Carbonaceous substances in Oklo reactors—Analogue for permanent deep geologic disposal of anthropogenic nuclear waste
Published:January 01, 2008
- PDF LinkChapter PDF
David J. Mossman, François Gauthier-Lafaye, Adriana Dutkiewicz, Ralf Brüning, 2008. "Carbonaceous substances in Oklo reactors—Analogue for permanent deep geologic disposal of anthropogenic nuclear waste", Deep Geologic Repositories, Norbert T. Rempe
Download citation file:
Seventeen known natural nuclear fission reactors sustained criticality in hydrothermally altered low-grade metasedimentary rocks of the Proterozoic Franceville Series ca. 1968 ± 50 Ma. About half of these reactors contain only traces of carbonaceous substances, and in these, fission products including strontium, cesium, rubidium, and boron migrated away from the reactors and were nearly completely lost. The others are rich in carbonaceous substances, particularly solid, partly graphitized bitumen and kerogen, as well as liquid oil in fluid inclusions. In these carbonaceous substance–rich reactors, uranium and fissiogenic isotopes are held in uraninite, which became enclosed in liquid bitumen during criticality and was subsequently fixed when the bitumen solidified. The preservation of liquid oil in fluid inclusions for over two billion years suggests that time is not a parameter that controls petroleum degradation. This is viewed as a potentially important aspect in engineered nuclear waste containment projects. Likewise, because of the hydrophobic qualities of solid bitumen, its inducible plasticity, and its capability of conversion to nonflammability, it deserves serious consideration for use in permanent deep geologic disposal sites. Indications are that at Oklo, Gabon, carbonaceous substances have combined to restrain the migration of radionuclides and limit the extent to which migration occurs. The strength of the Oklo analogue lies in the fact that it represents more extreme conditions than those likely to be met in a deep geologic repository. The carbonaceous substance–rich reactors of Oklo stand as time-tested analogues for anthropogenic nuclear waste containment strategies.
The Proterozoic Franceville Series of Gabon consists of a 1–4-km-thick succession of mainly siliciclastic sedimentary rocks deposited ca. 2.1 Ga (Gauthier-Lafaye and Weber, 1989). In this series, the uranium-bearing FA Formation and the organic-rich FB Formation have long been the focus of scientific interest. McKirdy and Imbus (1992) described the carbonaceous units in the FB Formation as the most impressive accumulation of reduced organic matter of its age. The maximum burial temperature of the FA and FB Formations is believed not to have exceeded 175 °C, and for the most part, the rocks are little deformed (Pourcelot and Gauthier-Lafaye, 1999; Gauthier-Lafaye and Weber, 1989) (Fig. 1).
During burial, almost 2100 m.y. ago, the FB Formation black shales generated enormous amounts of petroleum liquids (Mossman, 2001). These liquids subsequently encountered uranium-enriched fluids in deep oxidizing groundwater circulating through fractures in the underlying FA Formation (Gauthier-Lafaye and Weber, 1989). This resulted in the reduction of dissolved +6U to +4U and the precipitation of pitchblende and uraninite (Gauthier-Lafaye et al., 1989). Ca. 2.3 Ga, Earth's atmosphere underwent a major increase in O2 content (Bekker et al., 2004); this so-called Great Oxidation Event (Holland, 2002) helps to account for the uranium-bearing oxidizing solutions in the Francevillian rocks (Gauthier-Lafaye and Weber, 2003; Mossman et al., 2005). These phenomena all contributed to the formation of the natural nuclear fission reactors in uranium-rich pockets in the uranium ore bodies at the village of Oklo (Gauthier-Lafaye and Weber, 2003); common low-grade ore at Oklo contains 0.2%–1% UO2.
The Oklo reactors evolved from ore with 25%–60% uranium, and favorable factors included not only the presence of such high-grade ore pockets and the lack of neutron poisons (such as V, B, Li, and some rare earth elements [REEs]) but at the moment in geologic time (nearly two billion years ago) when they reached criticality, the abundance of 235U in natural uranium was 3.5%, about five times greater than it is today. Water served as moderator, regulating the fission reactions.
There are 16 reactors known within the Oklo and Okélobondo uranium mines and one at Bangombé, ∼35 km to the southeast. All of the reactors occur in hydrothermally altered siliciclastic sedimentary rocks containing uraninite and authigenic clay minerals. The alteration resulted from heat due to nuclear fission reactions. The earliest discovered reactors (1–6) at Oklo contain only traces of carbonaceous substances, whereas the others are enriched in carbonaceous substances to as much as 66 wt% organic carbon in heterogeneous distribution within a clayrich matrix. Reactors 7–9 contain abundant carbonaceous substances, and their total uranium content was ∼80 tonnes; they produced 1300 MW yr−1 of energy and 480 kg of fissioned 235U (Naudet, 1991). According to Naudet (1991), fission ceased after 105–106 yr when too little 235U remained; however, decreased permeability and porosity due to the formation of clay minerals also played important roles insofar as fission reactions in the natural reactors proceeded under hydrothermal conditions, whereas in manmade reactors, fission occurs under dry conditions. Strict comparison between the Oklo natural reactors and modern nuclear power reactors is difficult because, among other things, the natural reactors differ greatly in size, uranium concentration, and extent of depletion of 235U. It is estimated, however, that maximum burn-up of 235U in natural reactors during their operation was slightly less than that of nuclear fuel in modern reactors after three years (Janeczek, 2000). Results of a recent study of xenon isotopes (Meshik, 2005) indicate that reactor operation might have been switched on and off at intervals, similar to those typical of geyser activity. According to Gauthier-Lafaye et al. (1996), the temperature during criticality is estimated to have ranged between 400 °C and 500 °C in the reactors (cf. Holliger et al., 1978). Fluid inclusion studies hold great promise of revealing further details of physico-chemical conditions during criticality. Kerogen and bitumen as well as liquid oil in fluid inclusions (entrapped under various conditions) are preserved in reactor zones and the enclosing rock (Mossman et al., 1993; Volk et al., 2006). These carbonaceous substances had important effects on radionuclide containment (Fig. 2).
Carbonaceous substances such as bitumen are known to inhibit the migration movements of uraninite, uranium, and similar radiogenic decay products in uranium deposits (Nagy and Mossman, 1992). The same process occurred at Oklo, thereby providing important lessons for nuclear waste disposal (Nagy et al., 1993; Mossman et al., 2000). The Oklo analogue to waste containment provides numerous opportunities to investigate long-term behavior of deep geological disposal of nuclear wastes (Ewing et al., 2004). In this analogue, carbonaceous substances provide perhaps the most significant natural barriers—in addition to rock type, groundwater conditions, etc. In deep geologic disposal projects, multiple barriers to the escape of radionuclides often include engineered barriers in addition to natural barriers, and the emphasis differs according to country (Bodansky, 1997; Ewing et al., 2004). A wealth of diverse options exists in numerous effective mineralogical barriers for immobilization and disposal of radionuclides (Ewing, 2001; Burns et al., 2000). However, the focus in this review is to highlight the potential posed by carbonaceous substances, as exemplified by the Oklo analogue, bearing on safety issues in the permanent containment of anthropogenic nuclear waste.
Extent of Element Migration
Three main stages of actinide and radionuclide migration are recorded at Oklo (Gauthier-Lafaye et al., 1996). The first and major one corresponds to criticality under pressure-temperature (P-T) conditions respectively of 300 bars and 400–500 °C, conditions similar to those in the primary circuit of a pressurized water reactor. Compared with the organic matter–rich reactors (Gauthier-Lafaye et al., 1996; Hidaka and Holliger, 1998; Hidaka et al., 1993), the loss of various fission products, especially Sr, Cs, Rb, and Ba, is most noticeable in reactors where the uraninite was not enclosed in bitumen. Mass transfers were achieved by reactor-driven hydrothermal fluids. Above reactor zone 10, for example, constituent clays retaining only a few ppm of uranium show a high isotopic enrichment in 235U, at 235U/238U = 0.007682 ± 0.00034 compared to the normal ratio of 0.007254 (Bros et al., 1993). Naudet (1991) also gave instances where uranium has migrated as much as 20 m “downstream” away from the reactors. Migration of elements occurred due to the operation of hydrothermal cells surrounding the reactors, resulting in the extensive alteration of the enclosing rocks.
The second event was the regional intrusion ca. 800–1000 Ma of diabase dikes, several of which occur in close proximity to the reactors (Bonhomme et al., 1982). This resulted in the intense migration of radiogenic lead (Pb) and of platinum group metals, Mo and Te, which formed metallic and oxide inclusions in uranium oxides.
A later stage of element migration is recognized (Gauthier-Lafaye et al., 1996) that corresponds to a geologically recent interaction of the weathering profile with the Oklo-Okélobondo and Bangombé reactors. This surficial stage of element dispersal is currently under intensive investigation because it provides a unique opportunity to study reactors that have been subjected to supergene alteration.
Using the analogy with irradiated anthropogenic uranium dioxides, Gauthier-Lafaye et al. (1996) considered the behavior or four types of elements, namely: (1) fission product oxides soluble in the UO2 matrix (REEs [La, Ce, Pr, Nd, Sm, and Gd], Y, Nd, and Zr); (2) fission products insoluble in the UO2 matrix (Rb, Sr, and Ba); (3) metallic and oxide inclusions of Mo, Tc, Ru, Pd, and Te; and (4) volatile fission products (Cd and Cs, and noble gases).
Drawing on the results of a research program initiated in 1990 as a European Community project “Oklo, Natural Analogue,” and their own new isotopic analyses on Oklo uranium oxides, clays, and phosphates, Gauthier-Lafaye et al. (1996) were able to document that the UO2 matrix was indeed efficient at retaining rare earth elements (REEs), Y, and Zr, but not Rb, Sr, Ba, and the volatile fission products noted in the fourth category. The fissio-genic elements noted in the third category evidently could have formed as metallic or oxide inclusions in UO2. Most relevant to the present review, however, are Gauthier-Lafaye et al.'s (1996) conclusions concerning the main geological, mineralogical, and geochemical parameters bearing on the retention or migration of actinides and radionuclides in the system. They suggested that three factors controlled the preservation of the reactors, namely: the geological stability of the Franceville Basin, the low permeability of the rocks enclosing the reactors, and the preservation of the uraninites with their contained fission product oxides. This last and most important factor of all was realized due to the reducing conditions induced by the presence of carbonaceous substances.
According to Bros et al. (1993), the isotopic composition of uranium dissolved in the original hydrothermal fluids is indicated by the exchangeable uranium absorbed to clay mineral surfaces, which show substantial depletion in 235U/238U = 0.006805 ± 0.00068. The location of the enriched uranium lies within the crystalline chlorite lattice (Bros et al., 1993; Eberley et al., 1994). This resulted from the mobilization in reducing fluids and subsequent entrapment of 239Pu derived from 238U by neutron capture during criticality. Clays and phosphates are believed to have played an important role in retaining fissiogenic REEs and Pu (Bros et al., 1996; Gauthier-Lafaye et al., 1996); however, the extent of Pu migration into rocks surrounding the reactors is not yet known with certainty (Bros et al., 1993).
Investigation of trace-element composition of hydrothermal minerals in the vicinity of reactor zones continues in order to model mass transfer reactions. The identification of mechanisms involved in processes of alteration, mineral paragenesis, and trace-element migration poses interesting challenges (e.g., Ewing, 1991). A host of radionuclides and fission products were among the products of reactor operation (Hidaka et al., 1992). As a result of leaching from the uraninite host, many of the radionuclides, including Nb, Rb, Zn, Te, and Sn, were retained in the clay mantles surrounding the reactors (Janeczek, 2000; Hidaka and Holliger, 1998). A fraction of fissiogenic Zr is believed to have been transported by chloride-rich fluids (Mathieu et al., 2001).
Hidaka et al. (1994) reported that, with the exception of alkalis, alkaline earth elements, and rare gases, uranium and most fission products were retained in the reactor zones during criticality. However, this is not the case. According to Hidaka and Gauthier-Lafaye (2000), the results of isotopic analyses of REEs show that there has been a redistribution of fissiogenic and nonfissiogenic REEs depending on the depth of the reactor zone beneath the surface; specifically, they suggest that the shallow reactor zones 1–9 (see Fig. 2) may have undergone supergene alteration, in contrast to the more deeply buried reactor zones 10–16. Thus, for example, reactor 9 has higher amounts of nonfissiogenic REEs than reactors 10 and 13 (Hidaka and Gauthier-Lafaye, 2000). All fissiogenic REEs show high retention in the reactors, even in the Bangombé reactor located at the surface. They proposed that the nonfissiogenic REE may have been diminished by hydrothermal processes during criticality, although the exact local geochemical conditions, such as the temperature and chemical composition of each reactor, remain firmly linked to nuclear parameters (Hidaka and Gauthier-Lafaye, 2000).
The build up of trace elements in the reactor zones is doubtless in part due to compaction following pervasive silica dissolution during hydrothermal activity associated with reactor operation. Compared to the immediately surrounding rocks, a substantial enrichment of light rare earth elements (LREEs) over heavy rare earth elements (HREEs) exists in the reactor zones (Naudet, 1978; Menet-Dressayre et al., 1992), a situation rigorously confirmed by Hidaka and Gauthier-Lafaye (2000) in the case of reactor 9. Hidaka and Holliger (1998) suggested that perhaps organic matter somehow extracted REEs from fluids during formation of the uranium ores, thus accounting for a surprisingly low total REE content despite high uranium content. They further note that the low HREE content of Oklo uranium ores is chief among the geochemical characteristics that made criticality possible.
Mossman et al. (1998) recorded low total REE content and a slight enrichment of LREE (and a slight depletion of HREE) in samples of the FB Formation black shale closest to the reactor zone as opposed to those distant. These conditions mimic the overall low proportion of LREE to HREE reported in Oklo and Bangombé uraninites (Hidaka and Holliger, 1998). Although suggestive that fission products (or their precursors) may have escaped entrapment in reactor zone minerals (e.g., apatite, clays, carbonates, coffinite) and in carbonaceous substances, as a result of various hydrothermal events, their distant migration remains to be unequivocally demonstrated (Jackson and Mossman, 1999).
Carbonaceous Substances and Oil-Bearing Fluid Inclusions
Toward a standardization of the definitions of carbonaceous substances, kerogen is defined by Mossman and Thompson-Rizer (1993) as an autochthonous solid and random “polymer”-like organic material disseminated in sedimentary rocks that has not migrated since deposition. Bitumen, a soluble fraction derived from kerogen during oil generation, is an allochthonous organic random macromolecular substance that is mobile as a viscous fluid or that was once mobile but which has subsequently solidified to an immobile phase (Mossman and Nagy, 1996; Mossman and Thompson-Rizer, 1993; Mossman et al., 1993). As shown in Figure 3, graphite is the end product of thermal maturation of kerogen and bitumen.
Several types of organic matter, e.g., carbonaceous substances, are associated with the Oklo natural fission reactors, the principal types of which are kerogen and solid bitumen. Kerogen occurs in black shale of the FB Formation overlying the reactor zones, in pelites of the FA Formation, and in the reactors within the FA Formation. Several types of solid bitumen have been identified (Mossman et al., 2001). Various less common carbonaceous substances, as documented by Mossman et al. (1993) and by Nagy et al. (1991) are similarly associated within the FA and FB Formations (see Fig. 3).
Mossman et al. (1993) reported the presence of fluorescent organic-brown bitumenite in the groundmass of a rock sample (GL-3165) at the outer edge of reactor 9, and fluorescent bitumen in reactor 7, and they hypothesized that these carbonaceous substances were due to oil migration during the Paleozoic era. Results of research in progress on fluid inclusions (Dutkiewicz et al., 2007) reveal the widespread occurrence of oil-bearing fluid inclusions in quartz in some Oklo rocks. The inclusions, containing at least three fluid phases, are reminiscent of those reported from the 2.45 Ga Matinenda Formation in the Huronian Supergroup, Canada (Dutkiewicz et al., 2003, 2006a), and the Archean Witwatersrand Supergroup, South Africa (Dutkiewicz et al., 1998), both of which have experienced a thermal event and contain radioactive minerals.
Oil-bearing fluid inclusions are particularly abundant in one sample from a well-cemented quartz-pebble arenite characterized by syntaxial quartz overgrowths and located in the C1 sandstone of the FA Formation ∼50 m from the nearest reactor. These fluid inclusions display bright blue, white, yellow, and orange fluorescence under ultraviolet excitation and usually contain less than 5 vol% liquid hydrocarbon. Two oil-bearing fluid inclusion assemblages, as described next, are evident (see also (Figs. 4) and 5). Solid bitumen inclusions are absent here but have been found in one other sample that lacks oil-bearing fluid inclusions (Fig. 5D).
The first assemblage contains by far the most abundant type of oil-bearing fluid inclusions at Oklo. These most commonly occur as trails within intragranular microfractures restricted to individual detrital quartz grains and pebbles and transgranular microfractures that cut detrital quartz boundaries and syntaxial overgrowths (Fig. 4), and as clusters. The relationship between individual microfractures is extremely complex, with pervasive crosscutting fractures. The abundance of oil-bearing fluid inclusions in neighboring grains makes it extremely difficult to unravel their relative timing of formation and their intragranular versus transgranular nature (Fig. 4). These oil-bearing fluid inclusions also occur within syntaxial quartz overgrowths and at boundaries between quartz overgrowths and their detrital cores (Fig. 4). The fluid inclusions are usually around 10–20 µm in length and seldom exceed 30 µm. The smallest individuals occur within quartz overgrowths where they rarely exceed 5 µm in length. At room temperature, they consist of liquid water and a gas bubble with a thin rim of fluorescing oil around the bubble, one or multiple globules of fluorescing oil within the aqueous phase, or both (Fig. 5C). Multiple globules of oil within single fluid inclusions rarely display variable fluorescence colors. Their shapes range from relatively large, irregularly shaped inclusions to somewhat smaller spherical or oval-shaped individuals.
The second assemblage occurs as trails and clusters within microfractures and consists of fluid inclusions composed of a dominantly carbonic fluid, which commonly displays a mobile bubble of CO2 gas at room temperature within a CO2 liquid phase, surrounded by an aqueous phase. Oil-bearing inclusions contain a rim of white-, blue-, yellow-, or dull orange–fluorescing oil around the carbonic phase (Figs. 5C and 5D). These inclusions are generally larger than the H2O-dominated individuals and are typically 15–25 µm long.
At least two episodes of hydrocarbon migration and entrapment are evident from the textural relationships of the oil-bearing fluid inclusions. These appear to be consistent with the migration episodes described in Mossman (2001) and Mossman et al. (2001) based on evidence from carbonaceous substances. The earliest of these occurred during burial of the host sequence, when thermal maturation of the FB black shales caused expulsion and migration of hydrocarbons into the FA Formation sandstones and conglomerates. Liquid oil was trapped in intragranular fractures and within syntaxial quartz overgrowths that formed during diagenesis while the FA formation still retained some open pore space. The second major hydrocarbon migration episode is widely recognized as having occurred during mineralization, which was accompanied by criticality and tectonic reactivation ca. 1968 ± 50 Ma and involved hydrothermal solutions circulating around the nuclear reactors (Gauthier-Lafaye and Weber, 1989; Mossman et al., 2001). This episode may be recorded in some of the abundant oil-bearing fluid inclusions located in the transgranular microfractures that formed during or after destruction of visible pore space. These microfractures may have formed by hydrofracturing associated with either the flow of hydrothermal fluids and generation of hydrocarbons from residual kerogen and solid bitumen or as a result of uplift at this time. The presence of solid bitumens in the sample indicates that some of the hydrocarbon was affected by free radical polymerization. The survival of “pristine” oil inside fluid inclusions, apparently unaffected by ionizing radiation, may have been facilitated by its migration as an emulsion inside an aqueous and carbonic fluid and by its relatively rapid entrapment inside inclusion cavities.
A third episode of hydrocarbon migration, linked to the intrusion of dolerite dikes ca. 977–981 ± 27 Ma, has been described as relatively minor (Nagy et al., 1991). Oil-bearing fluid inclusions that may be related to this event are the relatively rare CO2-H2O–dominated types. In this scenario, the carbonic fluid may have originated through magmatic activity, and the hydrocarbons may have originated through local flash maturation of pre-existing bitumens. However, the sample is located 100 m from the nearest dike, and any thermal effects from its intrusion are likely to have been minimal. In addition, Dubessy et al. (1988) and Mathieu et al. (2000) ascribed traces of CO2 inside fluid inclusions to the functioning of the reactor. If this was the case, then the carbonic–hydrocarbon fluid mixture may have been the result of hydrous thermal alteration of organic matter and of petroleum in the pore space in the FA and FB Formations. This would also be consistent with the high total homogenization temperatures (250–300 °C) expected for such fluid inclusions (Dutkiewicz et al., 2007). Future work directed at synchrotron Fourier transform infrared microspectroscopy of these oil-bearing fluid inclusions may help to resolve some of these outstanding questions.
Both bitumen and the fluorescent oil inclusions contain pristine biomarkers such as hopanes, 2α-methylhopanes, terpanes, and steranes, indicating input from cyanobacteria and eukaryotes (Volk et al., 2006). Results of studies on the molecular composition of Paleoproterozoic oil-bearing fluid inclusions as old as 2.45 Ga highlight the fact that once trapped, inclusion oils can be shielded against contamination not only by microbes for billions of years, but also against debilitating influences that affect porosity-bound hydrocarbons such as bitumen, kerogen, and live oil (Dutkiewicz et al., 1998, 2003; George et al., 2006; Dutkiewicz et al., 2006b, 2007) (Figs. 4A, 4B, 4C, 4D, etc.).
The presence of biomarkers in the oil of fluid inclusions hosted by the Oklo fission reactors is a remarkable phenomenon, notwithstanding the fact that in high-pressure closed systems with inert hosts, the kinetics of hydrocarbon degradation are such as to permit oil to survive temperatures as high as 450 °C (George et al., 2006; Hoffmann et al., 1988; Price, 1993; Dutkiewicz et al., 2003). An additional factor that may facilitate the preservation of complex hydrocarbons inside fluid inclusions is the absence of organometallic complexes and clay minerals that provide catalytic sites for the secondary cracking of petroleum (Mango, 1990). The survival of “live” oil, i.e., active, containing dissolved gas oil inside fluid inclusions in rocks as old as 3 Ga (Dutkiewicz et al., 1998) suggests that time need no longer necessarily be considered a parameter that controls petroleum degradation (Buick et al., 1998). Judging from the Oklo analogue, where (solidified) hydrocarbon(s) have demonstrably served as a barrier to migrating radionuclides, this fact must be viewed as a potentially important aspect in engineered waste containment proposals.
Thermally matured to as high as metaanthracite rank, the Oklo solid bitumens consist of both aromatic (benzenes) and polyaromatic hydrocarbons (including napthalenes and anthracenes) together with a mixture of cryptocrystalline graphite (Nagy et al., 1991). In the characterization of Oklo bitumens, Nagy et al. (1993) identified a lower maturity of carbonaceous substances in and near the reactors compared with carbonaceous substances distant from the reactors. The lower maturity and less-well-developed cryptocrystalline graphite were attributed to radiation effects and hydrothermal alteration. They postulated also (Nagy et al., 1993, p. 658) that ionizing radiation led to the addition of free radicals to the carbonaceous substances (an organic free radical is any organic compound, not necessarily an aromatic moiety, possessing one or more unpaired electrons; Aizenstat et al., 1986).
It is believed that at Oklo, organic free radicals played important roles in the chemical and structural evolution of carbonaceous substances (Nagy et al., 1991; Gize, 1993), including the solidification of bitumen via free radical polymerization. Rigali and Nagy (1997) quantified organic free radical concentration in Oklo carbonaceous substances. Specifically, they determined by electron spin resonance (ESR) that organic free radicals, accessible to O2, are located at or very near the surfaces of carbonaceous substances, including internal surfaces of pore spaces.
According to Rigali and Nagy (1997), average pore sizes of Oklo solid graphitic bitumens range from 18.5 å to 38.1 å. This feature, together with the nonwater wetting of bitumen, means that the transport of radionuclides in aqueous solutions was prohibited (judging by the case for aqueous flow in water-wettable voids < 100 å between clay minerals; Dickey, 1986). One important mechanism responsible for the fact that uraninite enclosed in solid Oklo bitumen retains uranium and various fission products is thus envisaged as a free radical polymerization promoted by the irradiation of viscous bitumen by closely associated uranium minerals. Concerning this possibility, it will be important to compare diffusion coefficients through small-diameter spaces of ions dissolved in water versus their coefficients in oil.
The mesoscopic appearance of Oklo bitumens, their ESR characteristics, and the results of laser Raman spectroscopy, are similar to these same features in glassy carbon, which reportedly has a ribbon-like arrangement of graphitic moieties; this implies lots of edge carbons and hence a very strong D-band (disorder/defect) peak (Tice and Lowe, 2004). Oklo bitumens typically show weak D-bands, a feature consistent with elemental analyses, X-ray diffraction (XRD), and reflectance data, which show that uraninite-bearing carbonaceous substances are less well ordered than nonuraniferous carbonaceous substances at Oklo. Oklo solid bitumens commonly closely resemble bright shiny Karelian shungite (one of two end-member lithologies of shungite), which, at least by some workers, is considered to be a nongraphitizing glass-like carbon. All of the several samples investigated by XRD (see Fig. 6) qualify as nanocrystalline aggregates, well on their way to becoming glass; the Oklo and Congo specimens may even qualify as nongraphitizing, inhibited possibly by a sort of protofullerene structure. Other glassy carbonaceous substances discovered in Proterozoic pillow lavas at Mitrov in the Bohemian Massif reportedly contain fullerenes (Jehlicka et al., 2000) (Fig. 6). (J. Jehlicka [2005, personal commun.] suspects that the fullerenes may in this instance be hosted in the bitumen micropores.)
X-ray diffraction of two samples of powdered solid Oklo bitumen, and samples of shungite from Shunga (Russian Karelia) and Shinkolobwe (Congo), shows broad amorphous scattering. Oklo sample LN12, taken 8 km distant from the nearest reactor, shows a prominent peak at low angles, the position (18 å) of which is attributed to the micropores reported by Rigali and Nagy (1997) (see Fig. 6). In contrast, D75, a sample from reactor 16, shows amorphous scattering like shungite but little or no evidence of micropores. Thus, indications are that micropores did not facilitate transport of radionuclides through the solid bitumen matrix in the vicinity of the Oklo reactors.
A study of optical textures in various Oklo carbonaceous substances using transmitted light microscope (TEM) techniques revealed “local molecular orientations” (Cortial et al., 1990, p. 76). In some samples of Oklo carbonaceous substances, lattice fringe images show typical onion-skin structures characteristic of fullerenes. It was this feature, together with the general resemblance of Oklo bitumen to “shungite” (the highly carbonaceous natural material in which Buseck et al.  first reported the occurrence of fullerenes in geologic materials) that prompted a search for fullerenes in Oklo carbonaceous substances.
Fullerenes may find direct application in nuclear waste disposal. Experimentally, they have been proven capable of encapsulating various elements (Gadd et al., 2001), including isotopes of uranium. According to Anderson et al. (2000), encapsulated metal atoms exert a stabilizing influence on the fullerene/metallofullerene electronic structure. From an engineering perspective, use of fullerenes to encapsulate radiogenic nuclides might even improve on the reported high efficacy of solid graphitized Oklo bitumen at inhibiting transport of radionuclides through and out of the organic matrix. Mossman et al. (2003) investigated Oklo carbonaceous substances for the possible presence of fullerenes using laser desorption ionization and high-resolution mass spectroscopy (electron-impact mass spectroscopy) without success. Possibly fullerenes are limited in geological occurrences to higher-temperature events than those represented by the Oklo fission reactors. Conversely, if fullerenes were produced at Oklo during criticality, they are either present below the detection limit (∼100 fmol) or they have somehow been destabilized (Mossman et al., 2003).
The Oklo Analogue: Discussion and Conclusions
Côme and Chapman (1986, p. 19) define a natural analogue as “…an occurrence of materials or processes which resembles those expected in a proposed geological waste repository.” According to the International Atomic Energy Agency (IAEA) (1989), “Natural analogues are defined more by the methodology used to study and assess them than by any intrinsic physico-chemical properties they may possess.” Since it is increasingly clear that there are no alternative tests of long-term predictability, appreciation of natural analogues has increased to the point where natural analogue studies “…use information from the long-term behavior of materials and processes found in, or caused by, a repository to develop or test models appropriate to performance assessment work” (Miller et al., 1994, p. 9).
There are numerous issues constraining long-term predictions on the deep geologic disposal of nuclear wastes (e.g., Chapman and McKinley, 1987; Chapman and Smellie, 1986; Bodansky, 1997). Not least of these is the possibility that current concerns may be substantially alleviated by new developments in science and technology. According to Hannum et al. (2005), this will occur when a high-temperature method of recycling reactor waste into fuel becomes available and when fast neutron reactors capable of burning that fuel are built; these developments are expected to not only reduce the problem of long-term isolation of nuclear waste, but also to minimize the risks of weapon proliferation, and to yield vastly more energy from recycled nuclear waste.
Concerning the status quo, however, considerable efforts have been made to validate predictive radionuclide transport models, though consensus remains elusive. This is not surprising given the mix of assessors, regulators, and independent groups, all of whom seek, at one time or another, the geologist's advice (Ewing, 2000; Ewing et al., 1999). However, performance assessment models designed to test isolation of nuclear waste can no more be subject to proof than a scientific hypothesis.
Available data indicate that at Oklo, rock can strongly restrain the migration of radionuclides and limit the extent to which movement occurs. A strength of the Oklo analogue lies in the fact that it represents more extreme conditions than those likely to be experienced in an engineered deep geological repository (DGR). Thus, initial temperatures in the analogue are higher and the interval of isolation is far longer than that envisaged for deep geological repositories (Roxburgh, 1987). These aspects of the problem effectively argue for the principle of maximizing the geological barrier (Gibb, 1999), since it is the longest enduring barrier. Performance assessment will need to consider the long-term behavior of the combination of geological (geochemical and hydrological) barriers together with the engineered barriers in a deep geological repository (Ewing et al., 1999).
The extent of corrosion and/or alteration products within a repository is of special concern. A particular concern is the promotion of designs that minimize the amount of water that may access a repository. According to Miller et al. (1994), even bitumen can be destroyed under damp conditions with the possible involvement of microbes. However, efficient as this process might be under some conditions, it is worth noting that abundant solid bitumen occurs in the Bangombé reactor zone only ∼12 m below the surface, well within the profile of tropical weathering.
It has been said (Brookins, 1990) that none of the 17 reactors at Oklo would receive serious consideration for radioactive waste disposal today in view of their wide range in porosity and permeability. Nevertheless, they substantially retained many fission products, radiogenic nuclides, etc., for many millions of years during reactor operation and for millions of years thereafter. Even the Bangombé reactor (billed as the “last natural nuclear reactor on Earth”; Blanc et al., 1997, p. 337), buried a mere 12 m below the surface, merits recommendation as a natural analogue, if only for the fact that substantial amounts of carbonaceous substances (mostly solid bitumen) are preserved in the reactor zone. At Bangombé, carbonaceous substances and clays have served to create and continue to maintain reducing conditions highly favorable to the retention of radionuclides. This has occurred despite many variables—among them porosity, Eh, pH, water chemistry, sorption, colloids, and organic complexes—all of which can play important roles in the transport of radionuclides.
The Oklo reactors provide ideal subjects for study by workers concerned with the safe disposal of nuclear waste. For example, these reactors are analogues to modern pressurized-water reactors in several significant respects: (1) at the time of criticality, the isotopic abundance of 235U was 3.5%, which is very close to the enrichment used in modern light water reactors; (2) their cores consist of UO2, similar to spent fuel; (3) the bitumen of the Oklo reactors resembles the alpha-waste matrices proposed to confine wastes; (4) hydrothermal clays surround them, similar to the clay backfill proposed for repositories; and (5) fractured rocks surround the reactors, as at many proposed repositories.
Research continues to reveal fascinating details of reactor operation. A recent example relevant to furthering insights into the Oklo analogue is the discovery of xenon entrapped in aluminum phosphate minerals. Meshik (2005) documented how xenon isotopes (nine stable isotopes) are produced in varying proportions by different nuclear processes. Meshik and co-workers (2005) determined that the anomalous composition of Oklo xenon best fits a situation where the reactors behaved as open systems—in effect, they regulated themselves, pulse-like in operation through the action of groundwater, which served to moderate the nuclear reactions. Meshik (2005) suggested that the regulatory action was geyser-like and that it protected the reactors from destruction through overheating. Like some of the Oklo bitumens (Mossman, 2001), the aluminum phosphate was generated following reactor start-up. The process illustrates how some radioactive elements like 135Xe, 85Kr, and other inert gases released into the atmosphere through nuclear power generation could be captured and retained as part of a containment strategy, albeit, in this instance, not in a deep geological repository.
The technique employed by Meshik (2005) on a small sample of reactor material, presumably carbonaceous substance–free, could also be applied to establish more rigorously the paragenesis of Oklo carbonaceous substances. It remains to be determined whether xenon is retained in uraninite that is enclosed within solid bitumen. If this is the case, it would mean that the bitumen was present during nuclear reactions or immediately after they ceased, preventing fission product (xenon) loss. In another dimension, the presence of micropores in carbonaceous substances is relevant to the containment of gaseous radionuclides (226Ra, 129Xe) and fission products in a deep geological repository, and therefore this is also a subject that deserves further study.
Given then, that long-term (permanent?) deep geologic disposal of high-level nuclear waste appears to be the universal choice (Bodansky, 1997), the best natural analogue to a nuclear waste repository is one that supplies the basis of long-term prediction. As demonstrated by the results of research focused on the geochemistry of carbonaceous substances (particularly solid graphitized bitumen) associated with uranium ore at Oklo, Okélobondo, and Bangombé, the organic-rich natural reactors at these sites are unique and valuable time-tested analogues of radioactive waste containment.
The senior author gratefully acknowledges support received through Natural Sciences and Engineering Research Council operating grant no. 8295. The fluid inclusion work was supported by the Australian Research Council Grant DP0556493 to Dutkiewicz. We are pleased to acknowledge the helpful advice of Mark Rigali and Norbert Rempe. Tim Lowenstein and Sherilyn Williams-Stroud kindly suggested many improvements to the manuscript.
Figures & Tables
Deep Geologic Repositories
Deep Geologic Repositories reviews the success stories of underground waste isolation. It focuses on repositories that did, do, and will permanently and safely isolate dangerous materials from the near-surface biosphere. Complementary topics address the isolation capability of average crustal rock, investigations at one representative underground research laboratory, and the geologic preservation of fission products from Precambrian nuclear reactors. An international cast of contributors presents proven practical solutions to a formerly confounding issue in environmental and engineering geology: What do we do with wastes that retain their dangerous characteristics in human terms forever? The principal answer: Recycling into the lithosphere by “reverse” mining.
- carbonaceous composition
- Central Africa
- clastic rocks
- environmental analysis
- fluid dynamics
- fluid inclusions
- human activity
- metamorphic rocks
- metasedimentary rocks
- radioactive waste
- sedimentary rocks
- site exploration
- underground disposal
- upper Precambrian
- waste disposal