Curiosities / Episode 03

The Reactor Nobody Built

How did a uranium deposit run a nuclear chain reaction roughly two billion years before people could build one?

Oklo, Gabon. In 1972, a small discrepancy in an analysis opened a window into the rock’s distant past.

Examine the mismatch ↓
Water, rock and an ancient reactionOriginal cross-section schematic: groundwater passes through a uranium-bearing rock layer. Neutron paths branch within it. It is not the measured geometry of an Oklo reactor.OKLO / GABONA REACTION WRITTEN INTO ROCK≈2 BILLION YEARS AGO
Original schematic of water and uranium-rich rock, not a measured cross-section.
01 / A result that didn’t quite fit

A little uranium was missing.

In 1972, an analysis of uranium from Gabon came back with less uranium-235 than expected. U-235 is one form of uranium, especially useful for sustaining a nuclear chain reaction.

The difference was small. The IAEA’s account gives a rounded comparison of 0.720% for ordinary natural uranium and 0.717% for the anomalous material.

A tiny shortfall does not explain itself. The next clue was what else the rock contained.

The IAEA’s account of the discovery.

02 / The traces left behind

The rock had already hosted a reaction.

Researchers found isotope patterns characteristic of fission products: material left when heavy nuclei split. The uranium was natural, and the evidence pointed to fission that had occurred in the deposit.

A split can release energy and neutrons. Some of those neutrons can trigger more splits, continuing a chain.

The lab discrepancy had become evidence of something extraordinary: a natural reactor, roughly two billion years old.

03 / Water joins the story

The water could help it run, then help it stop.

Water in the rock helped slow neutrons, making them more effective at causing further U-235 fission under the conditions at Oklo.

The reaction released heat. Heating could drive water away; with less water to slow neutrons, activity would diminish. Cooling and returning water could allow another active period.

WATER SLOWSTHE NEUTRONSFISSION RELEASESHEATWATER LEAVES;ACTIVITY FALLSROCK COOLS;WATER CAN RETURNAN INFERRED FEEDBACK, NOT A CLOCK RECORD
An inferred feedback between water, heat and the reaction. The diagram is not a measured temperature curve or clock record.

The surroundings were part of the mechanism. The reaction could change the very conditions that helped sustain it.

04 / The right ingredients, long ago

An odd result became a record of deep time.

Ancient uranium contained a larger proportion of U-235 than uranium does today. Together with the right geology and water, that made conditions for a natural chain reaction possible.

The reaction ended long ago. The altered isotope proportions and fission products remained as evidence.

What began as a small mismatch in a sample became a way to read a process that no person had been there to watch.

A little further into the story

How small was the discrepancy?

The difference between the rounded 0.720% and 0.717% figures is 0.003 percentage points. The comparison is small enough that an ordinary full-scale chart would conceal it.

0.710%0.715%0.720%0.725%REFERENCE: 0.720%ANOMALY: 0.717%
Deliberately magnified, truncated scale. Rounded figures from the IAEA account, not a reproduction of the original assay data.

The figure deliberately magnifies a truncated range. It is not raw assay data, and the quoted anomalous value does not apply to every sample from the reactor zones. The fission-product evidence matters alongside the shortfall.

Why does slowing a neutron help?

Under the conditions at Oklo, slower neutrons were more effective at causing further U-235 fission. Slowing them is called moderation; it is different from absorbing them.

Some neutrons escape or are absorbed without continuing a chain. A reaction sustains itself when enough cause further fissions to replace the useful neutron population on average. A steady reaction can still release energy.

How could researchers tell it switched on and off?

A 2004 study examined trapped xenon isotopes and reconstructed active periods of about 30 minutes separated by approximately 2.5 hours of dormancy.

Those intervals are an inference for the studied material, not a directly recorded clock trace or a universal schedule for every reactor zone.

Meshik, Hohenberg and Pravdivtseva, Physical Review Letters (2004); read the abstract.

Could any uranium deposit do this?

No. The proportion of U-235 was higher in the distant past because it decays faster than U-238. Uranium concentration, the deposit’s physical conditions and water also mattered.

Oklo shows that self-sustaining fission occurred naturally under particular conditions. It does not make every uranium-bearing rock a reactor, or turn one geological case into a verdict about every modern nuclear-waste site.

Follow the evidence

Original explanatory diagrams. The reaction and inferred cycling are reconstructed from the cited evidence.