The 2-Billion-Year-Old Nuclear Reactor

In June 1972, an analyst named Francois Perrin at France's Commissariat a l'energie atomique was checking uranium hexafluoride processed from ore mined at Oklo, in Gabon. The U-235 reading came back at 0.7171 percent instead of the universal 0.7202 percent. A gap of three thousandths of a percent sounds trivial, but at that level of precision it is an alarm bell, because that number does not vary. Follow-up sampling made it stranger still: some zones in the ore ran as low as 0.440 percent U-235, far below anything found anywhere else on Earth. The uranium was depleted, in exactly the way a reactor depletes its fuel, and the only natural process that strips out U-235 like that is fission.

The proof was written in the ash. When researchers looked closely at the depleted ore, they found the unmistakable fingerprint of a chain reaction: isotopic anomalies in neodymium, samarium, cerium, ruthenium, zirconium and other elements, in precisely the proportions you get from splitting U-235 with slow neutrons. These were not ordinary rocks with an odd uranium reading. They were spent nuclear fuel, and they had been sitting in the ground since a time when life on Earth was nothing but microbes.

Start with the fuel. A modern power reactor runs on uranium enriched to between three and five percent U-235, because natural uranium at today's 0.72 percent is too dilute to sustain a chain reaction in ordinary conditions. But 0.72 percent is only today's figure. U-235 decays more than six times faster than U-238: its half-life is 704 million years against U-238's 4.47 billion. Run the clock backward and the mix gets steadily richer. Isochron dating of the fission products, using the samarium-neodymium system, pins the reactors' operation to about 1.78 billion years ago, and at that moment natural uranium was roughly 3.7 percent U-235. In other words, the ore was already enriched to reactor grade, for free, simply by being ancient. Dig up that same seam today and it would be far too lean to ever ignite.

Rich isotopes are not enough on their own; you also need the uranium packed tightly together. At Oklo the ore reaches over ten percent U3O8 in the reactor layers, an extraordinary concentration, and geochemistry did the packing. Uranium is soluble when it is oxidized, so oxygenated groundwater dissolved it, carried it through the rock, and then dropped it wherever the chemistry turned reducing, precipitating it as dense uraninite at that boundary. Organic matter and anaerobic microbes may have helped force that reaction. Water sorted and concentrated a fissile metal into ore bodies rich enough to burn.

And there is a deeper prerequisite still, one that ties Oklo to the history of the very air we breathe. Uranium only dissolves and travels when there is oxygen around to oxidize it, and for most of Earth's early history there was almost none. It was the Great Oxygenation Event, about 2.4 billion years ago, that first flooded the atmosphere and oceans with free oxygen, and only after that could uranium be mobilized, transported and reconcentrated into deposits rich enough to go critical. Oklo could not have existed on the older, oxygen-poor Earth. It is a direct consequence of the planet learning to breathe.

The last ingredient was ordinary water. Fission neutrons come off fast, too fast to reliably split the next U-235 nucleus, so a reactor needs a moderator to slow them down, and at Oklo the moderator was groundwater soaking the ore. The same water was also a mild neutron absorber, which turns out to matter enormously for how the reactors behaved. The rule was simple: no water, no reaction. The seams only ran while they were saturated.

Here is where Oklo stops being merely astonishing and becomes almost unbelievable, except that the evidence is airtight. Xenon isotopes, a telltale fission gas, are trapped inside aluminum phosphate grains in the reactor zones, and their pattern preserves the reactors' actual operating rhythm. Oklo did not burn steadily. It pulsed. The ore would go critical and run for roughly thirty minutes, heating the groundwater until it boiled away. With its moderator gone as steam, the reaction shut itself down. Then, over about two and a half hours, the rock cooled, water seeped back in, and the reactor switched on again. A thirty-minute pulse, a two-and-a-half-hour rest, over and over, governed by nothing more than the boiling point of water. Meshik and colleagues read that cycle straight out of the trapped xenon and published it in Physical Review Letters in 2004.

The output was modest and the endurance was staggering. Averaged over its lifetime the Oklo system put out around 100 kilowatts of thermal power, about what a small research reactor produces, nothing like a power station. But it kept that up, on and off, for somewhere between 150,000 and 300,000 years, and then fell silent well over a billion years before the first city was ever built.

The totals are just as concrete. Across all sixteen zones the reactors fissioned roughly five to six tonnes of U-235, and in doing so they produced about two tonnes of fission products and around six tonnes of plutonium-239. That plutonium is long gone now, decayed away through its own chain over the eons since, but it was really made, in the ground, by a natural reactor, more than a billion years before anyone knew plutonium existed.

For two billion years, the transuranic elements and fission products that Oklo created have mostly stayed put. In the clay-rich sedimentary rock around the reactor zones, most of the fission products migrated less than a few meters from where they were born, held in place by nothing more elaborate than ordinary geology. That is a remarkable result, because keeping radioactive material contained across deep geological time is the central engineering challenge of modern waste disposal, and Oklo is a case where the rock did exactly that, on its own, for two billion years.

The plutonium tells the cleanest version of the story. Oklo's roughly six tonnes of plutonium-239 did not leak away and spread. It stayed inside the ore body and decayed there, in place, exactly where a waste engineer would want it to stay. For anyone trying to design a repository that can hold its contents across geological time, Oklo is direct, physical evidence that a well-chosen geological barrier really can do the job. It is why the site is studied by the nuclear-waste community and not just by geologists.

But Oklo is honest about the limits too, and this is the part that makes it genuinely useful rather than just reassuring. Not everything stayed home. Some of the more mobile fission products, cesium, rubidium, iodine and xenon among them, did travel farther from the reactor zones than the rest. That is exactly the kind of element-by-element detail a modern repository has to plan for, because a barrier that traps one isotope beautifully may let another slip through. Oklo does not just say containment is possible. It shows which elements are the ones to watch.

There is a strange possibility folded into how the ore formed in the first place. Concentrating uranium out of groundwater and fixing it in place is exactly the sort of chemistry that certain microbes drive, by reducing dissolved uranium into an insoluble solid, and modern bacteria do precisely this today. The Francevillian sediments that host Oklo are rich in organic carbon, a signature of ancient life, and the idea that microbial activity helped pile up the uranium is debated but genuinely supported. If it is right, then living things had a hand in gathering the fuel, which means that nearly two billion years ago, life on Earth may have unknowingly helped assemble a working nuclear reactor. It would not learn it had done so until 1972.

The obvious question is whether Oklo was a one-off or just the one we happened to find. In principle it need not be unique: other high-grade uranium deposits from the same Paleoproterozoic era, when U-235 was still abundant, could in theory have met the same conditions. Researchers have gone looking. A handful of uranium deposits in Australia and Canada have shown mildly anomalous isotope ratios worth a second look, but none has ever been confirmed as a fossil reactor. After more than fifty years of searching, Oklo and its neighbor Okelobondo remain the only natural nuclear reactors known anywhere on Earth.

Oklo proved that nature can run a reactor near the surface, and that raised a bolder question, one a serious scientist took all the way to the planet's core. In 2001 the geochemist J. Marvin Herndon proposed, and in 2003 published in the Proceedings of the National Academy of Sciences, that a natural fission reactor might be burning at the very center of the Earth, helping to power the planet's internal heat and generate its magnetic field. This is not fringe pseudoscience. It is a real, peer-reviewed, minority hypothesis, argued carefully in a top journal. It is also, by the current weight of evidence, wrong. Mainstream geophysics expects far too little uranium to be concentrated in the core to sustain such a reactor, and direct measurements settle it: the KamLAND neutrino detector, which watches the faint stream of neutrinos coming out of Earth's interior, sees nothing like the signal a georeactor of the proposed size would produce. Herndon's idea is a clean example of how real science handles a real disagreement. Someone published a bold, testable claim, the data came in, and the data said no.

There is exactly one claim about Oklo that this file refuses outright, and it deserves a flat answer, because the internet loves it: the notion that the reactors were built or operated by some ancient or extraterrestrial civilization. They were not, and the evidence against it is total. The reactor zones are ordinary uranium ore, layered by groundwater chemistry, with no machined parts, no shaped structures, no engineering of any kind. The reactors ran about two billion years ago, more than a billion years before complex multicellular life existed, let alone anyone to design a power plant. And every feature of how they worked, the enrichment, the concentration, the water moderator, the pulsed cycle, is fully and precisely explained by ordinary geochemistry and nuclear physics, with nothing left over for a builder to account for. Oklo is more wonderful than the ancient-astronaut version, not less. It says the universe can build a reactor with no one in charge at all. [DEBUNKED]