Natural nuclear fission reactor

A natural nuclear fission reactor is a uranium deposit where self-sustaining nuclear chain reactions occur. The conditions under which a natural nuclear reactor could exist were predicted in 1956 by Paul Kuroda.a, chemist at the University of Arkansas, revealed three necessary conditions for the existence of a natural fission reactor. Firstly, the width of the uranium deposit must be greater than two-thirds of a meter. This allows neutrons to travel a sufficient distance to be absorbed and continue the nuclear chain reaction, before leaving the uranium deposit. Secondly, the percentage of ²³⁵U must be greater than 3% in uranium ore to properly fuel a nuclear reactor. ²³⁵U decays over time, so that two billion years ago, there was still enough ²³⁵U in the Oklo deposit for a natural fission reactor to exist (today, the concentration of 235U on Earth is too low). Thirdly, a neutron moderator (such as water) is needed to control the nuclear reaction, but one without too much boron, lithium or other elements that would absorb too many neutrons.

The remnants of an extinct or fossil nuclear fission reactor, where self-sustaining nuclear reactions have occurred in the past, are verified by analysis of isotope ratios of uranium and of the fission products (and the stable daughter nuclides of those fission products). This was first discovered in 1972 in Oklo, Gabon by researchers from French Commissariat à l'énergie atomique (CEA) under conditions very similar to Kuroda's predictions.

Oklo is the only location where this phenomenon is known to have occurred, and consists of 16 sites with patches of centimeter-sized ore layers. There, self-sustaining nuclear fission reactions are thought to have taken place approximately 1.7 billion years ago, during the Statherian period of the Paleoproterozoic, and continued for a few hundred thousand years, probably averaging less than 100 kW of thermal power during that time. During that era, life on Earth consisted of little more than aquatic single-cell organisms.

Gabon was a French colony when the first analyses of the subsoil were carried out by the CEA from the MABA base in Franceville, more precisely by its industrial direction which later became COGEMA, leading in 1956 to the discovery of uranium deposits in the region.

France almost immediately opened mines, managed by the “Compagnie des Mines d'Uranium de Franceville” (COMUF), to exploit the resources, near the village of Mounana. After independence in 1960, the state of Gabon received a small share of the company's profits.

Discovery of the Oklo fossil reactors

The "Oklo phenomenon" was discovered in June 1972 by the laboratory at the uranium enrichment plant in Pierrelatte, France. Routine analysis of a sample of natural uranium revealed a slight but abnormal deficit of ²³⁵U⁶. The normal proportion of ²³⁵U is 0.7202%, whereas this sample showed only 0.7171%. As the quantities of fissile isotopes are precisely catalogued, this discrepancy had to be explained, so an investigation was launched by CEA on samples from all the mines operated by CEA in France, Gabon and Niger, and at all stages of ore processing and uranium purification.

For uranium and ²³⁵U analyses, CEA's Production Division relies on the Analytical Laboratory at the Pierrelatte plant and on CEA's Central Analysis and Control Laboratory at Cadarache, headed by Michele Neuilly, where Jean François Dozol is in charge of mass spectrometry analyses.Analyses carried out at Pierrelatte and Cadarache showed that magnesium uranates (or yellow cakes) from Gabon had a variable but constant ²³⁵U depletion. On July 7, 1972, researchers at Cadarache discovered an anomaly in uranium ore from Oklo in Gabon. Its ²³⁵U content was much lower than usual. Isotopic analyses revealed the origin of the ²³⁵U depletion: the depleted uranium came from Oklo ore in Gabon, mined by COMUF. A systematic analysis campaign was then carried out in the Cadarache and Pierrelatte laboratories (uranium content measurements, isotopic content measurements). On Oklo samples, Cadarache analysts noted a ²³⁵U depletion for magnesium uranate from the Mounana plant (²³⁵U = 0.625%) and an even greater depletion for a magnesium uranate (Oklo M) (²³⁵U = 0.440%): Oklo 310 and 311 ores have uranium contents of 12% and 46% respectively, and ²³⁵U contents of 0.592% and 0.625%.

In this context, J.F. Dozol took the initiative of analyzing magnesium uranate and ore samples from Oklo on the AEI MS 702 Spark Source Mass Spectrometer (SSMS).

The advantage of the SSMS is its ability to produce substantial quantities of ions from all the elements present in the electrodes. The electrodes, between which a spark is generated, have to be conductive (to achieve this, Oklo samples were mixed with high-purity silver). All the isotopes in the sample, from lithium to uranium, are plotted on a photo plate. On examining the plate (see below), J.F. Dozol noted in particular the very high uranium content of Oklo 311 ore:

- elements present in significant quantities around masses 85-105 and 130-150, corresponding to the two bumps of ²³⁵U fission yields. (The mass distribution of fission products follows a "camel's hump" curve, with two maxima),

- the last lanthanides (holmium to lutetium) are not detected (beyond mass 166). In nature, all 14 lanthanides are found; in nuclear fuel, having undergone fission reactions, the isotopes of the last lanthanides are not detected.

Photo plate obtained for the OKLO 311 sample by spark mass spectrometry analysis and which reveals all the isotopes present in the sample.

The next step is isotopic analysis of certain elements on a thermal ionization mass spectrometer, after chemical separation of neodymium and samarium. From the first analyses of Oklo "M" uranate and "Oklo 311" ore, it is clear that neodymium and samarium have an isotopic composition much closer to that found in irradiated fuel than to that of the natural element. The detection of ¹⁴²Nd and ¹⁴⁴Sm isotopes not produced by fission indicates that these elements are also present in the natural state, from which their contribution can be subtracted.

These results were passed on to neutron scientist Jean Claude Nimal (CEA Saclay), who estimated the neutron flux received by the analyzed sample on the basis of its ²³⁵U deficit. This made it possible to estimate the neutron capture by the isotopes ¹⁴³Nd and ¹⁴⁵Nd, leading to the additional formation of ¹⁴⁴Nd and ¹⁴⁶Nd respectively. This excess must be subtracted to obtain fission yields for uranium 235. As can be seen from the table below, the fission yields (M) agree with the results corrected (C) for the presence of natural neodymium and neutron capture.

Nd	143	144	145	146	148	150
C/M	0.99	1.00	1.00	1.01	0.98	1.06

Fission product isotope signatures

Isotope signatures of natural neodymium and fission product neodymium from ²³⁵
U
which had been subjected to thermal neutrons.

Neodymium

The neodymium found at Oklo has a different isotopic composition to that of natural neodymium: the latter contains 27% ¹⁴²
Nd
, while that of Oklo contains less than 6%. The ¹⁴²
Nd
is not produced by fission; the ore contains both fission-produced and natural neodymium. From this ¹⁴²
Nd
content, we can subtract the natural neodymium and gain access to the isotopic composition of neodymium produced by the fission of ²³⁵
U
. The two isotopes ¹⁴³
Nd
and ¹⁴⁵
Nd
lead to the formation of ¹⁴⁴
Nd
and ¹⁴⁶
Nd
by neutron capture, and this excess must be corrected (see above) to obtain perfect agreement between this corrected isotopic composition and that deduced from fission yields.

Ruthenium

Isotope signatures of natural ruthenium and fission product ruthenium from ²³⁵
U
which had been subjected to thermal neutrons. The ¹⁰⁰
Mo
(an extremely long-lived double beta emitter) has not had time to decay to ¹⁰⁰
Ru
in more than trace quantities over the time since the reactors stopped working.

Similar investigations into the isotopic ratios of ruthenium at Oklo found a much higher ⁹⁹
Ru
concentration than otherwise naturally occurring (27–30% vs. 12.7%). This anomaly could be explained by the decay of ⁹⁹
Tc
to ⁹⁹
Ru
. In the bar chart the normal natural isotope signature of ruthenium is compared with that for fission product ruthenium which is the result of the fission of ²³⁵
U
with thermal neutrons. It is clear that the fission ruthenium has a different isotope signature. The level of ¹⁰⁰
Ru
in the fission product mixture is low because fission produces neutron rich isotopes which subsequently beta decay and ¹⁰⁰
Ru
would only be produced in appreciable quantities by double beta decay of the very long-lived (half life 7.1×10¹⁸ years) molybdenum isotope ¹⁰⁰
Mo. On the timescale of when the reactors were in operation, very little (about 0.17 ppb) decay to ¹⁰⁰
Ru
will have occurred. Other pathways of ¹⁰⁰
Ru production like neutron capture in ⁹⁹
Ru or ⁹⁹
Tc (quickly followed by beta decay) can only have occurred during high neutron flux and thus ceased when the fission chain reaction stopped.

Mechanism

The natural nuclear reactor at Oklo formed when a uranium-rich mineral deposit became inundated with groundwater, which could act as a moderator for the neutrons produced by nuclear fission. A chain reaction took place, producing heat that caused the groundwater to boil away; without a moderator that could slow the neutrons, however, the reaction slowed or stopped. The reactor thus had a negative void coefficient of reactivity, something employed as a safety mechanism in human-made light water reactors. After cooling of the mineral deposit, the water returned, and the reaction restarted, completing a full cycle every 3 hours. The fission reaction cycles continued for hundreds of thousands of years and ended when the ever-decreasing fissile materials, coupled with the build-up of neutron poisons, no longer could sustain a chain reaction.

Meshik and colleagues studied the xenon isotopic composition of the Oklo rock, to probe the nature of Gabon's reactors. The researchers made two surprising discoveries. Firstly, they observed that xenon was not mainly localized in uranium-rich mineral grains as they had expected, but rather in aluminum phosphate minerals. Secondly, they observed that the isotopic distribution differed from that of modern nuclear reactors. Of the nine stable xenon isotopes, the Oklo material was depleted in ¹³⁶Xe and ¹³⁴Xe^. Meshik and colleagues recognized that none of the xenon isotopes were produced directly by the fission of uranium itself, but rather were formed by the decay of other precursor fission products (in particular radioactive iodine and tellurium). Thus, they recognized that the formation of the various xenon isotopes would depend on the lifetime of their precursors. This would result in the formation of ¹³⁶Xe within a minute of the start of the self-sustaining fission reaction, the formation of ¹³⁴Xe after an hour, the formation of ¹³²Xe and ¹³¹Xe within a few days, and the formation of ¹²⁹Xe only, after millions of years.

A reminder of how fission products are formed: ²³⁵U fission produces two fission products, ⁹⁷Mo and ¹³⁷Sn for example, which have an excess of neutrons. As a result, the radionuclides are unstable, and the excess neutrons are transformed into a proton and an electron, expelled from the nucleus as beta minus radiation. Before reaching a stable state, the decay chain passes through the following stages: ¹³⁷Sn,¹³⁷Sb,¹³⁷Te,¹³⁷Sn,¹³⁷I, ¹³⁷Xe,¹³⁷Cs,¹³⁷Ba stable. With the exception of ¹³⁷Cs, the radioactive periods of these elements are extremely short (less than a few minutes, often a few seconds).

Meshik and his colleagues estimated that the Gabon reactors were probably operating in a series of "on/off" cycles. In the presence of water, neutrons are slowed and a fission chain reaction can occur. However, the heat generated by the reaction would eventually boil the water, stopping fission until the groundwater returned. This cycle explains the distribution of xenon isotopes: when the reactor was operating, the rapidly-forming ¹³⁴Xe and ¹³⁶Xe gases were expelled, but the precursors that would eventually give rise to ¹³²Xe, ¹³¹Xe, and ¹²⁹Xe were incorporated into aluminum phosphate minerals as the reactor cooled, during a "shutdown" cycle. This explains the absence of ¹³⁴Xe and ¹³⁶Xe in these aluminum phosphate minerals. Moreover, this dependence on a water moderator could also explain the absence of xenon in uranium-rich mineral grains: water would have carried away the water-soluble isotopes of tellurium and iodine that would have led to xenon.

Using the ¹³¹Xe /¹³⁴Xe and ¹³²Xe /¹³⁴Xe ratios of aluminum phosphate from the Oklo site, Meshik and colleagues calculated the operating schedule for the Oklo reactors. Their calculations indicated an "on" fission period of 30 minutes (with concomitant boiling of the water moderator) followed by an "off" cooling period of 2 hours and 30 minutes before the water returned and the reactor became self-sufficient again18.

The groundwater permeating the repository acts as a moderator, allowing ²³⁵U fission to begin. The decay of some radioactive fission products quickly gave rise to ¹³⁴Xe and ¹³⁶Xe, but these gas atoms tended to be driven out by the increasing heat of the reactor. The longer-lived xenon precursors, ¹³¹I (t_1/2 = 8.02 d) and ¹³²I (t_1/2 = 20.8 h), entered solution and were carried away as quickly as they had been created.

Around 30 minutes after the start of nuclear fission, the temperature reached the point where most of the groundwater boiled away, depriving the reactor of its moderator and interrupting fission. Some of the ¹³¹I and ¹³²I iodine atoms created during the previous half-hour were retained in the residual groundwater between the uranium mineral grains. In the absence of fission reactions to sustain them, they were removed from solution.

A few hours later, the temperature dropped sufficiently to allow the groundwater to return. The substances dissolved in the warm groundwater came out of solution, forming aluminum phosphate minerals that incorporated ¹³¹I and ¹³²I. These minerals also absorbed ¹²⁹I (t_1/2 = 16.14 10¹⁶ a) several million years later. In the presence of a moderator, fission resumed..

Xenon-135 is the strongest known neutron poison. However, it is not produced directly in appreciable amounts but rather as a decay product of Iodine-135 (or one of its parent nuclides). Xenon-135 itself is unstable and decays to caesium-135 if not allowed to absorb neutrons. While caesium-135 is relatively long lived, all caesium-135 produced by the Oklo reactor has since decayed further to stable barium-135. Meanwhile, Xenon-136, the product of neutron capture in xenon-135 only decays extremely slowly via double beta decay and thus scientists were able to determine the neutronics of this reactor by calculations based on those isotope ratios almost two billion years after it stopped fissioning uranium.

Change of content of Uranium-235 in natural uranium. The content was 3.65% 2 billion years ago.

A key factor that made the reaction possible was that, at the time the reactor went critical 1.7 billion years ago, the fissile isotope ²³⁵
U
made up about 3.1% of the natural uranium, which is comparable to the amount used in some of today's reactors. (The remaining 96.9% was non-fissile ²³⁸
U
and roughly 55 ppm ²³⁴
U.) Because ²³⁵
U
has a shorter half-life than ²³⁸
U
, and thus decays more rapidly, the current abundance of ²³⁵
U
in natural uranium is only 0.72%. A natural nuclear reactor is therefore no longer possible on Earth without heavy water or graphite.

The Oklo uranium ore deposits are the only known sites in which natural nuclear reactors existed. Other rich uranium ore bodies would also have had sufficient uranium to support nuclear reactions at that time, but the combination of uranium, water, and physical conditions needed to support the chain reaction was unique, as far as is currently known, to the Oklo ore bodies. It is also possible that other natural nuclear fission reactors were once operating but have since been geologically disturbed so much as to be unrecognizable, possibly even "diluting" the uranium so far that the isotope ratio would no longer serve as a "fingerprint". Only a small part of the continental crust and no part of the oceanic crust reaches the age of the deposits at Oklo or an age during which isotope ratios of natural uranium would have allowed a self sustaining chain reaction with water as a moderator.

It was only around 2.2 billion years ago that the patient work of photosynthesis carried out by the first algae released enough oxygen into our atmosphere for surface and groundwater to become oxidizing.  Only then could the uranium diluted in granite be leached and concentrated before mineralizing in places where redox would occur.  Rich deposits cannot be older than this.  On the other hand, over the last 1.5 billion years, the abundance of ²³⁵U has fallen below a level that makes spontaneous fission possible.  It took a great deal of study - in geology, chemistry and reactor physics - to narrow the time range down to the current estimate: reactions must have started 1,950 ± 30 million years ago.

The deposits were located in highly porous sandstones, where the concentration of groundwater may have been as high as 40%, probably due to the partial leaching of silica (quartz particles) by hot groundwater, at a time when, with the Earth's radioactivity higher than today, the thermal gradient of the subsoil was probably higher too. During operation of the reactor, the temperature of the water rose significantly, accelerating this process of "de-silicification" and, by difference, increasing the concentration of uranium, thus compensating for its depletion by fission.  Indeed, uranium concentration in the reaction zones is extremely high, sometimes over 50%, and the higher the uranium concentration, the lower the ²³⁵U content. In addition, as the surrounding sandstone has lost its silica, it has turned to clay, preventing excessive groundwater migration and holding the uranium in place.

Thanks to a detailed analysis of the spectrum of fission products, we know that a certain number of fissions occurred in plutonium, the product of neutron capture in ²³⁸U and now completely disintegrated into ²³⁵U, since its half-life is only 24,000 years.  This has enabled physicists to calculate that, varying from one area to another, reactions have indeed taken place over an enormous period of time ranging from 150,000 to 850,000 years!

The reactors were "controlled" by several mechanisms, the main one being temperature: as fission power was released, the temperature increased.  A higher temperature means both an increase in neutron absorption (without fission) by ²³⁸U and a decrease in the effectiveness of water as a moderator: at a given temperature level, which varies with time and the gradual depletion of fissile uranium, reactions stabilize, as is the case in our reactors.

Combining geological and temperature considerations, we now estimate that the reactors in the northern part of the deposit operated at a depth of several thousand meters, beneath deltaic and then marine sediments.  At such depths, pressure and temperature conditions were close to those of today's pressurized water reactors (350 to 400°C, 15 to 25 Mpa), while the southern zones operated at a depth of around 500 meters, with conditions more akin to those of a boiling water reactor (250°C, 5 Mpa).

It is estimated that nuclear reactions in the uranium in centimeter- to meter-sized veins consumed about five tons of ²³⁵
U
and elevated temperatures to a few hundred degrees Celsius. Most of the non-volatile fission products and actinides have only moved centimeters in the veins during the last 2 billion years. Studies have suggested this as a useful natural analogue for nuclear waste disposal. The overall mass defect from the fission of five tons of ²³⁵
U is about 4.6 kilograms (10 lb). Over its lifetime the reactor produced roughly 100 megatonnes of TNT (420 PJ) in thermal energy, including neutrinos. If one ignores fission of plutonium (which makes up roughly a third of fission events over the course of normal burnup in modern human-made light water reactors), then fission product yields amount to roughly 129 kilograms (284 lb) of technetium-99 (since decayed to ruthenium-99), 108 kilograms (238 lb) of zirconium-93 (since decayed to niobium-93), 198 kilograms (437 lb) of caesium-135 (since decayed to barium-135, but the real value is probably lower as its parent nuclide, xenon-135, is a strong neutron poison and will have absorbed neutrons before decaying to ¹³⁵
Cs
in some cases), 28 kilograms (62 lb) of palladium-107 (since decayed to silver), 86 kilograms (190 lb) of strontium-90 (long since decayed to zirconium), and 185 kilograms (408 lb) of caesium-137 (long since decayed to barium).

Relation to the atomic fine-structure constant

The natural reactor of Oklo has been used to check if the atomic fine-structure constant α might have changed over the past 2 billion years. That is because α influences the rate of various nuclear reactions. For example, ¹⁴⁹
Sm
captures a neutron to become ¹⁵⁰
Sm
, and since the rate of neutron capture depends on the value of α, the ratio of the two samarium isotopes in samples from Oklo can be used to calculate the value of α from 2 billion years ago.

Several studies have analysed the relative concentrations of radioactive isotopes left behind at Oklo, and most have concluded that nuclear reactions then were much the same as they are today, which implies α was the same too.

See also

• Deep geological repository
• Geology of Gabon
• Mounana

Sources

• Bentridi, S.E.; Gall, B.; Gauthier-Lafaye, F.; Seghour, A.; Medjadi, D. (2011). "Génèse et évolution des réacteurs naturels d'Oklo" [Inception and evolution of Oklo natural nuclear reactors]. Comptes Rendus Geoscience (in French). 343 (11–12): 738–748. Bibcode:2011CRGeo.343..738B. doi:10.1016/j.crte.2011.09.008.

External links

• The natural nuclear reactor at Oklo: A comparison with modern nuclear reactors, Radiation Information Network, April 2005
• Oklo Fossil Reactors
• NASA Astronomy Picture of the Day: NASA, Oklo, Fossile Reactor, Zone 15 (16 October 2002)
• הכור הגרעיני של הטבע (in Hebrew language)