Putting the squeeze on oxidation

The results of experiments conducted under the extreme temperature and pressure conditions of Earth's lower mantle suggest that, at these depths, low oxygen content is no barrier to iron oxidation.

The oxidation of iron is a process with which we are all familiar. The surface of Mars, the rust on our cars, the blood in our veins — all bear the conspicuous red colour that marks the oxidation of iron from its metallic (Fe), or divalent (Fe²⁺), form to a trivalent cationic species (Fe³⁺). At atmospheric pressure, the process of iron oxidation in the presence of oxygen depends on the partial pressure of gaseous oxygen (often referred to in geological literature as oxygen fugacity). Earth scientists have used the same thermodynamic formulations, with great success, to calibrate the oxidation state of iron in the rocks and minerals of Earth's upper mantle, at much higher pressures and temperatures¹. Now Frost et al.² (page 409 of this issue) show that, when it comes to the more extreme pressure conditions in the lower mantle, we must abandon our conventional thinking about oxidation state.

Geology students are taught that certain minerals — such as (Mg,Fe)Al₂O₄, or ‘spinel’ — serve as sensors of oxidation state because their Fe³⁺/Fe²⁺ ratios are known, experimentally and thermochemically, to depend on oxygen partial pressure. From such ‘geobarometry’, we have learnt that Earth's upper mantle is much too oxidized to be in chemical equilibrium with metallic iron¹ — the abundance of Fe³⁺ is too great for metallic iron to be stable. In contrast, we are quite confident that the mantle as a whole was in chemical equilibrium with metallic iron as the planet's core formed (when Earth was less than 50 million years old)³. How the uppermost mantle, the part we can sample, became so oxidized is unknown.

The portion of the mantle at depths greater than 660 km, referred to as the lower mantle, is thought to be dominated (70–80% by weight) by a dense, magnesium–iron silicate, (Mg,Fe,Al)(Si,Al)O₃. This mineral has a stable ‘perovskite’ crystal structure (a structure well known in ceramics to have an array of rich and unusual properties). The oxygen content of the lower mantle is unknown because we have no direct rock samples, but it is usually assumed to be similar to that of the upper mantle. In 1997, McCammon⁴ reported the surprising discovery that magnesium–silicate perovskites synthesized in the laboratory can have very high Fe³⁺/Fe2⁺ ratios — much higher, for example, than are preserved by spinel in the relatively oxidized upper mantle. However, the relationship between the oxygen partial pressure and the Fe³⁺/Fe²⁺ ratio in perovskite remained unclear because of the difficulty of knowing (let alone controlling) the oxygen content in experiments at extreme conditions.

Frost et al.² have performed experiments in which they brought magnesium perovskite into chemical equilibrium with metallic iron under pressure–temperature conditions typical of the uppermost lower mantle. They made the remarkable discovery that, with very low oxygen content in their samples, the perovskite synthesized can have surprisingly high Fe³⁺/Fe²⁺ ratios — similar, in fact, to the ratios measured in perovskites synthesized at very high oxygen contents. This decoupling of oxygen content from the Fe³⁺/Fe²⁺ ratio shatters our traditional view of oxidation state based on oxygen partial pressure.

This result leads Frost et al.² to predict that a free iron-metal phase should exist in Earth's lower mantle — only 1% by weight, but nevertheless an important component. Ironically, this metal phase forms because of the high Fe³⁺/Fe²⁺ ratio in perovskite, rather than being excluded by it. For a mineral to achieve a high Fe³⁺/Fe²⁺ ratio, oxygen must be available to convert FeO into Fe₂O₃ (Fe²⁺ to Fe³⁺). This could happen in the mantle if some chemical reaction serves as an oxygen pump: for example, the reduction of CO₂ to form C and O₂. An alternative mechanism proposed by Frost et al. is that magnesium perovskite has such an affinity for Fe³⁺ that it is energetically favourable for FeO in perovskite to ‘self-reduce’ by the reaction 3Fe²⁺O → Fe (metal) + 2Fe³⁺O_1.5. In this case, no external oxygen pump is needed and a free iron-metal phase forms out of the necessity for mass balance.

The lower mantle constitutes a large volume of the Earth, about 58%, so even a small amount of metallic iron represents a substantial reservoir — several per cent of all metallic iron in the planet. The lower mantle may first have formed because of the increase in pressure as the solid Earth grew, or it may have crystallized from a deep magma ocean. The core formed very early in Earth's history by the segregation of metal from silicate, a large-scale differentiation event that substantially decreased the Fe/O ratio of the silicate mantle. This ratio, dictated by metal–silicate equilibrium, would have been considerably higher than the current Fe/O ratio in the upper mantle. Frost et al.² calculate that if about 10% of the perovskite-manufactured free-metal phase were mechanically removed from the lower mantle through interaction with descending plumes of iron during core segregation, the mantle could have acquired its present Fe/O ratio. Thus, Frost and colleagues' discovery may explain the vexing question of why the upper mantle seems to have had such a high oxidation state for the past 4 billion years⁵.