Abiogenic Methane Formation and Isotopic Fractionation Under Hydrothermal Conditions

See allHide authors and affiliations

Science  13 Aug 1999:
Vol. 285, Issue 5430, pp. 1055-1057
DOI: 10.1126/science.285.5430.1055


Recently, methane (CH4) of possible abiogenic origin has been reported from many localities within Earth's crust. However, little is known about the mechanisms of abiogenic methane formation, or about isotopic fractionation during such processes. Here, a hydrothermally formed nickel-iron alloy was shown to catalyze the otherwise prohibitively slow formation of abiogenic CH4from dissolved bicarbonate (HCO3 ) under hydrothermal conditions. Isotopic fractionation by the catalyst resulted in δ13C values of the CH4 formed that are as low as those typically observed for microbial methane, with similarly high CH4/(C2H6 + C3H8) ratios. These results, combined with the increasing recognition of nickel-iron alloy occurrence in oceanic crusts, suggest that abiogenic methane may be more widespread than previously thought.

Methane, by far the most abundant natural gas, is formed largely by the digestion of organic compounds by microorganisms (microbial origin) and the thermal decomposition of organic matter (thermogenic origin). Although the occurrence of deep-Earth, mantle methane is still controversial (1, 2), reports of high-temperature fluids venting from sediment-poor mid-ocean ridges (3), seepages from regions associated with ultramafic rocks on land (4), fluids from Precambrian shields (5), and fluid inclusions in mantle and igneous rocks (6) suggest the inorganic (abiogenic) formation of methane in Earth's crust. Several geochemical indicators have been used to distinguish methane of different origins, primarily between microbial and thermogenic. Primary criteria applied to date include CH4/(C2H6 + C3H8) ratios and carbon and hydrogen isotope compositions of methane (7). However, there are very few indicators or criteria for distinguishing methane of abiogenic origin (8); the occurrence of unambiguously abiogenic methane has been documented on only a few occasions, and very little is known about the natural processes that produce it.

It is recognized that abiogenic methane formation is prohibitively slow in the absence of catalysts at low to moderately high temperatures even under reducing conditions, where methane is thermodynamically the most stable carbon-bearing form. However, recent experiments (9) demonstrated that dissolved HCO3 can be converted to CH4 and other hydrocarbons in the presence of ultramafic rocks under reducing hydrothermal conditions; this suggests that natural minerals with high iron or other transition metal content can catalyze the formation of abiogenic CH4. If this process is widely operative in nature, abiogenic methane could be more widespread than previously thought. Determining isotopic fractionation during abiogenic CH4 formation is of particular importance because isotopic composition is widely used as a tool for identifying the origin and mechanisms of CH4 formation. Here, we report experimental results on isotopic fractionation of CH4 abiotically formed from dissolved HCO3 under hydrothermal conditions.

Experiments were performed under conditions similar to those commonly encountered during the serpentinization of ultramafic rocks of the oceanic crust (temperatures of ≤400°C and pressures of ≤100 MPa), where olivine is converted to serpentine and magnetite, producing H2, while released Ni is deposited as awaruite (Ni3Fe). First, a mixture of magnetite and Ni-Fe alloy in an H2-rich aqueous solution (250 to 300 mmol/kg soln) was generated at 400°C and 50 MPa inside a closed, flexible gold-cell hydrothermal apparatus (10). After cooling the vessel to the study conditions, a NaHCO3 solution was injected once to reach an initial total dissolved carbonate species (ΣCO2) (11) of about 5 to 12 mmol/kg soln in the fluid. Samples were then collected periodically for analysis of dissolved gases to monitor CH4 formation (12). A series of experiments was conducted at 200° to 400°C and 50 MPa, including two in which isotopic fractionation was measured (13). Methane formed rapidly in the presence of the Ni-Fe alloy (Fig. 1) via the reactionEmbedded Image(1)In a control experiment where no Ni-Fe alloy was present at 400°C, no methane was formed (Fig. 1C). The amounts of methane produced were directly related to the amount of Ni-Fe alloy (8 to 49 mg in ∼46 g of water). These observations indicate that methane formation from HCO3 is catalyzed by the Ni-Fe alloy, not by magnetite. Reaction at 400°C appeared to slow markedly with time, whereas reaction at 300°C was much faster and continued to near completion. An experiment at 200°C required even longer reaction times.

Figure 1

Chemical composition of fluids during abiogenic hydrothermal CH4 (open symbols) formation from total dissolved CO2 (ΣCO2, closed symbols) in water (∼46 g) for (A) 200°C and 50 MPa with 26 mg of Ni-Fe alloy, (B) 300°C and 50 MPa with 17 mg (diamonds), 26 mg (squares), and 46 mg (circles) of Ni-Fe alloy, and (C) 400°C and 50 MPa with 0 mg (triangles), 8.4 mg (squares), 27 mg (diamonds), and 49 mg (circles) of Ni-Fe alloy. HCO3 was a predominant (>70%) CO2 species in the fluids.

The combined concentrations of ΣCO2 and CH4in the fluid decreased early in the experiments, but rebounded toward the end of each experiment. This indicated the presence of an intermediate carbon phase or species during the reduction of HCO3 to CH4. Thus, the overall reaction described by Eq. 1 is better described by a chain reaction via an intermediate product or products, HCO3 → intermediate(s) → CH4. Intermediate products such as carbon monoxide (CO), graphitic carbon, and carbide phases are commonly found in dry-gas CO2 methanation, although the precise role that each plays in the reactions is not always clearly understood (14). No CO was detected in the fluids during the experiments, and no carbon phases were detected in the solids after the experiments had been completed (10). One possibility for the intermediate species in the experiments is formate ion (HCOO), because formate, a hydrated form of CO, is thermodynamically more stable than CO under hydrothermal conditions (15). This species could not have been detected by the gas chromatography techniques used here (12), but it was observed as a common intermediate product in similar experiments using ion chromatography techniques at 200° to 300°C (16). However, because we did not measure formate concentrations in our Ni-Fe alloy experiments in Fig. 1, and because we cannot rule out the existence of other intermediate species, the concentration of the intermediate product or products was calculated as a single component using mass balance constraints (Table 1). The fact that this intermediate phase was present early in the experiments suggests that it forms from dissolved CO2 more rapidly than it is converted to CH4.

Table 1

Chemical and isotopic composition of fluids during abiogenic hydrothermal CH4 formation from dissolved CO2. Values for intermediates were calculated from carbon mass and isotope balances with propagated errors (1σ). For the dissolved gas concentrations, the overall analytical error is ±5%, except for the intermediate. Values of δ13C are with respect to the Vienna Pee Dee belemnite standard, ±0.1 per mil for ΣCO2 and ±0.5 per mil for CH4.

View this table:

In the two isotope experiments, the δ13C value of ΣCO2 steadily increased while HCO3 was reduced (Fig. 2). The δ13C values of the CH4 produced, on the other hand, were consistently much lower than those of ΣCO2, and increased slightly with reaction progress. The δ13C values of the intermediate product, calculated from an isotope mass balance of the system, were initially between those of ΣCO2 and CH4 at 200°C. However, at the end of the experiment, the calculated δ13C value of the intermediate product became higher than that of the residual ΣCO2 (Fig. 2A). At 300°C, the δ13C value of the intermediate product was only slightly lower than that of ΣCO2 by the time the first sample was collected, but much higher than that of CH4. The δ13C value of the intermediate eventually surpassed that of the coexisting ΣCO2 (Fig. 2B).

Figure 2

δ13C values of fluids during abiogenic hydrothermal CH4 formation from total dissolved CO2 (ΣCO2) in water for (A) 200°C and 50 MPa with 26 mg of Ni-Fe, and (B) 300°C and 50 MPa with 17 mg of Ni-Fe. The δ13C values of intermediate products were calculated from an isotope mass balance (Table 1).

Overall 13C/12C isotope fractionation during the formation of abiogenic CH4 from dissolved CO2 (largely HCO3 ) was evaluated as α = (13CH4/12CH4)/(13CO2/12CO2) = 0.940 to 0.950 (−50 to −60 per mil) at 200°C and α = 0.960 to 0.965 (−35 to −40 per mil) at 300°C (Fig. 2). Equilibrium13C/12C fractionations expected between ΣCO2 and CH4 in this study were calculated as α(equil) = 0.970 (−30 per mil) at 200°C and α(equil) = 0.982 (−18 per mil) at 300°C, with an uncertainty of ±5 per mil (17). The direction of isotope fractionation obtained from our experimental results is the same as those of the equilibrium fractionation factors, but their magnitudes are much larger. This suggests that the reduction of HCO3 via intermediate(s) to CH4was largely controlled by kinetic isotopic fractionation. The magnitude of overall kinetic isotope fractionations observed during abiogenic CH4 formation from HCO3 under hydrothermal conditions is as large as those of microbial reduction of CO2 to CH4 by methanogenic bacteria at ambient temperature (α = 0.93 to 0.96) (18). There is, to our knowledge, no experimental study on carbon isotope fractionation during hydrothermal CH4 formation from dissolved CO2, although several studies of the conversion of CO to CH4, CO2, and other hydrocarbons in the presence of Co or Fe catalysts under dry H2 atmosphere (Fischer-Tropsch reaction) showed large, variable kinetic carbon isotope fractionation (α = 0.94 to 0.99) (19).

Our experimental results show that abiogenic methane forms rapidly in the presence of even small amounts of hydrothermally generated Ni-Fe alloys under reducing conditions. In nature, Ni-Fe alloys are important phases in the paragenetic sequence of minerals that form during hydrothermal alteration of olivine-rich rocks (20). Olivine commonly contains several thousand parts per million of Ni, and it commonly converts to Ni-Fe alloys under the reducing conditions that typically prevail during serpentinization. Awaruite, Ni-Fe alloy, and Ni-Fe sulfide have been reported from many continental ultramafic settings (20), oceanic crustal environments (21), and meteorites (22). Indeed, the oceanic crust is largely composed of ultramafic rocks, and serpentinite containing awaruite and Ni-Fe alloy could make up a large fraction of the oceanic crust (23). Thus, abiogenic methane formation via an awaruite/Ni-Fe alloy–catalyzed reaction may occur more commonly in Earth's crust than is currently recognized.

As noted above, the criteria for identifying methane as abiogenic are very ambiguous (7, 8). One suggested geochemical criterion is δ13C values higher than −25 per mil, based on limited isotopic data of probable abiogenic CH4 (Fig. 3) (8). However, our experimental results reveal that δ13C values of abiogenic CH4 formed from dissolved CO2 with a mantle-like δ13C value (−4 per mil) can be very low, and may overlap values typical for microbial CH4 (Fig. 3). Moreover, the high CH4/(C2H6 + C3H8) ratios of the gases produced by the Ni-Fe alloy are also similar to those of microbial origin. This is in direct contrast to previous results, where abiogenic methane formation catalyzed by magnetite was accompanied by ethane and propane (9). The chemical compositions of abiogenic hydrocarbon gases clearly depend strongly on the catalysts and mechanisms involved (14). This may explain low CH4/(C2H6 + C3H8) ratios of probable abiogenic methane found in many localities (Fig. 3). Another suggested criterion for abiogenic CH4 is that δ13C value decreases in the order CH4, C2H6, and C3H8 (“isotopic reversal”) (8). However, limited experimental data (19) suggest that this is not necessarily the case for abiogenic hydrocarbons. It has been suggested that reduced carbonaceous materials from early Archaean sediments, which have low δ13C values (≤–20 per mil), are of biogenic origin (24). However, our results show that kinetic isotopic fractionation during abiotic processes can also produce reduced carbon-bearing species with very low δ13C values.

Figure 3

Plot of δ13C(CH4) and CH4/(C2H6 + C3H8) values of abiogenic methane hydrothermally formed at 200° and 300°C and at 50 MPa in this study (open rectangles), together with values for biogenic (microbial and thermogenic) methane (7) and several reported values for abiogenic methane (3–6). CH4/(C2H6 + C3H8) ratios of abiogenic CH4 in this study are lower limits, as shown by arrows, because both C2H6 and C3H8 were always below the detection limit (∼1 μmol/kg soln). δ13C values of ≥–25 per mil have been suggested for abiogenic CH4 (8).

The demonstrated hydrothermal synthesis of organic compounds (organic acids, hydrocarbons, alcohols) (25) under conditions resembling those in Earth's upper crust may have implications for the possible abiogenic formation not only of methane, but also of petroleum (26) and for the evolution of prebiotic organic molecules in the early oceans (27).

  • * To whom correspondence should be addressed. E-mail: horitaj{at}


View Abstract

Navigate This Article