Report

Nonequilibrium clumped isotope signals in microbial methane

See allHide authors and affiliations

Science  24 Apr 2015:
Vol. 348, Issue 6233, pp. 428-431
DOI: 10.1126/science.aaa4326

What controls clumped isotopes?

Stable isotopes of a molecule can clump together in several combinations, depending on their mass. Even for simple molecules such as O2, which can contain 16O, 17O, and 18O in various combinations, clumped isotopes can potentially reveal the temperatures at which molecules form. Away from equilibrium, however, the pattern of clumped isotopes may reflect a complex array of processes. Using high-resolution gas-phase mass spectrometry, Yeung et al. found that biological factors influence the clumped isotope signature of oxygen produced during photosynthesis (see the Perspective by Passey). Similarly, Wang et al. showed that away from equilibrium, kinetic effects causing isotope clumping can lead to overestimation of the temperature at which microbially produced methane forms.

Science, this issue p. 431; p. 428; see also p. 394

Abstract

Methane is a key component in the global carbon cycle, with a wide range of anthropogenic and natural sources. Although isotopic compositions of methane have traditionally aided source identification, the abundance of its multiply substituted “clumped” isotopologues (for example, 13CH3D) has recently emerged as a proxy for determining methane-formation temperatures. However, the effect of biological processes on methane’s clumped isotopologue signature is poorly constrained. We show that methanogenesis proceeding at relatively high rates in cattle, surface environments, and laboratory cultures exerts kinetic control on 13CH3D abundances and results in anomalously elevated formation-temperature estimates. We demonstrate quantitatively that H2 availability accounts for this effect. Clumped methane thermometry can therefore provide constraints on the generation of methane in diverse settings, including continental serpentinization sites and ancient, deep groundwaters.

Carbon (13C/12C) and hydrogen (D/H) isotope ratios of methane are widely applied for distinguishing microbial from thermogenic methane in the environment (17), as well as for apportioning pathways of microbial methane production (810). This bulk isotope approach, however, is largely based on empirical observations, and different origins of methane often yield overlapping characteristic isotope signals (3, 7, 1113). Beyond conventional bulk isotope ratios, it has become possible to precisely measure the abundance of multiply substituted “clumped” isotopologues (e.g., 13CH3D) (14, 15). In particular, the abundance of clumped isotopes makes it possible to obtain information about the temperature at which C–H bonds were formed or last equilibrated (14) (fig. S1). Formation temperatures of both thermogenic and microbial methane in natural gas reservoirs can be estimated on the basis of clumped isotopologues (16). The mechanisms by which isotopologues attain distributions consistent with thermodynamic equilibrium, however, remain unclear because bulk methane isotopes (δ13C and δD) often reflect kinetic isotope fractionations (13, 17), and H isotope exchange between methane and water is sluggish (18).

To test whether clumped methane thermometry can be widely applied for methane sources beyond natural gas reservoirs, we examined methane samples from diverse systems, including lakes, wetlands, cow rumen, laboratory cultures of methanogenic microbes, and geological settings that may support abiogenic methane production. We used a recently developed tunable laser spectroscopy technique (14, 19) to measure the relative abundances of four methane isotopologues (12CH4, 13CH4, 12CH3D, and 13CH3D).

Our measurements for dominantly thermogenic gases from the Marcellus and Utica shales (1, 20) yielded Δ13CH3D-based temperatures of Embedded Image°C and Embedded Image°C, respectively. The clumped isotope temperature for the Marcellus Shale sample is comparable to, although slightly lower than, estimates by Stolper et al. (16) of 179° to 207°C (Fig. 1). In addition, microbial methane in pore waters and gas hydrates from northern Cascadia margin sediments (3) and from wells producing from coal seams in the Powder River Basin (2, 21) yielded Δ13CH3D temperatures of 12° to 42°C and 35° to 52°C, respectively. These are consistent with their expected low formation temperatures. Furthermore, thermogenic methane sampled from a hydrothermal vent in the Guaymas Basin, Gulf of California (6), yielded a Δ13CH3D temperature of Embedded Image°C, within error of the measured vent temperature (299°C) (22). Therefore, our data provide independent support of the hypothesis that 13CH3D abundance reflects the temperature at which methane is generated in these sedimentary basins (16).

Fig. 1 Isotopologue compositions of methane samples.

(A) Δ13CH3D plotted against δD. The Δ13CH3D temperature scale corresponds to calibration in fig. S1. Error bars are 95% confidence intervals (table S1). Data from (16) were scaled to their corresponding Δ13CH3D values (15). The shaded area represents the temperature range within which microbial life has been demonstrated to date (35). The dotted line represents Δ13CH3D = 0‰ (temperature T → ∞); data plotting below this line cannot yield corresponding apparent temperatures. (B) δ13C plotted against δD, showing characteristic fields for different methane sources from (13).

In contrast, we found that methane sampled from lakes, a swamp, and the rumen of a cow carries 13CH3D signals that correspond to anomalously high Δ13CH3D temperatures (139° to 775°C) (Fig. 1A) that are well above the environmental temperatures (<40°C). Such signals are clearly not controlled by equilibrium. Notably, a positive correlation between Δ13CH3D and the extent of D/H fractionation between methane and environmental water [εmethane/water (23) (Fig. 2)] suggests a strong link between isotopologue (i.e., 13CH3D) and isotope (D/H) disequilibria. In contrast, the above-mentioned methane samples from sedimentary basins appear to have attained hydrogen isotope equilibrium with associated waters at or near the temperatures indicated by the Δ13CH3D data (Fig. 2).

Fig. 2 Extent of clumped and hydrogen isotopic disequilibria in methane.

Symbols and vertical error bars are the same as those in Fig. 1. Horizontal error bars represent uncertainties on estimates of εmethane/water (23) (table S4). The solid green curve represents isotopic equilibrium, with the εmethane/water calibration given by (36). Green shading represents ranges of εmethane/water calibrations from published reports (fig. S3). Gray shading represents model predictions from this study, for microbial methane formed between 0° and 40°C. Metabolic reversibility (φ) increases from bottom (φ = 0, fully kinetic) to top (φ → 1, equilibrium) within this field (19).

To confirm these observations from the natural environment, we demonstrated that strong disequilibrium 13CH3D signals are also produced by cultures of methanogenic archaea in the laboratory (Fig. 3). Thermophilic methanogens cultured at 40° to 85°C produced methane with Δ13CH3D values from +0.5 to +2.3 per mil (‰) (corresponding to Δ13CH3D temperatures of 216° to 620°C), and mesophilic methanogens cultured at ambient temperature produced methane with conspicuously “anticlumped” signatures (i.e., values of Δ13CH3D <0‰, for which no apparent temperature can be expressed) as low as –1.3‰ (Fig. 3). Methane from cultures is also characterized by large kinetic D/H fractionation with respect to water (17, 24). Because laboratory cultures are grown under optimal conditions (high H2 and high CO2), these anticlumped Δ13CH3D and low εmethane/water values are primarily expressions of kinetic isotope effects. Consequently, the distribution of samples with Δ13CH3D and εmethane/water values in Fig. 2 can be explained by microbial methanogenesis operating on a spectrum between fully kinetic (low Δ13CH3D and low εmethane/water) and equilibrium (high Δ13CH3D and high εmethane/water) end members.

Fig. 3 Δ13CH3D values of methane produced by hydrogenotrophic methanogens in batch cultures reflect kinetic effects.

Data and error bars are from table S2. The green line represents clumped isotopologue equilibrium (i.e., samples for which Δ13CH3D temperature is equal to growth temperature) (fig. S1).

We constructed a mathematical framework to describe the controls on the correlation of Δ13CH3D and εmethane/water signals from hydrogenotrophic methanogenesis. The model largely follows those developed for microbial sulfate reduction (25, 26) and predicts the isotopologue compositions of product methane as a result of a series of enzymatic reactions (fig. S4) (19). Using isotope fractionation factors estimated from theory, experiments, and observations as input parameters (table S3) (19), our model reproduces the observed correlation between Δ13CH3D and εmethane/water of natural samples (Fig. 2). The isotopologue compositions of product methane reflect the degree of metabolic reversibility. Fully reversible reactions yield equilibrium end members (27), whereas irreversible reactions result in kinetic (disequilibrium) end-member signals. In this model, the reversibility is linked to available free energy (26, 27), in this case expressed as H2 concentration ([H2]). The model can explain the relationship among [H2], εmethane/water (28), and Δ13CH3D via Michaelis-Menten kinetics and can predict the observed patterns in diverse settings, ranging from marine sediments (low [H2], high Δ13CH3D and εmethane/water) to bovine rumen (high [H2], low Δ13CH3D and εmethane/water) (Fig. 4). We note that mixing of methane sources with different δ13C and δD values or oxidation of methane could also alter the relationships over the primary signal of microbial methanogenesis (19). Likewise, inheritance of clumping signals from precursor organic substrates (e.g., via acetoclastic or methylotrophic methanogenesis) cannot be ruled out entirely and awaits experimental validation.

Fig. 4 Relationships between Δ13CH3D and H2 concentration for microbial methane.

Symbols and vertical error bars are the same as in Fig. 1. The H2 data are from table S4; when a range of [H2] values is given, points are plotted at the geometric mean of the maximum and minimum values. Dashed lines represent model predictions for microbial methane produced at 20°C, calculated using Michaelis-Menten constants (KM) of 0.3, 3.0, and 30 μM H2. Data for samples of dominantly nonmicrobial methane from Guaymas Basin and Kidd Creek are plotted for comparison.

We showed above that the combination of Δ13CH3D and εmethane/water values provides mechanistic constraints on whether methane was formed under kinetic versus near-equilibrium conditions. Next, we used this framework to place constraints on the origins of methane at two sites of present-day serpentinization in Phanerozoic ophiolites [The Cedars (29) and Coast Range Ophiolite Microbial Observatory (CROMO) (30)] in northern California, as well as in deep (>2 km below surface) fracture fluids with billion-year residence times in the Kidd Creek mine, Canada (5, 31).

Methane collected from groundwater springs associated with serpentinization at The Cedars yielded anticlumped Δ13CH3D signals (–3‰) with low εmethane/water values (Figs. 1A and 2). The data plot along the microbial (kinetic) trend defined in Fig. 2, supporting a previous hypothesis that methane at The Cedars is being produced by active microbial methanogenesis (29). The exceptionally high H2 concentration (up to 50% by volume in bubbles) at The Cedars indicate the massive excess of electron donors. This, along with severe inorganic carbon limitation [due to high pH (>11) and precipitation of carbonate minerals (29)], drives the formation of methane carrying strong kinetic imprints, consistent with the observed anticlumped Δ13CH3D signals (Fig. 4).

Despite the similarity in geologic setting, methane associated with serpentinization at CROMO (30) revealed very different Δ13CH3D values, which correspond to low apparent temperatures (42° to 76°C) and plot close to the equilibrium line (Fig. 2). Although the conventional δ13C and δD values of methane from CROMO are nearly identical to those of the Utica Shale sample (Fig. 1B), methane at CROMO carries much higher Δ13CH3D values (Fig. 1A). The origin of methane at the CROMO site remains unresolved (30), but the comparably high Δ13CH3D values at CROMO suggest that methane here could be sourced from a mixture of thermogenic and microbial methane. Alternatively, lower H2 availability at CROMO, compared with The Cedars (table S4), may support microbial methanogenesis under near-equilibrium conditions (Fig. 4). Regardless, the different isotopologue signatures in methane from CROMO versus The Cedars demonstrate that distinct processes contribute to methane formation in these two serpentinization systems.

Deep, ancient fracture fluids in the Kidd Creek mine in the Canadian Shield (31) contain copious quantities of both dissolved methane and hydrogen (5). The Kidd Creek methane occupies a distinct region in the diagram of Δ13CH3D versus εmethane/water (Fig. 2), due to strong D/H disequilibria between methane and water (4) and low–Δ13CH3D temperature signals of 56° to 90°C that are consistent with other temperature estimates for these groundwaters (4). Although the specific mechanisms by which the proposed abiotic hydrocarbons at Kidd Creek are generated remain under investigation (5, 32), the distinct isotopologue signals provide further support for the hypothesis that methane here is neither microbial nor thermogenic.

Our results demonstrate that measurements of 13CH3D provide information beyond the simple formation temperature of methane. The combination of methane and water hydrogen-isotope fractionation and 13CH3D abundance enables the differentiation of methane that has been formed at extremely low rates in the subsurface (3, 21, 27) from methane formed in cattle and surface environments in which methanogenesis proceeds at comparatively high rates (33, 34).

Supplementary Materials

www.sciencemag.org/content/348/6233/428/suppl/DC1

Materials and Methods

Supplementary Text

Figs. S1 to S5

Tables S1 to S6

References (3787)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. The abundance of 13CH3D is captured by a metric, Δ13CH3D, that quantifies its deviation from a random distribution of isotopic substitutions among all isotopologues in a sample of methane: Δ13CH3D = ln Q, where Q is the reaction quotient of the isotope exchange reaction Embedded Image. The reported δ values are conventional isotopic notation, e.g., δD = (D/H)sample/(D/H)reference – 1. Mass spectrometric measurements yield Δ18, a parameter that quantifies the combined abundance of 13CH3D and 12CH2D2. For most natural samples of methane, Δ18 temperature is expected to be directly relatable to Δ13CH3D temperature, as measured by laser spectroscopy. The D/H fractionation between methane and environmental water is defined as εmethane/water = (D/H)methane/(D/H)water – 1.
  3. Acknowledgments: We thank J. Hayes, R. Summons, A. Whitehill, S. Zaarur, C. Ruppel, L. T. Bryndzia, N. Blair, D. Vinson, K. Nealson, and M. Schrenk for discussions; W. Olszewski, D. Nelson, G. Lacrampe-Couloume, and B. Topçuoğlu for technical assistance; A. Whitehill, G. Luo, A. Apprill, K. Twing, W. Brazelton, A. Wray, J. Oh, A. Rowe, G. Chadwick, and A. Rietze for assistance in the field; R. Michener for the δDwater analyses; L. T. Bryndzia (Shell) for providing the shale gas samples; R. Dias (USGS) for sharing the NGS samples; and R. Raiche, D. McCrory, S. Moore (Homestake Mining Co.), the staff of the McLaughlin Natural Reserve, and the well operators for access to samples. Grants from the NSF (EAR-1250394 to S.O. and EAR-1322805 to J.C.M.), N. R. Braunsdorf and D. J. H. Smit of Shell PTI/EG (to S.O.), the Deep Carbon Observatory (to S.O., B.S.L., M.K., and K.-U.H.), the Natural Sciences and Engineering Research Council of Canada (to B.S.L.), and the Gottfried Wilhelm Leibniz Program of the Deutsche Forschungsgemeinschaft (HI 616-14-1 to K.-U.H. and M.K.) supported this study. D.T.W. was supported by a National Defense Science and Engineering Graduate Fellowship. D.S.G. was supported by the Neil and Anna Rasmussen Foundation Fund, the Grayce B. Kerr Fellowship, and a Shell-MITEI Graduate Fellowship. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. government. All data used to support the conclusions in this manuscript are provided in the supplementary materials. Author contributions: D.T.W. and S.O. developed the methods, analyzed data, and performed modeling. D.T.W. and D.S.G. performed isotopic analyses. D.S.G., L.C.S., J.F.H., M.K., K.-U.H., and S.O. designed and/or conducted microbiological experiments. D.T.W., D.S.G., B.S.L., P.L.M., K.B.D., A.N.H., C.N.S., M.D.K., D.J.R., J.C.M., D.C., and S.O. designed and/or executed the field-sampling campaigns. D.T.W. and S.O. wrote the manuscript with input from all authors.
View Abstract

Navigate This Article