C. R. Physique 3 (2002) 839–849
Physique appliquée/Applied physics
DOSSIER
DU COMBUSTIBLE NUCLÉAIRE AUX DÉCHETS :
RECHERCHES ACTUELLES
FROM NUCLEAR FUELS TO WASTE: CURRENT RESEARCH
2 billion year old natural analogs for nuclear waste
disposal: the natural nuclear ﬁssion reactors in Gabon
(Africa)
François Gauthier-Lafaye
Centre de géochimie de la surface, UMR7517-CNRS-ULP, 1, rue Blessig, 67084 Strasbourg cedex, France
Received 24 January 2002; accepted after revision 20 March 2002
Note presented by Édouard Brézin.
Abstract Two billion years ago, the increase of oxygen in atmosphere and the high 235U/238U
uranium ratio (> 3%) made possible the occurrence of natural nuclear reactors on Earth.
These reactors are considered to be a good natural analogue for nuclear waste disposal.
Their preservation during such a long period of time is mainly due to the geological
stability of the site, the occurrence of clays surrounding the reactors and acting as an
impermeable shield, and the occurrence of organic matter that maintained the environment
in reducing conditions, favourable for the stability of uraninite. Hydrogeochemical studies
and modelling have shown the complexity of the geochemical system at Oklo and
Bangombé (Gabon) and the lack of precise data about uranium and ﬁssion products
retention and migration mechanisms in geological environments. To cite this article:
F. Gauthier-Lafaye, C. R. Physique 3 (2002) 839–849.
 2002 Académie des sciences/Éditions scientiﬁques et médicales Elsevier SAS
natural nuclear reactors / natural analogue / Oklo / Bangombé / uranium / precam-
brian
Des analogues naturels de sites de stockage de déchets nucléaires
vieux de 2 milliards d’années : les réacteurs de ﬁssion nucléaire
naturels du Gabon (Afrique)
Résumé Il y a 2 milliard d’années, les réacteurs nucléaires naturels du Gabon ont fonctionné grâce à
l’augmentation de l’oxygène dans l’atmosphère et à un rapport 235U/238U de l’uranium
élevé (> 3%) à cette époque. Ces réacteurs sont considérés comme de bon analogues
naturels de site de stockage de déchets nucléaires. Leur préservation est due en grande partie
à la stabilité géologique du site, la présence d’argiles entourant les réacteurs et constituant
un bouclier imperméable et à la présence de matières organiques qui a maintenu un
milieu réducteur favorable à la stabilité de l’uraninite. Les études hydrogéochimiques et les
modélisations géochimiques mettent en évidence la complexité des systèmes géochimiques
à Oklo et Bangombé et le manque de données précises sur les mécanismes de rétention
E-mail address: gauthier@illite.u-strasbg.fr (F. Gauthier-Lafaye).
 2002 Académie des sciences/Éditions scientiﬁques et médicales Elsevier SAS. Tous droits réservés
S1631-0705(02)01351-8/FLA 839
F. Gauthier-Lafaye / C. R. Physique 3 (2002) 839–849
et de migration de U et des produits de ﬁssion dans des environnements géologiques.
Pour citer cet article : F. Gauthier-Lafaye, C. R. Physique 3 (2002) 839–849.
 2002 Académie des sciences/Éditions scientiﬁques et médicales Elsevier SAS
réacteurs nucléaires naturels / analogues naturels / Oklo / Bangombé / uranium /
précambrien
1. Introduction
Natural nuclear ﬁssion reactors (hereafter just called reactors) are unique places in the world were ﬁssion
reactions took place spontaneously 2000 million years ago and have been sustained for quite a long time
(100 000 to 500 000 years) in a natural environment. The reactors now contain high-grade uranium ores
with a high content of ﬁssion products and their end members’ decay. This makes these geological objects
similar to places where spent nuclear fuels have been stored in a geological environment over a very long
period of time i.e. 2.0 b.y.
Since 1972, the year when the ﬁrst reactor was discovered in the uranium deposit of Oklo [1], many
programs have been conducted in order to understand the processes by which ﬁssion reactions could
occur and the effects of these nuclear reactions on the surrounding rocks. The reactors are considered
as good natural analogues for nuclear waste disposal and research is now mainly focused on the behaviour
of actinides and ﬁssion products in a geological environment. This paper tries to brieﬂy synthesise our
knowledge of the geological history of the reactors and their hosted uranium deposits and to give an
overview of what we have learned on the geological and geochemical conditions that led to the preservation
and the conservation of the reactors.
2. Geological background
2.1. Uranium deposits
The Franceville basin contained ﬁve uranium deposits (Oklo, Mounana, Boyindzi, Mikouloungou and
Bangombé) which have been now mined out with the exception of the Bangombé deposit that contains
the last of the natural nuclear ﬁssion reactors (Fig. 1). All together, these deposits contained 28 000 ton of
uranium that was mined from 1961 to 1999.
The uranium deposits of the Franceville basin are the oldest known high-grade uranium deposits
in sedimentary rocks that were formed by dissolution–precipitation processes [2]. Franceville uranium
deposits are younger than those of Witwatersrand, South Africa (2.7–2.9 GA Ma) and Blind River–Elliot
Figure 1. Schematic geological map of the
Franceville basin showing the location of the
uranium ore deposits.
Figure 1. Carte géologique schématique du bassin
de Franceville montrant la situation des gisements
d’uranium.
840
To cite this article: F. Gauthier-Lafaye, C. R. Physique 3 (2002) 839–849
Figure 2. Stratigraphic Log of the Francevillian
Series in the Franceville basin.
Figure 2. Log stratigraphic du Francevillian dans le
bassin de Franceville.
Lake, Canada (2.4 Ga) where most of the uranium has been concentrated by sedimentary (mechanical)
processes [3]. The high grade of uranium in the Franceville deposits implies uranium oxidation–reduction
processes during their formation and therefore a high oxygen concentration in the contemporaneous
hydrosphere and atmosphere. The age of the uranium deposit is 1950 ± 40 Ma as determined using the
classical U–Pb method on uraninites [4] and by dating the ﬁssion reactions themselves using the ﬁssiogenic
REE (Rare Earth elements) and uranium system [5–7].
Uranium deposits are located at the top of a 400 to 1000 thick FA formation mainly made of sandstones
and conglomerates (Fig. 2). This formation is underlain by the FB formation made of black shales with
organic carbon content ranging between 1 to 15% and few carbonate (dolomite) layers. In the central part
of the basin the FB formation is 1000 meters thick [8].
Most of the uranium is of the form of uraninite and is closely associated with migrated hydrocarbons that
ﬁll the secondary porosity of the sandstone [9]. Uranium mineralisation occurred during diagenesis, when
the mineralised FA formation was 3000 m deep and when oxidised uranium bearing ﬂuids met reduced
ﬂuids associated with hydrocarbons in tectonic traps [9]. Uranium bearing ﬂuids are hot (100–170 ◦C),
saline waters with concentrations of Ca up to 30% eq. wt% and Na up to 30 eq. wt.% [10,11]. The sources
of uranium are believed to be detrital uranium bearing minerals such as monazite and thorite which are
concentrated in several 100 m thick conglomerates in the lower FA formation [2,8].
841
F. Gauthier-Lafaye / C. R. Physique 3 (2002) 839–849
2.2. Uniqueness of the reactors of the Franceville basin
The conditions for the occurrence of nuclear ﬁssion reactions in a geological system have been discussed
by Naudet [7]. Fission reactions occur spontaneously in 238U resulting in the production of fast neutrons.
If these neutrons are slowed down and thermalised, they may induce ﬁssion in 235U or 239Pu. The nuclear
reaction can then be sustained. This can happen in a geological system if three conditions are met:
1. The uranium ore must have high uranium content.
2. The fast neutrons produced by the spontaneous ﬁssion of 238U must be thermalised by migration
through water or graphite.
Naudet [7] has shown that water should have been the main moderator for neutrons in the Oklo reactors.
Water is present in the porosity of the sandstones and in clays that were important mineralogical phases in
the reactors. Naudet suggests that carbon may also have played an important role in some organic matter
rich reactors.
3. The 235U content of the uranium should be high.
At Oklo this condition was met because the radioactive decay constant for 235U (99.8485 · 10−10) is
smaller than that for 238U (1.55125 · 10−10). Fig. 3 shows the variation of the 235U/238U ratio versus
time. 2.0 b.y. ago the ratio was similar to that of fuel in PWR (Pressured Water Reactor) nuclear plants.
There is little chance of ﬁnding natural ﬁssion reactors younger than 1.9 Ma because the amount of
ﬁssile 235U was then too small to sustain the ﬁssion reaction. At ages > 2.0 Ga, the 235U/238U was
sufﬁcient to sustain ﬁssion in natural reactors and even very high at ages > 2.5 Ga. This is the main
argument for predicting that ﬁssion reactors should have been very common prior to 2.0 b.y. However,
no traces of nuclear ﬁssion reactions have been found in pre-2.0 Ga uranium deposits such as those in the
Witwatersrand or at Elliot lake. This means that other conditions must be important as well. Of course, the
concentration of uranium is of particular importance and this may be achieved only in certain geochemical
conditions.
Fulﬁllment of conditions 1 and 3 allows a uranium deposit to reach the critical mass required to sustain
a ﬁssion reaction. If one takes into account the various physical (porosity) and chemical (content of various
minerals and neutron poisons such as REE, B, and V) properties of the Oklo sandstones, Naudet [7] has
shown that a sustained ﬁssion reaction would have occurred at Oklo where the uranium content was 10% in
one cubic meter. Such high uranium concentrations in a sedimentary rock imply that the uranium had been
subjected to one or more dissolution–precipitation episodes and implies that enough oxygen was available
in water to allow oxidation–reduction reactions of U to occur. Such a condition was not fulﬁlled before
2.0 Ga, the oxygen content of the atmosphere being low. Holland [12] has suggested that the PO2 level
Figure 3. Evolution of the 235U/238U ration versus
time.
Figure 3. Évolution du rapport 235U/238U en
fonction du temps.
842
Pour citer cet article : F. Gauthier-Lafaye, C. R. Physique 3 (2002) 839–849
in the atmosphere increased rapidly from < 1% PAL to > 15% PAL during the early Proterozoic (2.2–
< 1.9 Ga).
Therefore, it appears that 2.0 Ga is a crucial age that allows, at the same time, a high 235U/238U ratio
and, for the ﬁrst time in the Earth’s history, a high uranium content in the rocks.
3. Description of the reactors
Fifteen natural nuclear ﬁssion reactors were discovered in two deposits of the Franceville basin. Fourteen
are located in the Oklo deposit and one in the small deposit of Bangombé which is 30 km from Oklo. The
reactors of Oklo have been totally or partially mined out. Because the open pit and the underground mines
have been ﬂooded, the remaining parts of the reactors are no longer accessible. The reactor of Bangombé,
which is the smallest, has been preserved, together with the overall uranium deposit for further scientiﬁc
study on the behavior of ﬁssion products and actinides in a geological environment. The Bangombé reactor
is located very close to the surface (12 m for Bangombé and 100 to 260 m for the Oklo–Okelonbondo
reactors) and is therefore much more affected by weathering than the other one. Extensive geological,
mineralogical and geochemical descriptions of the reactors are given by [8,13,14] and in a European report
that synthesised the results obtained during the last 3-year European program on the behavior of ﬁssion
products in the water–rock systems of the various reactors [15].
A typical reactor is shown in Fig. 4 (reactor 9). The size of the reactors is quite variable. The biggest
reactor (reactor 2) is a 12 m long, 18 m deep and 20 to 50 cm thick lens. At Bangombé the reactor is
only 5 m long, 1 m wide and a few centimeters thick. Recent weathering has dissolved a large part of this
reactor [15].
The core of the reactors consists of a 5 to 20 cm thick layer of uraninite embedded in clays (illite and
chlorites). The uranium content of the core ranges between 40 and 60%. Accessory minerals are mainly
sulﬁdes (pyrite and galena), hematite and phosphates (mainly hydroxyapatite). The core of the reactors is
usually located in a clayey layer. The newly crystallised clay minerals reﬂect the thermal gradient during
the operation of the reactors [16].
Fission reactions started in high-grade uranium ore, when the uranium content of the sandstone
reached 10%. When ﬁssion reactions started, the temperature in the core of the reactor increased and a
convective hydrothermal system around the reactor started. The temperature in the core reached 400 ◦C
and decreased rapidly at the edges (geothermal gradient around 100 ◦C·m−1). Heat was transferred largely
Figure 4. Cross section of the
reactor 9, in the Oklo deposit.
From Gauthier-Lafaye et al. [14].
Figure 4. Coupe du réacteur 9
d’Oklo.
843
F. Gauthier-Lafaye / C. R. Physique 3 (2002) 839–849
by conduction [17]. The hydrothermal circulations led to the dissolution and migration of 80% of the silica
from the mineralised sandstone. This decreased the volume of the mineralised sandstone and increased
the uranium content of the residual layer which then consisted mainly of clays and uranium dioxide. Such
processes allow the formation of a reactor core containing 40 to 60% uranium.
In order to start and sustain a ﬁssion reaction, the concentration of ‘neutron’ poisons in the system
must be low. In geological systems B, V and REE are the most important poisons. Differences in the V
concentration of the various deposits in the Franceville basin may explain the distribution of the reactors.
The deposits of Mounana, Boyindzi and Mikouloungou have no reactors and a V content ranging from
> 100 ppm to 1%. Vanadium has even been mined at Mounana. Oklo has a V content that is always below
100 ppm. The concentration of B and REE is low in all of the uranium deposits. For B, this is due to the
meteoric origin of the mineralising ﬂuids, which did not migrate through the basement, a common source
of this element. If we consider that uranium was released by the leaching of detrital monazites then REE
and U fractionated during the migration of the ﬂuids from the conglomerates to the deposits due to varying
pH–Eh conditions [11].
4. Behaviour of ﬁssion products and actinides
Two main episodes of migration of actinides and ﬁssion products must be considered. The ﬁrst episode,
responsible for the so-called ‘old migrations’, took place during the ﬁssion reactions, or soon after, as the
reactors were still deeply buried (3000 m). The second episode of migration occurred more recently and
is still active. It took place when the reactors rose to the surface and went into contact with groundwater.
These migrations are dominant in the reactor of Bangombé whereas they are less important in the deeper
reactors of Oklo.
4.1. Old migrations
From the early studies on the Oklo reactors it has been recognised that some ﬁssion products or their
decay products had been retained almost completely, while others had been almost totally lost from the
vicinity of the reactors (Table 1) [14]. Elements that have been retained essentially within the core of the
reactors were Pu, Th, REE, Zr, Ru, Rh and Pd. All of these elements are soluble in the uraninite crystals
and have low solubility in natural groundwater, especially under reducing conditions that pertained in the
uranium deposits of Oklo. Other ﬁssion product elements, such as rare gases, I, Cd, alkali metals and
alkaline earths (e.g. Sr), appeared to have been totally removed from the reactors. These elements either are
gaseous or have high solubility in natural groundwater, even under reducing conditions.
Detailed analyses of speciﬁc isotopes allowed one to obtain information on the behaviour of their
precursors which were formed during the ﬁssion reactions. Analyses of isotopes such as 235U [18],
99Ru [19–23], 90Zr [24] and Bi [14], whose precursor are 239Pu, 99Tc, 90Sr and 237Np, respectively, allowed
one to show that these ﬁssion elements and actinides migrated during the ﬁssion reaction or soon after. The
processes that allowed their retention in the core of the reactors or its vicinity are highly variable and have
different efﬁciency.
In the reactor it has been recognised that the most efﬁcient matrix for the retention of actinides and
ﬁssion products is the uraninite mineral. However, it has been shown that other matrices such as apatite,
clay minerals, zirconium silicates and oxides (Fe, Mn) may have played a role in the retention of ﬁssion
products and actinides. As an example, Pu was stored in apatites [25] and chlorites [26] in the core of the
reactor 10. In the core of the reactors, between uraninites grains, 20 to 200 µm size metallic aggregates
containing ﬁssiogenic Ru, Rh and Te associated with As, Pb and S were found. These aggregates also exist
in artiﬁcial spent fuels of PWR plant suggesting their analogy with spent fuels.
It appears therefore that the reactors are similar to spent commercial reactor fuels in that for both
uraninites, the 235U/238U ratio at the time of reactor operation was similar and that there are solid solutions
844
To cite this article: F. Gauthier-Lafaye, C. R. Physique 3 (2002) 839–849
Table 1. Synthesis of the behavior of the ﬁssion products and actinides at Oklo; from Gauthier-Lafaye et al. [14]
Tableau 1. Inventaire des actinides et produits de ﬁssion à Oklo
Retention
Element Core of the reactor Clays of the reactor
Uraninites Inclusions Migration Clays Others
Cs +
Rb +
Sr +
Ba +
Mo + +
Tc + +
Ru + + ?
Rh + + ?
Pd + + ?
Y + +
Nb ?
Zr + + heavy.miner/oxide
Te
REE light + + apatite
Ce
Pb
Pb + galena
Bi
Th + + heavy miner.
U + + apatite
Np ?
Pu + + apatite
+ + Close relation to no relation between the minerals and the chemical elements.
between UO2 and the original ﬁssion products or the isotopes produced after decay. In all cases, and before
the ascent of the reactors, the redox conditions have been such that uraninite remained stable.
Uraninites at Oklo and Bangombé have a very constant chemical composition [27]. Their ‘impurities’
contents, when normalised to lead content are: SiO2: 0.7–1%; FeO: 0.3–0.7%; CaO: 1.5–1.7%. Lead
content, however, ranges between 0 in the very altered uraninites (affected by recent alterations) to 25%
in well-preserved uraninite of the deepest reactor (reactor 10). Most of uraninites at Oklo and Bangombé
indicate that they have been subjected to cofﬁnitisation, that is, the transformation of U(IV)O2−x to U(IV)silicate.
It has been shown that in old uraninites, this phenomenon did not change drastically the chemical
composition of uraninites [27]. It is possible that cofﬁnitisation occurred under anoxic conditions and is
related to the process of oxidative alteration of uraninites by water radiolysis and the later precipitation of
U(IV)-silicate as the reducing conditions are restored.
4.2. Recent migrations
Recent migrations of actinides and ﬁssion products were mainly studied in the Bangombé reactor which
is located close to the surface. The Bangombé reactor is located at a depth of around 12 m within an area
of regional discharge and is subjected therefore to weathering and chemical exchange with groundwater.
Present time equatorial climate originates the formation of a laterite proﬁle affecting the FB black shales
845
F. Gauthier-Lafaye / C. R. Physique 3 (2002) 839–849
Figure 5. Schematic Log of the bore hole BAX 3 at Bangombé.
From Pourcelot and Gauthier-Lafaye [16].
Figure 5. Log schématique du sondage BAX 3 à Bangombé.
underlying the reactor and the oxidation of the mineralised sandstone that host the reactor. An idealised
cross-section of the reactor is given in Fig. 5.
The main effect of migration of the surﬁcial oxidising ﬂuids into the reactor is the dissolution of uraninite
crystals and the resulting liberation of ﬁssion products in the environment. Therefore one of the main
questions that has been studied during the last European programme concerned the behaviour of these
elements in the geological environment and their possibilities to be retained in the core of the reactor or its
close vicinity.
Dissolution of uraninites has been observed either in the reactor core where secondary U(VI) minerals
(idiomorphic U, CA, Fe, Al, LREE phosphates, U (Al, Pb) and U (Fe, Ti) silicates) were observed [28,29]
and in the groundwater where uranium with 235U/238U ratio lower than 0.72% has been measured [30].
It is assumed that the present time shape and size of the reactor were greatly modiﬁed due to such
alteration processes. However, the fact that uraninite is still present after a long period of time suggests
that some geochemical processes should maintain the dissolution phenomenon at a very low rate. It has
been shown [31] that the stability of uraninite results from the buffer effects due to FeII/FeIII equilibrium
controlled by the iron bearing paragenesis and by the hydrolysis of graphite which is encountered within
the ‘argiles de pile’ in the close vicinity of the reactor. The reducing environment, which is generally
maintained inside the reactor can explain the conservation of the uranium in spite of the leaching by
oxidising groundwater. These apparent redox-buffering conditions happen to vary with time, i.e. changes in
the water table level, explaining the occurrence in the same samples of minerals representative of reducing
(uraninite, sulphides) and oxidising (uranophosphates, Fe-oxides, P-rich cofﬁnites) conditions.
Mechanisms of retention of uranium during weathering include the precipitation of new U-bearing
minerals and adsorption processes [29,32]. The vertical and lateral ﬂow of groundwater is responsible for
the oxidation and dissolution of primary sulphides originating acidic solution which facilitated the oxidation
and dissolution of uraninite. The uranyl cations so formed migrated and precipitated as uranyl minerals,
mainly phosphates, silicates, silico-phosphates and some of them, especially cofﬁnite, with high content of
P and LREE. Adsorption of uranium, together with P, mainly occurs on Fe-oxyhydroxides,but this uranium
retention process seems less efﬁcient than the precipitation one, at least in the close vicinity of the reactor.
On the other hand, in the weathered FB formation, U adsorption on clays and (Fe–Mn)-oxyhydroxides
minerals may be the dominant process [32].
846
Pour citer cet article : F. Gauthier-Lafaye, C. R. Physique 3 (2002) 839–849
Figure 6. Distribution of ﬁssiogenic Sm and Nd
in the bore hole BAX 8 which cross-cut the
reactor of Bangombé. Norm.Sm and norm.Nd
refer to Sm and Nd having normal isotopic
composition, not inﬂuenced by the ﬁssion
products. From Bracke et al. [33].
Figure 6. Distribution de Sm et Nd de ﬁssion
dans le sondage BAX 8 qui recoupe le réacteur
de Bangombé. Norm.Sm et norm.Nd
correspondent aux compositions isotopiques
normales de Sm et Nd, non perturbées par les
produits de ﬁssion.
Fissiogenic REE were liberated during the dissolution of uraninite and have migrated in the vicinity of
the reactor (Fig. 6) as it is suggested by their distribution along the bore-hole BAX8 [33] and the occurrence
in groundwater samples of Sm and Nd having isotopic composition reﬂecting ﬁssiogenic contribution [34].
Detailed studies on the mineralogy of the reactor and on the elemental and isotopic compositions of U,
Sm and Nd in groundwater samples show that new precipitation of P-LREE-cofﬁnites in the core of the
reactor allows the retention of most of the dissolved uranium and ﬁssiogenic LREE that were liberated by
the dissolution of primary uraninites. This observation is in agreement with the experimental results that
have shown that the UO2 stability in the solid phase and U(VI) predominance in the aqueous phase are
thermodynamically compatible, particularly in slightly acidic to neutral pH ranges and in transitional redox
waters [35].
5. Conclusions
Natural nuclear ﬁssion reactors occurrence is restricted to a very short period of time during the history
of the Earth, i.e. around 2000 Ma, when the oxygen content in the atmosphere and hydrosphere allowed
the oxidation of U(IV) to U(VI) whereas the 235U/238U ratio was still high, slightly higher than 3%.
Other ﬁssion reactors may have existed at that time. Recently, Meshik et al. [36] suggested that the 136Xe
content (part of which is of ﬁssiogenic origin) of the atmosphere was enhanced by precambrian nuclear
reactors similar to those of Oklo and which have involved 1% of the precambrian crustal uranium. However,
these reactors should have been destroyed during the various orogenesis that affected most of the ancient
sedimentary basins. The very unique preservation of the reactors of the Franceville basin is due to the
location of the basin in the centre part of the very stable western African craton. The occurrence of clays
around the reactor created an impermeable shield that protected the core of the reactor against the aggressive
diagenetic and weathering ﬂuids. The reduced potential around the reactor was maintained both by Fe2+
minerals (chlorite, pyrite, siderite) and by organic matter that played an important role in the preservation
of the reactors.
Uraninite crystals have shown good capacity to retain most of the actinides produced by the ﬁssion
reactions and most of the ﬁssion products that have ionic radii close to that of uranium. When uraninite
has been dissolved, i.e. in the weathered zone of the Bangombé reactor, the reduced conditions in the close
847
F. Gauthier-Lafaye / C. R. Physique 3 (2002) 839–849
vicinity of the reactor allow its precipitation in newly formed Si–P–REE–uranium minerals. Around the
reactor, when more oxidising conditions prevail, adsorption of U on clays and Fe-oxyhydroxides represent
an important mechanism for its retention. Hydrogeochemicalstudies have shown that at the metric scale, the
chemical conditions are highly variable around the reactor of Bangombé involving therefore many different
mechanisms for the retention of uranium and REE. It is of great importance to be able to detect each of
these processes and estimate their importance in the overall uranium and other elements (REE, platinum
metals, etc.) migrations.
The geochemical models that study and predict the behaviour of uranium around the reactors at
Bangombé and Oklo are good examples which illustrate our capability to understand such a complicate
natural system and our knowledge of the mechanism of retention and migration of uranium and
ﬁssion products in a natural system. It appears that the basic information regarding the kinetics and
thermodynamics of these processes are scarce [37] and in the future, a great effort must be put in that
direction in order to make the models more precise and to increase our conﬁdence in them.
Acknowledgement. I would like to very much thank Francis WEBER for the fruitful discussions and his helpful
comments on the manuscript.
References
[1] R. Bodu, H. Bouzigues, N. Morin, J.P. Pﬁffelmann, C. R. Acad. Sci. Paris 275 (1972) 1731–1734.
[2] F. Gauthier-Lafaye, F. Weber, Econ. Geol. 84 (1989) 2267–2285.
[3] F.J. Dahlkamp, Uranium Ore Deposits, Springer-Verlag, Berlin, 1993.
[4] A.J. Gancarz, in: Proc. The Natural Fission Reactors, IAEA, 1978, pp. 513–520.
[5] J.C. Ruffenach, in: Proc. The Natural Fission Reactors, IAEA, 1978, pp. 441–471.
[6] P. Holliger, C. R. Acad. Sci. Paris 307 (1988) 367–373.
[7] R. Naudet, Oklo : des réacteurs nucléaires fossiles, Collection du Commissariat à l’Energie Atomique, Paris, 1991,
695 p.
[8] F. Gauthier-Lafaye, in: Mémoire Sciences Géologiques, Vol. 78, 1986.
[9] F. Gauthier-Lafaye, F. Weber, C. R. Acad. Sci. Paris 292 (1981) 69–74.
[10] R. Mathieu, Thesis, Institut National de Polytechnique de Lorraine, Nancy, 1999, 518 p.
[11] R. Mathieu, L. Zetterström, M. Cuney, F. Gauthier-Lafaye, H. Hidaka, Chem. Geol. 171 (2000) 147–171.
[12] H.D. Holland, in: S. Bengtson (Ed.), Early Life on Earth, Nobel Symposium N◦84, Columbia University Press,
New York, 1994, pp. 237–244.
[13] F. Gauthier-Lafaye, F. Weber, H. Ohmoto, Econ. Geol. 84 (1989) 2286–2295.
[14] F. Gauthier-Lafaye, P. Holliger, P.-L. Blanc, Geochim. Cosmochim. Acta 60 (1996) 4831–4852.
[15] F. Gauthier-Lafaye, E. Ledoux, J. Smellie, D. Louvat, V. Michaud, L. Pérez del Villar, V. Oversby, J. Bruno,
OKLO-Natural Analogue Phase II. Behaviour of nuclear reaction products in a natural environment. European
Commission. Nuclear Science and Technology Serie. Contract: FI4W-CT96-0020. Final report. EUR 19139 EN.,
2000, 116 p.
[16] L. Pourcelot, F. Gauthier-Lafaye, Chem. Geol. 157 (1999) 155–174.
[17] B. Gerard, J.J. Royer, C. le Carlier, M. Pagel, H. Scius, F. Gauthier-Lafaye, in: Louvat, von Maravic (Eds.), Proc.
Sitges Meeting, 19–20 June 1997. Oklo Phase II Workshop, in: EUR Report Series, Vol. 18314, 1997, pp. 301–
308.
[18] R. Bros, L. Turpin, F. Gauthier-Lafaye, Ph. Holliger, P. Stille, Geochim. Cosmochim. Acta. 57 (1993) 1351–1356.
[19] R. Hagemann, M. Lucas, G. Nief, E. Roth, Earth Planetary Sci. Lett. 23 (1974) 170–176.
[20] A. Gancarz, G. Cowan, D. Curtis, W. Maek, in: Scientiﬁc Basis for Nuclear Waste Management, Vol. 2, 1980,
pp. 601–608.
[21] H. Hidaka, K. Shinotsuka, P. Holliger, Radiochim. Acta 63 (1993) 19–22.
[22] H. Hidaka, P. Holliger, F. Gauthier-Lafaye, Chem. Geol. 155 (1999) 323–333.
[23] D. Curtis, T. Benjamin, A. Gancarz, R. Loss, K. Rosman, J. DeLaeter, J.E. Demore, W.J. Maek, Appl. Geochem. 4
(1989) 49–62.
[24] H. Hidaka, T. Sugiyama, M. Ebihara, P. Holliger, Earth Planetary Sci. Lett. 122 (1994) 173–182.
[25] R. Bros, J. Carpena, V. Sere, A. Beltritti, Radiochim. Acta 74 (1996) 277–282.
[26] R. Bros, L. Turpin, F. Gauthier-Lafaye, Ph. Holliger, P. Stille, Geochim. Cosmochim. Acta 57 (1993) 1351–1356.
[27] K.A. Jensen, R.C. Ewing, F. Gauthier-Lafaye, in: W. Gray, I. Triay (Eds.), Scientiﬁc Basis for Nuclear Waste
Management XX, in: Materials Research Society Symposium Proceedings, Vol. 465, 1997, pp. 1209–1218.
848
To cite this article: F. Gauthier-Lafaye, C. R. Physique 3 (2002) 839–849
[28] S. Salah, F. Gauthier-Lafaye, M. Del Nero, G. Bracke, in: European Commission, Nuclear Science and
Technology, OKLO Working Group Proceedings of the Final Meeting OKLO-Natural Analogue Phase II Project
held in Cadarache, France, from 20 to 21 May 1999, EUR 19137EN, 2000, pp. 75–90.
[29] L. Pérez del Villar, J.S. Cozar, J. Pardillo, M.A. Labajos, in: European Commission, Nuclear Science and
Technology, EUR 19137 EN, 2000, pp. 45–74.
[30] D. Louvat, K. Lot, J.-L. Michelot, J. Smellie, C. Tuniz, in: European Commission, Nuclear Science and
Technology, EUR19116 EN, 2000, pp. 391–398.
[31] B. Madé, E. Ledoux, A.-L. Salignac, B. Le Boursicaud, I. Gurban, C. R. Acad. Sci. Paris 331 (2000) 587–594.
[32] S. Salah, Thèse, Univ. Louis Pasteur, Strasbourg, December 2000.
[33] G. Bracke, S. Salah, F. Gauthier-Lafaye, Environmental Geol. 40 (2001) 403–408.
[34] P. Stille, F. Gauthier-Lafaye, K.A. Jensen, S. Salah, G. Bracke, P. Gomez, R. Exing, D. Louvat, Chem. Geol.
(2001), submitted.
[35] I. Casas, J. de Pablo, M.E. Torrero, J. Bruno, E. Cera, R.J. Finch, R.J. Ewing, Geochim. Cosmochim. Acta 62
(1998) 2223–2231.
[36] A.P. Meshik, K. Kehm, C.M. Hohenberg, Geochim. Cosmochim. Acta 64 (2000) 1651–1661.
[37] V. Michaud, D. Louvat, in: European Commission, Nuclear Science and Technology, EUR 19137EN, 2000,
pp. 401–408.
Discussion
Question de É. Brézin
Est-ce qu’aux époques encore plus reculées où la proportion d’235U était encore bien supérieure on peut
imaginer qu’il y a eu des phénomènes de nature explosive?
Réponse de F. Gauthier-Lafay
On pense que avant 2000 Ma d’années il n’y avait pas de processus géochimiques permettant d’obtenir de
fortes teneurs en uranium dans la nature. Aussi, même si les proportions en 235U par rapport 238U étaient
très fortes avant 2000 Ma il y a peu de chance d’avoir pu obtenir une masse critique. De plus, Naudet
a montré que dans les terrains géologiques et en particulier à Oklo, la chaleur produite par les réactions
de ﬁssion s’évacue très mal ce qui assure la stabilité des réactions nucléaires et interdit que les réacteurs
puissent « s’emballer».
849