Magma May Have Been a Hidden Oxygen Carrier in Earth’s Early Atmosphere

October 23 — Ask where the oxygen in Earth’s atmosphere came from and most people will say photosynthesis. But a new study proposes that, at a key moment in Earth’s history, magma itself may also have acted as an important carrier of oxygen.

A high-pressure experimental and modelling study by Chinese researchers, published in National Science Review, shows that silicate melts under high pressure can hold far more “excess oxygen” than traditional stoichiometry predicts. This oxygen is gradually released as magma rises and pressure falls, offering a previously unconsidered deep pathway for understanding the first big rise in atmospheric oxygen roughly 2.5 to 2.0 billion years ago.

Traditionally, geoscientists calculate the oxygen content of magma from the abundance and valence of major elements such as silicon, magnesium and iron. Oxygen makes up more than half of all atoms in the silicate mantle, but most of it is locked into chemical bonds, and its amount has been thought to be entirely determined by the “recipe” of those cations.

Unlike water and carbon dioxide, oxygen itself has long not been seen as a volatile component that can dissolve independently in magma. A team led by researcher Lin Yanhao and Academician Mao Ho-kwang of the Center for High Pressure Science and Technology Advanced Research (HPSTAR), joined by Professor Niu Haiyang of Northwestern Polytechnical University and Professor Wim van Westrenen of VU Amsterdam, directly tested that default assumption.

In the lab the team synthesised mid-ocean-ridge basalt and ran melting experiments at 1,200 to 1,800 degrees Celsius and from 1 atmosphere to 7 gigapascals. Seven gigapascals, roughly a million pounds per square inch, approaches conditions in parts of the upper mantle. To create an oxygen-rich environment they added platinum oxide to some samples, then rapidly cooled the melt into glass and measured its elemental composition, water content and carbon dioxide content, separating out the “stoichiometrically required oxygen” and the oxygen in volatiles; the remainder was defined as “excess oxygen.”

The results show a clear dependence on pressure: at 1 atmosphere excess oxygen is close to zero; at 7 gigapascals it exceeds 5% of the sample’s weight, at precisely 5.2 ± 0.2 wt.%. As a control, equivalent temperature and pressure experiments without platinum oxide produced only 0.2% to 0.5 wt.% excess oxygen.

A key question is whether this extra oxygen enters the melt’s structure or is merely physically trapped as bubbles. Raman spectroscopy found no gaseous oxygen signal in the glass, suggesting that the excess oxygen is incorporated into the silicate network.

Simulations by Niu Haiyang’s team based on a deep-neural-network interatomic potential further explain the storage mechanism: as oxygen content rises, the charge distribution of oxygen atoms broadens, and a small fraction of them form peroxide-type bonds with each other. The simplified magnesium-silicon-oxygen system used in the simulations gives lower absolute oxygen contents than the basalt experiments, but reproduces the same trend of rising oxygen storage with pressure. Because some oxygen may escape during quenching, the measured 5% figure is very likely only a lower bound on the melt’s actual storage capacity.

The finding’s significance extends to the surface. Magma formed in relatively oxidised regions of the mantle can load large amounts of excess oxygen at depth; as it moves upward and pressure drops, the melt’s capacity to hold oxygen falls, and the surplus is released, oxidising the shallow mantle, the crust and the atmosphere.

The authors note that this mechanism echoes an overlap in the geological record: atmospheric oxygen rose markedly during the Great Oxidation Event (GOE) roughly 2.5 to 2.0 billion years ago, and some of the largest large igneous provinces (LIPs) also erupted on a massive scale in a similar time window.

A rough estimate based on the experimental data suggests that if the magma erupted in that period had released just 0.27% of its weight in excess oxygen, that would be enough to supply all the oxygen in today’s atmosphere. That number is not demanding compared with the storage capacity measured at 7 gigapascals — but note that 0.27% is an order-of-magnitude inference for global-scale magmatism, not a direct measurement of any single eruption, and how well it matches real geological processes remains to be tested.