Atmospheric oxygenation and volcanism

Abstract

Arising from F. Gaillard, B. Scaillet & N. T. Arndt Nature 478, 229–232(2011)10.1038/nature10460

Around 2.5 billion years ago, Earth’s atmosphere turned from anoxic to oxic in what is known as the Great Oxidation Event. Gaillard et al.¹ suggest that this oxygenation was caused by the emergence of the continents and a shift in volcanism from predominantly submarine to primarily subaerial conditions. Because the ratio of volcanic SO₂ to H₂S in their model increases with this shift, they argue that the atmosphere became more oxidized. But their model also predicts that outgassing of CO₂ decreases at the lower pressure of continental volcanism, and that this should act against atmospheric oxygenation because CO₂ is the substrate for the production of oxygen from photosynthesis. Hence, their mechanism may not trigger a rise in atmospheric O₂.

Main

There are three problems to consider. First, Gaillard et al.¹ argue that the release of SO₂ comes at the expense of H₂S, according to:

But the extra H₂ released by this reaction exactly compensates for the switch from H₂S to SO₂. The mixture of SO₂ and H₂ has the same redox potential as H₂S and water². Fortunately, what happens in their model is different. The H₂S to SO₂ ratio is buffered by magma, which has a fixed oxidation state, described in terms of a fixed oxygen fugacity. The relevant reaction is:

where ‘O₂’ represents oxidants (mostly ferric iron) in the melt. Reaction (2) produces SO₂ and H₂ in equimolar proportions, in agreement with results shown in figure 2 of ref. 1. When SO₂ is reduced to pyrite, FeS₂, it consumes hydrogen as follows:

Because reaction (3) consumes more hydrogen per mole of SO₂ than is produced by reaction (2), the net effect is to oxidize the atmosphere. Thus, this part of their proposed atmospheric oxidation mechanism appears to work, but not for the reasons they stated.

The second issue is that sulphur-containing gas fluxes are not the only ones that control O₂. CO₂ is crucial because some CO₂ is reduced to organic carbon, CH₂O, which is then buried. A mole of photosynthetic O₂ is released for every mole of buried CH₂O. However, O₂ can be lost in atmospheric reaction with H₂, so the net effect is:

The carbon isotope record suggests that ∼20% of outgassed CO₂ is reduced and buried in this manner^3,4. Holland⁴ quantified the effect of volcanic gases on the hydrogen budget by defining a parameter denoted f. Gases for which f > 1 can reduce 20% of the CO₂ to organic matter and all of the SO₂ to pyrite, leaving excess hydrogen to support an anoxic atmosphere; gases with f < 1 contain too little hydrogen to do this. Atmospheric O₂ ‘wins’ over hydrogen when f drops below unity. This analysis neglects other sources of hydrogen that might have been important (for example, serpentinization of ultramafic rocks on the continents and seafloor, or precipitation of ferric iron in banded iron formations), but it is useful for analysing the volcanic gases of ref. 1.

In the model of Gaillard et al.¹, the proportion of CO₂ in released volcanic gases decreases by a factor of ∼3 when volcanoes switch from submarine to subaerial conditions because the pressure decreases from 100 bar to 1 bar. This change should tend to oppose atmospheric oxygenation, as less organic matter is buried (given a constant burial fraction), and thus less H₂ is consumed. To estimate the net effect on atmospheric redox balance, we calculated f values for cases listed in the Supplementary Information of ref. 1. Gaillard et al.¹ include S₂ as a species, which requires us to modify Holland’s f parameter, as follows:

Here, m(i) is the mole fraction of species i in the released gases. As shown in Table 1, f is less than unity in the cases shown in Supplementary Tables 1 and 2 of Gaillard et al.¹, and so would not allow a reduced atmosphere before 2.4 Gyr ago. This means that no rise of O₂ is possible because the initial atmosphere is already oxidized. This could be because other hydrogen sources are neglected, as mentioned above, or because model parameters—for example, the outgassing temperature—are incorrect. In any case, the model, as presented in ref. 1, appears incapable of triggering a rise in atmospheric oxygen if Holland’s method of analysis is correct.

Table 1 f values for different cases from ref. 1

Last, Gaillard et al.¹ assume that gases released from a submarine magma at 1,300 °C do not re-equilibrate with the surrounding basalts at lower temperatures as they emerge from depth; this assumption is at odds with previous work^5,6. Such re-equilibration might have allowed Archaean volcanic gases to be more reduced; hence, a switch to subaerial volcanism around 2.5 Gyr ago might then result in atmospheric oxidation. But this would be a different oxidation mechanism from the one described in ref. 1. There is also a broader caveat. The reduced oxidation state of volcanic gases comes at the expense of oxidizing the source rocks; thus, erosion and subsequent reduction of the source rocks may cancel out surficial redox change in the long term.