived from abiotic reactions. Although thermodynamics are favorable and extensive experimental work has been done on Fischer-Tropsch-type reactions, for instance, nothing is clear yet about their origin and formation mechanism from actual data. Since chemolithotrophic microbial communities commonly colonize hydrothermal vents, biogenic and thermogenic processes are likely to contribute to the production of H2, CH4, and other organic compounds. There seems to be a consensus toward a mixed origin (both sources and processes) that is consistent with the ambiguous nature of the isotopic data. But the question that remains is, to what proportions? More systematic experiments as well as integrated geochemical approaches are needed to disentangle hydrothermal geochemistry. This understanding is of prime importance considering the implications of hydrothermal H2, CH4, and organic compounds for the ocean global budget, global cycles, and the origin of life. Key Words: Hydrogen—Methane—Organics—MAR—Abiotic synthesis—Serpentinization—Ultramafic-hosted hydrothermal vents. Astrobiology 15, 381–399.


Introduction
H ydrothermal circulation is a common process along oceanic spreading centers. The visible expression of this subsurface circulation is hydrothermal fields at the seafloor that are the foci of submarine oases of life. More than 500 sites have been located or inferred along the Mid-Ocean Ridge (MOR) system, including fast-, ultrafast-, slow-, and ultraslow-spreading ridges, as well as in back-arc basins, and many more are to be discovered based on the assumption of about 1 field every 100 km of ridge. More than 250 fields have been confirmed active and studied during oceanographic cruises with submersibles and/or remotely operated vehicles (Fig. 1).
Hydrothermal circulation occurs when seawater percolates downward through fractured ocean crust. The heated seawater is transformed into hydrothermal fluid through reaction with the host rock at temperatures that can exceed 400°C and exhaled forming thick smoke-like plumes of black metal sulfides often termed ''black smokers.'' During its transit through the oceanic crust, seawater composition is mainly modified by phase separation and water-rock interactions but is also influenced by (micro)biological processes and magmatic degassing. As a result, fluids become enriched in a variety of compounds and depleted in some others. In spite of their similar appearance, high-temperature hydrothermal vent fluids exhibit a wide range of temperatures and chemical compositions depending on subsurface reaction conditions and nature of the leached rocks (basalts, ultramafic rocks, felsic rocks). Notably, alteration of ultramafic rocks is associated with hydrogen release, which leads to high reducing conditions in these environments. In turn, it has been suggested that these reducing conditions would be favorable for the abiogenic production of methane and other organic molecules. One of the major implications for abiogenic synthesis is the origin of life. In this paper, we will discuss the origin of H 2 , CH 4 , and organic compounds in the deep sea and in ultramafic-hosted vents on the Mid-Atlantic Ridge (MAR) where most of these particular systems have been reported (Fig. 2). Particular attention is given to the use and helpfulness of stable isotopes in addressing the question of origin. We also present estimations of fluxes of hydrothermal H 2 and CH 4 entering the ocean and show that hydrothermal inputs are significant and should be considered in ocean global cycles studies. Finally, organic abiotic synthesis feasibility and implications for our understanding of the origin of life are the focus of the second part of the paper.

H 2 and CH 4 Origins in the Deep-Sea Environment
Molecular hydrogen is produced by biotic and abiotic processes, but its concentration in natural systems is usually very low due to the activity of hydrogen-consuming bacteria (Libert et al., 2011;Petersen et al., 2011). Heterotrophic bacteria produce hydrogen by fermentation of organic matter or by anaerobic oxidation of carbon monoxide (Silva et al., 2000;Sokolova et al., 2004;Pawar and van Niel, 2013). The evidence of a hydrogen-based subsurface microbial ecosystem was brought up by Takai et al. (2004a). More recently, the presence of communities capable of producing and oxidizing H 2 has been reported at the Lost City vent field (Brazelton et al., 2012). Hydrogen production by abiogenic processes includes deep crustal outgassing (Okuchi, 1997;Karato, 2006), crystallization of the basaltic magma (Christie et al., 1986;Holloway andO'Day, 1999, 2000), and low-temperature reactions that occur in the shallow crust, and it is often related to active faulting (Wakita et al., 1980;Sugisaki et al., 1983;Ware et al., 1985;Ito et al., 1999). Active volcanic or seismic activities generate volatile compounds such as H 2 , CO, and H 2 S, although CO 2 generally dominates, sometimes spectacularly (Sarda and Graham, 1990;Javoy and Pineau, 1991;Dixon et al., 1995;Soule et al., 2012). In addition, the radiolytic dissociation of water during the radioactive decay of natural U, Th, and K radionuclides in the host rock is another potential source of H 2 (Lin et al., 2005). Hydrogen may also be derived from small amounts of water included in minerals in the form of hydroxyls or peroxides (Freund et al., 2002). Finally, and more to the focus of this paper, the production of H 2 may be caused by the interaction of water at temperatures ranging from below 100°C up to 500°C with highly reduced FIG. 1. The MOR system showing the presently known and sampled hydrothermal sites. (Color graphics available at www.liebertonline.com/ast) iron-containing minerals occurring in ultramafic diapirs present in continental or submarine environments (Mevel, 2003). These so-called serpentinization reactions are likely to represent the dominant process for abiogenic formation of H 2 in ultramafic-hosted hydrothermal systems and most efficiently at temperatures around 300-350°C (Klein et al., 2013). Hydrogen-rich fluids issued from the serpentinization process are discussed in Section 3 and by Sleep et al. (2004).
Methane in deep-sea environments has many sources, which are discussed in detail in a review by Martin Schoell (1988). Most of the commercial methane is thermally derived from petroleum and has a biogenic origin (Rice and Claypool, 1981). This methane is often called thermogenic. Under pressure and low temperature, methane forms a thermodynamically stable association with water. These solid compounds are called methane clathrates and are typically found in permafrost or associated with rocks and mud in deep seafloor environments. Besides, microbial methane is produced by bacteria and archaea in sediments, subsurface and hydrothermal vents via CO 2 reduction and/or fermentation (e.g., Whiticar et al., 1986;Takai et al., 2004b;Amend and Teske, 2005;Roussel et al., 2011). The observations made along the MAR together with the methane-rich fluids found in gabbroic inclusions from the South West Indian Ridge (SWIR) indicate that plutonic rocks represent a potentially immense reservoir for abiogenic methane (Kelley et al., 1993;Kelley, 1996;Evans, 1996;Kelley and Früh-Green, 1999). A prevalent hypothesis for the abiotic production of significant amounts of CH 4 in the deep sea at high pressure and temperature is via catalytic reduction of certain carbon oxides in the presence of H 2 . Field data associated with laboratory experiments (McCollom and Seewald, 2001) show that CH 4 , together with H 2 , is a major emission by-product of serpentinization. CH 4 outgassing associated with intense H 2 output has consistently been observed in ultramafic settings such as in peridotites of the Oman ophiolite (Neal and Stanger, 1983), in serpentinized rocks of the Zambales ophiolite, Philippines (Abrajano et al., 1988), in serpentine seamount drilled during ODP Leg 125 in the Mariana Forearc (Haggerty, 1991), and along MORs (Charlou et al., 1991(Charlou et al., , 1993a(Charlou et al., , 1996a(Charlou et al., , 2010Charlou and Donval, 1993;Kelley, 1996;Früh-Green and Kelley, 1998). Alternatively, reasonable processes of abiogenic CH 4 formation at high T and P include the reduction of bicarbonate to graphite and methane (Holloway, 1984;Berndt et al., 1996), the thermal decomposition of siderite (McCollom, 2003), and clay mineral-catalyzed reactions (Williams et al., 2005).  Finally, low-temperature ( < 150°C) production of CH 4 would be possible by hydration of olivine without H 2 mediation and would be more common than previously thought (Miura et al., 2011;Suda et al., 2014). For a complete examination of potential abiotic CH 4 sources on Earth, the reader is directed toward the review by Etiope and Sherwood Lollar (2013).
3. Abiotic H 2 and CH 4 Production in Ultramafic Hydrothermal Systems on the MAR All the ultramafic-hosted, hot-temperature hydrothermal fields discovered along the MAR are characterized by strong enrichment of dissolved H 2 and CH 4 , with end-member concentrations covering a range of 10-16 mmol/kg for H 2 and 1.7-2.5 mmol/kg for CH 4 . While either H 2 or CH 4 may also be enriched in basaltic-hosted hydrothermal systems in unsedimented settings under certain conditions (Von Damm, 1995;Lilley et al., 1993), only serpentinization of mantle rocks produces the characteristic coupled constant enrichment of both gases observed in ultramafic-hosted hydrothermal systems (Charlou et al., , 2010Douville et al., 2002).
Along the MAR, enhanced permeability at the intersections of the rift valley with the fracture zones favors seawater circulation and serpentinization of lower-crustal and upper-mantle ultramafic rocks. Close to transform-ridge intersections, structural settings enhance fluid circulation and wall-rock reactions (Bougault et al., 1990(Bougault et al., , 1993Gracia et al., 2000), generating ultramafic rock exposures and methane outputs (Bougault et al., 1993;Charlou et al., 1996aCharlou et al., , 1998. Faulting facilitates hydrothermal circulation through ultramafic outcrops, amplifies serpentinization reactions, and accelerates hydrogen and hydrocarbon degassing, as predicted by theoretical calculations (McCollom and Shock, 1998;Wetzel and Shock, 2000).
Serpentinization is an ongoing process at depth in the seafloor that leads to significant changes in topography, the occurrence of diapiric serpentinite bodies, focused microseismic activity as a result of continuous cracking, and significant heat generation (Fyfe, 1974;Allen and Seyfried, 2004). Chemically, serpentinization is the hydration of the olivine and orthopyroxene minerals that mainly constitute the upper mantle. Highly reducing conditions can be generated during serpentinization as a result of, on the one hand, the oxidation of Fe(II) in olivine, pyroxene, and pyrrhotite to Fe(III) in magnetite and serpentine, and on the other hand, the reduction of hydrogen from water to H 2 (Allen and Seyfried, 2003;Klein et al., 2009;McCollom and Bach, 2009). Hydrogen production by serpentinization proceeds most effectively in ultramafic rocks because the minerals that form in these silica-poor rocks during alteration tend to exclude Fe(II) from their metal sites and to partly oxidize and precipitate iron in magnetite (McCollom and Seewald, 2001 and serpentine (Andreani et al., 2013;Evans et al., 2013). One of the serpentinization equations is (Moody, 1976;Neal and Stanger, 1983;Mevel, 2003): Phase equilibrium and mass balance calculations indicate that the production of H 2 -rich fluids during hydrothermal alteration is primarily controlled by the alteration of olivine and the formation of magnetite. It occurs when the lower crustal and shallow mantle sequences have cooled to temperatures below 400°C where serpentine and brucite are thermodynamically stable (Früh-Green and Kelley, 1998). Recent observations have revealed that this production is also widely controlled by the activity of aqueous silica in the interacting fluid, which is in turn controlled by the specific phase assemblages that develop during serpentinization (Frost and Beard, 2007;Ogasawara et al., 2013). Depending on mineral assemblages, pressure, temperature, reducing power, and equilibrium conditions, the fluid may be supersaturated with respect to hydrogen. As the fluids ascend upward to lower pressures and temperatures, gas bubbles may be generated (Sleep et al., 2004).
A consequence of this H 2 production is the possible formation of abiogenic CH 4 . Among the different plausible reaction pathways, reduction of gaseous or dissolved carbon mono-and dioxides using catalysts such as Ni-Fe alloys and/or oxides has been by far the most studied and referred to (e.g., Berndt et al., 1996;Horita and Berndt, 1999;Chen and Bahnemann, 2000;Foustoukos and Seyfried, 2004;Taran et al., 2010a). The latter is best described by Reactions 1 and 2: (1) the Sabatier reaction Although CH 4 and other hydrocarbons have been synthesized by Fischer-Tropsch-type (FTT) reaction in the gas phase from CO for more than 100 years (Anderson, 1984;Steynberg and Dry, 2004), these reactions can also proceed under aqueous hydrothermal conditions with CO 2 as the carbon source (Berndt et al., 1996;Horita and Berndt, 1999;Horita, 2001;McCollom and Seewald, 2001Foustoukos and Seyfried, 2004;Seewald et al., 2006). Experiments carried out at temperatures lower than 500°C combined with thermodynamically favorable conditions (Shock, 1990(Shock, , 1994Schulte, 1990, 1998;Amend and Shock, 1998) confirm that MORs' ultramafic hydrothermal systems may provide conditions that allow for abiogenic generation of CH 4 from CO 2 (Berndt et al., 1996;Holm and Andersson, 1998;Holm and Charlou, 2001;Kelley et al., 2002).

Fluxes of H 2 , CH 4 along the MAR
It is now demonstrated that the concentrations of methane and hydrogen in hydrothermal vent fluids are very variable. High concentrations of these gases are found in ultramafichosted vent fields (Table 1), whereas relatively lower concentrations are found in basalt-hosted vent fluids. Exceptions are fluids from the Menez Gwen, Lucky Strike, and more recently discovered Piccard vent field usually referred to as basalt-hosted vents. Their concentrations of CH 4 are similar to, and sometimes exceed, the ones observed in ultramafichosted vents. Lucky Strike has consistently vented high CH 4 but low H 2 fluids for almost three decades, which could be related to magmatic events (Von Damm et al., 1998;Charlou et al., 2000;Pester et al., 2012;Crawford et al., 2013). As for the Menez Gwen and Piccard vent fields, the occurrence of ultramafic rocks at depth or in the vicinity cannot be excluded (Stroup and Fox, 1981;Charlou et al., 2002). However, if one discards this option, high production of CH 4 and H 2 in those basaltic environments could be due to the reaction of water with CO 2 from direct magmatic degassing or leaching from CH 4 -rich inclusions in gabbroic rocks (Kelley et al., 1993;Kelley, 1996), with respect to CH 4 , and to water cracking related to dike formation or to the alteration of troctolites, with respect to H 2 (Elthon, 1987;Nakamura et al., 2009). Mid-Ocean Ridge fluxes of hydrogen and methane at oceanic spreading centers appear nonetheless to be mainly supplied by serpentinization of ultramafic rocks. The total hydrogen and methane produced by serpentinization is estimated to be 133 · 10 9 mol/yr and 14 · 10 9 mol/yr, respectively (Keir, 2010), representing about 70% of the total ocean ridges and rises flux of these gases. These fluxes are slightly lower than those calculated by Cannat et al. (2010) according to the rate of mantle rock exhumation and the stoichiometry of olivine hydration (167 · 10 9 mol/yr and 25 · 10 9 mol/yr for H 2 and CH 4 , respectively). Emmanuel and Ague (2007) obtained even higher values for the flux of CH 4 generated by serpentinized lithosphere based on estimates of the rate of seafloor spreading and the degree of serpentinization within the oceanic crust. They calculated it to be 84 · 10 9 mol/yr. In comparison, volcanic and geothermal sources are estimated to contribute for *6.2 · 10 9 mol/yr and *56 · 10 9 mol/yr, respectively. Based on Rainbow H 2 / 3 He and 3 He/heat ratios ( 3 He/heat = 9.3 · 10 8 mol J -1 in Jean-Baptiste et al., 2004), global H 2 and CH 4 fluxes for slow spreading ridges have been calculated to be 89 · 10 9 mol/yr and 9 · 10 9 mol/yr, respectively (Charlou et al., 2010). All these results are within the uncertainty of plus/minus a factor of 2 inherent to these estimates, but they indicate the necessity of taking hydrothermal gas inputs into account in global cycles studies.

Origin of CH 4 in the Fluids of Ultramafic-Hosted Vents on the MAR from H and C Stable Isotopes
The possible contribution to the methane budget in fluids from ultramafic-hosted vents may be multiple, as mentioned before and discussed in a review by McCollom (2008). Isotopic considerations appear compulsory to unravel the origin of CH 4 . Several authors have suggested that abiogenic methane is, or was, produced in various geological settings such as crystalline rocks, ophiolites, gas seepages, fumarolic discharges, and ultramafic-hosted hydrothermal systems on land, based on carbon and hydrogen isotope measurements (Sherwood Lollar et al., 1993Fiebig et al., 2004Fiebig et al., , 2007Fiebig et al., , 2009Hosgormez et al., 2008;Taran et al., 2010b;Etiope et al., 2011;Suda et al., 2014). The consensus is that it is formed via the Sabatier and FTT pathways (Reactions 1 and 2). In ultramafic-hosted fluids collected up to now along the MAR, the d 13 C (CH4) value is found to be -11.9 & at Lost City, -17.8 & at Rainbow, -10.2 & at Logatchev 1, -6.1 & at Logatchev 2, -12.3 & at Ashadze 1, and -8.7 & at Ashadze 2 ( Table 2). These results combined with the dD (CH4) fall into a range that is neither thermogenic nor biogenic as previously found in other hydrothermal fluids and shown in Fig. 3 (Schoell, 1988;Welhan, 1988;Charlou et al., 1993bCharlou et al., , 1996bCharlou et al., , 2000. The same trend is observed when considering d 13 C (CH4) versus the CH 4 -to-higher-hydrocarbons ratio (C 1 /C 2 + ) ( Table 2). The results for Rainbow, Lost City, Logatchev 1 and 2, and Ashadze 1 are consistent with typical values for unsedimented MOR systems, whereas Ashadze 2 has a lower C 1 /C 2 + ratio and is closer to the Zambales ophiolite values (McCollom, 2008). Only an abiogenic contribution may account for these observations. The variability in d 13 C (CH4) observed here and also reported in the literature is probably due to fractionation at different T and P conditions. Nevertheless, there is a clear abiogenic contribution to the CH 4 budget in hydrothermal fluids issued from ultramafic environments, which is supported by experiments (Horita and Berndt, 1999;Horita, 2001;Lazar et al., 2012;Cao et al., 2014).
The number of ultramafic-hosted sites discovered along the MAR associated with the detection of numerous CH 4rich plumes associated with mantle rock alteration indicates that ultramafic-hosted sites are more widespread than previously thought (Charlou and Donval 1993;Charlou et al., 1993a;Keir et al., 2005). This points out their significant contribution to the global abiogenic methane flux along the MAR (Charlou et al., 1991(Charlou et al., , 1996a(Charlou et al., , 1996b(Charlou et al., , 1998Bougault et al., 1993). This is also supported by Keir et al. (2005), who estimated the amount of methane escaping from the MAR to be equivalent to about 0.06 nmol per liter of middepth water and per year. Extrapolated to the whole Atlantic, this comes to about 1 · 10 9 mol/yr. The methane production required to support this escape rate ( > 3.6 · 10 9 mol/yr) is significantly greater than the maximum input by basalt degassing ( < 0.6 · 10 9 mol/yr)-evidence that serpentinization of ultramafic rocks generates most of the methane and is the main active process producing isotopically heavy methane along the MAR.

Organic Geochemistry-Pathways
Organic compounds have been found in hydrothermal fluids and investigated since the 2000s. Semivolatile ones ( > 6 carbon chain length) such as aliphatic hydrocarbons, mono-and polyaromatic hydrocarbons, and carboxylic acids have been reported on the MAR (Holm and Charlou, 2001;Konn et al., 2009;McCollom et al., 2015). Formate (36-158 lM) and acetate (1-35 lM) have been observed by Lang et al. (2010) in fluids from the Lost City hydrothermal field. Reeves et al. (2014) measured concentrations of 10 -9 to 10 -6 M of methanethiol in fluids from hydrothermal vents in various geological settings including the MAR. Amino acids were found in fluids from various sites of the MAR. The total dissolved free amino acid concentrations were up to 377 nM versus < 50 nM for deep seawater (Sumoondur et al., 2006;Klevenz et al., 2010). Organic geochemistry of hydrothermal fluids is a brand new field that raises a lot of questions, in particular regarding the origin of organic compounds and their potential importance for the origin of life on Earth. The processes that control the organic composition of the fluids are not yet fully constrained or understood. Both the sources of the building units composing organic molecules (C, H, O, and N) and possible reaction pathways leading to organic compounds are multiple; therefore, determining the origin of organic compounds and understanding their formation appears a real challenge. On the one hand, rocks, minerals, and seawater are potential C, H, O, and N sources; in a simplified view, CO 2 and carbonates are sources of C, H 2 and water sources of H, N 2 and NH 3 sources of N, water and oxygen-bearing minerals sources of O. On the other hand, microorganisms that inhabit hydrothermal vents and the subsurface may provide C, H, O, and N by two mechanisms: (i) direct production of Bradley and Summons (2010). Ranges of d 13 C and dD detected in methane produced by a variety of sources. ''Autotrophic'' and ''heterotrophic'' is microbial methane. ''Thermogenic'' refers to cracking of biologically derived oils, while ''geothermal'' refers to cracking of high-molecular-weight organic compounds. The remaining are values observed at several locations where abiotic methane formation has been suggested: Canadian Shield gases (including Kidd Creek), the Oman ophiolite, Zambales ophiolite. The dark surface represents the range of values (from Table 2) measured for methane in fluids from the Rainbow, Lost City, Logatchev 1 and 2, Ashadze 1 and 2 ultramafic-hosted vent fields. simple molecules (e.g., CH 4 , H 2 , acetate, CO 2 ), (ii) thermal degradation of the microorganisms themselves if exposed to high-temperature fluids. In addition, macroorganisms may also undergo thermal degradation after death and sedimentation. Whether C, H, O, N are derived from gas, minerals, and seawater or from organisms, they will be referred to as abiogenic or biogenic, respectively. Multiple processes possibly leading to the formation of organic molecules from those C, H, O, N sources in hydrothermal environments include abiogenic processes that represent any purely chemical reactions; biogenic processes that encompass all reactions driven by microorganisms; and thermogenic processes that refer to both thermal degradation of large organic molecules (e.g., proteins, lipids, DNA) to smaller and simpler ones as well as rearrangement of compounds under high temperature and pressure conditions such as condensation, cleavage, cyclization, hydrolysis, oxidation, hydrogenation, and hydroformylation Simoneit, 2004, 2006;Loison et al., 2010).

Abiogenic processes
Abiotic synthesis will only occur if thermodynamics are favorable. This has been shown to be the case for a wide range of organic compounds under conditions found at modern subseafloor hydrothermal systems (e.g., Shock, 1990;McCollom and Seewald, 2001;Lemke, 2013). The hypothesis is that organic compounds would occur in metastable equilibrium due to kinetic barriers that would prevent the inherently sluggish stable equilibria CO 2 /CH 4 and N 2 /NH 3 to be reached in hydrothermal solutions. We refer to the work of Shock (1990Shock ( , 1992 and Konn et al. (2009) for a more in-depth discussion. Reaction pathways are not well constrained, although mechanisms have been proposed for hydrocarbons and amino acids.
6.1.1. Hydrocarbons-FTT reactions. Abiogenic origin of hydrocarbons was first brought up by Mendeleev in 1877 and has been highly debated since 1940 (e.g., Mendeleev, 1877;Kudryavtsev, 1951;Hedberg, 1969;Szatmari, 1989;Gold, 1993;Kutcherov and Krayushkin, 2010). The two recent reviews published in 2013 by Sephton and Hazen, on the one hand, and by Etiope and Sherwood Lollar, on the other hand, are evidence that this debate is still fueled. Today, technological advances have helped clarify the various controversial theories that were elaborated on. A fairly recent review by Glasby (2006) highlights the lack of strong evidence to support the abiogenic petroleum theories and rules out the abiogenic production of oil in commercial quantities. Notably, this does not exclude the possibility of a minor abiotic contribution ( Jenden, 1993;Sherwood Lollar et al., 2002). Apps and van de Kamp (1993) concluded that, even though commercial hydrocarbon deposits appear to be exclusively biogenic in origin, this may be at the exception of deposits associated with serpentinization. In 1964, a mixed origin of hydrocarbons was implied (Sylvester-Bradley). With the advent of modern analytical tools, the co-occurrence of both biogenic and abiogenic signatures in most hydrocarbon fields has been confidently revealed (Mello and Moldowan, 2005;Scalera, 2011). In that respect, Scalera (2011) developed a possible new harmonic scenario of hydrocarbon formation combining both thermogenic and abiogenic processes. It seems that scientists are moving toward a consensus that hydrocarbons may be produced by different pathways on Earth, but the question that remains is what the contribution of each process is both globally and in specific geological contexts. In ultramafic-hosted hydrothermal systems, several processes are capable of generating reduced carbon species . These include FTT reactions, methane polymerization, carbonate decomposition, organosulfur pathways, and clay-catalyzed reactions. They are presented in detail in a review by McCollom (2013b). Among them, FTT reactions have been addressed by extensive experimental work under hydrothermal conditions in the past decades. FTT processes are considered prime candidates to account for the generation of abiogenic hydrocarbons in ultramafic-hosted hydrothermal systems (Holm and Charlou, 2001;Konn et al., 2009).
The Fischer-Tropsch reaction (3) was a common industrial process used to produce hydrocarbons from CO and H 2 (Fischer and Tropsch, 1923). It was invented by two German scientists, Franz Fischer and Hans Tropsch, in the 1920s and was largely developed during World War II to generate substitute fuels. The original process takes place in the gas phase at high pressure and temperature according to the following mass balance equation: Numerous laboratory experiments conducted under hydrothermal conditions have demonstrated the feasibility of the abiogenic production of hydrocarbons (e.g., Rushdi and Simoneit, 2001;Foustoukos and Seyfried 2004;McCollom and Seewald, 2006). Thermodynamic calculations have shown that saturated hydrocarbons can be abiotically produced via FTT reactions under hydrothermal conditions from dissolved CO 2 (Shock, 1990(Shock, , 1992. The reactions involved in FTT reduction of aqueous CO 2 can be expressed as follows: The distribution of the observed hydrocarbons makes it quite clear that they are the product of FTT reactions, but isotopically labeled experiments have shown that CO is a much more effective precursor than CO 2 (McCollom and Seewald, 2001;McCollom et al., 2010). It is still unclear whether the reactions occur in the gas phase or in the water phase, which could favor one or the other oxidized form of C to react preferentially. The exact pathways along which these hydrocarbons are formed are not fully constrained yet, either, and the actual occurrence of FTT processes at deepsea hydrothermal conditions (where CO 2 , CO, H 2 are dissolved) is still uncertain. Although every single experiment has been a great step forward, their results are in most cases not intercomparable because of crucial differences in the experimental conditions. Parameters such as P, T, redox state, presence/absence of a catalyst, and carbon source significantly impact the resulting products and most likely the involved processes. Unraveling the reaction pathways is a real challenge considering this unfortunate inconsistency in the overall set of published experiments. We urge the reader to refer to the work of McCollom and Seewald (2007) and McCollom (2013b) for a detailed review.
6.1.2. Amino acids-Strecker synthesis. The abiotic synthesis of amino acids is of particular interest in the origin of life question, as they represent the fundamental building blocks of proteins that are required for the development of living organisms. The amino acid synthesis has been generally proposed to occur via a Strecker-type mechanism under hydrothermal conditions (e.g., Hennet et al., 1992;Schulte and Shock, 1995;Islam et al., 2001;Aubrey et al., 2009). The original Strecker amino acid synthesis, devised by Adolph Strecker in 1850, is a series of chemical reactions that synthesize an amino acid from an aldehyde (or ketone) according to Reaction 5 (Strecker, 1850): The Strecker synthesis has been shown to be thermodynamically favorable over all ranges of temperatures appropriate for a hydrothermal system at 300 bar (Brandes et al., 1998). Nevertheless, for this reaction pathway to proceed in hydrothermal systems the formation of the required reactants (HCN and aldehyde or ketone) by reduction of inorganic carbon (CO or CO 2 ) and nitrogen (N 2 ) must first occur. In that respect, highly reducing conditions encountered in ultramafic-hosted hydrothermal environments are very favorable. Formation of HCN from N 2 and CO 2 in the presence of H 2 , possibly with CH 4 as an intermediate, is both thermodynamically and experimentally strongly supported (Shock, 1992;Holm and Neubeck, 2009). Experimental works have shown that amino acids are likely formed under hydrothermal conditions and more favorably under high hydrogen concentrations (Hennet et al., 1992;Islam et al., 2001;Huber and Wächtershäuser, 2006;Simoneit et al., 2007). Unless they are protected by mineral surfaces or undergo polymerization or cyclization (Ito et al., 2006;Cox and Seward, 2007), amino acids are likely to be destroyed by deamination, decarboxylation, and dehydration at temperatures above 240-260°C and even at *170°C in the presence of certain mineral assemblages (Bada et al., 1995;Faisal et al., 2007;McCollom, 2013a). The possible occurrence and persistence of amino acids in hydrothermal fluids thus depends at least on redox conditions, mineral assemblages, temperature, and pressure.
To date, there is no evidence that abiotic amino acid synthesis occurs in natural environments. Undeniably, amino acids have been found in fluids of hydrothermal systems in various geological settings, including ultramafic-hosted hydrothermal vents, but unanimously the authors reporting these amino acids have concluded that they are most likely derived from microorganisms living on the surface of the chimney (Horiuchi et al., 2004;Takano et al., 2004;Sumoondur et al., 2006;Klevenz et al., 2010;Lang et al., 2013;Fuchida et al., 2014). To complement the set of data on amino acids, Table 3 (from Konn et al., 2015) gives some preliminary results of fluids from other vents on the MAR. Consistent with the above-cited works, only a portion of the entire set of amino acids was detected. It is probably due to different limits of detection, degradation rates, as well as different abilities to polymerize and to adsorb on mineral surfaces (Gupta et al., 1983;Henrichs and Sugai, 1993).

Biogenic processes
Chemolithotrophic microbial communities commonly colonize hydrothermal vents and may represent analogues for life on early Earth and other planets. Chemolithotrophic organisms, by definition, utilize only inorganic and/or abiotic simple molecules for their carbon and energy sources so that they do not rely on other living organisms to feed, develop, and multiply (Lang et al., 2012). To date, the maximum temperature for some of such organisms to grow is 122°C (Takai et al., 2008).
In our case, the archaea methanogens, which are one of the most common microorganism groups found at hydrothermal vents, are of particular interest, as they synthesize CH 4 from CO 2 and H 2 (Schoell, 1988;Takai et al., 2004b;Brazelton et al., 2011;Nishizawa et al., 2014). The consumption of methane leading to the production of CO 2 by methanotrophic bacteria occurs to a lesser extent because methanotrophs are less abundant in hydrothermal environments. Also, from CO 2 and H 2 , acetogenic bacteria are able to generate acetate which can, in turn, be used as substrate by heterotrophic methanogens (e.g., Chapelle and Bradley, 1996). As mentioned earlier, the majority of amino acids detected in hydrothermal fluids is thought to be microbially derived partly because the autotrophic synthesis of several amino acids from CO 2 (aq), NH 4 + , and H 2 is thermodynamically favorable at hydrothermal conditions (Amend and Shock, 1998).

Thermogenic processes
Typically, thermogenic processes occur in sedimentary basins and are associated with maturation of petroleum, which is defined as the oil and gas generated during thermolysis from the former (e.g., Demaison and Murris, 1984;Tissot and Welte, 1984). Hydrothermal systems definitely meet the condition of high temperature required for thermal degradation (see reviews in McCollom and Seewald, 2007;McCollom, 2008). Organic matter is present in the form of macroorganisms and microorganisms that thrive both around the chimneys and in the subsurface. Macroorganisms from the surface ocean will inevitably die and fall to the seafloor. Degradation products may be taken up by seawater and penetrate the crust in the recharge zone of hydrothermal systems and thus undergo thermogenesis deeper in the crust (Brault et al., 1988). Similarly, microbial organisms growing in the subsurface may be either flushed by a cold fluid and carried away to a place where temperature would be high enough to degrade the very durable lipids that form the membranes of the bacteria and archaea, or burned off as a very hot fluid would encounter these communities (Reeves et al., 2014).

Biogenic versus abiogenic-the use of carbon stable isotopes
Organic compounds in hydrothermal systems are likely to result from co-occurring abiogenic, biogenic, and thermogenic processes using both biogenic and abiogenic C, H, O, and N and eventually leading to extensive mixing of biogenic and abiogenic C, H, O, and N elements within organic molecules. For example, biogenic methane, carbon dioxide, and acetate could well then be involved in abiotic processes,  As seawater mixing occurred in some samples, the % of pure fluid is given in the Fluid column. The Pre-C column gives how much the samples were concentrated before analyses (in fold).
Concentrations are given in ppt, i.e., in ng L -1 , and refer to the concentration in the natural hydrothermal fluid originally, before preconcentration.
*Spiked samples with 50 lL of the 1000-fold diluted standard solution.
nd: not detected. nm: not measured. such as the previously described FTT and Strecker reactions. A more extensive discussion can be found in the work of McCollom (2008). This raises two issues: (i) How do we discriminate? (ii) What do we call those resulting organic compounds? They are neither biogenic nor abiogenic nor thermogenic. Table 4 is an attempt to illustrate this dilemma. We dare to suggest that, as long as terminology has not been agreed upon, distinction might be made between sources and processes. In the above-mentioned example, compounds might be called biogenic with respect to their source and abiogenic with respect to the process they were generated along. Our ability to classify organic compounds with more than two carbon atoms into the biogenic or abiogenic category might be very challenging (Horita, 2005). Even for the simplest organic molecule that is methane, the sole use of isotopic composition is sometimes insufficient, and complementary techniques have been used to determine its origin (e.g., Bradley and Summons, 2010). Elsewhere, Etiope and Sherwood Lollar (2013) described the importance of integrated geochemical techniques to confirm the occurrence of abiogenic methane. It has been generally presumed that thermogenic, biogenic, and abiogenic hydrocarbons should differ in their carbon isotopic composition. Typical reference values of d 13 C (CH4) are -70& to -60& for a biological production, -60& to -40& for a thermogenic origin, -30& to -20& for geothermal hydrocarbons, and -20& to -5& for MORs (Schoell, 1988;Bradley and Summons, 2010), but this division is being debated (e.g., Sherwood Lollar and McCollom, 2006;Ueno et al., 2006aUeno et al., , 2006b. As for hydrocarbon gases (C 1 -C 4 ), it has been suggested that a slight decrease in d 13 C with increasing carbon number could be an indication of an abiotic catalytic formation, while a thermogenic origin has always shown a strongly positive correlation (Des Marais et al., 1981;Sherwood Lollar et al., 2002;Pan et al., 2006). This isotope reversal trend has been attributed to kinetic isotope fractionation effects during surface-catalyzed polymerization reactions of methylene units (e.g., Schoell, 1983;Jenden, 1993;Fu et al., 2011). As the trend is weak to almost flat, it was even suggested that no fractionation occurs during polymerization. However, hydrocarbon gases produced experimentally via abiogenic reactions do not consistently produce inverse or flat trends, and results are rather heterogeneous (e.g., McCollom and See-wald, 2006;Fu et al., 2007;Taran et al., 2007Taran et al., , 2010aMcCollom et al., 2010). Experiments reported in the literature to date were carried out under various physical and chemical conditions. This strongly indicates that carbon isotope fractionation of hydrocarbons is controlled by their formation processes and kinetics, which in turn may differ according to temperature, pressure, and redox conditions Fu et al., 2011). Whether the experiments were conducted in a gas and/or water phase and in a closed or flow-through reactor are other possible influencing factors. This is discussed more in depth elsewhere (McCollom, 2013b). In addition, several thermogenic gases do show reversals of the kind attributable to abiotic reactions (e.g., Burruss and Laughrey, 2010). The reverse or flat trend has generally been observed for hydrocarbon gases in ultramafichosted hydrothermal systems, but no clear evidence of their abiogenic origin has been brought forth (Proskurowski et al., 2008;Charlou et al., 2010). Clearly, it will be even more difficult to determine the origin of longer n-alkanes and other organic compounds detected in fluids from ultramafichosted vents. Semivolatile organic compounds ( > 6 C atoms chain length) have rarely been reported as products of FTT reaction experiments probably due to their low concentration (e.g., Taran et al., 2007Taran et al., , 2010a. And, as far as the authors know, there are only two examples of an FTT experiment in which d 13 C values of the heavy products ( > C12) have been measured McCollom et al., 2010). These results indicate that different d 13 C trends may be expected depending on the carbon source. Another influential parameter may be pressure . Carbon isotopic ratios for the n-alkane series C 9 -C 20 seem to show a different pattern (Fig. 4) from the one produced in McCollom's experiments (2010). Despite new experimental work and field data, the conclusion on the origin of hydrocarbons from various sources and/or processes drawn by Konn et al. (2009) remains.
Whereas Lang and coworkers (2010) were able to conclude with some confidence on the origin of formate (abiogenic) and acetate (microbially derived) using stable C isotopes, the origin of heavier fatty acids cannot be unraveled by using their individual d 13 C value. Konn et al. (2009) showed that fatty acids were slightly enriched in 13 C For example, a molecule resulting from FTT reaction using mantle CO 2 will be called abiogenic. If the same process uses CO 2 from respiration (although we do not know if this kind of reaction can actually occur, so we beg the reader to take this as an illustration), we currently do not know what to call the resulting product, which would be biogenic with respect to the source and abiogenic with respect to the process. Boxes left with a ? point out the word missing.
compared to the alkanes of the same chain length and inferred that this fractionation could possibly be created by biological processes.
It was generally admitted that amino acids were of biogenic origin in hydrothermal systems. Based on their carbon isotope measurements, Lang et al. (2013) were able to demonstrate that the amino acids in the hydrothermal fluids at Lost City were derived from chemolithoautotrophs living on the surface and subsurface of the chimneys.
In conclusion, stable carbon isotope ratios are very useful and may give indication of the origin of organic compounds (both gas and semivolatiles) in hydrothermal contexts, but they should be complemented with other approaches.
To cite a few examples in the literature: position-specific methods have been shown to be efficient in determining the origin of deep gases (Corso and Brenna, 1997); combined analyses with noble gases (Sherwood Lollar and Ballentine, 2009); clumped isotopes is a rather new tool that would be worth trying; thermodynamic calculations could help discrimination (Reeves et al., 2014); the use of radiogenic carbon has proven to be efficient in determining the source of methane (Proskurowski et al., 2008); the thermal degradation experiments of biomass carried out by Konn et al. (2011) brought additional lines of evidence toward a plausible abiogenic origin of a portion of n-alkanes versus a likely biogenic origin of a portion of aromatic hydrocarbons and n-carboxylic acids (all > C 8 ) detected in fluids from ultramafic-hosted systems.

Implications for the origin of life
Life may have appeared on Earth in the earliest Archean or even before in the Hadean (Russell and Hall, 1997;Rosing, 1999;Korenaga, 2013). Hydrothermal activity is relevant to Hadean and Archean Earth, as it began as soon as water condensed to form oceans, and some kind of plate tectonics (corresponding to crust formation) appeared 4.4 billion years ago (Wilde et al., 2001). Also, hydrothermal systems as well as ultramafic rocks were much more abundant on primitive Earth than today (Russell et al., 1988).
Although the composition, oxidation state, temperature, and pressure of the early atmosphere after the bombardment is unknown (e.g., Marshall, 1994;Schoonen et al., 1999), a proposed composition on which most of the scientific community agrees is domination by CO 2 in a dense state, N 2 and H 2 O; little amount of H 2 S, HCl, SO 2 , and elemental sulfur S 0 ; and minor amounts of H 2 and Ar (Chen and Chen, 2005;Russell and Arndt, 2005). Different lines of evidence indicate the presence of significant levels of CH 4 (100-1000 ppm) in the atmosphere in the Archean (Pavlov et al., 2000;Kasting, 2005). Fiebig et al. (2007) proposed an abiogenic origin of this CH 4 . Magnesium (Mg) as well as transition metals such as iron (Fe) and nickel (Ni) must have been abundant in the early ocean (Mloszewska et al., 2012). Mg 2 + together with Ca 2 + would have been the prevalent divalent cation, while the prevalent monovalent cation was Na + (Pontes- Buarque et al., 2000). The ocean is considered to have been fairly acidic with a pH *5-6 ( Russell and Arndt, 2005). Finally, almost uncontested to date, is the view that both atmosphere and ocean would have remained anoxic (oxygen-free) until the great oxidizing event postulated at 2.4 billion years ago. However, several controversial lines of evidence, including the sulfur isotopic composition of pyrites and the elemental compositions of ancient soil horizons, have been put forth to instead support the presence of appreciable amounts, or at least whiffs, of oceanic and atmospheric oxygen long before (Anbar et al., 2007;Konhauser, 2009). Moreover, a recent paper by Hoashi et al. (2009) reports on the observation of hematite crystals in marine sediments of 3.46 Ga, which indicates that free oxygen would have existed at least locally in the oceans at that time.
As a conclusion, conditions at modern seafloor hydrothermal systems seem to be similar, to some extent, to early FIG. 4. Carbon stable isotope ratios versus carbon number (Nc) for the n-alkane series detected in fluids from the Lost City field (triangles) and the Rainbow field (squares). C 9 -C 14 (Konn et al., 2009) and C 16 -C 20 (Konn et al., unpublished results). Analytical and sample collection methods are described by Konn et al. (2009).
Earth's conditions and thus can be considered a place of primary focus in the search for the origin of life. Moreover, hydrothermal vents constitute very favorable environments for the start of life, as much in terms of protection against the sterilizing effect of giant impacts as in terms of scale. Microenvironments such as mineral surfaces favor adsorption, concentration of organics, and subsequent reactions. In addition, a hydrothermal mound provides some kind of protection (niches), physicochemical gradients, and nonequilibrium conditions that are required for the majority of macromolecules typical of the cell organization to persist as well as for the emergence of a living organism (Russell and Hall, 1997;Kompanichenko, 2009). The serpentinization process is emerging as an increasingly likely source of the energy essential for life to have emerged from CO 2 , rocks, and water on early Earth (Russell et al., 1989(Russell et al., , 2010McCollom and Seewald, 2013). Alkaline (high pH) hydrothermal systems are thought to be even more relevant to Archean hydrothermal vents, and the Lost City hydrothermal field could provide particular insights into past mantle geochemistry and present a better understanding of the chemical constraints that existed during the evolutionary transition from geochemical to biochemical processes. In parallel and in quest of the origin of life, experimental work simulating alkaline hydrothermal vents and/or Hadean conditions is being done (Herschy et al., 2014;Yamaguchi et al., 2014).

Acknowledgments
This research was supported and funded by IFREMER through the Theme 5-Exploration et Exploitation des Fonds Océaniques-Program PGE01-Géologie des Environnnements extremes Océaniques, conducted by the Marines Geosciences Department. We are grateful to the captains and crews of the R/V L'Atalante and Pourquoi Pas? as well as to the Nautile submersible and Victor-6000 ROV technical teams. They all helped us with fluid sampling through their splendid support and their ship-handling capabilities during surveys and operations at sea. The authors gratefully acknowledge support from the International Space Science Institute (ISSI) for our team ''The Methane Balance-Formation and Destruction Processes on Planets, their Satellites and in the Interstellar Medium.'' We also thank Muriel Andreani, Fabrice Brunet, Valérie Chavagnac, Yves Fouquet, Benjamin Malvoisin, and Ewan Pelleter for their valuable discussions.

Disclosure Statement
No competing financial interests exist.