Evidence of Uranium and Associated Trace Element Mobilization and

tions of some of the rare earth elements, i.e., Nd and Yb. In addition, we have ... the various waste management strategies proposed (1, 2). In this c...
0 downloads 0 Views 85KB Size
Download PDF
Environ. Sci. Technol. 2004, 38, 3310-3315

Evidence of Uranium and Associated Trace Element Mobilization and Retention Processes at Oklo (Gabon), a Naturally Radioactive Site IGNASI CASAS,† JOAN DE PABLO,† I S A B E L P EÄ R E Z , † J A V I E R G I M EÄ N E Z , † L A R A D U R O , ‡ A N D J O R D I B R U N O * ,†,§ Department of Chemical Engineering, UPC, Avenue Diagonal 647, H-4, 08028 Barcelona, Enviros Spain S.L., Passeig de Rubı´ 29-31, 08197 Valldoreix, Barcelona, Spain, and Enresa-Enviros Environmental Science and Waste Management, UPC, Jordi Girona 1-3, B2, 08034 Barcelona, Spain

The processes that affect the mobility of uranium and other radionuclides in the environment have been largely studied at both the laboratory and the field scales. The natural reactors found at the Oklo uranium mine in Gabon constitute a unique investigation setting as spontaneous fission reactions occurred two billion years ago. Oklo uraninites contain a large amount of other radionuclides as a result of the fission process. We have investigated the dissolution behavior of four uraninite samples from Oklo as a function of temperature (25 and 60 °C) and bicarbonate concentration (2.7-30 mmol/L). We have also investigated the dissolution behavior of minor components of the uraninites (i.e., Nd, Cs, Mo, Yb, and Sb) in relation to the dissolution of uranium. The results of the reported work are in good agreement with the kinetic rate laws derived from other uranium(IV) dioxide studies. Some of the minor components are found to be congruently released from the uraninite phase, while it is postulated that dissolution from segregated phases might affect the final concentrations of some of the rare earth elements, i.e., Nd and Yb. In addition, we have performed dissolution studies at 60 °C with two uraninites representative of different geochemical environments at Oklo, to study the uranium dissolution rates as a function of the temperature. This has allowed derivation of apparent activation energies for the bicarbonate-promoted oxidative dissolution of the Oklo uraninites. The dissolution behavior of the minor components of the uraninites at 60 °C was found to closely follow the behavior observed at 25 °C. This indicates that similar codissolution mechanisms operate in the temperature range studied. The implications for the mobility of uranium and other radionuclides in natural and anthropogenic environments are discussed.

* Corresponding author (Enviros Spain S.L.) phone: +34 93.583. 05.00; fax: +34 93.589.00.91; e-mail: [email protected]. † Department of Chemical Engineering, UPC. ‡ Enviros Spain S.L. § Enresa-Enviros Environmental Science and Waste Management, UPC. 3310

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 12, 2004

Introduction Activities associated with the anthropogenic use of uranium have generated substantial amounts of waste materials contaminated with uranium and other radionuclides. A large body of investigations have been performed in the past decade to understand the key processes controlling the mobility of uranium and other radionuclides in natural environments to assess the environmental implications of the various waste management strategies proposed (1, 2). In this context, uranium ore deposits have been largely studied as natural analogues to understand and quantify some of the key retention/mobilization processes for uranium and other radionuclides in naturally occurring environments (3, 4). The natural reactors found at the Oklo uranium mine in Gabon are of particular importance, as spontaneous fission reactions occurred two billion years ago (5). In this respect, they provide a unique opportunity to study the environmental behavior of the fission products that grew in the uranium dioxide matrix during the long-lasting fission episodes. The uraninites from the Oklo site constitute an actual fossil spent nuclear fuel, and the study of their behavior may complement information for the long-term dissolution behavior of radioactive wastes obtained from other uranium ore deposits and from synthetic unirradiated uranium dioxide dissolution studies (6-14). Moreover, the study of a complex matrix such as the Oklo uraninites, which contains substantial amounts of other radionuclides and fission products (i.e., Cs, Nd, Mo, Ba, Sb, Sr, etc.) (15-17), implies the use of an experimental methodology able to determine the potential congruent release of the minor elements. Bicarbonate being one of the main groundwater constituents in most natural systems, it is necessary to include this anion as one of the experimental variables of the studies. The main reason for that is the highly stable uranium(VI)carbonate aqueous complexes formed, which are prone to facilitate uranium release and mobility in natural systems. Recently, de Pablo et al. (6) have shown the effect of the bicarbonate-promoted oxidative dissolution of uranium dioxide and have proposed a mechanistic model to rationalize both experimental and field observations. The mechanism proposed by those authors is constituted by three consecutive steps: (i) initial surface oxidation of >U(IV) sites to >U(VI), (ii) formation of the >U(VI)-carbonate surface complexes, (iii) detachment of the >U(VI)-carbonate surface complexes (dissolution). These studies have also shown that the dependence of the dissolution rate on the bicarbonate concentration varies with temperature (ref 6 and references therein), with apparent activation energy values ranging between 20 and 80 kJ/mol. Hence, in the present work we have studied the kinetics of dissolution of four uraninite samples from different locations of the Oklo site as a function of temperature (25 and 60 °C) and bicarbonate concentration (2.7-30 mmol/ L). We have also investigated the dissolution behavior of some minor elements of the uraninites (Nd, Cs, Mo, Yb, and Sb) in relation to the dissolution behavior of uranium, the main component. The experimental results will be discussed in the frame of the microstructural investigations of the minor elements on the selected samples (15, 18, 19). We will as well try to show the implications of these findings for the mobility of uranium and other radionuclides in anthropogenically affected environments.

Experimental Section Sample Characterization. Uraninite samples were selected from borehole cores extracted from different depths and for 10.1021/es0353863 CCC: $27.50

 2004 American Chemical Society Published on Web 05/01/2004

TABLE 1. Concentrations (mg/kg) Determined after Complete Dissolution of the Four Solid Samples Studieda Nd Cs Mo Sb Yb U

RZ9

AP9

RZ10

RZ13

xi (RZ9)

xi (AP9)

xi (RZ10)

xi (RZ13)

2687 0.1 519 12.8 36.4 6.9 × 105

590 13.8 156 24.4 12.1 2.45 × 105

608 13.1 134 3.4 3.1 3.09 × 105

2855 2.0 309 4.5 0.9 6.9 × 105

6.42 × 10-3 2.59 × 10-7 1.87 × 10-3 3.61 × 10-5 7.25 × 10-5 1

3.97 × 10-3 1.01 × 10-4 1.58 × 10-3 1.94 × 10-4 6.79 × 10-5 1

3.24 × 10-3 7.56 × 10-5 1.08 × 10-3 2.14 × 10-5 1.39 × 10-5 1

6.83 × 10-3 5.11 × 10-6 1.11 × 10-3 1.27 × 10-5 1.73 × 10-6 1

a Columns 2-5. Molar ratios (x ) between corresponding minor element and uranium, needed to normalize the dissolution rates obtained in i the flow-through experiments to be performed (columns 6-9).

different reactor zones. The solid samples selected correspond to reactor zones 9 (sample Afft GL13-19, hereafter named RZ9), 13 (sample SD37-S2-10/1, hereafter named RZ13), and 10 (sample SF42, hereafter named RZ10), located at 120, 250, and 450 m depth, respectively. A fourth sample studied comes from the surroundings of reactor zone 9 (sample Afft GL1363, hereafter named AP9), which is known as argyle de pile (20). The solid samples were characterized by X-ray powder diffraction (XPD), X-ray photoelectron spectroscopy (XPS), and scanning electron microscopy (SEM). The elemental composition of the samples was determined by inductively coupled plasma mass spectrometry (ICP-MS), following dissolution of a known weight of sample by acid mixtures at high temperature. Samples were crushed and sieved, and the fraction between 100 and 300 µm was used in the dissolution studies. The specific surface area of these powdered samples was determined by means of the BET method (21). Experimental Methodology. With the final goal of determining the dissolution rates of the species dissolved from the solid samples selected, we used a series of continuously stirred tank flow-through reactors. In these experiments the attainment of a steady-state situation readily allows determination of dissolution rates by multiplying the flow rate used times the steady-state concentration of the dissolved species. Different experimental conditions can be imposed simply by changing the test solution. In this sense, our intention is to check the effect of different bicarbonate concentrations on the dissolution rates as well as the effect of the temperature. The reactors were built with polymethacrylate with the idea of being able to check for any malfunctioning inside the reactor through the transparent walls. Previous static tests were performed with these reactors using different uranium concentrations and for periods of time longer than the ones expected for the experiments which did not show significant uranyl sorption on the reactor walls. The reactors had an inner volume of 25 mL. A small quantity of powdered solid (between 0.3 and 0.5 g) was introduced inside the reactor. The solid was suspended on a holder of 30 µm pore size, to keep it away from the magnetic stirrer. A filter of 0.45 µm eliminated the possibility of solid particles in the outflow solution. See the detailed experimental procedure described by Bruno et al. (22). We performed the experiments at a selected flow rate value of 0.1 mL/min, which ensures the attainment of steady state. This value was experimentally determined as described in previous works (22). The flow rate was carefully remeasured at each sampling time. The pH was continuously monitored in the outflow solution by means of an on-line combined glass electrode calibrated with commercial buffer solutions. The test solution selected contained constant sulfate (1480 mg/L) and chloride (300 mg/L) concentrations together with a varying bicarbonate concentration (in the range 2.7-30 mmol/L). In all cases the test solutions were maintained in contact with air.

TABLE 2. Specific Surface Areas Determined by BET Method for Four Solid Samples Studied fraction 100-300 µm

specific surface area (m2/g)

fraction 100-300 µm

specific surface area (m2/g)

RZ9 AP9

0.821 ( 0.007 0.409 ( 0.007

RZ10 RZ13

2.398 ( 0.019 0.031 ( 0.004

The concentration of the dissolved species present in the outflow solution was measured using an inductively coupled plasma mass spectrometer (Elan 6000, Perkin-Elmer). The detection limit was practically the same for all the elements measured in the present work and was around 10 ppb.

Results and Discussion Sample Characterization. In Table 1 we present the measured concentrations (mg/kg) as well as the calculated molar ratios (xi) of the minor components with respect to uranium determined for the uraninite samples studied. These values are obtained from the analysis performed after complete dissolution of a known amount of solid, as explained above. Only the values of the minor elements determined in the subsequent experiments are included. In Table 2 we present the specific surface areas determined by the BET method for the four powdered samples studied, measured in a FLOWSORB II 2300 Micromeritics instrument. Besides, the main observations extracted from the solid characterization of the four samples can be summarized as follows. SEM-Energy-Dispersive Spectrometry (SEM-EDS). SEMEDS was performed with a JEOL 6450, EDX-LINK-LZ5 instrument. We tried to identify the presence of the minor elements, which was not possible due to the low amounts of them present in the different samples studied. XPS. By using this technique, we tried to get an idea of the relative amounts of U(IV) and U(VI) in the solid surface, in an average depth of about 10 monolayers. The apparatus used was a PHI-Perkin-Elmer ESCA system multianalyzer 5500. In summary, we can say that the different solids showed a similar degree of surface oxidation, ranging in all the samples from 20% to 30% U(VI) of the total uranium in the surface. XPD. In this case we used a Siemmens D-500 instrument. In general, the very first observation is that samples RZ9 and AP9 show broader peaks with a number of accessory minerals, while samples RZ10 and RZ13 show very thin peaks indicative of a high degree of crystallinity. For all the samples the bulk of the solid matches the reference of a uraninite with a percentage of U(VI) in the range of values determined by XPS for the surface composition. Also, all samples show the presence of galena, which is assumed to be of radiogenic origin. More in particular, the XPD of sample RZ13 shows only the presence of uraninite and galena, and on the basis of this, together with the results obtained by the elementary analysis, we calculate for this sample a percentage of uraninite of about 70%. The same percentage is calculated for sample VOL. 38, NO. 12, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3311

FIGURE 1. Normalized uranium and minor element dissolution rates (mol/(s‚m2)) determined at 25 °C for uraninite sample RZ13 as a function of the total bicarbonate concentration. The full line corresponds to calculated uranium dissolution rates (see the text).

FIGURE 2. Normalized uranium and minor element dissolution rates (mol/(s‚m2)) determined at 25 °C for uraninite sample RZ10 as a function of the total bicarbonate concentration. The full line corresponds to calculated uranium dissolution rates (see the text).

RZ10, while for samples RZ9 and AP9 we determined uraninite percentages of 30% and 25%, respectively. In addition to uraninite and galena, from the XPD we could unambiguously determine chlorite in sample RZ10, small amounts of quartz in sample RZ9, and quartz, illite, chlorite, and coffinite in sample AP9. We could not see any pure phase related to any of the minor elements (Nd, Mo, Cs, Sb, Yb), which either were not present or were present in small amounts, below the detection limit of the technique (only detects solid phases present in a percentage of about 10% of the total). Dissolution Experiments. Dissolution rates (mol/(s‚m2)) were determined from eq 1,

rdissolution )

[i]ssQ Sxi

(1)

where Q is the flow rate (L/s), [i]ss is the steady-state concentration of each element (mol/L), S is the total surface area of the solid inside the reactor (m2), and xi is the trace element fraction in the initial solid phase (see Table 1). The concentrations measured in the outflow solution for all the experiments performed showed initially an erratic behavior which evolved in less than 48 h to a constant value that in some cases was followed for as much as 800 h and in the shortest experiment for about 400 h. These constant values were the ones used as [i]ss (eq 1) to calculate the dissolution rate for all the elements measured for the four samples and for the five bicarbonate concentrations studied. We also ensured that we were readily measuring steadystate concentrations by repeating the experiments at 30 mmol/L bicarbonate concentration at higher flow rates (0.3 mL/min). The effect of this on the different samples will be discussed below. Experiments at 25 °C. The normalized uranium dissolution rates (mol/(s‚m2)) determined for samples RZ13, RZ10, RZ9, and AP9 are plotted in Figures 1-4, respectively. The numerical values are presented in Table 3 together with their associated experimental errors, which cannot be seen in the figures because the error bars are smaller than the size of the points. In all cases a similar trend of the uranium dissolution rate on the total bicarbonate concentration is observed (except for the values at 30 mmol/L bicarbonate concentration for samples AP9 and RZ9, which will be discussed separately below). We have applied the model previously developed (6) based on the mechanism of UO2 bicarbonate-promoted oxidative dissolution that was discussed in the Introduction. This mechanism is summarized, under our experimental condi3312

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 12, 2004

FIGURE 3. Normalized uranium and minor element dissolution rates (mol/(s‚m2)) determined at 25 °C for uraninite sample RZ9 as a function of the total bicarbonate concentration. The full line corresponds to calculated uranium dissolution rates (see the text).

FIGURE 4. Normalized uranium and minor element dissolution rates (mol/(s‚m2)) determined at 25 °C for uraninite sample AP9 as a function of the total bicarbonate concentration. The full line corresponds to calculated uranium dissolution rates (see the text). tions, in the following equation:

r)

(3.726 × 1010)k-1k2[HCO3-] k-1 + k2[HCO3-] + (3.726 × 1016)k-1

(2)

with k-1 and k2 being the rate constants of the different steps. The fitting of eq 2 to our data is presented as full lines in

TABLE 3. Logarithm of Normalized Dissolution Rates (mol/(s‚m2)) Experimentally Determined for Four Samples Studied at 25 °C and for Five Bicarbonate Concentrations Used in Test Solutionsa log r

[HCO3-] (mM)

U

Nd

2.7 8 15 20 30

-10.67 ( 0.03 -10.11 ( 0.02 -10.07 ( 0.04 -10.08 ( 0.04 -10.22 ( 0.04

-11.56 ( 0.09 -11.07 ( 0.03 -10.75 ( 0.03 -10.59 ( 0.03 -10.55 ( 0.09

2.7 8 15 20 30

-10.87 ( 0.02 -10.34 ( 0.02 -10.22 ( 0.02 -10.24 ( 0.04 -10.47 ( 0.03

-11.65 ( 0.08 -11.33 ( 0.08 -11.03 ( 0.05 -10.87 ( 0.06 -10.85 ( 0.09

2.7 8 15 20 30

-11.23 ( 0.02 -10.81 ( 0.03 -10.69 ( 0.02 -10.68 ( 0.03 -10.60 ( 0.02

-12.40 ( 0.14 -12.02 ( 0.09 -11.59 ( 0.10 -11.41 ( 0.06 -11.06 ( 0.03

2.7 8 15 20 30

-10.02 ( 0.01 -9.49 ( 0.02 -9.39 ( 0.04 -9.26 ( 0.03 -9.20 ( 0.02

-10.75 ( 0.23 -10.17 ( 0.03 -9.92 ( 0.05 -9.64 ( 0.02 -9.44 ( 0.10

a

Cs

Mo

Argile de Pile Reactor Zone 9 -10.22 ( 0.16 -9.97 ( 0.07 -9.88 ( 0.14 -9.97 ( 0.14 -10.18 ( 0.13 Reactor Zone 9

Reactor Zone 10 -10.81 ( 0.06 -10.92 ( 0.06 -10.78 ( 0.13 -10.65 ( 0.08 -10.50 ( 0.02 Reactor Zone 13

-10.33 ( 0.08 -9.82 ( 0.05 -9.90 ( 0.05 -9.97 ( 0.08 -10.25 ( 0.18

Sb

Yb

-10.39 ( 0.16 -10.13 ( 0.02 -10.08 ( 0.08 -10.08 ( 0.17 -10.20 ( 0.11

-10.37 ( 0.04 -10.16 ( 0.05 -10.08 ( 0.07

-10.17 ( 0.08 -9.79 ( 0.07 -9.70 ( 0.12 -9.72 ( 0.14 -9.79 ( 0.08

-11.01 ( 0.41 -10.46 ( 0.05 -10.28 ( 0.04 -10.22 ( 0.11

-10.82 ( 0.07 -10.53 ( 0.06 -10.53 ( 0.03 -10.44 ( 0.07 -10.46 ( 0.02 -9.27 ( 0.05 -9.11 ( 0.05 -8.52 ( 0.34 -8.77 ( 0.07 -8.97 ( 0.07

Empty spaces indicate that concentrations were below detection limits and no dissolution rate could be determined.

Figures 1-4, showing the good agreement obtained between measured and calculated values. The unexpected decrease of the dissolution rate at 30 mmol/L bicarbonate concentration observed for samples RZ9 and AP9 (see Figures 3 and 4) was attributed to the possibility of a secondary uranium phase formation, as these two experiments were the ones where the highest uranium concentrations were measured. The experiments performed at 30 mmol/L total bicarbonate concentration and at faster flow rates (0.3 mL/min) showed a clear dependence of the dissolution rate on the flow rate. This was interpreted as a possibility of precipitation of a secondary phase, which was not suppressed in the residence time range of the experiments, though we have no clear evidence of such a secondary solid phase. For this reason, the dissolution rates determined at 30 mmol/L total bicarbonate concentration were not included in the modeling exercise of these two samples. On the other hand, we did not see any significant effect of the flow rate on the dissolution rate determined for samples RZ10 and RZ13, indicating the attainment of a steady-state situation with no secondary phase formation. The uranium dissolution rates obtained for the Oklo uraninite samples were compared with values obtained under similar experimental conditions for other uraninite samples (7-9) as well as for synthetic uranium dioxide (6) and for spent nuclear fuel (10). The experimental values collected are presented in Figure 5. As seen in this figure, the uranium dissolution rates for most of the Oklo samples are lower than the ones corresponding to synthetic uranium dioxide and spent nuclear fuel. The effect of the surface area on the dissolution rates calculated in the present work must be taken into account. For samples RZ9, RZ10, and AP9, the sample characterization showed the presence of accessory minerals that will be accounted for in the BET specific surface area measured, not all of which will correspond to uraninite. However, sample RZ13 gives a BET specific surface area that will mostly correspond to uraninite. Because of that, the normalized dissolution rates of this sample are more clearly

FIGURE 5. Normalized uranium dissolution rates (mol/(s‚m2)) determined in this work for Oklo uraninite samples, compared with the results found in the literature for uranium dioxide, spent fuel, and uraninites under similar experimental conditions. attributed to this phase and can be more readily compared with the synthetic uranium dioxide determinations, while for the rest of the samples, the presence of accessory phases might introduce a nonconservative factor in the calculation of dissolution rates, giving values lower than those in the case of a pure uraninite phase. It is interesting to see the excellent agreement between the value determined in a batch system for a uraninite sample from the Cigar Lake site (7) and the dissolution rate determined for Oklo sample RZ10, at the same total bicarbonate and oxygen concentrations. In both cases, the uraninite sample corresponds to a relatively deep and unaltered location. However, the value determined by Grandstaff (8) is much larger than any other value given in Figure 5. According to the description of this sample given by the author, it corresponds to a highly altered, oxidized uraninite. This would explain the relatively high dissolution rate measured experimentally, and shows, together with the other values presented in Figure 5, the significant implication that initial solid alteration may have in uranium mobilization. VOL. 38, NO. 12, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3313

TABLE 4. Normalized Dissolution Rates (mol/(s‚m2)), Given as log r, Obtained at 60 °C as a Function of the Total Bicarbonate Concentration in the Test Solutiona log r

a

[HCO3-] (mM)

U

2.7 8 15

-10.82 ( 0.03 -9.78 ( 0.05 -9.27 ( 0.06

2.7 8 15

-9.69 ( 0.04 -9.00 ( 0.05 -8.49 ( 0.05

Nd

Mo

Sb

Argile de Pile Reactor Zone 9 -11.35 ( 0.07 -10.86 ( 0.06 -10.32 ( 0.04

-10.39 ( 0.11 -9.41 ( 0.06 -9.28 ( 0.07

-10.67 ( 0.09 -9.87 ( 0.11 -9.15 ( 0.02

Reactor Zone 13 -10.44 ( 0.04 -10.19 ( 0.05 -9.38 ( 0.06

-9.17 ( 0.06 -8.19 ( 0.07 -7.89 ( 0.04

Empty spaces indicate that concentrations were below detection limits and no dissolution rate could be determined.

The minor element dissolution rate values determined for the different uraninite samples in the experiments detailed above are collected in Figures 1-4 as well as in Table 3. The main observations can be summarized as follows. Neodymium. The behavior of neodymium was found consistent throughout all experiments performed on the samples studied. The results indicate a noncongruent dissolution of neodymium with respect to the uraninite matrix. Neodymium is one of the REEs considered to be stable within the uraninite structure and, consequently, should be expected to be retained in the uraninite matrix (23). However, results obtained in the leaching experiments performed in this study provide contrary results. Experimental results point to a source other than uraninite as responsible for the main Nd contribution in solution. In this sense, careful characterization of the minerals surrounding the uraninite grains proved that phosphate minerals as well as clays were very effective at retaining neodymium (24). In addition, the characterization of these minerals proved that most of the Nd was of fissiogenic origin. Therefore, on the basis of the experimental results presented, our hypothesis is that these accessory minerals may be the main source of the neodymium found in solution rather than uraninite, though they could not be identified in the solid samples studied. Alternatively, the dissolution behavior of the lanthanides could be controlled by the secondary precipitation of sparingly soluble phosphates, which have a tendency to control REE concentrations in low-temperature geochemical environments. Molybdenum. The dissolution rate observed for molybdenum at different bicarbonate concentrations showed a trend very similar to that obtained for uraninite, mostly for samples RZ10 and AP9 (Figures 2 and 4, respectively). Because of this similarity, it is postulated that the Mo fraction present as MoO2 and structurally related to the uraninite might be the one actually being dissolved. Similar observations have been made in some spent nuclear fuel dissolution studies (23, 25). Cesium. This element was only detected in measurable quantities in the experiments performed with sample RZ10 (Figure 2). This is not surprising because of the mobility of this element, and therefore, it cannot be retained in the uraninite structure. In fact, elements such as Rb, Sr, Ba, and Cs were almost totally removed in practically all the natural reactors. However, Hidaka et al. (15) have estimated that roughly 5% of these elements were retained in the core of reactor 10. For this sample, the average U/Cs dissolution rate ratio was calculated to be 0.8 ( 0.3, showing a very close congruency between these two elements, as would be expected since cesium is of fissiogenic origin and considering that the remaining cesium found in the solid sample RZ10 is assumed to be retained in the uraninite structure. Antimony. This element was found in the leaching solutions of samples RZ9 and AP9, which also showed the 3314

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 12, 2004

highest concentration of this element in the chemical characterization of the solid samples. It is observed that the normalized dissolution rate of Sb is not far away from congruency with respect to uranium (especially for sample AP9, Figure 4). Unfortunately, we could not find any information about the distribution of this element in the uraninite samples of Oklo, but on the basis of the experimental observations, our hypothesis is that this element would be found retained in the uraninite structure. Ytterbium. This rare earth element was only found in detectable quantities for samples RZ9 and AP9, and only for the highest total bicarbonate concentrations studied. Although its dissolution rates are quite close to those of uranium, ytterbium dissolution data do not seem to follow the same trend observed for the uraninite matrix dissolution. Rather, they seem to follow the same dependence on bicarbonate as neodymium, though with higher dissolution rate values. With the few data available it is very difficult to extract any definitive conclusion, but we tend to believe that the ytterbium source/sink is more likely to be the same as that of neodymium, that is, accessory minerals as phosphates or clays. Experiments at 60 °C. One of the variables of interest in the present study was temperature because of the expected heat radiation of radioactive wastes in a hypothetical future repository. Also, it is of interest to see how differences in the dissolution rate due to this variable might affect the release of the minor components and/or affect the mechanism of dissolution of the uraninite matrix. The leaching studies at this temperature were performed with two uraninite samples selected from the shallowest and the deepest reactor zones studied, the argyle de pile of reactor 9 and the uraninite from reactor 13. Uranium Dissolution Rates. Results determined at 60 °C are collected in Table 4 as a function of the total bicarbonate concentration in the test solution. From the experiments performed at both 25 and 60 °C, we calculated the apparent activation energies, Eap, corresponding to the overall dissolution process, using the Arrhenius equation. The values obtained (Table 5) are similar to those reported in the literature for spent nuclear fuel (11), uranium dioxide (6, 12-14), and uraninite (8), which range between 20 and 80 kJ/mol. According to Lasaga (26), apparent energies lower than 40 kJ/mol suggest diffusion-controlled processes, while values ranging from 40 to 85 kJ/mol seem to indicate a surfacecontrolled mechanism. The range of values of Ea derived from our measurements would indicate that the ratecontrolling step of the dissolution process is mainly diffusioncontrolled for the lower bicarbonate concentrations. Once the bicarbonate concentration is sufficiently large to ensure its presence at the surface, then the process becomes surfacecontrolled. This is, again, in good agreement with the

TABLE 5. Apparent Activation Energies Calculated from Arrhenius Equation, Using the Dissolution Rates Experimentally Determined at 25 and 60 °C for Samples AP9 and RZ13a Eap (kJ/mol) -]

[HCO3 (mM)

AP9

RZ13

2.7 8 15

17.8 43.1

17.0 26.5 49.1

a For sample AP9, the value at 2.7 mM total bicarbonate concentration is not included because the dissolution rate determined at 60 °C was slightly lower than that obtained at 25 °C, giving a negative Eap value.

mechanism previously postulated for the oxidative dissolution of unirradiated UO2 (6). Minor Element Dissolution Rates. The results obtained for the minor elements determined in the samples studied are shown in Table 4 as a function of the total bicarbonate concentration in the test solution. The overall trend observed at 60 °C is similar to that obtained at 25 °C. In the case of Nd, as was the case at 25 °C, a lower dissolution rate was observed as compared to that corresponding to uranium. This fact confirmed once again that this element is most probably segregated from the uraninite matrix, in accessory minerals associated with the uraninite sample. The dissolution rate of Mo was consistently found to be slightly higher than the rate of dissolution of U, and the dependence on the bicarbonate concentration appears to follow a trend similar to that observed for U. Again this behavior is very similar to that obtained at 25 °C. Finally, the good agreement between Sb and U at 60 °C, not only in the rate values but also in the dependence of the dissolution rates on the bicarbonate concentration, would be a clear indication of congruent dissolution behavior. By and large, the results of the dissolution studies indicate that the release and mobility of some of the minor radionuclides from the Oklo samples follow the same mechanism that controls the dissolution behavior of uranium, the main component of the solid matrix, as has also been observed in some synthetic uranium dioxide and spent nuclear fuel dissolution studies. Only in the case of Nd and Yb, the minor components appear to be controlled by their individual solubility. The same behavior has been found for REE in other natural analogue studies (27).

Acknowledgments This work was financially supported by the European Commission (Contract FI4W-CT96-0020) and ENRESA (Spanish Radioactive Waste Management Co).

Literature Cited (1) Murphy, W. M.; Shock, E. L. Environmental Aqueous Geochemistry. In Uranium, Mineralogy, Geochemistry and the Environment; Burns, P. C., Finch, R., Eds.; Reviews in Mineralogy Vol. 38; Mineralogical Society of America: Washington, DC, 1999. (2) Abdelouas, A.; Lutze, W.; Nuttall, E. Uranium Contamination in the Subsurface Characterization and Remediation. Geochemistry. In Uranium, Mineralogy, Geochemistry and the Environment; Burns, P. C., Finch, R., Eds.; Reviews in Mineralogy Vol. 38; Mineralogical Society of America: Washington, DC, 1999. (3) Miller, W.; Allexander, R.; Chapman, N.; McKinley, I.; Smellie, J. Geological Disposal of Radioactive Wastes and Natural Analogues; Pergamon: New York, 2000. (4) Cramer, J., Smellie, J., Eds. SKB Technical Report TR-94-04; 1994. (5) Neuilly, M.; Bussac, C.; Fre´jacques, G.; Nief, G.; Vendrye`s, G.; Yvon, J. C. R. Acad. Sci. 1972, 275, 1847. (6) de Pablo, J.; Casas, I.; Gime´nez, J.; Molera, M.; Rovira, M.; Duro, L.; Bruno, J. Geochim. Cosmochim. Acta 1999, 63, 3097. (7) Bruno, J.; Casas, I.; Cera, E.; Ewing, R. C.; Finch, R. J.; Werme, L. O. Mater. Res. Soc. Symp. Proc. 1995, 353, 633. (8) Grandstaff, D. E. Econ. Geol. 1976, 8, 1493. (9) Blomqvist; et al. EUR 19611E; 2000. (10) Gray, W. J.; Wilson, C. N. Report PNL-11895; 1995. (11) Gray, W. J.; Leider, H. R.; Steward, S. A. J. Nucl. Mater. 1992, 190, 46. (12) Aronson, S.; Roof, R. B., Jr.; Belle, J. J. Chem. Phys. 1957, 27, 137. (13) Shortmann, W. E.; De Sesa, M. A. Proc. Second United Nations Int. Conf. Peaceful Uses At. Energy 1958, 3, 333-341. (14) Hiskey, J. B. Trans. Inst. Min. Metall., Sect. C. 1979, 88, C145. (15) Hidaka, H.; Takahashi, K.; Holliger, P. Radiochim. Acta 1994, 66/7, 463. (16) Gauthier-Lafaye, F.; Weber, F. Econ. Geol. 1989, 84, 2286. (17) Janeczek, J. Mineralogy and Geochemistry of Natural Fission Reactors in Gabon. In Uranium, Mineralogy, Geochemistry and the Environment; Burns, P. C., Finch, R., Eds.; Reviews in Mineralogy Vol. 38; Mineralogical Society of America: Washington, DC, 1999. (18) Curtis, D.; Benjamin, T.; Gancarz, A.; Loss, R.; Rosman, K.; De Laeter, J.; Delmore, J. E.; Maeck, W. J. Appl. Geochem. 1989, 4, 49. (19) Brookins, D. G. Waste Manage. 1990, 10, 285. (20) Gauthier-Lafaye, F.; Ledoux, E.; Smellie, J.; Louvat, D.; Michaud, V.; Pe´rez del Villar, L.; Oversby, V.; Bruno, J. Report EUR 19139; 2000. (21) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309. (22) Bruno, J.; Casas, I.; Puigdome`nech, I. Geochim. Cosmochim. Acta 1991, 55, 647. (23) Kleykamp, H. Nucl. Technol. 1988, 80, 412. (24) Gauthier-Lafaye, F.; Holliger, P.; Blanc, P. L. Geochim. Cosmochim. Acta 1996, 60, 4831. (25) Hanson, B. D. Report PNL-10540; 1998. (26) Lasaga, A. C. J. Geophys. Res. 1984, 89, 4009. (27) Bruno, J.; Duro, L.; Grive´, M. Chem. Geol. 2002, 190, 371.

Received for review December 12, 2003. Revised manuscript received March 12, 2004. Accepted March 30, 2004. ES0353863

VOL. 38, NO. 12, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3315