Introduction
For the first two billion years of Earth's history, the atmosphere contained essentially no free molecular oxygen. The oceans were rich in dissolved iron (Fe²⁺), hydrogen sulfide (H₂S), and other reduced compounds; any organism that could exploit these electron donors for photosynthesis had an enormous metabolic advantage. The earliest photosynthesisers — appearing by at least ~3.5 billion years ago (Ga) — practised anoxygenic photosynthesis, stripping electrons from H₂S, Fe²⁺, or organic molecules rather than water. Crucially, these reactions produce no oxygen byproduct: the electron donors are consumed silently, leaving behind sulfur, oxidised iron, or CO₂, but not O₂. Life had been harvesting sunlight for hundreds of millions of years without altering the redox state of the atmosphere.
That changed with the evolution of oxygenic photosynthesis in cyanobacteria, estimated to have first appeared by ~2.7 Ga. Cyanobacteria evolved a unique two-photosystem reaction centre — Photosystem I coupled with Photosystem II — capable of splitting water (H₂O) as the electron donor. The reaction is thermodynamically demanding: water is a far more stable, less reactive molecule than H₂S or Fe²⁺, requiring a powerful oxidant (the oxygen-evolving complex in PSII, centred on a manganese cluster) to extract electrons from it. But water is essentially unlimited in abundance, freeing oxygenic photosynthesisers from dependence on scarcer reduced substrates. The byproduct of water-splitting is molecular oxygen — O₂ — released as a waste gas with every photosynthetic cycle.
Initially, the O₂ produced by early cyanobacteria was immediately consumed by chemical reactions with the abundant reduced species in the oceans and atmosphere. Dissolved Fe²⁺ oxidised to Fe³⁺ and precipitated as iron oxides; H₂S was oxidised to sulfate; methane was oxidised to CO₂ and water. The geological record captures this buffering phase in banded iron formations (BIFs) — laminated sedimentary rocks containing alternating iron-rich (magnetite, hematite) and silica-rich layers. BIFs are found globally in rocks from ~3.5 Ga to ~1.8 Ga, with the richest deposits concentrated between ~2.6 and ~1.8 Ga. Their pattern of deposition records the ongoing interaction between photosynthetic O₂ production and dissolved ocean iron: as cyanobacterial blooms produced O₂, Fe²⁺ oxidised and precipitated as iron oxides, forming the iron-rich laminae; when blooms waned or iron supply increased, silica-rich layers were deposited. BIFs are thus a geological archive of the oxygen revolution unfolding in slow motion.
By approximately 2.4 Ga, the accumulated production of O₂ by cyanobacteria finally overwhelmed Earth's geochemical buffering capacity. The oceanic iron and sulfide sinks were saturated; crustal oxidation could no longer keep pace. Atmospheric O₂ began to rise above a few parts per million for the first time in Earth's history — an event called the Great Oxidation Event (GOE). The evidence for the GOE timing is multi-pronged and compelling. The most precise indicator is the disappearance of mass-independent sulfur isotope fractionation (MIF-S) in the rock record at ~2.4 Ga. Photochemical reactions of SO₂ in an anoxic atmosphere — by UV photolysis — produce sulfur compounds with anomalous isotopic compositions (departures from mass-dependent fractionation in ³³S) that are preserved in sedimentary sulfides and sulfates. Once atmospheric O₂ rose above ~10⁻⁵ of present atmospheric level, the UV-shielding ozone layer began to form, shutting down these photochemical reactions and erasing MIF-S from the record. Its disappearance is a precise chemical clock for the onset of the GOE. Additional evidence includes: the disappearance of detrital pyrite and uraninite (minerals that oxidise rapidly in an oxic atmosphere) from riverbeds after 2.4 Ga; the first appearance of continental red beds (iron-oxide-cemented sedimentary rocks indicating oxidising weathering conditions); and oxidised paleosol horizons showing that soil surfaces were exposed to free oxygen. The BIF record itself largely ends by ~1.8 Ga, when the deep ocean was sufficiently oxygenated that dissolved Fe²⁺ could no longer accumulate.
The GOE did not simply add oxygen to an otherwise unchanged world — it transformed the entire Earth system. Methane (CH₄), which had been a significant atmospheric component produced by methanogenic archaea in the anoxic early ocean, was chemically destroyed by reaction with O₂. Methane is a potent greenhouse gas, and its drawdown triggered a severe planetary cooling: the Huronian glaciation (~2.4–2.1 Ga), one of the most intense ice ages in Earth's history, likely extending glaciation to equatorial latitudes. This episode may represent Earth's first approach to a global "Snowball Earth" state.
Oxygenation was not a single event. Oxygen levels after the GOE rose only to perhaps 1–10% of present atmospheric level, then fluctuated over billions of years. A second major oxygenation occurred during the Neoproterozoic Oxygenation Event (NOE) around ~600–800 Ma, when O₂ rose to levels sufficient to support active animal metabolism (~10–20% of present atmospheric level). It is not coincidental that the first complex multicellular animals (metazoans) appear in the fossil record shortly afterward: aerobic respiration is far more energetically efficient than anaerobic metabolism, yielding ~18× more ATP per glucose molecule, enabling the high metabolic rates required by mobile, complex organisms.
The long-term persistence of atmospheric O₂ depends on the carbon cycle. Photosynthesis alone does not produce a net increase in atmospheric O₂: if all organic matter were immediately remineralised by respiration or decomposition, O₂ consumed would equal O₂ produced and no net change would occur. Atmospheric O₂ accumulates only when a fraction of photosynthetically fixed organic carbon is buried in sediments before it can be oxidised back to CO₂. Each mole of organic carbon buried represents one mole of O₂ permanently released to the atmosphere. Over geological time, the rate of organic carbon burial versus remineralisation has controlled O₂ levels. The Carboniferous period (~300 Ma) saw unusually high O₂ concentrations (~30–35%) linked to massive burial of organic carbon from the first lignin-rich forests, before fungi and bacteria that could decompose lignin became widespread.
The endosymbiotic origin of chloroplasts was the second great step in the photosynthetic revolution. All eukaryotic photosynthesisers — algae, land plants — possess chloroplasts that are phylogenetically cyanobacterial, acquired by engulfment of a free-living cyanobacterium by an ancestral eukaryotic host cell approximately 1.5 Ga. The host and endosymbiont became metabolically integrated; the cyanobacterial genes were progressively transferred to the host nucleus, and the cyanobacterium became a semi-autonomous organelle incapable of independent existence. This event democratised oxygenic photosynthesis across the domain Eukarya and dramatically expanded the diversity and productivity of photosynthetic life.
For astrobiology, O₂ occupies a special position as a biosignature gas. Abiotic processes can produce small amounts of O₂ (e.g., photolysis of CO₂ or H₂O), but the concentrations and isotopic compositions found in Earth's atmosphere overwhelmingly require a biological source sustained over geological time. The James Webb Space Telescope and future missions aim to detect O₂ and its photochemical product ozone (O₃) in the atmospheres of rocky exoplanets in habitable zones. A detection of O₂ or O₃ — particularly in combination with a reducing gas like CH₄ (which would be rapidly destroyed by O₂ in the absence of a continuous biological source) — would constitute extraordinarily strong evidence for life.
Key Terms
The geologically abrupt rise of free molecular oxygen (O₂) in Earth's atmosphere approximately 2.4 billion years ago, driven by cyanobacterial oxygenic photosynthesis overwhelming the planet's geochemical oxygen sinks (dissolved iron, sulfide, crustal minerals). Evidenced by the disappearance of mass-independent sulfur isotope fractionation (MIF-S), the end of detrital pyrite and uraninite in riverbeds, the first appearance of continental red beds and oxidised paleosols, and the decline of banded iron formations. The GOE permanently transformed Earth's surface chemistry, atmosphere, and biosphere.
A phylum of gram-negative bacteria that evolved oxygenic photosynthesis — the ability to use water (H₂O) as an electron donor, splitting it via Photosystem II and releasing O₂ as a byproduct. The oldest unambiguous fossil cyanobacteria date to ~2.1 Ga, with molecular clock and biomarker evidence extending their origin to ~2.7 Ga. Cyanobacteria were the primary producers responsible for the Great Oxidation Event and are the prokaryotic ancestors of all eukaryotic chloroplasts via endosymbiosis.
Laminated Precambrian sedimentary rocks consisting of alternating iron-rich (magnetite, hematite, siderite) and silica-rich layers, typically deposited in marine basins between ~3.5 and ~1.8 Ga with a peak between ~2.6 and ~1.8 Ga. BIFs record the episodic oxidation of dissolved ferrous iron (Fe²⁺) by photosynthetically produced O₂: Fe²⁺ + O₂ → Fe³⁺ oxides that precipitate. They constitute the world's largest iron ore deposits (e.g., Hamersley Basin, Australia; Transvaal, South Africa) and are a direct geological archive of the oxygenation of the early ocean.
Anomalous departures from mass-dependent fractionation in the ratios of the four stable sulfur isotopes (³²S, ³³S, ³⁴S, ³⁶S), specifically in ³³S (reported as Δ³³S ≠ 0), caused by UV photochemical reactions of SO₂ in an anoxic atmosphere. MIF-S signals are preserved in sedimentary sulfides and sulfates older than ~2.4 Ga but are absent in younger rocks. Their disappearance marks the onset of the GOE: once atmospheric O₂ rose above ~10⁻⁵ of present levels, an ozone layer formed that shielded the troposphere from UV, shutting down MIF-S-producing photochemistry. MIF-S is therefore a precise geochemical proxy for atmospheric anoxia.
An evolutionary process in which one organism lives inside another in a mutually beneficial intracellular relationship, eventually becoming an organelle. The primary endosymbiotic event relevant to photosynthesis occurred approximately 1.5 Ga when an ancestral eukaryotic cell engulfed a cyanobacterium that, rather than being digested, became integrated as the chloroplast. Evidence: chloroplasts have a double membrane (the inner from the cyanobacterium, the outer from the engulfment vesicle), remnant cyanobacterial DNA, ribosomes similar to bacterial 70S type, and phylogenetic analyses showing chloroplast genes cluster with cyanobacteria. Secondary endosymbiosis, in which a non-photosynthetic eukaryote engulfs a photosynthetic one, gave rise to many algal lineages.