Biogenic Methane, Hydrogen Escape, and the Irreversible Oxidation of Early Earth

See allHide authors and affiliations

Science  03 Aug 2001:
Vol. 293, Issue 5531, pp. 839-843
DOI: 10.1126/science.1061976


The low O2 content of the Archean atmosphere implies that methane should have been present at levels ∼102 to 103 parts per million volume (ppmv) (compared with 1.7 ppmv today) given a plausible biogenic source. CH4 is favored as the greenhouse gas that countered the lower luminosity of the early Sun. But abundant CH4 implies that hydrogen escapes to space (↑space) orders of magnitude faster than today. Such reductant loss oxidizes the Earth. Photosynthesis splits water into O2 and H, and methanogenesis transfers the H into CH4. Hydrogen escape after CH4 photolysis, therefore, causes a net gain of oxygen [CO2 + 2H2O → CH4 + 2O2 → CO2 + O2 + 4H(↑space)]. Expected irreversible oxidation (∼1012 to 1013 moles oxygen per year) may help explain how Earth's surface environment became irreversibly oxidized.

The rise of atmospheric O2 about 2.4 to 2.2 billion years ago (Ga) (1,2) changed the course of biological evolution. Yet explaining why O2 rose at that time has remained elusive, given that bacterial oxygenic photosynthesis was present hundreds of millions of years earlier, before 2.7 Ga (3) and possibly since 3.8 to 3.5 Ga (2, 4, 5). Oxygenic photosynthesis splits water into O2 and a reductant, H. Hydrogen is used to reduce CO2 for biosynthesis of organic matter. Nearly all photosynthesized organic matter (today, ∼99.9% of ∼9000 × 1012 mol C year−1) recombines with O2 via decay or respiration (6, 7). Conventional thinking has focused on the burial of organic carbon as the means of separating photosynthetic reductant from O2, thereby enabling O2 to accumulate at the surface. However, the small flux of organic carbon that escapes oxidation through burial in sediments [currently ∼1013 mol C year−1(6)] would only cause atmospheric O2 to rise if the burial rate exceeded the rate of O2 consumption by reductants supplied to the atmosphere and ocean by geologic processes. Today, these rates appear balanced, with no atmospheric O2 increase (6). Moreover, atmospheric O2 only increases if reductant that is buried at a preferential rate relative to oxidized material does not later return to the atmosphere or ocean, canceling the O2 gains (e.g., by reduced metamorphic gases or dissolution of uplifted, reduced continental sediments). The early environment was sufficiently reducing to scavenge O2 (2), so reductant had to be removed preferentially relative to oxidized species and irreversibly to oxygenate the environment permanently. However, no consensus theory has yet emerged to explain why O2 rose long after oxygenic photosynthesis evolved (5), and all current hypotheses are problematic (8).

We describe an overlooked biogeochemical mechanism relevant to Earth's redox history: the coupling of early oxygenic photosynthesis to the escape of H to space. H escape provides an alternative to organic burial for removing photosynthetic reductant; H escape is irreversible, whereas metamorphism and continental erosion recycle the reducing power of buried organic matter. In the biosphere, H is transferred from photosynthetic organics to CH4 by methanogenesis. When CH4 is decomposed in the upper atmosphere by ultraviolet (UV) radiation, H escapes to space forever. The overall chemistry is CO2 + 2H2O → CH4 + 2O2 → CO2 + O2 + 4H(↑space), where the first reaction sums photosynthesis and methanogenesis. Currently, Earth gains oxygen by CH4-induced H escape at a negligible rate ∼1010 mol O2 year−1 because the rate depends on the magnitude of the atmospheric mixing ratio of CH4 (f CH4), which today is only 1.7 ppmv.

However, CH4 would have been an important trace atmospheric constituent before the rise of O2. Today, the large biogenic flux of CH4 to the atmosphere is oxidized, limiting f CH4 (9). But in the low-O2 Archean, the kinetic fates of biogenic O2 and CH4 would have been reversed. O2 would have been rapidly consumed and CH4, long-lived. Rapid reaction of O2 with reduced metamorphic and volcanic gases and with upwelling oceanic cations like Fe2+ would have buffered O2 to trace levels (10). Also, organic carbon uplifted onto continents and washed to the ocean would have been consumed aerobically to produce CO2 or anaerobically to make CH4 plus CO2, given that Archean elemental carbon is found in biologically mediated fine-grained shales (fixed from CO2) rather than in detrital form (11). An Archean methanogen biosphere is suggested by biochemistry (12) and carbon isotope evidence (13–15). Photochemical models suggest Archeanf CH4 ∼200 to 3000 ppmv (16–18) if the biogenic CH4production rate were 0.1 to 1 times that of the present.

Abundant atmospheric CH4 is also the most plausible explanation for Archean greenhouse warming (17). A large greenhouse effect is needed to explain the temperate Archean climate when solar luminosity was 20 to 30% lower than today (19). A partial pressure of carbon dioxide (pCO2) a few hundred to 1000 times larger than today has been postulated (20) but is improbable for several reasons. Paleosols indicate that pCO2 was an order of magnitude too low to counter a fainter Sun at 2.75 Ga (21). The mineralogy of banded iron formations also suggests that pCO2 < 0.15 bar at 3.5 Ga (22). Abundant Archean marine limestone indicates calcite supersaturation then, as now (7). If pCO2 were high, oceanic Ca2+ should have been depleted, but evaporitic gypsum (23) suggests otherwise. Also, carbonatization of the seafloor should have lowered pCO2 to levels inconsistent with a dominant greenhouse role (24). Further, Archean geochemical data do not indicate levels of acid weathering expected for pCO2 > 100 times present (25). Consequently, the theory of Archean CH4greenhouse warming (Fig. 1A) has become favored (15–18, 21,24). High CH4 is consistent with relatively low pCO2 because if a large greenhouse enhancement by CO2 were added to warming dominated by CH4, CO2 would be consumed in negative feedback by temperature-dependent weathering of continental silicates. A CH4-mediated climate can be stabilized in negative feedback with O2; e.g., increasingf CH4 causes greenhouse warming, which increases weathering, sedimentation, and, ultimately, organic burial rates. The latter, in turn, increases O2, which lowers f CH4.

Figure 1

(A) The calculated mixing ratio of CH4 (left ordinate axis) needed to maintain a surface temperature of 290 K on early Earth against the lower luminosity of the young Sun. We used the radiative modeling of (17,18). The mean global temperature in the Archean is assumed to be similar to that of the present day, given the absence of extensive glaciation in the Archean and constraints from Archean evaporites (7, 23). CH4 mixing ratios are calculated at three fixed levels of pCO2 as indicated, where PAL indicates present atmospheric level ≈ 0.0003 bar. The upper pCO2 limit, pCO2 = 0.01 bar, is derived from paleosols for 2.2 to 2.8 Ga at 290 K (21) and yields a lower limit on CH4. The irreversible oxidation fluxes due to escape of hydrogen, corresponding to particular levels of CH4, are expressed as molar O2 equivalents per year (right ordinate axis). We end calculations at 2.4 to 2.3 Ga, assuming that CH4 levels collapsed upon the rise of atmospheric O2. (B) Integrated oxidation due to CH4-induced H escape to space, using the three atmospheric CH4 levels from (A), shown with matching labels. Cumulative oxidation and the observed molar oxygen inventory in the continental crust (Table 2) are comparable.

Climatologically important CH4 (Fig. 1A) induces rapid escape of H to space. H escapes from the base of Earth's exosphere (∼300- to 500-km altitude), where H atoms are the only H species (26, 27). Several processes rapidly depopulate H atoms from the exosphere so that diffusive supply of H from lower levels is the rate-limiting step (27). The total concentration of all H-bearing compounds in the lower stratosphere, f total (=f H2O +f H2 + 2f CH4…) (expressed as H2 molecules for these calculations) determines the diffusion-limited H escape rate, φescape, given by (27):Embedded Image Embedded Image(1)Today, φescape is trivial (Table 1) because f totalis small, given only 3 ppmv water vapor, 1.7 ppmv CH4, and 0.55 ppmv H2 in the lower stratosphere. Upward transport of H in H2O, in particular, is limited by a “cold trap” at the tropopause where water condenses. Because such a cold trap is a general feature of paleoatmospheres, oxygen production by abiotic H escape from water vapor can be neglected (7). However, CH4is not cold trapped and increased H escape is unavoidable with increased f CH4. If Archeanf CH4 were ∼1000 ppmv (16–18), φescape would be ∼300 times higher than today's flux (Table 1). Large CH4-induced H escape rates and significant global oxidation rates are general consequences of a high Archeanf CH4 (28).

Table 1

Earth's oxygen fluxes.

View this table:

Escape of H to space oxidizes Earth as a whole. Oxidation is expressed in the geochemical reservoir where the H originates, although the resultant oxidized species may subsequently be transported to other reservoirs. We explain how oxidation results from CH4-induced H escape in three cases: (i) when CH4 originates from organic matter produced by oxygenic photosynthesis, (ii) when CH4 derives from organic matter produced by anoxygenic autotrophic metabolisms, and (iii) when CH4 derives from mantle H.

In (i), oxygen is gained irreversibly because photosynthetic splitting of water produces O2 and H, and CH4-mediated escape removes the H forever. This process is schematically represented in Eqs. 2 through 5. Oxygenic photosynthesis can be summarized asEmbedded Image(2)where CH2O represents organic matter. Production of CH4 mainly derives from symbiotic communities of heterotrophs and methanogens that decompose organic matter (13).Embedded Image(3)H escape to space via CH4 can be represented as follows, noting that the detailed photochemistry (16) is rather more complex.Embedded Image(4a) Embedded Image(4b)Thus, the combined effect of the early biosphere, using the processes of oxygenic photosynthesis (Eq. 2), methanogenesis (Eq. 3), and H escape (Eq. 4), is described by the sequential sum of these processes [(2 × Eq. 2) + Eq. 3 + Eq. 4]. This gives the overall chemical transformation of the crustal system:Embedded Image Embedded Image(5)Consequently, the irreversible gain of oxygen from CH4-induced H escape derives from water split by oxygenic photosynthesis. A more circuitous route to oxygen gain occurs when buried organic matter devolatilizes by diagenesis or metamorphism to produce H2 (e.g., via CH2O + H2O = 2H2 + CO2) or CH4 (via 2CH2O = CH4 + CO2). During the Archean, if methanogens produced CH4 by consuming metamorphic H2, or if CH4 or H2 fluxed directly from decomposed buried organic matter, the net effect of Eq. 5 would still apply.

Case (ii) concerns CH4 originating from anoxygenic photoautotrophs or chemoautotrophs. Such prokaryotes use H2, reduced sulfur, or Fe2+ as electron donors in biosynthesis (e.g., H2S + CO2 + hν → CH2O + H2O + S). If CH4 were derived from such organic matter, H escape would leave behind oxidized S or Fe, contributing to net crustal oxidation (though free O2 is not produced), provided that the electron donor originated from the crust (e.g., metamorphic H2S). If the electron donor fluxed from the mantle, case (iii) would apply.

Case (iii) concerns methanogenic CH4 derived from mantle hydrogen in volcanic gases. Volatile fluxes to the atmosphere have probably been dominated by recycling of crustal sedimentary rocks since the early Proterozoic or earlier via metamorphism or volcanism (29, 24). Volcanic gases derive from magma, whereas metamorphic gases are not directly associated with a silicate melt. Mantle minerals buffer the redox state of volcanic gases, and when H is exported mantle minerals are oxidized to satisfy redox balance (30). Redox-sensitive elements in igneous rocks show that the oxygen fugacity of volcanic gases has not changed by more than 0.5 log10 units since 3.6 Ga (31), presumably because of effective mantle buffering. These data rule out the suggestion that mantle oxidation was an important factor in the rise of O2 (30, 32). Although mantle H can escape directly to space, biogenic CH4 may have helped prevent sequestration of mantle reductant into the Archean crust. If mantle H had transferred reducing power to solids in the crust (e.g., if H were efficiently scavenged by bacteria to reduce CO2 to organic matter), the crust could have become gradually more reduced. However, fermentation of organic matter to CH4 and resultant H escape would allow mantle H to be lost to space.

In all cases discussed above, Earth's overall oxidation state increases. Case (iii) oxidizes the mantle. Cases (i) and (ii) oxidize the crust (e.g., as Fe2O3 or SO4 2–), which, in the long-term, must shift kinetics to favor the survival of free O2. Free O2 is only produced in case (i), which includes oxygenic photosynthesis. Oxygenic photosynthetic bacteria extract H from water, making them independent of abiotic sources of reductants; they would have dominated global productivity once they evolved (33), rendering the other cases inconsequential for effecting significant crustal redox changes. Biogenic CH4 would be the major H-bearing species in the Archean stratosphere (16–18), so Eq. 1 can be rewritten withf total ≈ 2f CH4 Embedded Image(6)Thus, iff CH4 in the Archean atmosphere were ∼100 to 2000 ppmv (Fig. 1A), the effective flux of O2 into the crust due to CH4-induced H escape would be (0.7 to 14) × 1012 mol year−1. This rate is comparable in magnitude to the (reversible) modern O2 flux due to organic burial of ∼1013moles O2 year−1 (Table 1) and would produce (0.7 to 14) × 1021 mol O2 in ∼109 year, comparable to the continental crustal reservoir of excess oxygen (Table 2).

Table 2

Oxidized and reduced reservoirs in Earth's continental crust. The Earth's exterior contains Fe2O3 and SO4 2−, which arose via oxidation, and atmospheric O2. Oxidized species are expressed in terms of the O2 moles required for their production; e.g., each mole of Fe3+ needed 1/4 mole O2 to be produced from Fe2+. Reduced species are expressed in terms of O2 moles required for their consumption. By, billion years; R OX, oxygen in the continental crust; AOS, atmosphere, ocean, and sedimentary;R AOS, oxygen in the AOS system;R redC, reduced carbon in the continental crust.

View this table:

Large oxygen inventories include the continental crust (Table 2) and mantle. The continental crust's excess oxygen mostly resides in altered and metamorphosed igneous rocks. Archean basalts have a weight ratio Fe3+/ΣFe several times greater than fresh basalt, for which Fe3+/ΣFe ∼ 0.07. Metamorphic oxidation of crustal ferrous minerals by water alone requires extreme volumes of water (e.g., ∼1500-g water per 1-g magnetite to oxidize magnetite to hematite at 5 kbar and 630°C), so SO4 2– or O2 are often implicated as oxidants whenever Fe3+ is observed to increase (34). Transfer of the oxidizing power of SO4 2– to Fe3+ is consistent with low SO4 2– in Archean oceans relative to today's oceans (35). In the ocean, continuous oxidation facilitated by CH4-induced H escape would have produced Fe3+ from oceanic Fe2+. Thus, O2would have been exported to the mantle through past subduction of Fe3+ (Table 2). Ferric oxides are denser than mantle material with a refractory tendency for deep subduction (36), so it is probable that surviving Archean iron formations are a mere fraction of those originally deposited.

The oxidation caused by H escape for greenhouse CH4levels (Fig. 1A) can be integrated over time and compared with the crustal oxidized inventory (Table 2). For thef CH4 required for warming the early Earth, cumulative oxidation [Fig. 1B, curves (i) and (ii)] is consistent with estimates of the continental crust's inventory of oxygen (Table 2).

Net oxidation of crustal rocks in the past would have increasingly enhanced the kinetic stability of atmospheric O2. Today, most degassed carbon volatiles are recycled via metamorphism rather than volcanism. The ratio of metamorphic to volcanic gas fluxes has likely increased through time (7, 24). Thus, models that equate the early Earth's H escape flux to fluxes of reductant from the mantle (30, 32) are incorrect. These models neglect metamorphic and continental sources of reductant, providing no explanation for the net oxidized state of crustal reservoirs in Table 2. If crustal volatile recycling dominates, to first order CH4-induced H escape to space would oxidize the crust by Eqs. 5 and 6. Because the residence time of Archean O2 would be small (10, 16), O2 would be sequestered into oxides (e.g., Fe2O3, SO4 2–, CO3 2−). Unlike volcanic gases, the average oxidation state of metamorphic gases is independent of mantle buffering and is controlled largely by the oxidation state of the original sediments (34, 37). Thus, the oxidation state of Archean metamorphic gases would have increased over time as crustal rocks became more oxidized (37). Reductants released by metamorphism (H2, CO, H2S, etc.) would remove atmospheric O2, enabling high f CH4 (10) and rapid H escape. Oxidation resulting from such H escape would be expressed inside the crust where the reductant originated. The surface would remain weakly reducing, although river fluxes of reductant to the ocean from weathering would presumably have declined with increasing oxidation of uplifted rocks. However, the details of metamorphic or weathering redox changes are superfluous: Le Châtelier's principle demands that atmospheric and oceanic O2 sinks decrease as the crust is irreversibly oxidized via CH4-induced H escape. This is consistent with the prevalence of methanotrophs in the late Archean using increasing levels of dissolved SO4 2– or O2(12). Then, in the early Proterozoic, peak iron formation deposition occurred (7) and sulfate-reducing bacteria became increasingly ubiquitous (35).

That the crust is at a higher oxidation state than the mantle from which it was derived suggests irreversible oxidation. Crustal oxygen fugacity varies by orders of magnitude from fayalite-magnetite-quartz (FMQ) to hematite-magnetite (HM) buffer levels (38), whereas the upper mantle is near FMQ (39). Furthermore, photosynthesis produces organic carbon balanced by oxides of sulfur and iron (after loss of O2), so buried organic carbon should balance oxidized materials in the crust if no H escape occurred. However, estimates of the continental crustal inventory show that oxidized species exceed reduced carbon by a factor of 1.5 to 2.2 (Table 2). This budget excludes oxidized carbon, some of which may have started out reduced; i.e., when reduced carbon delivered by impact bombardment during 4.4 to 3.8 Ga is subtracted from the reduced inventory, the dominance of oxidized species increases. Time-integrated subducted losses of Fe3+ to the mantle may also further increase the redox imbalance (Table 2). A greater oxidized versus reduced inventory can be reconciled only by H escape, or preferential subduction of organic carbon relative to oxidized species, or both. Subduction of 12C-enriched graphitic carbon relative to carbonate is unlikely because marine carbonates do not become increasingly 12C-depleted with geologic time (40). But we cannot discount enhanced subduction of graphitic carbon relative to subduction of oxidized species as a whole. Any irreversible, preferential loss of reductant into the mantle would be identical in its crustal redox effect to H escape to space. Nonetheless, oxidation due to CH4-induced H escape is chemically expected and can reconcile the observed redox inventory on its own. Other geochemical evidence of CH4-induced H escape may reside in low values of D/H (deuterium/hydrogen) inferred for Archean seawater (41).

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


View Abstract

Stay Connected to Science

Navigate This Article