← all systems · see it in The System
Oklo natural nuclear reactors
Uranium-rich zones of the Oklo, Okelobondo and Bangombe deposits in Gabon's Francevillian basin where, about two billion years ago, groundwater-moderated self-sustaining nuclear chain reactions ran, in pulses, for some hundreds of thousands of years: 'a uranium deposit where self-sustaining nuclear chain reactions occur through natural geological processes rather than human construction' (encyclopedia).
Properties published scienceour proposal
Sourced fields, and our own readings.
| aliases | Oklo natural nuclear reactor |
| born at | 1950 +/- 40 Myr ago (contested: 1.7-2.05 Gyr, C-1260) |
| born precision | +/- 40 Myr (Davis et al.) |
| boundary | Inside: the reaction zones (lens-shaped uraninite-rich layers about 10 m by 10 m and up to 0.8 m thick), their pore water, the neutrons and the fission products and actinides they made, through the zones' operation and the record they left. Outside: the host sandstone and clays beyond the reflector, the groundwater system, the atmosphere whose oxygen made the ore, and the people who mined and read it (src:davis-2014-oklo). |
| end precision | contested |
| ended at | about 1e5-1e6 years after start (C-1261) |
| nested level | Built from uraninite ore, water, neutrons and fission; a part of Earth's crust in the Francevillian basin; one known case of the natural nuclear fission reactor. |
| scale | planetary |
Measures (49) published science
| Cs/Ba capture closure after shutdown | 4.98 +/- 0.56 years |
| operating cycle (zone RZ13) | on about 0.5 h, off about 2.5 h (about 3 h period) hours |
| operating duration (Davis et al.) | 100 to 300 thousand (zones 10, 13, Bangombe) years |
| operating duration per zone (Uzan) | RZ2 8.5e5; RZ10 1.6e5 years |
| operating duration (Ngouba Madinda et al.) | 0.6 to 1.5 million years |
| operating duration (heat-transfer model) | 1e5 to 1e6 years |
| Oklo age (Davis et al.) | 1950 +/- 40 Myr ago |
| Oklo age (Bauer et al.) | 2050 +/- 30 Ma |
| Oklo age (Gauthier-Lafaye et al. 2023) | 1.95 Gyr ago |
| Oklo age (Uzan) | about 1.8e9 years ago |
| Oklo age (encyclopedia) | approximately 1.7 billion years ago |
| age of the Francevillian sediments | 2.265 +/- 0.15 Gyr |
| U-235 half-life | 710 Myr |
| U-238 half-life | 4.51 Gyr |
| U-235 enrichment in natural uranium over time | 17% at solar-system formation; 4.0% 2.10 Gyr ago; 2.3% 1.40 Gyr ago; 1.3% 700 Myr ago; 0.7202% today atom % |
| U-235 share of natural uranium about two billion years ago | about 3.68 % |
| U-235 abundance in Oklo ore | 0.600 (first anomaly); 0.440 (sample M) % |
| most depleted natural uranium (RZ13) | 235U/238U = 0.3655 +/- 0.0007 (2 sigma) % |
| reactor zone size | about 10 m x 10 m x up to 0.8 m (lens-shaped) |
| uraninite content | up to 80 wt % |
| uranium in the rich ores | up to 60 % |
| modelled multiplication factor k_eff | RZ2 1.033; RZ10 1.036 dimensionless |
| poison tolerance | RZ2 10 ppm; RZ10 under 1 ppm (0.8 ppm in the model) boron-10 equivalent |
| spectral index r0 | RZ2 0.20-0.25; RZ10 0.15 +/- 0.02 dimensionless |
| thermal power | 10 to 50 (Uzan); never above about 100 (encyclopedia); 18 in the RZ10 gamma model kW |
| U-235 fissioned | RZ2 1800; RZ10 650 kg |
| average neutron flux density | about 1e8 (five orders below power reactors) cm^-2 s^-1 |
| neutron fluence | about 1e21 cm^-2 |
| prompt gamma flux (RZ10 model) | about 3e9 cm^-2 s^-1 |
| core temperature and pressure (PWR-like reading) | about 300 C, about 150 atm C; atm |
| core temperature (lutetium thermometry, RZ10) | 100 +/- 30 (revised); earlier RZ2 260, RZ3 280 C |
| pore-water temperature pulses (model) | pulses 70-100; plateau 30-70 C |
| water's critical point (upper limit for the moderator) | 374 C, 218 atm C; atm |
| reactor zone depths | RZ10 400 m; Okelobondo OK84 about 450 m; Bangombe 10-12 m (11 m) |
| 149Sm/147Sm ratio | about 0.9 natural; about 0.02 in Oklo ores (0.003 in sample M) ratio |
| Sm-149 resonance energy today | 97.3 meV |
| RZ10 effective Sm-149 capture cross section | 85.0 +/- 6.8 kb |
| allowed resonance shift (RZ10) | right branch -11.6 to 26.0; left branch -101.9 to -79.6 meV |
| fission Xe and Kr in Oklo phosphates | up to 0.03 cm3 STP per g |
| Oklo bound on alpha change (Davis and Hamdan 2015) | |alpha_Oklo - alpha_now| < 1.1e-8 alpha_now (95%) fraction |
| Oklo bound on alpha change (Gould et al. 2006) | -0.24e-7 to 0.11e-7 (2 sigma) fraction |
| Oklo bound on alpha rate (Petrov et al. 2006) | -4e-17 to 3e-17 per year |
| Oklo bound on alpha change (Damour and Dyson 1996) | -0.9e-7 to 1.2e-7 (95%) fraction |
| Oklo non-zero alpha change (Lamoreaux 2004) | >= 4.5e-8 (6 sigma) fraction |
| Oklo alpha change over 1.9 Gyr (Davis 2023) | < 0.01 ppm (95%) fraction |
| Shlyakhter's limits on constant variation | strong ~2e-19; electromagnetic ~5e-18; weak ~1e-12 per year |
| formamide from an Oklo-like fission zone (estimate) | 0.1 to 1 mol m^-2 yr^-1 |
| oxygen rise at the Great Oxidation Event | about a factor of a hundred, about 2.2 Gyr ago ratio |
| papers on Oklo since 2000 | about 140 papers |
Open contentions (6) open contention
Where sources or readings disagree. Both stay on the map.
- C-1260 How old are the natural nuclear reactors at Oklo?
booked: The map uses 1,950 ± 40 million years (2014), and keeps the estimates of about 1.7 and 2 billion years alongside.
the other reading: 2050 +/- 30 Ma (Bauer et al. 2025); 1.95 Gyr (Gauthier-Lafaye et al. 2023); about 1.8e9 yr (Uzan 2011); about 1.7 Gyr (encyclopedia) - C-1261 How long did the Oklo natural nuclear reactors run?
booked: The map uses 100,000 to 300,000 years, and keeps the longer estimates of up to 1.5 million years alongside.
the other reading: per zone RZ2 8.5e5 yr, RZ10 1.6e5 yr; 'shut a few million years later' (Uzan 2011); 1e5-1e6 yr (Adam 2016) - C-1262 How hot did the Oklo natural reactors get?
booked: The map keeps both: near 300 °C, and pulses of 70-100 °C, with some models reaching about 550 °C.
the other reading: up to about 550 C in some criticality models (cited by Davis et al.) - C-1445 Has the strength of electromagnetism changed over two billion years?
booked: The map follows the analyses of Oklo that find no change, and keeps the one that found a change alongside.
the other reading: a non-zero change, delta alpha/alpha >= 4.5e-8 at 6 sigma (Lamoreaux 2004), and the non-zero left-branch solution for the Sm resonance shift, 'not ruled out on experimental grounds' (Davis et al. 2014); also open: how reliable the Damour-Dyson sensitivity is (Davis et al. 2014 found an unnoticed assumption) - C-1446 Why did the Oklo natural reactors shut down?
booked: The map keeps both: fuel ran low and by-products built up, and the water ran out.
the other reading: a mix varying zone by zone (no source read rules on it) - C-1447 How many natural reactor zones were there at Oklo?
booked: The map uses sixteen, the count most sources give, and keeps the other counts alongside.
the other reading: fifteen (Gauthier-Lafaye as cited by Davis et al.: fourteen at Oklo-Okelobondo plus Bangombe); seventeen ('other authors', Davis et al.)
Valence: what it trades, and with whom (17) published science
Each exchange with a neighbour, read from both sides: who gains, who pays, and what carries it.
- Nuclear fission via sys:residual-strong-force · Oklo natural nuclear reactors gain
- Atomic nucleus via sys:residual-strong-force · Oklo natural nuclear reactors gain
- Water molecule via sys:water-molecule · Oklo natural nuclear reactors mixed
- Electromagnetic field via sys:electromagnetic-field · Oklo natural nuclear reactors mixed
- Neutron via real:neutron · Oklo natural nuclear reactors gain
- Earth's crust (primordial, oceanic and continental) via sys:earth-crust · Oklo natural nuclear reactors gain
- Formamide via sys:electromagnetic-field · Oklo natural nuclear reactors loss
- Cyanobacterium (Cyanobacteria, phylum Cyanobacteriota) via sys:atmosphere-of-earth · Oklo natural nuclear reactors gain
- Thermal energy (heat, temperature and heat flow) via sys:thermal-energy · Oklo natural nuclear reactors loss
- Great Oxidation Event (the oxygen crisis and what it released) via sys:groundwater · Oklo natural nuclear reactors gain
- Human (Homo sapiens) via sys:earth-crust · Oklo natural nuclear reactors loss
- Science (as a human social and cultural system) via sys:electromagnetic-field · Oklo natural nuclear reactors even
- Local position invariance via sys:gravitational-field · Oklo natural nuclear reactors even
- Infiltration, groundwater and aquifers via sys:groundwater · Oklo natural nuclear reactors gain
- Buried organic matter via sys:groundwater · Oklo natural nuclear reactors gain
- Uranium-235 via sys:uraninite · Oklo natural nuclear reactors gain
- Plutonium-239 via sys:residual-strong-force · Oklo natural nuclear reactors gain
Partners and parts published science
Built from / builds (6)
- Water molecule
- Neutron
- Nuclear fission
- Clay minerals (secondary, charged sheets)
- Uraninite ore
- Uranium mining and science over the Oklo reaction zones
Sits inside / holds (4)
- Nuclear fission
- Earth's crust (primordial, oceanic and continental)
- Paleoproterozoic era (2500-1600 Ma)
- Francevillian basin
expressed by (6)
- Generation/Synthesis
- Templating / Standardization
- Niche Construction (alter environment)
- Resource Denial/Attrition
- Physical/Virtual Barrier
- Quality Control & Error Correction
floor filled by (17)
- Nuclear fission
- Water molecule
- Neutron
- Earth's crust (primordial, oceanic and continental)
- Thermal energy (heat, temperature and heat flow)
- Science (as a human social and cultural system)
- Thermal neutron
- Xenon
- Ruthenium (Ru, Z = 44)
- Neodymium (Nd, Z = 60)
- Samarium (Sm, Z = 62)
- Infiltration, groundwater and aquifers
- Uranium-238
- Uranium-235
- Plutonium-239
- Neutron poison (absorber)
- Uraninite ore
Kinds (6)
- Bangombe natural reactor
- Okelobondo natural reactor
- Deep, little-weathered reaction zones (RZ10, RZ13, RZ16)
- Shallow open-pit reaction zones (RZ1-RZ9)
- Nuclear fission reactor
- Natural nuclear fission reactor
wired (5)
- Water molecule
- Neutron
- Earth's crust (primordial, oceanic and continental)
- Human (Homo sapiens)
- Science (as a human social and cultural system)
Notes published science
- Life built a nuclear reactor without meaning to: cyanobacteria raised oxygen about a hundredfold, oxygen made uranium dissolve, groundwater carried it into organic-rich sediments where it piled up as ore up to 60% uranium, and about two billion years ago that ore went critical. The reviewers call the reactors 'basically formed by cyanobacteria' (src:davis-2014-oklo).
- The reactor ran on a three-hour heartbeat and wrote it down: about 30 minutes critical until its water boiled away, about 2.5 hours dormant until water seeped back, read two billion years later from xenon trapped in phosphate grains (src:meshik-2004-oklo-cycling).
- Oklo is a fossil tuning fork for physics: samarium-149 has a capture resonance at 97.3 meV that sits where electric repulsion nearly cancels the strong force, so if the fine-structure constant had drifted, Oklo would have burned a different amount of samarium. Most analyses say alpha has changed by less than about one part in a hundred million in two billion years; one analysis claimed a change, and a non-zero solution is still 'not ruled out on experimental grounds' (src:davis-hamdan-2015; src:davis-2014-oklo; src:lamoreaux-2004-oklo).
Layers
The ways it takes in its world, where they meet, and how it acts back, each from L1 (where it meets the world) up.
the Oklo natural reactors as reactor physicists and geologists describe them (reader's view)
- L1: The ore body: uraninite-rich lenses in Francevillian sandstone, soaked in groundwater, with almost no neutron poisons
- L2: Moderation: water slows fission neutrons to thermal energies; the zone large enough to keep them
- L3: Neutron events: fission of U-235 (some U-238), capture by U-238 and by absorbers
- L4: Criticality: the neutron economy, k_eff just above one
- L5: The burn: power of tens of kW, fuel burned, products made, poisons burned, plutonium bred
- L6: The self-regulating cycle: 30 min on, 2.5 h off, held by boiling and returning water
- L7: The reactor's life and what it leaves: hundreds of thousands of years of pulses, shutdown, and an isotopic record held in the rock for two billion years
the boardroom (shared floors)
- L4: The neutron economy: multiplication factor k_eff (modelled 1.033 for RZ2, 1.036 for RZ10); critical when the zone's size, water, fuel and poisons let it pass one
- L5: The burn: fission rate and power (10-100 kW), U-235 burned (RZ2 1800 kg over 8.5e5 yr), fission products made (neodymium matching fission yields), absorbers burned (Sm-149), Pu-239 bred
- L6: The self-regulating cycle: about 30 minutes critical, about 2.5 hours dormant, held by the negative void coefficient (Petrov 1977; Meshik et al. 2004)
- L7: The reactor's life as a geological event: start (1950 +/- 40 Myr ago), 1e5-1e6 years of pulses, shutdown, then two billion years of keeping its products in place
way in: Burial: depth, pressure and the water's boiling point
- L1: Overburden: RZ10 at 400 m, Okelobondo OK84 about 450 m, Bangombe 10-12 m
- L2: Pressure keeps the pore water liquid (about 150 atm; water's critical point 374 C and 218 atm sets a plausible upper limit)
- L3: Uzan: RZ2 and RZ10 worked at several thousand metres, water pressure and temperature close to PWR conditions (20 MPa, 300 C)
way in: Fuel: uranium-235 in uraninite ore
- L1: Uraninite (UO2) ore in lens-shaped layers about 10 m by 10 m and up to 0.8 m thick; uraninite up to 80 wt% in some samples
- L2: The U-235 share then: about 3.7-4% (4.0% 2.10 Gyr ago; 3.68% about two billion years ago), set by U-235's faster decay (710 Myr against 4.51 Gyr for U-238)
- L3: Thermal neutrons captured by U-235 and fission (sigma_5 about 0.6 kb)
way in: Heat: the core's own temperature
- L1: Core and pore-water temperature: near 300 C at about 150 atm (PWR-like) or lower (lutetium reanalysis 100 +/- 30 C for RZ10; model pulses up to 70-100 C), contested
- L2: With rising temperature the moderating water turns to steam and can leave the active core
- L3: Negative void coefficient of reactivity: less moderation, less reactivity
way in: Moderator: groundwater in the pores and cracks
- L1: A large amount of water in the pores and cracks of minerals (condition ii); the Oklo reactor formed when the ore was flooded with groundwater
- L2: Hydrogen slows fission neutrons: a thermal component below about 0.5-1 eV and an epithermal 1/E tail
- L3: The amount of water sets the shape of the neutron spectrum (spectral index r0 0.20-0.25 for RZ2, 0.15 for RZ10)
way in: Neutrons: the chain's own inner channel
- L1: Free neutrons from fission in the ore: an integrated fluence of about 10^21 per cm^2; mean flux about 10^8 per cm^2 per s, far higher in a pulse
- L2: Thermal and epithermal populations, set by the water and the temperature
- L3: Fate of each neutron: fission of U-235 (and, markedly in RZ13, fast fission of U-238), capture by U-238 into Pu-239, or capture by absorbers
way in: Ore supply: oxidised groundwater bringing uranium (deep time)
- L1: Oxidised groundwater carrying soluble uranium(VI) after the Great Oxidation Event (oxygen up about a hundredfold, about 2.2 Gyr ago)
- L2: Reductive precipitation in organic-rich rocks: by carbon, methane or iron, or by microbes
- L3: High-grade ore: uranium up to 60% of the ore
way in: Poisons: neutron absorbers in and around the ore
- L1: Boron, lithium and the rare earths in the ore and gangue (nearly absent at Oklo: condition iii); later fission-product absorbers (Sm-149, Gd, Xe-135, the strongest known)
- L2: Capture without fission: Sm-149 (sigma >= 70 kb) burned down during operation
- L3: Tolerance: RZ10 critical only below 1 ppm boron-10 equivalent, RZ2 up to 10 ppm
way in: Size and reflector: the rock around the zone
- L1: Water-saturated sandstone around the zone (modelled as a 1 m reflector around a 6 m by 70 cm cylinder)
- L2: The zone large enough that neutrons did not escape faster than they were produced
- L3: Thickness, radius, porosity and uranium content set the multiplication factor (two-group diffusion, RZ9)
way out: Fission products and actinides into the ore and minerals (the record)
- L1: Held in the rock: uranium and most rare earths barely moved in two billion years; non-volatile products and actinides moved centimetres
- L3: Captured in minerals: Xe and Kr in aluminous phosphates (up to 0.03 cm3 STP/g), Cs and Ba in Ru metal/sulfide aggregates about 5 years after shutdown; elsewhere Cs and Ba nearly all lost
- L5: Fission products and actinides made: rare earths, Zr, Ru (from Tc-99), Xe and Kr, Cs and Ba; Pu-239 bred
way out: Gamma rays and neutrons into the surroundings
- L1: Ionising radiation in the surrounding rock and water; modelled radiolysis of acetonitrile to formamide at 0.1-1 mol per m^2 per year
- L3: Prompt gamma flux about 3e9 per cm^2 per s in the RZ10 model (18 kW, on 0.5 h, off 2.5 h)
- L5: Prompt gammas from fission and capture (about 7 MeV each per event) and delayed gammas from fission products (about 6.3 MeV)
way out: Heat into the rock and pore water
- L1: Heat in the host rock and groundwater; convective cooling
- L3: Temperature oscillations in the pore water, PCR-like: pulses up to 70-100 C, plateau 30-70 C, relaxation (heat-transfer model)
- L5: Fission energy appears as heat: 10-50 kW (Uzan), never above about 100 kW
way out: Steam: the moderator driven off
- L1: Water away from the core until it cools and seeps back
- L2: Water converted to steam, which can even leave the active core
- L6: The cycle turns the reactor off by boiling its moderator
Sources (27) published science
Uzan J-P (2011) Varying constants, gravitation and cosmology. Living Rev. Relativ. 14, 2 (doi:10.12942/lrr-2011-2), section 'The Oklo phenomenon' (behind its links)
Hidaka H, Bentridi SE, Gall B (2023) Evidence of fast neutron operating in Oklo. Radiat. Prot. Dosimetry (doi:10.1093/rpd/ncad015) (behind its links)
Bauer AM, Li et al. (2025) Subaerial oxidative uranium mobilization at the culmination of the Great Oxidation Event. PNAS (doi:10.1073/pnas.2510289122), CC BY-NC-ND (behind its links)
Gauthier-Lafaye F, Gall B, Bentridi SE (2023) The Oklo phenomenon, the discovery, first questions and first answers. Radiat. Prot. Dosimetry (doi:10.1093/rpd/ncad044) (behind its links)
Gould CR, Sharapov EI, Lamoreaux SK (2006) Time-variability of alpha from realistic models of Oklo reactors. Phys. Rev. C 74, 024607; arXiv:nucl-ex/0701019 (behind its links)
Lamoreaux SK (2004) Neutron moderation in the Oklo natural reactor and the time variation of alpha. Phys. Rev. D 69, 121701; arXiv:nucl-th/0309048 (behind its links)
Davis ED (2023) Oklo natural fission reactors and dynamical models of dark energy. Radiat. Prot. Dosimetry (doi:10.1093/rpd/ncad020) (behind its links)
Shlyakhter AI (1983 preprint, posted 2003) Direct test of the time-independence of fundamental nuclear constants using the Oklo natural reactor. arXiv:physics/0307023 (behind its links)
Cherif KM, Seghour A, Dehimi FZ (2019) Criticality of the reaction zone 9 of Oklo reactors revisited. Appl. Radiat. Isot. (doi:10.1016/j.apradiso.2019.04.018) (behind its links)
De Laeter JR, Hidaka H (2007) The role of mass spectrometry to study the Oklo-Bangombe natural reactors. Mass Spectrom. Rev. 26 (doi:10.1002/mas.20141) (behind its links)
Gall BJ (2023) Parallels between natural Oklo reaction zones and industrial reactors dynamics. Radiat. Prot. Dosimetry (doi:10.1093/rpd/ncad034) (behind its links)
Nimal JC (2023) Oklo: historic and lessons learned. Radiat. Prot. Dosimetry (doi:10.1093/rpd/ncad043) (behind its links)
El Khoury A et al. (2025) A battle against arsenic toxicity by Earth's earliest complex life forms. Nat. Commun. (doi:10.1038/s41467-025-59760-9) (behind its links)