Abiogenic Hydrocarbon Production at Lost City Hydrothermal Field

See allHide authors and affiliations

Science  01 Feb 2008:
Vol. 319, Issue 5863, pp. 604-607
DOI: 10.1126/science.1151194


Low-molecular-weight hydrocarbons in natural hydrothermal fluids have been attributed to abiogenic production by Fischer-Tropsch type (FTT) reactions, although clear evidence for such a process has been elusive. Here, we present concentration, and stable and radiocarbon isotope, data from hydrocarbons dissolved in hydrogen-rich fluids venting at the ultramafic-hosted Lost City Hydrothermal Field. A distinct “inverse” trend in the stable carbon and hydrogen isotopic composition of C1 to C4 hydrocarbons is compatible with FTT genesis. Radiocarbon evidence rules out seawater bicarbonate as the carbon source for FTT reactions, suggesting that a mantle-derived inorganic carbon source is leached from the host rocks. Our findings illustrate that the abiotic synthesis of hydrocarbons in nature may occur in the presence of ultramafic rocks, water, and moderate amounts of heat.

Fischer-Tropsch type (FTT) reactions involve the surface-catalyzed reduction of oxidized carbon to CH4 and low-molecular-weight hydrocarbons under conditions of excess H2. This set of reactions has been commonly invoked to explain elevated hydrocarbon concentrations in hydrothermal fluids venting from submarine ultramafic-hosted systems (1) and in springs issuing from ophiolites (2); however, whether naturally occurring FTT reactions are an important source of hydrocarbons to the biosphere remains unclear. Although CH4 and higher hydrocarbons have been synthesized by FTT in the gas phase from CO for more than 100 years (3), only recently were FTT reactions shown to proceed, albeit with low yields, under aqueous hydrothermal conditions, with dissolved CO2 as the carbon source (4, 5). The reactions involved in Fischer-Tropsch reduction of aqueous CO2 can be expressed in general terms by the reaction Embedded Image(1)

Here, we show that low-molecular-weight hydrocarbons in high-pH vent fluids from the ultramafic-hosted Lost City Hydrothermal Field (LCHF) at 30°N on the Mid-Atlantic Ridge (MAR) are likely produced abiotically through FTT reactions.

The LCHF is situated near the summit of the Atlantis Massif, ∼15 km west of the MAR axis. Towering carbonate chimneys (up to 60 m tall) diffusely vent high-pH (9 to 11), moderate-temperature (28° to 90°C) fluids, produced by reaction of seawater with rocks originating from the mantle (6, 7). The basement directly beneath this system consists of highly serpentinized peridotites (dominated by depleted mantle harz-burgites) with lesser talc schists and metagabbros exposed by long-lived detachment faulting (8, 9). Fluid circulation is driven by cooling of the underlying rocks (10), perhaps supported by minor contributions from exothermic serpentinization reactions (7, 11) or a yet undetected magmatic source.

Serpentinization is the hydration of olivine and orthopyroxene minerals, the main constituents of ultramafic rocks, and creates a reducing chemical environment characterized by high H2 concentrations that is well suited to abiotic hydrocarbon production. The general reaction is Embedded Image(2)

At Lost City, vent fluids have end-member H2 concentrations of 0.5 to 14.4 mmol/kg (12), greater than the highest H2 concentrations in fluids sampled within basaltic-hosted environments that are unperturbed by magmatic and eruptive events (6, 13). The highest H2 concentrations within the LCHF approach those measured at Rainbow (16 mmol/kg) and surpass those measured at Logatchev (12 mmol/kg); both are peridotite-and-gabbro–hosted systems on the MAR (1). Rainbow and Logatchev vent fluids differ from those at the LCHF in that they have elevated concentrations of CO2 and dissolved metals, low pH, and substantially higher temperatures (350° to 360°C): characteristics typical of high-temperature black smoker vents hosted in mid-ocean ridge basalts (MORBs) (1). In contrast, the moderate-temperature, high-pH, sulfate-bearing fluids at the LCHF are enriched in H2, CH4, and low-molecular-weight volatile hydrocarbons, but are highly depleted in CO2 and dissolved metal contents.

The chemistry of LCHF fluids implies that fluid temperatures beneath the seafloor may be 200° ± 50°C (10), although oxygen isotope data from carbonate veins and low D/H ratios of H2 argue for temperatures <150°C (8, 12). In general, H2 concentrations in LCHF fluids are lower than values observed during laboratory experiments (14) and predicted by equilibrium models (10) and likely reflect a more limited extent of serpentinization at the lower temperatures associated with the natural system (15). The low and variable H2 concentrations in the LCHF fluids may also reflect some removal and use of H2 by microbes. The porous carbonate structures of active vents at Lost City, created by the mixing of vent fluid with seawater, have high microbial cell counts, and phylogenetic studies indicate the presence of H2-utilizing microbes (7, 16, 17). Additionally, 16S ribosomal RNA genes corresponding to those of methanogens, aerobic and anaerobic methanotrophs, sulfate reducers, and sulfur oxidizers were detected in these energy-rich environments (16).

Lost City fluids have CH4 concentrations (1 to 2 mmol/kg) greater than values from unsedimented basalt-hosted hydrothermal systems, but similar to values in fluids from the serpentine-hosted Rainbow and Logatchev vent fields (1, 16). The concentrations are low relative to those of hydrothermal fluids from sediment-hosted environments, where CH4 is produced by the thermal decomposition of organic matter within the sediment (16, 17). Thermogenic CH4 in such sedimented systems is commonly characterized by δ13C values of –50 to –30 per mil (‰) (18, 19). Methane δ13C values at LCHF range from –14 to –9‰ [all carbon and hydrogen isotopic values are given in standard δ notation and ‰ units, referenced to the Vienna Pee Dee belemnite (vPDB) and Vienna standard mean ocean water (vSMOW) standards, respectively]. Lost City δ13C values of CH4 are similar to values of CH4 hypothesized to be of abiogenic origin from Rainbow (–16‰), Logatchev (–14‰), and the ultramafic-hosted Zambales ophiolite seeps (–7‰) (1, 16). The high δ13C values of CH4 from Lost City fluids, as well as the lack of a sediment source rich in organic matter along the reaction path, suggest that CH4 is not appreciably derived from a thermogenic source.

The isotopic composition of short-chain hydrocarbons at LCHF suggests that abiotic synthesis is responsible for their formation. The carbon isotope compositions of C1 to C4 hydrocarbons from LCHF fluids are increasingly negative (δ13C ranges from –9 to –16‰) with increasing chain length (Table 1 and Fig. 1A). This isotopic pattern is opposite to that for hydrocarbons produced thermogenically (20). Such an “inverse” isotopic trend (δ13C1 > δ13C2 > δ13C3 >...) has been shown experimentally to be a possible indicator of abiotic synthesis, and specifically of FTT reactions under conditions of incomplete (35 to 75%) conversion of CO (21). Abiotic synthesis has also been invoked to account for decreases in δ13C values of C2 to C4 alkanes with respect to CH4 in other natural systems (22). The hydrogen isotopic composition of Lost City C1 to C3 hydrocarbons shows a similar, although less defined, trend in which molecules of longer chain length have similar or slightly lower δD values (–120 to –170‰) relative to shorter-chain alkanes (Fig. 1B). The pattern is, again, opposite to the trend expected for thermogenic gases (23), and also opposite to the trends for gases from Precambrian rocks in the Canadian Shield where an enrichment in D with increasing chain length (–425 to –250‰) was attributed to formation by polymerization of CH4 (22). The differences between LCHF and the Canadian Shield D/H trends likely reflect different formation mechanisms and require further investigation through laboratory experiments.

Fig. 1.

13C and D trends. (A) δ13C versus carbon number for C1 to C4 LCHF alkanes. (B) δD versus carbon number for C1 to C3 LCHF alkanes. Trend of increasing isotopic depletion with increasing carbon number is opposite to the observed isotopic trend for thermogenically produced alkanes.

Table 1.

Carbon and hydrogen isotope data from Lost City vents. All isotope values are in ‰ units; δ13C is reported as vPDB, and δD as vSMOW.

View this table:

In addition to their isotopic composition, the abundance and relative distribution of short-chain hydrocarbons are consistent with an abiotic formation mechanism. At Lost City, the alkanes and ethene make up most of the volatile-gas composition, although acetylene, propene, and propyne are present in several samples (table S1). A plot of the log of the n-alkane concentration against carbon number shows a strong linear correlation with carbon number for C2 to C4 alkanes, with elevated CH4 concentrations (fig. S1D). This trend is consistent with an Anderson-Schulz-Flory distribution predicted for FTT synthesis (3). However, such a trend is also consistent with thermogenic production of low-molecular-weight n-alkanes (24).

The elevated CH4 concentrations, relative to those of C2 to C4 n-alkanes, suggests that FTT synthesis may not reflect the only abiotic source of hydrocarbons to LCHF fluids. Laboratory experiments have demonstrated that reduction of CO2 to CH4 under hydrothermal conditions in the presence of Fe-Ni alloys is rapid and does not result in the production of C2+ hydrocarbons (25). Ultramafic rocks from the Atlantis Massif have Ni concentrations of 1700 to 2400 parts per million (ppm), a range that is representative of dredged ultramafic samples from numerous locations along the MAR (26). Although most of the Ni in ultramafic rocks is incorporated into silicate minerals and not catalytically available, Fe-Ni alloys are viable catalysts and stable under highly reducing conditions associated with early stages of serpentinization and lower water/rock ratios (W/Rs) (27).

Radiocarbon measurements represent a powerful tool to constrain the origin of carbon involved in abiotic synthesis of hydrocarbons. Accelerator mass spectrometric measurements on six aqueous CH4 samples from LCHF consistently yielded 14C contents near the detectable limit (Table 1), i.e., the CH4 at LCHF is radiocarbon dead (28). Thus, the carbon source of CH4 cannot be seawater bicarbonate (14C-seawaterDIC measured as modern) that has been microbially or abiogenically reduced. The absence of a modern seawater bicarbonate signal requires that bicarbonate be removed before production of the hydrocarbons in the vent fluids. This likely occurs by precipitation as CaCO3, either in the more permeable zones of seawater recharge, as observed in the recharge limbs of circulation cells at black smoker systems (29) and in ridge-flank environments (30), or in the serpentinizing basement, where increasing pH leads to carbonate precipitation within the serpentinites (8, 10).

The 14C content of short-chain hydrocarbons suggests that the requisite carbon for abiotic synthesis is derived by leaching of primordial radiocarbon-dead carbon from mantle host rocks. Mantle rocks and hydrothermal vent fluids typically have CO2/3He ratios of ∼1 × 109 (31). In contrast, the ratios in the LCHF fluids are much lower, ranging from 3 × 104 to 1 × 107. Lost City 4He concentrations are one order of magnitude less than those of MOR samples, and CO2 concentrations (0.1 to 26 μmol/kg) are up to six orders of magnitude less (Fig. 2 and table S2). The low CO2/3He ratios at Lost City thus primarily reflect extremely low CO2 concentrations, suggesting that mantle CO2 has been removed from the fluids before venting. If a standard MOR value for CO2/3He is assumed, an average CO2 concentration of 2.9 ± 0.7 mmol/kg is predicted from the He content of the fluids (table S2). On an individual sample basis, total hydrocarbon concentrations can account for 35 to 56% of predicted CO2 concentrations. These yields are high relative to the 1% yields typical of aqueous experimental studies (4, 5), but are consistent with the 35 to 75% yields of the FTT experiments that resulted in “inverse” carbon isotope trends (21). Even higher yields are predicted if, before abiotic reduction, mantle CO2 were removed from the system by carbonate precipitation under alkaline conditions created by serpentinization reactions.

Fig. 2.

Measured 4He and CO2 concentrations. LCHF 4He and CO2 concentrations are depleted relative to those of other hydrothermal vent sites (table S2). The plotted line represents CO2/3He = 1 × 109 (assuming a constant 3He/4He, where R/Ra = 8.7), a value that typifies MORB glasses and MOR hydrothermal fluids (31). LCHF data plot well to the left of the line, indicating a loss of CO2 at LCHF.

Near-quantitative reduction of mantle CO2 to hydrocarbons is consistent with the similar 13C content of CO2 and hydrocarbons at the LCHF. The δ13C value of CO2 at LCHF is estimated to be –9‰ (32). This value is within the –10 to –5‰ range observed for δ13C of CO2 from olivine gabbro norites from the South West Indian Ridge (33). Although laboratory experiments conducted at 250°C indicate that fractionation between CO2 and CH4 is ∼35‰ during low-yield FTT synthesis (5) and equilibrium fractionation at this temperature is of similar magnitude, quantitative conversion of CO2 to hydrocarbons will result in the reduced carbon species inheriting the isotopic composition of the initial CO2. Thus, δ13C values for CH4 that range from –13.6 to –9.4‰ in LCHF fluids are consistent with the near-complete reduction of mantle-derived CO2 with a δ13C value of –9‰. Mass balance constraints (assuming a closed system, and a fractionation factor ϵ = 35‰) suggest that a –14‰ δ13C value of CH4 can be explained by an 85% conversion of a –9‰ CO2 source. Extensive reduction of mantle-derived CO2 is supported by a positive correlation between δ13C values of CH4 and percent conversion of mantle CO2 calculated from 3He content of the LCHF fluids (Fig. 3).

Fig. 3.

13C composition of CH4 varies with modeled conversion efficiency. Percent conversion is based on predicted CO2 concentrations (as calculated from the LCHF 3He value and a standard MOR value for CO2/3He). Higher yields correspond with δ13C values of CH4 closer to the approximately –9‰ δ13C values of CO2 at Lost City. Although calculated yields are below quantitative levels, if CO2 is removed by non–CH4-forming mechanisms before FTT reactions (e.g., precipitation of CaCO3 at high-pH conditions), these yields would increase.

The possibility exists that microbial metabolic activity is partially responsible for the near-quantitative conversion of mantle-derived CO2 to CH4. Indeed, the presence of methanogens is indicated in the porous carbonate structures formed by diffuse flow (7, 34, 35). Our measurements do not counterindicate methanogenesis as a source of CH4 to Lost City fluids; however, microbial processes cannot explain the high concentrations and distinctive isotopic signature of C2+ hydrocarbons.

Carbon in mantle rocks occurs in a variety of forms: primarily as inorganic carbon dissolved within the mineral matrix or trapped in fluid inclusions as graphite, or amorphous carbon residing along mineral grain boundaries (36). During hydrothermal circulation at mid-ocean ridges, there is a net transfer of carbon from the host rocks to the circulating fluid during fluid-rock reactions or through magmatic degassing. If all hydrocarbons in LCHF fluids are derived from CO2 originally stored in the underlying rocks (or CO2 produced from carbon in them), the W/R describing the fluid reaction history can be constrained.

For hydrocarbon (simplified in this model to just CH4) concentrations of 1.4 mmol/kg, the maximum W/R is 64, assuming 100% conversion of CO2 to CH4 and an initial (but high) CO2 concentration of ∼4000 ppm in the basement rocks (37) (fig. S2). This W/R is at the low end of those predicted from the Sr and Nd isotopic compositions of LCHF serpentinites (37); however, the samples from seafloor outcrops almost certainly have a reaction history different from that of the rocks directly supplying the present-day fluids at Lost City. More typical and lower initial basement rock CO2 concentrations would yield lower W/Rs. On the basis of a system constrained by a 400-ppm CO2 concentration in the basement rocks (27) and a conversion of ∼50% (as suggested by the He and CO2 data), we posit that the fluids feeding the LCHF have reacted with rocks in a W/R of less than 5 (fig. S2).

Lost City may be just one of many, as yet undiscovered, off-axis hydrothermal systems. Hydrocarbon production by FTT could be a common means for producing precursors of life-essential building blocks in ocean-floor environments or wherever warm ultramafic rocks are in contact with water.

Supporting Online Material

Materials and Methods

Figs. S1 and S2

Tables S1 and S2


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article