Nearly two billion years before engineers built the first nuclear reactor, water seeped through uranium-rich rock beneath what is now Gabon and helped start a self-sustaining chain reaction. The deposit heated, boiled away much of its own moderator and shut itself down. As the rock cooled, groundwater returned and the reaction began again.
This was not a single underground event. Isotopes preserved in minerals at Oklo record a reactor operating in a repeating rhythm: roughly 30 minutes active, followed by about two and a half hours dormant. The cycle may have continued over immense stretches of time, turning an ordinary ore body into the only known natural nuclear fission reactor system on Earth.
“Switched itself on” is a useful shorthand, but no external trigger was thrown. A rare combination of isotope abundance, ore concentration, geological geometry and groundwater allowed enough uranium nuclei to split and release neutrons that caused further fissions. Once conditions crossed the threshold known as criticality, the chain reaction sustained itself.
The discovery began with missing uranium
The Oklo phenomenon was uncovered in 1972 after technicians examining uranium ore from Gabon noticed that its isotopic composition was slightly wrong. Natural uranium found almost everywhere today contains about 0.72 per cent uranium-235, the isotope that readily sustains a fission chain reaction. Some Oklo material contained less.
The discrepancy was small in percentage points but enormous in meaning. The International Atomic Energy Agency’s history of the discovery describes one high-grade sample containing 0.717 per cent uranium-235 instead of the expected 0.720 per cent. Later measurements found much stronger depletion in particular reactor zones.
Missing uranium-235 could have suggested contamination or artificial processing. Instead, complementary analysis revealed the characteristic products left by nuclear fission. The ore was natural, but part of its fissile uranium had been consumed long before the mine opened.
A 1976 US Geological Survey measurement found a uranium-235 abundance of 0.54 atomic per cent in an Oklo sample. Decades later, researchers reported an even lower value of about 0.366 per cent in reactor zone 13. Patterns among neodymium, xenon, ruthenium and other isotopes supplied further evidence that sustained fission had occurred.
Natural uranium was different two billion years ago
The decisive ingredient was time. Uranium-235 and uranium-238 are both radioactive, but uranium-235 decays much faster. Its half-life is about 704 million years, compared with roughly 4.5 billion years for uranium-238. Looking backwards nearly two billion years, uranium-235 therefore formed a much larger fraction of natural uranium than it does today.
At the time the Oklo reactors operated, uranium-235 made up around 3 per cent of natural uranium, similar to the enrichment level used by some modern light-water reactor fuels. The isotope study of Oklo reactor zone 13 places the ancient abundance near 3 per cent and dates the broader phenomenon to about two billion years ago.
High abundance alone was not enough. Uranium also had to be concentrated into deposits thick and rich enough that too many neutrons would not escape. The rocks needed relatively few elements that strongly absorb neutrons without fissioning. Finally, water had to enter the ore.
Geological research has identified multiple natural reaction zones in the Oklo and nearby Okelobondo deposits, with another at Bangombé. A review of their formation and evolution describes how geological and chemical processes concentrated uranium and created the conditions for criticality around 1.95 billion years ago.
Water made the chain reaction possible
When a uranium-235 nucleus fissions, it releases energy and fast neutrons. Those neutrons are more likely to trigger further fissions after they have been slowed. In reactor physics, a material that reduces neutron speed is called a moderator. At Oklo, groundwater filling pores and fractures in the ore performed that role.
With enough water present, more neutrons were slowed into an energy range where uranium-235 captured them and split. Each fission released additional neutrons, allowing the chain to continue. The water did not supply the nuclear energy. It controlled whether the geometry and neutron energies were favourable enough for fission to sustain itself.
Fission heated the ore and groundwater. As the temperature rose, liquid water turned to steam and left the active zone. With less water available to moderate neutrons, the reactor became subcritical and the chain reaction largely stopped. Heat then dissipated into the surrounding rock. Groundwater slowly returned, restored neutron moderation and brought the reactor back towards criticality.
This negative feedback made the system self-regulating. More fission produced more heat, which removed the water needed to maintain fission. The result was not a bomb or a runaway molten core. It was a low-power geological reactor repeatedly throttled by its own temperature.
Xenon preserved the reactor’s clock
The strongest evidence for the pulse length came from xenon isotopes produced through radioactive decay chains following fission. Xenon is a noble gas and might be expected to escape. At Oklo, however, researchers found unusual isotope patterns preserved in aluminium phosphate minerals associated with the reactor rock.
In 2004, Alex Meshik and colleagues used laser extraction and mass spectrometry to analyse those patterns. Their paper in Physical Review Letters concluded that the reactor ran in active pulses lasting about 30 minutes, separated by dormant periods of approximately two and a half hours.
The interpretation resembles a geyser cycle. During the active phase, rising heat converted unbound groundwater into steam and reduced the thermal-neutron flux. During the dormant phase, the rock cooled and water re-entered. A contemporary report in Nature described how the xenon distribution supported this water-controlled cycling rather than continuous operation.
The three-hour rhythm is a reconstruction from isotope ratios, mineral behaviour and reactor physics. Scientists did not find a fossil clock or a direct temperature record. Models of particular zones and later operating stages do not require every part of the Oklo system to have followed exactly the same cycle throughout its lifetime. The 30-minute and two-and-a-half-hour values describe the best-supported pulsed mode recorded in the analysed material.
A reactor that ran for geological timescales
Individual reactor zones appear to have operated over periods ranging from tens of thousands of years to perhaps a million years, gradually consuming uranium-235 and accumulating fission products that affected the neutron balance. Average power estimates were modest, often compared with a small research reactor rather than an electricity station.
Yet modest power sustained over such long periods produced a substantial nuclear fingerprint. Researchers have used Oklo to study how fission products and elements related to nuclear waste migrate through rock. The site is not a perfect substitute for a modern engineered repository, because its geology and chemical history are unique, but it provides a rare natural experiment extending across almost two billion years.
A 2018 investigation of caesium and barium found evidence that some fission products moved and were captured in surrounding minerals after reactor shutdown. Other elements remained surprisingly local. The complicated record is more informative than a simple claim that all waste stayed perfectly fixed.
Oklo has also been used to test whether fundamental physical constants could have changed over cosmic time. Nuclear reaction energies depend sensitively on the strength of interactions between particles. Comparing ancient isotope ratios with modern nuclear data can therefore place tight limits on possible variation, although the result depends on modelling the reactor’s neutron spectrum and temperature.
The deeper lesson is how narrowly nature had to thread the conditions. The uranium had the right ancient isotope ratio. Geological processes concentrated it into suitable ore bodies. Groundwater arrived to moderate neutrons, and heating removed that same water before the reaction could intensify indefinitely. Today, natural uranium contains too little uranium-235 for the same arrangement to work readily with ordinary water.
What remains at Oklo is not a functioning reactor but a mineral archive. Its depleted uranium and rearranged fission products reveal a chain reaction that began without machinery, valves or human design. For long spans of the distant Proterozoic, the deposit repeatedly warmed, dried, cooled and flooded, pulsing to a rhythm that scientists recovered nearly two billion years later from atoms locked inside stone.
