The Missing Uranium: Nature's Own Nuclear Reactor

Bölüm metni

In a French enrichment plant, a sample of ore came up a fraction short of uranium-235 — a shortfall of six hundredths of a percent that should have been impossible, because every rock in the solar system carries the same ratio. The explanation was that someone had already burned the missing atoms. Two billion years before anyone built a reactor, a seam of rock in Gabon had been running one.

In the spring of nineteen seventy-two, in the south of France, at the uranium enrichment plant in the town of Pierrelatte — the French enrichment plant, for short — a measurement was under way that nobody expected any excitement from. The raw uranium arriving at the plant moved through the pipes as a gas, uranium hexafluoride, and the proportion of uranium two thirty-five inside it was checked routinely with a mass spectrometer. This check was less a quality test than an accounting operation. Because the natural abundance of uranium two thirty-five was held to be as dependable as a constant of nature.

That constant was zero point seven two zero percent. Ore dug in Australia gave it, so did the shield rocks of Canada, so did the Czech mines, and so did stones brought back from the Moon and meteorites fallen out of the sky: out of every thousand uranium atoms, roughly seven point two were uranium two thirty-five. There was an explanation for this. In the cloud that formed the solar system, the two isotopes of uranium had been blended in a certain ratio, and ever since, each had been decaying at its own pace. Whichever planet, whichever continent, whichever rock you happened to be standing on, the clock read the same. Geologists treated that ratio as so solid that when a sample showed a deviation, the first thing that came to mind was not nature but a laboratory error.

That day the screen did not read zero point seven two zero percent. It read zero point seven one seven.

The difference was zero point zero zero three percent. On paper, nothing at all. But the instrument taking the measurement was sensitive enough to separate that difference from noise, and more importantly, the number had not shifted in years. Henri Bouzigues, on the team that took the reading, first did the expected thing: he checked the instrument, prepared a fresh sample, went over the calibration. The number would not move. Then the shipment records were examined, and a far more uncomfortable picture emerged. Spread across the whole mass of uranium arriving at the plant, that minuscule deviation amounted to a shortfall of about two hundred kilograms of uranium two thirty-five that was simply not there.

In nineteen seventy-two, in the very middle of the Cold War, it is not hard to imagine what the sentence "uranium two thirty-five is missing" meant inside a nuclear fuel facility. The first scenarios anyone reached for were not nature's but humanity's. Had someone mixed up the material? Had uranium that had already been through the fuel cycle, uranium burned in a reactor, been blended somewhere with fresh ore? Because only one known thing depletes uranium two thirty-five: fission. A reactor. A chain reaction consumes uranium two thirty-five and leaves behind a residue enriched in uranium two thirty-eight. What was showing on the screen was precisely that — the signature of spent fuel.

The investigation traced that batch back to where it had come from. In the Franceville basin, in the southeast of Gabon, an open-pit mine called Oklo. The French Atomic Energy Commission sent a team there, and samples began to be taken from the ore veins one by one. The results coming out of the pit made the deviation at the French enrichment plant look almost trivial. At some points the ratio dropped to zero point six zero percent. In one vein it was measured at zero point four four. That is to say, nearly forty percent of the uranium two thirty-five in that rock was gone. Nothing like it had ever been seen anywhere in the world, in any ore.

Could a geological process pick out one isotope of uranium from the other and carry it away selectively? Chemically, uranium two thirty-five and uranium two thirty-eight are identical twins; they form the same compounds, dissolve in the same water, sit in the same mineral. The only difference between them is the weight of three neutrons, and nature cannot use that difference to achieve a forty percent separation. To perform that separation, human beings were spending enormous facilities, thousands of stages, kilometres of pipework and as much electricity as a city. That, after all, was the entire reason the plant existed.

The decisive answer came not from the uranium but from the elements sitting beside it. The researchers looked at the neodymium in the Oklo ore. Neodymium in nature is a particular mixture of several isotopes, and about a quarter of that mixture is neodymium one forty-two. Fission, however, produces neodymium without producing neodymium one forty-two — there is no decay chain leading to that isotope. In the samples from Oklo, neodymium one forty-two was almost entirely absent. The remaining isotopes, meanwhile, matched the yield distribution of uranium two thirty-five fission, measured thousands of times in laboratories, almost line for line. Ruthenium told the same story. There was reactor ash inside the rock.

On the twenty-fifth of September, nineteen seventy-two, the physicist Francis Perrin announced the result at the French Academy of Sciences. Nobody had stolen the missing uranium. At Oklo, long before there were people, a nuclear reactor had run of its own accord.

The strangest thing about this finding was that it was not entirely a surprise. In nineteen fifty-six, the chemist Paul Kuroda, working at the University of Arkansas, had sat down and done a simple calculation: what conditions would be needed for a chain reaction to start by itself in nature? He had boiled the answer down to a handful of items, and the conclusion was this — there is a period in the Earth's past when it would have been possible. Kuroda's paper sat for sixteen years as an elegant but academic curiosity. Then rock from Gabon turned that calculation from a prediction into a record.

Kuroda's first condition had to do with time, and it is a condition that cannot be met today. The half-life of uranium two thirty-five is about seven hundred and four million years; that of uranium two thirty-eight is four billion four hundred and seventy million years. Both are dwindling, but uranium two thirty-five far faster. Which means that the ratio that stands at zero point seven two percent today grows the further back you go. A billion years ago it was around one and a half percent. Roughly two billion years ago it was above three percent. Note that figure: most commercial power reactors in the world today run on fuel enriched to between three and five percent. So two billion years ago every uranium vein in the Earth's crust was ready-made reactor fuel, without passing through any process at all. The world was once made of enriched uranium, and that age closed for good.

But enriched fuel alone is not enough. The neutrons released in fission are very fast, and fast neutrons do not split uranium two thirty-five efficiently; they have to be slowed down first. The substance that does this job is called a moderator, and the most common moderator on Earth is water. When a neutron strikes a hydrogen nucleus of almost exactly its own mass, it gives up most of its energy in a single collision, the way a billiard ball does when it hits its own twin. The rock at Oklo was porous, and those pores were filled with groundwater. The vein itself was a reactor core: fuel and moderator, interleaved with one another.

The third condition was cleanliness. Some elements swallow neutrons and kill the chain before it can even begin; boron, cadmium and certain rare earth elements are foremost among them. In human-built reactors these substances are deliberately placed in control rods. In the Oklo veins, as luck would have it, these neutron poisons were scarce.

The fourth condition, and perhaps the most astonishing, concerned how the uranium came to be piled up there in the first place. For a chain reaction, it is not enough for uranium to be sprinkled through the rock; it has to be concentrated. In some of Oklo's reactor veins the uranium content of the ore exceeded twenty percent. What produced that concentration was water, but for water to carry uranium there had to be oxygen in the atmosphere. In an oxygen-free world uranium does not dissolve; it stays where it is. About two billion four hundred million years ago, when photosynthesising microbes permanently altered the atmosphere — during the period known as the Great Oxidation Event — uranium became soluble. Groundwater began stripping it out of the rock, carrying it, and precipitating it again in sediments rich in organic matter. In other words, one of the things that made the Oklo reactor possible was living creatures. Bacteria had stockpiled the fuel for the world's first nuclear plant.

These four conditions came together roughly one billion seven hundred million to two billion years ago, in those few veins of the Oklo basin, and the reaction began on its own. In the same region, sixteen reactor zones were found at Oklo itself, one at the immediately adjacent Okélobondo deposit, and one more at Bangombé, about thirty kilometres to the south: seventeen separate natural reactors in all.

So why did this thing never run away? The most dangerous property of a chain reaction is that it feeds itself and grows. Oklo had its own brake, and that brake was the water inside the fuel. When the reaction started, the rock heated up, and the water in the pores boiled and turned to vapour. With the water gone, the moderator went too; the neutrons stayed fast, the chain broke, the reactor fell silent. As the rock cooled, groundwater seeped back into the pores, the moderator returned to its place, and the reaction reignited. A reactor that breathed in and out, like a geyser erupting and resting.

Even the rhythm of that cycle has been measured. In two thousand and four, Alex Meshik and his colleagues used a laser to release, one at a time, the xenon isotopes trapped inside aluminium phosphate minerals in the reactor veins, and studied them. Xenon is a product of fission, and different xenon isotopes form over different timescales; so the isotope pattern recorded inside the mineral gives away how long the reactor stayed on and how long it stayed off. The result was this: roughly thirty minutes of operation, followed by roughly two and a half hours of silence. Then thirty minutes again. A three-hour pulse.

That pulse beat, on and off, for several hundred thousand years. Its average power was modest; thermally it probably never exceeded one hundred kilowatts, enough to feed a few dozen toasters. But the span was so long that in total about five tonnes of uranium two thirty-five underwent fission, and the rock heated to several hundred degrees. That deviation of a few parts per thousand at the French enrichment plant was the leftover arithmetic of those five tonnes.

The real value of Oklo lies not in how strange a coincidence it was, but in what it left behind.

Every reactor produces waste. Oklo produced it too: strontium, caesium, technetium, iodine, rare earth elements and actinides. Some of the neutrons generated in the veins were absorbed by uranium two thirty-eight and turned into plutonium two thirty-nine; part of that plutonium fissioned where it stood, and the rest decayed, with its twenty-four-thousand-year half-life, back into uranium two thirty-five. Oklo, in other words, was also a plutonium factory.

Now the real question: where did all that material go? There was no sealed steel vessel over it. No concrete. No glass matrix, no copper canister, no engineered barrier. There was only rock, and groundwater flowed through that rock for two billion years.

The answer is one that everybody debating nuclear waste storage ought to stop and think about: most of the non-volatile fission products and the actinides barely stirred at all. Looking at how they are distributed through the veins, their movement is measured in centimetres, a few metres at most. Plutonium decayed where it was produced. The rare earth elements worked their way into the mineral structure of the ore and locked themselves there. We know that even a noble gas like xenon was held inside phosphate minerals for billions of years, because it was from that xenon that we read the reactor's pulse.

This is why Oklo is invoked today, in discussions of geological disposal, as a natural analogue. In a laboratory you can test how well a mineral holds something for a few years at most; from there you make a projection reaching a hundred thousand years into the future, and nobody will ever be alive to see whether the projection was right. Oklo has already run that experiment and written the result into the rock. It would be wrong, however, to draw from this the conclusion that nuclear waste simply stays put by itself, because the outcome is not the same everywhere. At Bangombé, the zone closest to the surface, uranium and some of the products can be seen to have migrated a certain distance into the surroundings. The difference comes less from chemistry than from earth science: depth, the impermeability of the overlying rock, the chemistry of the water, the presence of clay. What Oklo says is not "waste is safe"; it is "you can measure which geology holds and which one leaks."

The record in that rock also found an entirely different kind of reader. Whether the fundamental constants of physics are truly constant has long been argued over; there are theories proposing that the number known as the fine-structure constant, which sets the strength of the electromagnetic interaction, may have changed very slowly across the ages. To test this you need a sensitive nuclear apparatus that has been running for billions of years. Oklo is exactly that. The probability that the isotope samarium one forty-nine absorbs a neutron rises to a sharp peak across a very narrow band of energy, and the position of that peak is sensitive to even the slightest shift in the fine-structure constant. Had the constant been a little different, the proportions of the samarium isotopes remaining in the Oklo ore would have come out noticeably different from what they are. The measurements were made, and the result showed that the constant has stayed the same, within an extraordinarily tight limit, for two billion years. A mine in Gabon has given us one of the firmest pieces of evidence we currently hold that the rules of the universe do not change.

The Oklo pit itself no longer operates; the ore was extracted, and most of the zones disappeared along with the mining. A few veins have been placed under protection by Gabon for their scientific value. Looking back, we can see that the first controlled chain reaction, assembled in December of nineteen forty-two beneath a sports hall in Chicago out of graphite bricks and cubes of uranium, was not really an invention but a rediscovery. The same physics, the same logic of moderation, the same self-braking balance had been set up spontaneously by the Earth's crust in Gabon two billion years earlier. The difference is that nature did not do it on purpose; the right rock, the right water and the right age simply came together.

And we know all of this thanks to a deviation of three parts in a thousand that appeared on a screen in a laboratory. The missing uranium was the mark not of a theft, but of the world's own past. Had the team taking that measurement shrugged that day and said the instrument must be broken, the rock would have gone on keeping its silence.

Nöron Science'de bugünkü bölümün sonuna geldik. Yeni bir hikâyede yeniden buluşmak üzere.