In the early- to mid-1950s, Dr. Paul Kuroda from the University of Arkansas described the possibility of naturally occurring nuclear reactors lurking in the crust of ancient Earth. The key is an isotope of Uranium called U-235, which occurs naturally in small amounts. If enough of this isotope were pooled together under specific circumstances, Kuroda theorized, the natural reactor would go critical, and self-sustaining fission would occur. Such a reactor could not exist today, because too much of the Earth’s natural U-235 has decayed⁠—but a billion and a half years ago, there was enough of it around to make the idea plausible. In point of fact, it has since been discovered that it actually happened.

In 1972, workers in the Oklo mines of Gabon in West Africa found the radioactive remains of such a nuclear reactor. Uranium extracted from that mine was found to be abnormally short in U-235 isotopes, and upon examination, French scientists found that the uranium isotope levels had an uncanny resemblance to those in spent nuclear fuel from modern nuclear power plants. The evidence was strong enough to suggest a natural reactor, and further exploration confirmed it.

At the time of discovery, scientists were uncertain exactly how the Oklo reactor had operated without exploding or melting down. Base don the evidence left in he rocks, the reactor ran like clockwork for 150 million years, with a 30 minute reaction cycle, followed by a 2.5 hour cool-down cycle, putting out an average of 100 kilowatts of power. And it was consistently 30 minutes per cycle, without significant variation, which was baffling. Subsequent studies finally solved the mystery by discovering the regulating mechanism: Water.

Under normal conditions, radioactive atoms like U-235 cast off neutron particles at speeds so high that most of the neutrons skip off the surface of other atoms and fly away. But if you put enough of the radioactive material together, the cast-off neutrons bounce around inside the mass, some slowing down enough to be absorbed into another atom’s nucleus. The extra neutron causes the nucleus to become unstable and immediately split, which releases a large amount of energy. If one has enough radioactive material in sufficient density that a lot of nuclei split very rapidly⁠—a state known as critical mass⁠—the reaction increases exponentially, and results in an atomic explosion. Any less than that⁠—subcritical mass⁠—and it causes a sustained fission reaction, giving off energy as heat and radiation.

But researchers have determined that Oklo didn’t even have an appreciable, consolidated subcritical mass of Uranium⁠—it was too spread out. Instead, water would seep down through crevices to fill up the gaps between the uranium deposits, and act as a “neutron moderator,” slowing down the neutrons enough to allow them to hit-and-split other nuclei. When the reaction caused a sufficient heat increase, the water would boil off, removing the neutron moderator, and stop the process. The cavity would then slowly refill with water during the cooling period, starting the cycle again.

Fifteen such natural reactors have been found in the Oklo area, and they are now collectively referred to as the “Oklo Fossil Reactors.” These natural reactors are providing useful data on long-term storage of spent nuclear fuel, as well as some insights into possible improvements in man-made reactors.