The uranium ore from Mina Fe (Salamanca, Spain) as a natural analogue of processes in a spent fuel repository

Chemical Geology 190 (2002) 395 – 415 www.elsevier.com/locate/chemgeo

The uranium ore from Mina Fe (Salamanca, Spain) as a natural analogue of processes in a spent fuel repository L. Pe´rez del Villar a,*, J. Bruno b, R. Campos a, P. Go´mez a, J.S. Co´zar a, A. Garralo´n a, B. Buil a, D. Arcos b, G. Carretero c, J. Ruiz Sa´nchez-Porro d, P. Herna´n e a

b

CIEMAT, DIAE, CHE, Ed. 20, Avda. Complutense 22, 28040 Madrid, Spain Enviros-QuantiSci, Parc Tecnolo`gic del Valle´s, 08290 Cerdanyola del Valle´s, Barcelona, Spain c AITEMIN, Rı´o Valdemarı´as s.n. 45007 Toledo, Spain d ENUSA, Ctra. Ciudad Rodrigo-Lumbrales, Km 8, 37500 Ciudad Rodrigo, Salamanca, Spain e ENRESA, Emilio Vargas 7, 28043 Madrid, Spain

Abstract In the frame of the ENRESA natural analogue programme, the uranium ore from the ‘‘Mina Fe’’ (Salamanca, Spain) has been studied as a natural analogue of radioactive spent fuel behaviour. This uranium mine is hosted in highly fractured schistose rocks, a geological setting that has not been envisaged in the Spanish options for radioactive waste burial. However, some analogies with the processes that might be involved in the evolution of these geological repositories suggested this investigation. The pitchblende – pyrite – carbonate paragenesis has been studied ‘‘in situ’’ as natural analogue of the nuclear spent fuel behaviour under extremely oxidative dissolution conditions. Similarly, secondary Fe oxyhydroxides and clay minerals have also been considered as relevant analogue materials for the retention of uranium and other analogous trace metals. A multidisciplinary characterisation of the site has been performed in order to study these processes. Though the intense mining activities in the site hindered precise determination of the original hydrogeological and hydrochemical features of the investigated zone (Boa fault zone), the mineralogy and geochemistry of fracture fillings, mineralisation and associated clayey materials have allowed the geochemical evolution of the system to be established. Three geochemical zones have been clearly differentiated: (i) the oxidised zone, from the surface to approximately 20 m depth, (ii) the redox transition zone, from 20 to 50 m depth, and (iii) the reduced zone, located below the transition zone. The oxidised zone is characterised by the presence of the typical mineral association resulting from the strong acid conditions caused by the total oxidation of pyrite and other sulphides. The total oxidation, dissolution and leaching of U(IV), as uranyl – sulphate aqueous complexes, prevailed in this oxidised zone. The redox transition zone is characterised by the coexistence of the primary uranium paragenesis, oxidised minerals, as well as numerous secondary solid phases as a result of the physico-chemical changes in the environment. The optimal physico-chemical conditions for the coffinitisation of pitchblende and the coprecipitation of Fe(III) – U(VI) took place in this zone. In the reduced zone, where the primary uranium paragenesis is present, we currently find the necessary physico-chemical conditions to stabilise pitchblende, pyrite and carbonates. The physico-chemical conditions of the oxidised zone are not relevant to disposal conditions. In the transition zone, two main geochemical processes take place: (i) the coffinitisation of pitchblende, which may be an important process for the stability of

*

Corresponding author. Fax: +34-1-3466542. E-mail address: [email protected] (L. Pe´rez del Villar).

0009-2541/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 9 - 2 5 4 1 ( 0 2 ) 0 0 1 2 7 - 4

396

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

spent fuel in reducing conditions, and (ii) the co-precipitation of the Fe(III) and U(VI) as oxyhydroxides, another relevant mechanism for the retention of uranium. The physico-chemical conditions that prevail below 50 m depth should be sufficient to stabilise a spent nuclear fuel repository, in the same way as they have been able to preserve the 34-Ma-old uranium deposit of the Mina Fe. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Mina Fe uranium deposit; Fracture fillings; Hydrogeology; Hydrochemistry; Natural analogue

1. Introduction Uranium ore deposits have been extensively studied as natural analogues to the deep geological disposal of radioactive waste (Chapman et al., 1990; Murphy and Pearcy, 1992; Blanc, 1996; Rivas et al., 1997; Bruno et al., 1997; Gauthier-Lafaye et al., 2000; Blomqvist et al., 2000). These investigations constitute an essential element in both national and international research programmes applied to the assessment of the long-term evolution of geological repositories for spent nuclear fuel in particular. In this context and in the frame of the ENRESA natural analogue programme, the uranium ore deposit of ‘‘Mina Fe’’ has been studied as a natural analogue of radioactive spent fuel behaviour. The uranium ore deposit of ‘‘Mina Fe’’ (Arribas, 1962; 1985; Mangas and Arribas, 1984; Martı´n-Izard, 1989; Both et al., 1994) is hosted in highly fractured schistose rocks, a geological setting that has not been envisaged in the Spanish options for radioactive waste burial. However, their similarities with some features of the repository, as well as analogies with processes that could be involved in repository evolution, suggested the study of this uranium deposit as a natural analogue. The most important analogue features are the following. . The existence of large uranium concentrations as pitchblende (UO2 + x), which is chemically analogous to the main component of spent nuclear fuel, which has an oxidation degree of 2.25 < x < 2.66 as a result of radiolytic oxidation (Shoesmith and Sunder, 1992; Casas et al., 1994a; Bruno et al., 1995). . The solubility behaviour of pitchblende as a result of interaction with groundwaters of varying chemical composition can be used to validate predictive models for spent fuel stability under severe alteration conditions (Bruno et al., 2000).

. Some of the weathering products of pitchblende are similar to those that have been identified during the experimental oxidative dissolution of UO 2 , SimfuelR (Bruno et al., 1997), as well as natural uraninite and pitchblende (Casas et al., 1994b). . Fe(III)-oxyhydroxides in the oxidised zone of the deposit could be similar to the spent fuel steel container corrosion products that could be formed under redox transition conditions. These corrosion products may act as radionuclide and trace metal scavengers. . The hydrothermal alteration and weathering of the schistose rocks have formed clays, which are partially similar to the ones to be used in the engineering barriers of the repository. The understanding of the radionuclide diffusion/retention processes in these clays is of relevance in testing the current models used in the performance assessment of the engineered barriers in deep geological repositories. According to these characteristics, some key geochemical objectives of this investigation are the following. (1) ‘‘In situ’’ investigation of the oxidative dissolution of the pitchblende as the source term for mobilisation of uranium and other trace elements analogous to the radionuclides of interest. In addition, the potential retention of U(VI) as a result of the precipitation of secondary minerals in the wide range of chemical conditions represented by the local groundwaters (from acid and oxidising sulphate waters to alkaline and reducing bicarbonate ones) could be studied. (2) Determination of the retention properties and parameters of uranium and other trace metals in Fe(III)-oxyhydroxides and clayey alteration products. (3) Integration of all these processes in the current models for the geochemical evolution of a repository, including their quantification and comparison with present safety assessment models.

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

Pursuing these objectives will progress the understanding and quantification of the following aspects: (i) The longevity of the geochemical environment in the engineered barriers and rocks immediately surrounding spent fuel in a repository.

397

(ii) Radionuclide solubility and speciation in groundwaters of quite diverse physico-chemical characteristics. (iii) Radionuclide sorption in the corrosion products of the steel spent fuel container under oxic conditions.

Fig. 1. Geographical location of the uranium ore deposit of ‘‘Mina Fe’’ and ‘‘Mina D, Hole-01’’, in which investigations were performed. ESPERANZA, ALAMEDA and SAGERAS are other non-mined uranium zones in the region.

398

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

(iv) Irreversibility of sorption processes in the repository and surrounding rocks. This work summarises the main geochemical results obtained during a multidisciplinary, 3-year project, known as Matrix-Phase I, which was supported by ENRESA.

2. Geological background The ‘‘Mina Fe’’ uranium-ore deposit is the most important in the Spanish Iberian massif. It is located some 10 km northeast of Ciudad Rodrigo (Salamanca) (Fig. 1). The uranium mineralisation either fills open fractures or cements the fault breccia affecting metasediments of the Upper Proterozoic – Lower Cambrian schist – graywacke complex, known as ‘‘Complejo Esquisto –Grauva´quico’’ (CEG). In the ‘‘Mina Fe’’ area the schist – graywacke complex rocks consist mainly of a metamorphosed sequence, in part turbiditic, of carbonaceous pelitic and fine-grained psammitic rocks, in which sedimentary textures are frequently observed. The grade of metamorphism is mainly greenschist facies, the main rock types being slates, quartzites, conglomerates, and sericitic and chloritic phyllites and schists, with some interbedded calc-silicate rocks, which represent metamorphosed layers of impure carbonate sediments (Martı´n-Izard, 1989). The uranium mineralisation is the result of a hydrothermal process in which three main steps have been distinguished. The first occurred after a brecciation process that caused chloritisation of the host rocks and breccia fragments, as well as the formation of small ankerite – pyrite bearing veins, with minor galena, sphalerite and chalcopyrite. The second step took place after another brecciation process and was the most productive in uranium ores. During this step, adularia, pyrite, pitchblende, coffinite, dolomite and calcite were formed. The third step was characterised by the episodic, laminated and repeated precipitation of pyrite, carbonates and collophormic pitchblende. This type of uranium mineralisation fills open fractures and breccia voids (Arribas, 1985). The absolute age of the ‘‘Mina Fe’’ uranium deposit, determined by using the 207Pb/204Pb ‘‘vs’’ 235U/204Pb iso-

chrone, is 34.8 F 1.6 Ma; the same age as of the Pyrenean tectonic phase of the Alpine orogeny. This tectonic phase is also responsible for the formation of the Ciudad Rodrigo continental basin (Both et al., 1994). Based on fluid inclusion data and the chemical composition of the early chlorite, the temperature of the uranium mineralisation varies from 230 to 60 jC during the two first steps, while the third mineralisation step took place at a temperature lower than 60 jC (Mangas and Arribas, 1984). The uranium mineralisation has been intensively eroded and oxidised, as well as covered in places by continental Tertiary and Quaternary sediments. The thickness of these sediments in the mining area varies between 5 and 20 m. Numerous secondary uranium minerals such as yellow gummites, ianthinite, epi-ianthinite, alpha uranotyle, autunite, metaautunite, torbernite, saleeite and uranopilite were formed as a result of the weathering processes (Arribas, 1962, 1975).

3. Methodology Existing documentation concerning the geology and mining activities at ‘‘Mina Fe’’ was analysed in order to: (i) select a specific site for the study: the Boa fault; (ii) ensure the representativity of core and groundwater sampling and adequate planning of boreholes; and (iii) prepare a conceptual geochemical model of the Boa fault zone (Pe´rez del Villar, 1998). The intensively fractured exposure of the Boa fault, revealed in the open-pit mine, exhibits substantial uranium mineralisation that is largely oxidised and leached in the upper part of the fault due to weathering processes (Fig. 2). This zone was used to address the first and second objectives of this investigation. In order to achieve a consistent geochemical model of the system, we have applied a methodology that integrates structural, mineralogical, hydrogeological and hydrochemical information. This methodology is summarised as follows. . Geological characterisation of the site by means of: (i) lithological, fracture, radiometric and weathering mapping of the mine exposure of the Boa fault,

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

399

Fig. 2. Intensively fractured and oxidised mine exposure of the Boa fault. FS: sericitic phyllites; FB: banded phyllites; FN: quartz-sericitic phyllites; Qz: Hercynian – late Hercynian sulphide-bearing quartz veins; T: continental Tertiary sediments; PQ: Plio-Quaternary continental materials; Q: Quaternary materials.

(ii) structural analysis, and (iii) mineralogical and geochemical characterisation of the uranium ore from the Boa fault. . In-depth geological characterisation of the site by means of: (i) four boreholes (SM-1 to SM-4) with continuous core recovery and low contamination potential (Fig. 3), (ii) lithological, mineralogical and radiometric logs, and (iii) the quantification and analysis of the fractures as identified in the borehole-cores. . The mineralogical and geochemical characterisation of the fracture filling materials and the uranium mineralisation hosted by the Boa fault, with special emphasis on: (i) the mineral transformations due to secondary processes involving redox and precipitation reactions, (ii) secondary mineral phases that indicate the specific pH and Eh conditions under which they were formed, (iii) oxidative dissolution processes affecting of the original pitchblende, and (iv) quanti-

fication of the extent of trace element incorporation into secondary minerals. . Hydrogeological characterisation of the system by: (i) transmissivity studies by using pulse and slug tests of relevant zones, particularly in the Boa fault and (ii) piezometric study of these relevant zones. . Groundwater characterisation and modelling, particularly in the Boa fault zone, including: (i) ‘‘in situ’’ determination of the physico-chemical parameters, (ii) chemical and isotopic analysis of the groundwaters in conventional labs, (iii) geochemical modelling of the groundwater mixing and water/rock reactions by using the NETPATH (Plummer et al., 1991) and EQ3/6 (Wolery and Daveler, 1992) geochemical code packages. The data obtained have allowed definition of the dominant geochemical processes, which has allowed us to establish a consistent and integrated model of the geochemical evolution of the system, in spite of

400

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

Fig. 3. Topographic map of ‘‘Mina D, Hole H-01’’ in which the Boa fault, the location of boreholes SM-1 to SM-4 and the orientation of crosssections A – AV, B – BV and C – CV in Fig. 4 are shown.

extensive perturbation of the investigated zone, mainly due to the mining activities.

4. Results 4.1. Geology The Boa fault zone, known as ‘‘Mina D’’, Hole H01, was partially covered by continental Tertiary and Quaternary sediments from the Ciudad Rodrigo basin. There are some remains of these sedimentary materials in the upper part of the western zone of the quarry. The mine exposure of the Boa fault breccia exhibits, from the bottom to the top, a black band of carbonaceous clayey material and the mineralised breccia formed by fragments of schistose rocks, mainly cemented by Fe-oxyhydroxides from the oxi-

dation of pyrite and some residual carbonates. This fault zone has an apparent width of 30 to 70 cm and intersects the different types of schistose rocks from the CEG, some of which are rich in organic matter. Small hydrothermal sulphide-bearing quartz veins formed during the Hercynian and/or late Hercynian tectonic activity also intersect these schistose rocks. Geometrical and statistical analysis of the Boa fault and other fractures in the site (Campos et al., 1999a,b) indicated the following. The Boa fault, with variable direction and dip, N60-80E/30-60N, originated as an Hercynian inverse fault, which was reactivated as a normal fault during the Alpine orogeny, when the uranium-ore deposit was formed. Fracture frequency and the number of fracture intersections indicate that the area is intensively fractured. The fracture intensity is larger in the down-

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

thrown block (N) than in the uplifted block (S) of the Boa fault. This fact would explain the greater weathering and more intensive uranium leaching in the downthrown block, revealed by radiometric data.

401

4.2. Mineralogy and geochemistry of fracture fillings The mineralogical and geochemical analyses of the primary and secondary fracture fill minerals have

Fig. 4. Schematic cross-sections of ‘‘Mina D’’, Hole H-01, showing the vertical redox zoning, based on fracture filling mineralogy of coresamples from boreholes SM-1, 2, 3 and 4. The location of profiles is shown in Fig. 3.

402

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

allowed us to establish the following vertical zoning: (i) an upper oxidised zone, with the lower limit around 18 F 2 m depth; (ii) the transition redox zone, with a lower limit at 50 m depth; and (iii) the reduced zone, below about 50 m depth (Pe´rez del Villar et al., 1999, 2000) (Fig. 4). The oxidised zone shows a large abundance of Fe(III) and Mn – Fe oxyhydroxides, sometimes including Mn – Al –Ni and Mn –Pb oxides. The Fe(III)-oxyhydroxides contain significant amounts of Ti, Cu, Zn, with lesser concentrations of Cr, U and P. The Mn –Fe

oxyhydroxides have variable amounts of Co, Ni and Ba (Fig. 5a and b). All these metals are at trace level. Other mineral phases that indicate acid conditions are jarosite (Fig. 5a), a double sulphate of Fe(III) and K; allophane (amorphous SiO2
Al2O3
nH2O) (Fig. 5c) and kaolinite (Fig. 5d). The oxidised zone also contains some secondary mineral phases formed at neutral pH, such as: autunite (Ca – U(VI) phosphate) (Fig. 6a), cerium oxides (Fig. 6b), Ce phosphate, La –Nd phosphates, and rhodochrosite (Mn carbonate) (Fig. 6c). Halloysite and/or

Fig. 5. Electron images showing some minerals neoformed in the oxidised zone under acid conditions. a: Jarosite (1) coated by Feoxyhydroxides (2), with colloformic texture. b: Mn-oxyhydroxides with minor Al, Si, Fe, Ca, Ni, Co and Ba (1), closely associated with Feoxyhydroxides (2) and coated by smectite (3). c: Allophane-like material showing its typical colloidal texture. d: Idiomorphic kaolinite showing its typical face to face texture.

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

403

Fig. 6. Electron images showing some minerals neoformed in the oxidised zone when the neutrality is restored. a: Subidiomorphic autunite (1) precipitated on inherited biotite (2) from the host rock. b: Ce oxides (2) precipitated on smectite coated by Fe-oxyhydroxides (1). c: Spheroidal secondary rhodochrosite (2) precipitated on Mn-oxyhydroxides (1) and closely associated with a mixture of allophane-like material and halloysite – metahalloysite. d: Fibrous-radiating halloysite – metahalloysite formed from allophane-like material by silication processes.

metahalloysite and smectite are frequent in the oxidised zone (Fig. 6d). Multivariate analysis of major and trace elements analysed in oxidised samples, by using a tree-like, two dimensional correlation diagram (dendrogram), shows that Fe2O3, W and P2O5 determine a cluster indicating that Fe oxyhydroxides control, either by sorption and/ or coprecipitation, W and P2O5. Similarly, MnO, Ni, Co, Zn and Cu form another cluster that reveals the role played by Mn-oxyhydroxides in the retention of Ni, Co, Zn and Cu. The cluster formed by Al2O3, MgO, TiO2, Na2O and CaO is connected to the clay

fraction of the samples, while uranium is slightly associated with the La –Y cluster. This last association must be explained taking into account that residual uranium in the samples can be linked either to some inherited and resistant minerals from the host rocks, such as monazite and xenotime, or to the abovementioned secondary REE phosphates, with uranium as trace element (Fig. 7). The transition redox zone is characterised by: (i) primary phases, which originally filled the fractures; (ii) phases described above characteristic of the oxidised zone; and (iii) neoformed secondary phases

404

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

Fig. 7. A tree-like, two-dimensional diagram (dendrogram) depicting the mutual relationships among the chemical variables of fracture filling samples. Notice the clusters formed by MgO – TiO2 – Na2O – CaO – Al2O3; Fe2O3 – W – P2O5 – Be; MnO – Ni – Co – Zn – Cu and La – Y – U – Sr.

which are a result either of the transition pH and Eh conditions, or the partial or total transformation of the primary minerals. The primary minerals are: quartz, sulphides (molibdenite, pyrite, chalcopyrite, sphalerite and galena), lead selenides (clausthalite), carbonates (mainly ankerite, dolomite and calcite) and U(IV)/(VI)-oxides (pitchblende). The secondary minerals formed as a result of changes in redox and alkalinity conditions are: Fe – Cu sulphoselenides (Fig. 8a), Ag –Pb – Fe selenides, secondary Cu sulphides (covellite – chalcocite), undetermined Cu – Fe; Fe – Cu – Ag; Pb – Cu – Fe; Fe – Sn sulphides (Fig. 8b). These are clearly representative of the slightly reducing environment. Elemental Se or selenolite (Fig. 8c) indicates slightly oxidising conditions, while uranium (VI) hydroxides coprecipitated with Fe(III)-oxyhydroxides (Fig. 8d) are more representative of oxidised and slightly acid conditions (pH>4.2). Gypsum (Fig. 8e), barite and clay minerals such as smectite and corrensite indicate a neutral or slightly alkaline environment.

Among the phases that are the result of the total or partial transformation of the primary minerals we find partially coffinitised pitchblende (Fig. 9a and b) and pitchblende partially transformed into zippeite (poorly soluble uranyl sulphate) (Fig. 9c). In addition, secondary siderite (Fig. 9d) has formed from ankerite at low temperature (T < 25 jC) by ferrous metasomatism. Iron (III) and Zn oxyhydroxides are the result of the partial oxidation of sphalerite– marmatite and pyrite, and therefore closely related to these primary sulphides. From the geochemical point of view, the transition zone is enriched in Fe, Cu, Ag, Ni, Zn, Se and U, as expected from the supergenic enrichment zones. All these mineralogical and geochemical characteristics are consistent with neutral to slightly alkaline and intermediate redox conditions. The reduced zone is characterised by the presence of the original mineral associations previously described. The physico-chemical conditions of this zone are reducing and slightly alkaline since pitchblende, sulphides and carbonates are stable.

Fig. 8. Electron images showing secondary minerals formed by precipitation in the redox transition zone, as a result of changes in redox and alkalinity conditions. a: Secondary Fe – Cu sulphoselenides (1) precipitated on Fe-oxyhydroxides (2). b: Secondary Cu – Fe – Ag sulphides (1 and 2) precipitated on inherited muscovite from the host rock (3). c: Idiomorphic-acicular elemental selenium or selenolite (SeO2) (1) precipitated on massive Fe-oxyhydroxides (2). d: Probable uranium (VI) hydroxides (2) co-precipitated with Fe-oxyhydroxides (1). e: Radiating aggregates of idiomorphic gypsum (1) precipitated on massive Fe-oxyhydroxides.

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

405

406

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

Fig. 9. Electron images showing secondary minerals formed in the redox transition zone, as a result of partial transformation of the primary minerals. a: Botryoidal and colloformic pitchblende transformed into coffinite-like material (1); (2) Fe-oxyhydroxides. b: Massive pitchblende (1) partially transformed into coffinite (2); (3) secondary siderite. Notice the distribution, in irregular patches, of coffinite. c: Massive and residual pitchblende (1) almost totally transformed into probable zippeite (2). d: Primary ankeritic carbonates (3) partially transformed into secondary siderite (2) by ferrous metasomatism, at low temperature ( < 25 jC); (1) spheroidal pitchblende closely associated with chlorite (4) from the primary uranium paragenesis.

4.3. Hydrogeology and water chemistry of the system Mining activities have largely disturbed the hydrogeological system and have made it impossible to determine the original hydrogeological characteristics. Mining investigation boreholes have created a network of artificial, large transmissivity features, greater than the original transmissivity of the rock matrix. In addition, the excavation of Hole H-01 significantly modified the local drainage of the zone and drew down the water table to the level of the bottom of this

Hole. Blasting has also heavily modified the hydraulic properties of the abundant fractures present in the investigated zone (Carretero et al., 2000). Nevertheless, the present hydrogeological investigation has allowed us to reach the following conclusions: 

The Boa fault is not a preferential flow path, but is part of a larger fault and fracture set that works as a whole hydraulic system.  The fractures associated with the Boa fault do not have a substantial influence on groundwater circu-

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

lation. Other fractures that have no relation with this fault have a similar or larger transmissivity.  The hydraulics of the fracture system are quite uneven, with zones where the permeability of the fractures is quite low, due to their size and/or fracture fillings, together with areas where the permeability is quite high.  In general, according to the measured hydraulic heads, surface flow velocities are quite low (0.01 m/d), directed towards the bottom of Hole H-01. The deep groundwater flow is controlled by the regional flow system, which is conditioned by the ´ gueda River (Fig. 10). A The aquatic geochemistry of the system is affected by the disturbance of the hydrogeology of the site. Some of the investigation boreholes were affected by proximity to the sump formed by Hole H-01, which is continuously flooded with high acidity sulphate waters (pH 2.9) and consequently with a large concentration of dissolved metal ions, including uranium. In addition, the hydraulic connections between the boreholes and Hole H-01 hampered any possibility of understanding the original water chemistry of the system prior to mining activities (Go´mez et al., 2000).

407

The resulting waters from sections of the Boa fault intersected by boreholes drilled in the site may be classified as Ca – Mg-sulphate, with some larger bicarbonate concentration in borehole SM-2. The conductivity increases in the waters sampled from boreholes SM-2 to SM-4 (1220 to 6060 AS/cm), while the alkalinity decreases, with pH values from 7.1 to 4.8. The reducing capacity of the waters decreases in the same sense, with Eh values from  300 to + 8 mV (Table 1). This would indicate that there is an increasing portion of groundwater contamination from Hole H-01 water as we move from borehole SM-2 to SM-4. This is in agreement with the general flow trend. This would indicate that groundwaters from borehole SM-2 have the least influence from the acid drainage tube waters of Hole H-01, since they have lower Eh, more alkaline pH values, less sulphate content together with a larger bicarbonate concentration. According to this, the metal ion content is in general lower, except for uranium with concentrations in the order of 272 Ag/l, compared to 51.3 Ag/l measured in SM-4 waters (see Table 1). In spite of the intensive oxic disturbance of the system, the large reducing capacity of the SM-2 waters, which shows a measured Eh value around  300 mV, indicates that

Fig. 10. Idealised cross-section showing the hydrogeological model of the Boa fault zone. This cross-section is sub-parallel to the trace of the Boa fault. For location, see Fig. 4.

408

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

Table 1 Chemical and physicochemical characteristics of groundwater from the Boa fault packed of sections intersected by boreholes SM-2 to SM-4

Lower pack depth (m) Upper pack depth (m) O2 (mg/l) CE (AS/cm) in situ pH ‘‘in situ’’ Eh (mV) HCO 3 (mg/l) F  (mg/l) I  (mg/l) Cl  (mg/l) SO24  (mg/l) Br  (mg/l) NH4 + (mg/l) Ca2 + (mg/l) Mg2 + (mg/l) Na + (mg/l) K + (mg/l) Fe2 + (Ag/l) Fe3 + (Ag/l) TOC (mgC/l) SiO2 (mg/l) Li (mg/l) U (Ag/l) Al (mg/l) As (Ag/l) Cd (Ag/l) Co (Ag/l) Ni (Ag/l) Sr (Ag/l) V (Ag/l) Zn (Ag/l) Mn (mg/l) 226 Ra (Bq/m3) 241 Am (Bq/m3) dO18 dH2 Balance (%)

SM-2

SM-3

SM-4

45 36 0 1220 7.1  300 235.7 0.34 1.1 10.3 294.3 < 0.1 0.39 83.3 64.7 28 3.4 4667 100 29.6 22.7 0.06 272 < 0.05 7.9 < 30 < 50 53 310 < 50 1013 1.27 710 7440  7.01  47.89 4.0

51 38 0 2970 6.4  90 54.2 0.11 0.12 11 2267 < 0.1 0.19 240 285 44.7 3.5 171,000 1500 5.1 21.3 0.12 48.7 0.07 18.7 47 70 843 1467 < 50 8700 20.7 5670

31 21 0 6060 4.8 8 18.3 0.12 < 0.02 14.3 5133 < 0.1 0.93 503 645 48 7.8 297,000 2333 11.4 40.1 0.6 51.3 21.3 16.3 197 3000 9100 2667 360 8133 171.7 23,100 91,300  2.03  25.01  6.4

 7.29  49.44  6.5

Drilling water

Water Hole 01

176 7.11 440 22.38 < 0.1 < 0.02 8.16 55.88 < 0.1 0.14 20.13 5.16 5.41 2.69 < 50 < 50 2.21 5.75 0.01 4.88 < 0.05

4927.5 2.9 790 <3 < 0.1 < 0.02 15.5 4975 <1 1.14 510 561.25 41.75 7.72 4075 5975 1.4 107.75 0.89 102,250 67

<1 < 0.05 < 10 0.07

130 3.75 7850 2.02

< 0.03

96.75

2

 14.2

Characteristics of drilling water and water from Hole H-01 are also listed for comparison. Water from borehole SM-1 was not sampled since the Boa fault was intersected in the unsaturated zone.

there is a large redox buffer in these groundwaters. This could be the result of large concentrations of the redox pairs in the aqueous system, such as Fe(II)/Fe(III), Mn(II)/Mn(IV) and S2  /SO42  , and/or due to the presence of large amounts of TOC in this water (29.6 mg/l). Though the origin of this organic carbon is still not clear, it could come from the upper soil layer or as a result of the microbiological degradation and activation of the organic matter from the host rock.

The speciation calculated by using the EQ3/6 code indicates that in SM-2 waters all divalent cations such as Fe(II), Mn(II) and Ni(II) are mainly present as free ions and, in less extent, as carbonated species. The dissolved Fe(III) is present as Fe(OH)3(aq), while U, according to the speciation calculations, would also be present as U(OH)4(aq). The large concentration of uranium measured in these apparently reducing waters is surprising. The presence of U(OH)4(aq) could be

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

409

the result of the complexation of U(IV) by the large amount of organic matter present in SM-2. According to the NETPATH mixing calculations activities (Go´mez et al., 2000), the groundwaters collected in SM-3 could be the result of a 50/50 mix between the ground waters from the formation and the acid waters of Hole H-01 (the actual calculations give a 56 to 44 proportion). Similarly, the water sampled in SM-4 appears to be the result of the neutralisation of Hole H-01 waters by water/rock interaction processes. Finally, the tritium content of all the sampled waters (SM-2 = 2.6394 T.U.; SM-3 = 4.814 T.U.; SM4 = 3.1872 T.U.) indicates juvenile age (later than 1953), but the 14C content of SM-3 water (37.84 pMC) would indicate that this water is older than 5000 years. This contradictory dating could be the result either of the organic carbon dissolved in these waters or rock/water interaction processes with the fracture filling materials, mainly with older carbonates.

ering. The differences arise from the fact that, in this case, the primary mineralisation is a mixture of pyrite, carbonates and pitchblende.

5. Discussion

Consequently, the initial oxidation of pyrite would increase the acidity of the system and would produce large amounts of sulphate and dissolved Fe(III). Iron (III) constitutes a very efficient pyrite oxidant and, according to reaction (3), produces a large amount of dissolved Fe(II) which in turn can be oxidised to Fe(III) by the presence of oxygen and consequently would produce a feedback effect.

The most relevant geochemical processes in the system are related to the effect of the continuous oxidation of the site after the formation and burial of the uranium-ore deposit. However, while past geochemical processes can be deduced from the fracture filling mineralogy, the present weathering processes have been more difficult to identify due to the intense anthropogenic disturbance to the current groundwater system (Pe´rez del Villar et al., 2001). However, as the system has been continuously emerged and exposed to weathering since the Pyrenean tectonic phase of the Alpine orogeny, we assume that the present weathering processes are quite similar to those that affected the uranium mineralisation since its formation 34 Ma ago. Changes in the intensity and kinetics of the water/rock interactions (mainly oxidation and supergenic enrichment) could be given by tectonic and climatic changes that occurred in the last 34 Ma. With these assumptions, the mineralogical characteristics of the system indicate that the main geochemical driving forces have been the oxidation and supergenic enrichment processes that occur in sulphide-rich deposits when they are exposed to weath-

5.1. Geochemical processes in the oxidised zone The mineralogy and geochemistry of the fracture fillings in this zone can be explained by assuming: (i) the oxidation of pyrite triggered by the dissolved oxygen in rainwater and (ii) the initial ratio of pyrite and carbonate in the fracture filling materials, which is difficult to establish due to the large variability among fractures and in the various portions of a fracture. The sequence of reactions (1) and (2) that initiate the geochemical evolution of the system is: þ 2FeS2 ðsÞ þ 7O2 ðgÞ þ 2H2 O Z 2Fe2þ þ 4SO2 4 þ 4H

ð1Þ þ 4Fe2þ þ 6H2 O þ O2 Z 4FeðOHÞþ 2 þ 4H

þ 2þ FeS2 þ 14FeðOHÞþ þ 2SO2 4 2 þ 12H Z 15Fe þ 20H2 O

ð2Þ

ð3Þ

Iron (III) is quite soluble under strong acid conditions, but at pH values higher than 2.5 is hydrolysed forming Fe(OH)2 + (Garrels and Christ, 1965). In the presence of sulphate, this iron complex is the precursor for the precipitation of jarosite (Limpo et al., 1976; Pe´rez del Villar et al., 1979), according to reaction (4). This mineral phase has been frequently identified in the fracture fillings of the oxidised zone, either before or contemporaneous with goethite formation. þ þ 6FeðOHÞ2þ þ 4SO2 4 þ 2K ðNa Þ þ 6H2 O þ 6Hþ Z½ðK; NaÞ2 SO4
3Fe2 ðOHÞ4 SO4 ðsÞ þ 12Hþ

ð4Þ

410

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

As the alkalinity of the host rock increases, the large acidity produced by the pyrite oxidation reactions is progressively buffered, mainly by the carbonates in the rock. This results in the formation of gypsum according to reaction (5). 2Hþ þ SO2 4 þ CaCO3 ðsÞ þ 2H2 O þ Z CaSO4
2H2 OðsÞ þ HCO 3 þH

ð5Þ

Gypsum has been detected in fracture fillings and the Boa fault mineralisation, particularly in the deeper zones of the oxidised zone. The bicarbonate generated by neutralisation reaction (5) increases the alkalinity of the waters. This give arise to groundwaters such as those sampled in SM-2. In these more alkaline groundwaters, Fe(III) hydroxide gels precipitate (6) and age to produce goethite. þ FeðOHÞþ 2 þ H2 O Z FeðOHÞ3 ðsÞ þ H

ð6Þ

Acidity generated by the sulphide oxidation attacks the phyllosilicates, releasing some Al and Si, which precipitate as allophane. This is a solid phase that precipitates in acid media (pH f 4.8) (Parfitt and Kimble, 1989; Pe´rez del Villar et al., 1992) and has been identified in the oxidised zone. The presence of quite well crystallised kaolinite in the deeper parts of the oxidised zone indicates the acidity of the medium (pH c 6) where it precipitated, together with a well drained hydrological system, with a stable water table. Under these acid conditions, phosphates (apatite and monazite) inherited from the original host rock would partially dissolve, releasing, in the case of monazite, Ce(III) in solution. The released phosphate would, at least partially, be sorbed onto Fe(III)-oxyhydroxides, as may be inferred from the relatively high geochemical correlation between Fe2O3 and P2O5. In redox conditions under which both Fe(III) and Mn(IV) dominate, Ce(III) will be oxidised to Ce(IV), precipitating as CeO2. Consequently, this explains the close textural relation observed among these three oxides in the fracture fillings from the oxidised zone.

The behaviour of uranium in the oxidised zone of the system can be summarised as follows. Uranium, initially present as tetravalent in pitchblende, would be rapidly oxidised and dissolved as a result of pyrite oxidation. The predominant speciation of U(VI) under this acid and sulphate rich conditions is as UO2(SO4)22  and UO2(SO4)43  according to the general reaction (7). The released uranium would migrate away from the source or it would precipitate as uranopilite (hydrate uranyl sulphate), when the access of water would be low (drought periods).

þ 2UO2 ðsÞ þ 4SO2 4 þ 4H þ O2

Z 2UO2 ðSO4 Þ2 2 þ 2H2 O

ð7Þ

The geochemical behaviour of Fe, U, Mn and Ce under acid and oxidising conditions is summarised in the predominance diagrams of Fig. 11A, B and C. When the pH of the contacting solutions is increased, due to the increased alkalinity of the contacting rock away from the pyrite oxidation front, then uranium (VI) is hydrolysed and the dominant aqueous complex is UO2(OH)2(aq) in neutral media. The same applies for the dissolved Fe(III) which hydrolyses and precipitates under similar conditions. This would contribute to the coprecipitation of both cations under these conditions. However, the actual correlation between U and Fe2O3 is quite low. This could indicate that, in the oxidised zone, U mobilisation processes dominated retardation processes. Finally, the presence in this zone of secondary smectite and halloysite –metahalloysite from the silication of allophane (Parfitt et al., 1984; Parfitt and Wilson, 1985), and sometime autunite, Nd – La phosphates and rhodochrosite suggests the neutralisation of the environment once pyrite was totally consumed. Summarising, in the oxidised zone the conditions were dominantly acid and the water became abundant in sulphate until all the pyrite was consumed. As a consequence, uranium was mobilised, Fe(III) and Mn(IV) partially precipitated as oxyhydroxides and some significant amounts of trace metals (W, Cu, Ni, Co, Zn and Cu) were retained.

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

411

Fig. 11. Predominance diagrams in which the coexistence of Fe and U (A), Fe and Mn (B), Fe and Ce (C), and Fe and Se (D) species are shown. Dashed lines in the four diagrams represent the boundaries for the iron species dominance areas.

5.2. Geochemical processes in the transition redox zone The presence of jarosite and secondary Fe(III) oxyhydroxides indicate that this zone is affected by

oxidised, acid and sulphate containing fluids. The presence of U(VI) co-precipitated with Fe(III) oxyhydroxides could be explained by the process previously discussed and/or alternatively by the reducing effect of the remaining Fe(II) in this zone on the

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

412

U(VI) sulphate waters, according to the general reaction (8): 4Fe2þ þ 4UO2 ðSO4 Þ2 2 þ 18H2 O þ O2 Z4FeðOHÞ3  UO2 ðOHÞ2 ðsÞ þ 16Hþ þ 8SO2 4

ð8Þ

mineral phases and aqueous species that are redoxsensitive, once the pH and Eh conditions are restored. The geochemical behaviour of Fe, Mn, Ce and Se in the transitional redox zone is also summarised in the predominance diagrams of Fig. 11A, B, C and D. 5.3. Geochemical processes in the reduced zone

The existence of Fe(II) is shown by evidence of the transformation of ankeritic carbonates into siderite, according to the general process (9): Fe2þ þ SO2 4 þ ðFe; Mn; Mg; CaÞCO3 Z ðFeFMnÞCO3 þ ðCaFMgÞSO4 ðþH2 O ¼ gypsumÞ

The reduced zone is characterised by the stability of all primary minerals of the system. They are only weakly affected by the Fe(II)/Fe(III) content that has reached this depth as a function of the transmissivity of some fractures.

ð9Þ

The oxidative action of the Fe(III) coming from the upper oxidised zone is indicated by the partial oxidation of pyrite, which usually contains some adsorbed uranium, and marmatite, closely associated to Zn and Fe oxides. The presence of secondary metallic sulphides (Cu; Cu –Fe; Cu –Ag – Fe; Pb – Cu –Fe; Fe – Sn;), sulphoselenides (Cu and Fe) and elemental Se or SeO2(s), all products of the supergenic enrichment processes, indicates that the redox potential is transitional or slightly reducing. All these metallic cations, which come from the oxidation of sulphides and sulphoselenides, were transported from the oxidised zone as sulphates and selenates to the transitional zone were they precipitated, as the redox conditions became more reducing. The most relevant processes affecting the uranium primary minerals are: (i) the oxidation and partial transformation of pitchblende to zippeite, which is the result of the interaction between pitchblende and the sulphate containing waters, and (ii) the coffinitisation of pitchblende due to the larger H4SiO4 content of the waters. Finally, the scavenging of U by the slightly oxidised pyrite would indicate that U(VI) is reduced to U(IV) and subsequently precipitated, according to reaction (10). UO2 SO4ðaq:Þ þ FeS2 Z Fe2þ þ SO2 4 þ 2S þ UO2 ð10Þ Hence, this transitional redox zone results in the secondary formation and transformation of all the

6. Conclusions and implications for the performance assessment of a deep geological repository The key geochemical objectives of this study were to investigate: (i) the composition, long-term stability and the alteration/dissolution of uraninite and pitchblende as analogues to spent fuel matrix stability; (ii) the role of geochemical discontinuities (pH and Eh) in the solubility and mobilisation of natural radionuclides; and (iii) the identification of radionuclide retention processes in the ore alteration halo. 6.1. Pitchblende stability The intense downwards oxic and acid alteration in the upper zone of the system is of no relevance in assessing repository behaviour. In the transition zone, once the conditions of neutrality/alkalinity and reducing potential are restored, the following processes occur: (i) the U(VI) co-precipitation with Fe(III) oxyhydroxides, (ii) the coffinitisation of pitchblende, and (iii) the stabilisation of this later phase. The low temperature coffinitisation of pitchblende, as a result of the interaction with Si(OH)4 produced by the supergenic alteration, may be a significant process for the longterm stability of the spent fuel matrix. We cannot rule out the formation of colloidal and/or dissolved silica in the vicinity of the bentonite buffer material, in close contact with the spent fuel. The initial radiolytic oxidation of the spent fuel matrix could trigger the formation of USiO4(s) or coffinite.

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415

The study also illustrates the well-known stability of natural UO2 in the reduced zone. Overall, we could establish that there are no indications in the Mina Fe analogue of any basic process that would contradict the basic hypotheses that are used in the current models for spent fuel stability (Cera et al., 2000). 6.2. The role of geochemical discontinuities In spite of the difficulties in performing a reasonable chemical characterisation of the groundwaters of the site, the measured redox potentials indicated the possibility of redox buffering capacity controlled by the abundant organic carbon, particularly in borehole SM-2. This organic carbon, in addition to the abundant redox capacity arising from the Fe(II), Mn(II) and U(IV) containing mineral phases, restores reducing conditions to the site at some 20 m depth, regardless of fracture abundance. The origin of the TOC is still an open question. It could originate in the soil, and/or be the result of the microbiological degradation of organic matter in the rock, activated by the intense acidity of the percolating waters. The relevance of edaphic organic matter in restoring redox conditions of groundwater in fractured rocks has been shown in the REX experiment at the ¨ spo¨ hard rock laboratory in Sweden (Banwart, A 1995). Similarly, the role of the abundant organic matter around the fossil nuclear reactors at Oklo has also contributed to their stabilisation (Gauthier-Lafaye et al., 2000). It is important to establish the role of the ‘‘in situ’’ and introduced organic matter together with the residual oxygen in evaluating the post-operational redox conditions of a deep geological repository. These initial redox conditions will control the early evolution of redox condition in the repository. Further to this, the redox behaviour of relevant trace elements that may occur as radionuclides in some long-lived radioactive wastes, such as Se, Ce, Ni and obviously U, have some implications for the performance assessment of a deep geological repository. For instance, the mineralogical characterisation of the system has given evidence of the secondary formation of either elemental Se or SeO2 (selenolite) in the transition zone. This has consequences for the determination of solubility limits for this radionuclide. If we could identify this as elemental Se, which is the

413

thermodynamically favourable phase in the transition zone, this would give some additional basis to select elemental Se as the solubility-limiting phase in the waste. Determination of selenium concentrations in the waters of the transition zone, in equilibrium with this secondary phase, would be thus valuable. The existence of Ce(IV) in the oxidised zone is predicted from a thermodynamic point of view. Hence, this observation gives additional support to the basic hypothesis used when establishing the solubility limits of this element in repository conditions. The fact that Ni appears to be associated to Fe(III) and Mn(IV) oxides indicates the relevance of these associations in defining the behaviour of this radionuclide in redox discontinuities. Finally, the redox state induced by organic matter oxidation seems to control the redox speciation of uranium in the system. According to the thermodynamic calculations performed, uranium solubility is controlled by a slightly oxidised pitchblende (U4O9) and the aqueous speciation is dominated by the U(OH)4(aq). 6.3. Identification of the radionuclide retardation processes The mineralogical investigations have identified the existence of U(VI) (probably as UO2(OH)2)/Fe(III) coprecipitation phenomena in the transition zone. The analysis of the solid phases would indicate that the association of these elements goes beyond surface sorption. This observation confirms and strengthens the multiple observations, laboratory experiments and model calculation, which have been performed previously in other redox alteration halos of uranium ores. This evidence was particularly important in the ‘‘El Berrocal’’ system (Pe´rez del Villar et al., 1996; Bruno et al., 1996, 2000). Hence, we can reasonably establish that U(VI) co-precipitation with Fe(III) oxyhydroxides is a key retention process in redox transition zones of a geological repository. Further work at the site (Phase II of the project) aims: 

To determine the natural physico-chemical conditions of the stepwise oxidation of pitchblende in the site, following the initial observations regarding the formation of phases like ianthinite.

414

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415



To determine the mechanism of transformation of pitchblende to coffinite in reducing conditions by establishing the pH, Eh and SiO2 concentrations required and to verify the low temperature coffinitisation regime of pitchblende in the reducing environments of the site.  To determine the main water/rock interaction processes that controls the master hydrochemical variables of the system (pH and Eh) and consequently the relevant trace element solubility.  To establish the origin and role of organic matter in controlling the redox potential of the system, as well as the cycling of Fe and Mn and their influence in the redox conditions and trace element scavenging.

Acknowledgements Financial support for this work was provided by ENRESA. We are grateful to J. Pardillo, M.J. Turrero, M. Pelayo, B. Ruiz, P. Rivas, M.A. Labajo, J.M. Dura´n, M. Grive´, F. Marı´n A. Izquierdo, F.Ortun˜o and E. Floria´ for their collaboration in this research project. Dr. N. Chapman and another anonymous reviewer are also thanked for their useful comments and suggestions that improved the manuscript. [EO]

References Arribas, A., 1962. Mineralogı´a y metalogenia de los yacimientos espan˜oles de uranio; Las pizarras uranı´feras de la provincia de Salamanca. Estud. Geol. XVIII, 155 – 172. Arribas, A., 1975. Caracteres geolo´gicos de los yacimientos espan˜oles de uranio. Su importancia econo´mica e intere´s en el desarrollo energe´tico del paı´s. Stvd. Geol. IX, 7 – 63. Arribas, A., 1985. Origen, transporte y deposicio´n del uranio en los yacimientos en pizarras de la provincia de Salamanca. Estud. Geol. 4, 301 – 321. Banwart, S., 1995. The Aspo redox investigations in block scale. Project summary and implications for repository performance assessment. SKB Technical Report 95-26, 47 pp. (Stockholm). Blanc, P.-L., 1996. Oklo-Natural analogue for a radioactive waste repository. Acquirements of the project. E. C. Nuclear Science and Technology Series, Luxembourg. Final Report. EUR 16857/ 1 EN, 123 pp. Blomqvist, R., Ruskeeniemi, T., Kaija, J., Ahonen, L., Paananen, M., Smellie, J., Grundfelt, K., Bruno, J., Pe´rez del Villar, L., Cera, E., Rasilainen, K., Pitka¨nen, P., Suksi, J., Casanova, J.,

Read, D., Frape, S., 2000. The Palmottu natural analogue project. Phase II: Transport of Radionuclides in a Natural Flow System at Palmottu. E. C. Nuclear Science and Technology Series, Luxembourg. EUR 19611 EN, 174 pp. Both, R.A., Arribas, A., De Saint Andre´, B., 1994. The origin of breccia-hosted uranium deposits in carbonaceous metasediments of the Iberian Peninsula: U – Pb geochronology and stable isotope studies of the Fe deposit, Salamanca Province, Spain. Econ. Geol. 89, 584 – 601. Bruno, J., de Pablo, J., Duro, L., Figuerola, E., 1995. Experimental study and modelling of the U(VI) – Fe(OH)3 surface precipitation/coprecipitation equilibria. Geochim. Cosmochim. Acta 59 (20), 4113 – 4123. Bruno, J., Duro, L., Linklater, C.M., Goldberg, J.E., Gimeno, M.J., Pen˜a, J., Ayora, C., Delgado, J., Casas, I., 1996. Testing of geochemical models for trace elements mobility. El Berrocal Project: Characterisation and Validation of Natural Radionuclide Migration Processes Under Real Conditions on the Fissured Granitic Environment. European Commission contract no. FI2W/CT91/0080. Vol. II (Hydrogeochemistry), Topical Report 7. Enresa, Madrid, pp. 443 – 557. Bruno, J., Casas, I., Cera, E., Duro, L., 1997. Development and application of a model for the long-term alteration of UO2 spent nuclear fuel. Test of equilibrium and kinetic mass transfer models in the Cigar Lake ore deposit. J. Contam. Hydrol. 26, 19 – 26. Bruno, J., Duro, L., Grive´, M., 2000. Abilities and limitations of the geochemical models and tools used in the modelling of the behaviour of trace metals in natural waters. Lessons learned from the BPM exercises performed up to December 1999 (QuantiSci internal report). Campos, R., Ruiz, J., Pe´rez del Villar, L., Pardillo, J., 1999a. Caracterizacio´n de la fracturacio´n en torno a la Falla Boa, Zona D, Mina Fe, Ciudad Rodrigo (Salamanca) CIEMAT/ DIAE/54331/1/98, 66 pp. (Madrid). Campos, R., Pardillo, J., Dura´n, J.M., Pe´rez del Villar, L., 1999b. Caracterizacio´n de la fracturacio´n en los sondeos: zona D (Mina Fe, Ciudad Rodrigo, Salamanca) CIEMAT/DIAE/54331/4/99, 78 pp. (Madrid). Carretero, G., Ortun˜o, F., Florı´a, E., Suso, J., 2000. Testificacio´n hidra´ulica. Funcionamiento hidra´ulico de la Falla Boa. Informes AITEMIM, 16 pp. (Toledo). Casas, I., Gime´nez, J., Martı´, V., Torrero, M.E., de Pablo, J., 1994a. Kinetic studies of unirradiated UO2 dissolution under oxidising conditions in batch and flow experiments. Radiochim. Acta 66/ 67, 23 – 27. Casas I., Bruno J., Cera E., Finch R.J., Ewing, R.C., 1994b. Kinetic and thermodynamic studies of uranium minerals. Assessment of the long-term evolution of spent nuclear fuel. SKB Technical Report 94-16, 73 pp. (Stockholm). Cera, E., Merino, J., Bruno, J., 2000. Liberacio´n de los radionucleidos e iso´topos estables contenidos en la matriz del combustible. Modelo conceptual y modelo matema´tico del comportamiento del residuo. ENRESA. Publicacio´n Te´cnica 03/00, 110 pp. (Madrid). Chapman, N.A., McKinley, I.G., Shea, M.E., Smellie, J.A.T., 1990. The Pocßos de Caldas project: Summary and implications for radioactive waste management. SKB Technical Report 90-24, 147 pp. (Stockholm).

L. Pe´rez del Villar et al. / Chemical Geology 190 (2002) 395–415 Garrels, R.M., Christ, C., 1965. In: Croneis, C. (Ed.), Solutions, Minerals, and Equilibria. Harper’s Geoscience Series. Harper and Row Publishers, New York, 450 pp. Gauthier-Lafaye, F., Ladoux, E., Smellie, J., Louvat, D., Michaud, V., Pe´rez del Villar, L., Oversby, V., Bruno, J., 2000. OKLOnatural analogue Phase II. Behaviour of nuclear reaction products in a natural environment E. C. Nuclear Science and Technology Series, Luxembourg. EUR 19139 EN, 116 pp. Go´mez, P., Garralo´n, A., Turrero, Ma.J., Sa´nchez, L., Ferna´ndez, F., Ruiz, B., 2000. Caracterizacio´n Hidrogeoquı´mica de la Falla Boa (Mina Fe´, Salamanca). CIEMAT/DIAE/54331/2/2000, 46 pp. (Madrid). Limpo, J.L., Luis, A., Siguin, D., Hena´ndez, A., 1976. Cine´tica y mecanismo de la precipitacio´n del hierro como jarosita. Rev. Metal. CENIM 12 (3), 123 – 134. Mangas, J., Arribas, A., 1984. Caracterı´sticas fisico-quı´micas de los fluidos asociados con las mineralizaciones de uranio de mina Fe (Salamanca). 7th Congreso Internacional de Minerı´a y Metalurgia, Barcelona. Proccedings, vol. 1, pp. 435 – 451. Martı´n-Izard, A., 1989. Los yacimientos de uranio en pizarras: PhD Thesis. Serie Ciencias 4, Diputacio´n de Salamanca, 238 pp. Murphy, W.M., Pearcy, E.C., 1992. Source-term constraints for the proposed repository at Yucca Mountain, Nevada, derived from the natural analogue at Pen˜a Blanca, Mexico. Mater. Res. Soc. Symp. Proc. 257, 521 – 527. Parfitt, R.L., Saigusa, M., Eden, D.N., 1984. Soil development processes in an Aqualf – Ochrept sequence from loess with admixtures of tephra, New Zealand. J. Soil Sci. 35, 625 – 640. Parfitt, R.L., Wilson, A.D., 1985. Estimation of allophane and halloysite in three sequences of volcanic soils, New Zealand. Catena (Suppl. 7), 1 – 8. Parfitt, R.L., Kimble, J.M., 1989. Conditions for formation of allophane in soils. Soil Sci. Soc. Am. J. 53, 971 – 977. Pe´rez del Villar, L., 1998. Modelo Conceptual de funcionamiento geoquı´mico de la mineralizacio´n uranı´fera de la Falla Boa, Mina Fe´ (Ciudad Rodrigo, Salamanca). CIEMAT/DIAE/54331/6/98, 11 pp. (Madrid). Pe´rez del Villar, L., Gordillo, J., Fenoll, P., 1979. Natrojarosita en la base de la formacio´n Utrillas en Nuevalos (Provincia de Zaragoza). Clay Miner. 14, 139 – 153. Pe´rez del Villar, L., Moro, M.C., Cembranos, M.L., 1992. Allophane in weathered zones of barite ore deposits (Vide de Alba y San Blas, Zamora, Spain): mineralogy and genesis. Clay Miner. 27, 309 – 323. Pe´rez del Villar, L., Pelayo, M., Co´zar, J.S., de La Cruz, B., Pardillo, J., Rivas, P., Reyes, E., Delgado, A., Nun˜ez, R., Crespo, M.T.,

415

Jime´nez, A., Quejido, A., Sa´nchez, M., 1996. Mineralogical a geochemical characterisation of the El Berrocal fracture fillings in relation to the migration/retention of natural radionuclides. El Berrocal Project: Characterisation and validation of natural radionuclide migration processes under real conditions on the fissured granitic environment. European Commission contract no. FI2W/CT91/0080. Vol. 1 (Geological Studies), Topical Report 3. ENRESA, Madrid, pp. 490 – 595. Pe´rez del Villar, L., Co´zar, J.S., Pardillo, J., Labajo, M.A., 1999. La mineralizacio´n de uranio encajada en la Falla Boa (Mina Fe, Ciudad Rodrigo): mineralogı´a y geoquı´mica de las muestras de los sondeos SM-1, SM-2, SM-3 y SM-4, en relacio´n con los procesos meteo´ricos de oxidacio´n. CIEMAT/DIAE/54351/9/99, 31 pp. (Madrid). Pe´rez del Villar, L., Co´zar, J.S., Pardillo, J., Buil, B., Pelayo, M., Labajo, M.A., 2000. Mineralogı´a y geoquı´mica de los rellenos fisurales del Proyecto Matrix: implicaciones para los procesos de disolucio´n, transporte y retencio´n de los radionucleidos naturales y otros elementos ana´logos (mina Fe, Ciudad Rodrigo). CIEMAT/DIAE/54331/1/00, 58 pp. (Madrid). Pe´rez del Villar, L., Campos, R., Go´mez, P., Co´zar, J.S., Pardillo, J., Garralo´n, A., Turrero, M.J., Buil, B., Pelayo, M., Ruiz, B., Rivas, P., Arcos, D., Bruno, J., Grive´, M., Ruiz Sa´nchez-Porro, J., Marı´n, F., Izquierdo, A., Carretero, G., Ortun˜o, F., Floria´, E., Suso, J., 2001. La pechblenda de mina Fe (Ciudad Rodrigo, Salamanca), como ana´logo natural del comportamiento del combustible gastado (Proyecto Matrix I) (compiled by: L. Pe´rez del Villar, J. Bruno and J. Ruiz) ENRESA, Madrid Publicacio´n Te´cnica. 05/2001, 175 pp. Plummer, L.N., Prestemon, E.C., Parkhurst, D.L., 1991. An interactive code (NETPATH) for modelling NET geochemical reactions along a flow PATH. U.S. Geol. Surv., Water-Resour. Invest. Rep. 91-4078, 130 pp. Rivas, P., Herna´n, P., Bruno, J., Carrera, J., Go´mez, P., Guimera´, J., Marı´n, C., Pe´rez del Villar, L., 1997. El Berrocal Project: Characterization and Validation of Natural Radionuclide Migration Processes Under Real Conditions on the Fissured Granitic Environment E. C. Nuclear Science and Technology Series, Luxembourg. EUR 17478 EN, 522 pp. Shoesmith, D.W., Sunder, S., 1992. The prediction of nuclear fuel (UO2) dissolution rates under waste disposal conditions. J. Nucl. Mater. 190, 20. Wolery, T.J., Daveler, S.A., 1992. EQ3/6, a computer program for reaction path modelling of aqueous geochemical systems: Theoretical manual, User’s guide and related documentation (v7.0). LLNL Report UCRL-MA-110662 PT II, 337 pp.