THE CLAYEY FISSURAL FILLINGS ASSOCIATED WITH NIOO ° · IIooE FRACTURES AT THE EL BERROCAL URANIUM MINE ( SIERRA DE GREDOS , SPAIN ) : CHARACTERIZATION , GENESIS AND RETENTION CAPACITY OF RADIOACTIVE AND OTHER ELEMENTS

The clayey fissural fillings « 60!-lm and < 2!-lm fractions) associated with the NlOoo-1100E fractures, at the El Berrocal V mine, have been studied in relation to the natural radionuclide migration/retention processes in a fissured granitic environment. The fracture filled with pyrite, chalcopyrite, sphalerite, galena-bearing quartz, later mineralized by pyrite, pitchblende, carbonate and barite also belongs to that fracture seto According to the data obtained by X-ray diffraction, thermal and thermogravimetric analyses, EDX coupled to both scanning and transmission electron microscopes and cation exchange methods, the clayey fissural fillings are essentially composed of quartz, sericite, illite, Ca/Mg beidellite, K-feldspar, albite, apatite and occasional kaolinite. Traces of monazite, torbernite, autunite, other unidentified Ca, Y, Ce, La and REE phosphates, ferric sulfates, jarosite, gypsum, barite, calcite, Pb silicates, Fe-oxyhydroxides, gibbsite, Al gels and REE silicate complex gels have been also detected in the < 2 !-lm fraction. From a geochemical point ofview, the < 2 !-lm fractions, in relation to the < 60 !-lm fractions, are very positively anomalous in V, Th, Y, Cu, Zn, Sn, Ni, As, Ba, Ca and organic C. Furthermore, Cl is present, in important amounts, in almost all of the essential and accessory minerals and mineraloids in the < 2 !-lm fractions. These data suggest that the clayey materials analyzed were mainly produced by hydrothermal alteration of the granitic gouges during the NlOoo-llooE fracturation phase. Based on the b 0 data obtained from the quartz vein and the < 2 !-lm fractions, the argillitization process was probably produced by two phases. The first, of sericitic-illitic nature, was caused by the interaction between acid, KCI-rich hydrothermal solutions and the granitic fault gouges, at a temperature range between 70-120° C. The second, of smectitic nature, was probbly produced during ancient and/or present weathering processes by strong K leaching of the pre-existing sericite-illite, at room temperature. During the weathering phases, illite and probbly smectite were transformed to kaolinite in those fractures with sulfide mineralizations. Jarosite, ferric sulfates, gypsum, torbernite, autunite, gibbsite and Al gels were also probbly formed. The new formed accessory minerals and mineraloids play an important role, either by adsorption or precipitation, in the retention of the radioactive elements, V and Th, and other analogous elements, Y, Ce, La, REE, released during the hydrothermal alteration of the granitic fault gouges and weathering of the uraniferous mineralization. Th is mainly heId in inherited monazite. The V sequential leaching indicates that Fe oxyhydroxides and carbonates are important mineral phases for V retention. On the contrary, the clay minerals do not seem to play an important role in the V retention by adsorption, but they do work as a physico-chemical barrier for the VOz++ phosphate precipitation. The high organic C content in the clay fractions could be due to present biological activity in the clayey fissural filling or to organic acids transported from the topographic surface by the percolating water. In this last case, organie-clay complexes would probbly be formed.

El alto contenido en C orgánico de la fracción arcilla de las muestras puede explicarse por la actividad bacteriana actual observada en los rellenos fisurales, o por la presencia de

Introduction
The El Berrocal pluton is located at the central part of the Centro-Iberian Zone (Julivert el al., 1972), near the contact between the Tajo River Tertiary basin and the Sierra de Gredos (fig. la).The main granitic facies of this pluton, the so called El Berrocal facies, is the host rock of an U mineralization, mined in the 1960's and named El Berrocal U mine (fig. lb).
The ciayey materials associated with the NlOoo-1100E and N155°E fractures in this U mine have been studied in relation to the project «Characterization and validation of natural radionuclide migration processes under real conditions in a fissured granitic environment.El Berrocal experimental site».
The fracture filled with pyrite, chalcopyrite, sphalerite, galena-bearing quartz, later mineralized by pyrite, pitchblende, carbonate and barite (Arribas, 1965;Pérez del Villar and Pardillo, 1992) belongs to the NlOoo-110oE fracture set, and it is also flanked by clayey walls (fig.1c).Not all the fractures sampled con- :;::':':: ...    where: 1 is the reflection intensity used to quantify each mineral and RF is the reflecting factor of each mineral for the chosen reflection. 1, ... , n are the identified and quantified minerals in the sample.The sampling has been carried out in the U mine gallery, between 43.7 and 68 m from the entrance.The situation and characteristics of the 13 samples studied are shown in figure 2, table 1.The most important fractures where the c1ayey samples were obtained are drawn in figure 3 and the flowchart used to separate the < 60 11m and < 2 11m fractions is represented in figure 4.

N-1OO0/62°N
The granulometric control of the < 2 11m fraction has been carried out by photon spectroscopy correlation (PSC), using a Malvero 4700 system, equipped with a 3W argon ion laser.The experimental conditions were a laser angle at 90°and a laser power of 500 mW.
The mineralogical characterization of the < 60 11m and < 2 11m fractions has been carried out by X-ray diffraction (XRD), como bining the powder and oriented aggregates methods.The former has been applied to both fractions, using a rotating sample holder.The latter, with ethylene glycol and sulfoside dimethyl and thermic treatments, has been only applied to the < 2 11m fraction.
According to Pérez del Villar el al. (1990) the El Berrocal facies is affected by a deuteric and/or latemagmatic alteration processes, represented by a pervasive chloritization of biotite, ilmenite-anatase transformation, muscovitization, incipient albite sericitization and interstitial albitization.When this facies is affected by the N100o-110oE fractures, a strong tectonization and a hydrothermal alteration is observed, represented by a general sericitization which releases free quartz.After the fissures were fi11ed by this type of quartz, Mg-Fe chlorites, carbonates and minor and disseminated pitchblende were formed.
In the V mine ga11ery, the weathering processes affecting the hydrotherma11y-altered granite are represented by a general oxidation, dissolution of almost a11 the carbonates and formation of V0 2 ++ minerals, mainly autunite and torbernite.
These hydrothermal and supergenic mineralogical transformations produce an important loss of Na20, FeO and Zn, and a remarkable enrichment in K 2 0, Fe203, MgO, H 2 0+, Ba, Rb, Li, V and, less important, in Ce, Sr, Y, La and Th.
The aims of this work are: a) to determine whether the clayey materials associated with NlOoo-llooE fractures are fissure fillings or the hydrothermal alteration products of the granite; b) to define the later alteration processes, either hypogenic or supergenic, which could modify the original nature of the clay minerals; and c) to establish the behaviour of the clayey materials in relation to the mobilization/retention processes of natural analogue radionuclides (V and Th) and other heavy elements, analogous (La, Ce, Ti and V) or not, in the sense given by Chapman el al. (1984 and 1986).

72
:>  Due to the mineralogical composition of the c1ay samples, the following supplementary criteria have been introduced: i) the quartz reflection intensity at 3.34Á has been obtained by subtrac-ti~g the muscovite or i1iite reflection intensity at '" lOÁ, since 10 I'1,u '" 3.33ÁIMu (Bradley and Grim, 1961); ii) in sample DAR-3, the Yo of apatite has been calculated by stoichiometry, assuming that the ratio P 2 0 5 /CaO = 0.73 in the apatite of the granite (Pérez del Villar, el al., 1990).The first correction has only been applied to the non-oriented powder patterns on both fraetions and, only in sample DAR-3, equation (1) has been referred to 100% apatite.
For the identification of the c1ay and aceessory minerals and mineraloids in the < 2 11m fraction, a Zeiss scanning and a Jeol transmission e1ectron microscopes (SEM and TEM) eoupled to an energy dispersive X-ray analysis system (EDX) have been used.Electron images have also been taken.
The thermal eharacteristics of the < 2 11m fraetion have been determined using a Perkin-Elmer 1700 DTA equipment, at a sean rate of 100 Clmin, under air dynamic atmosphere (40 ce/min), at temperatures between 400 and '" 980 0 C.
Labile Si02 has been dissolved aeeording to Ross and Hendrick's (1945) method and determined by speetrophotometry.Labile AI20 3 and Fe203 were dissolved according to Quejido's et al. (1988) method and determined by ICP speetroscopy.
The eation exchange capacity (CEe) of the < 2 11m fraction has been determined by the NH 4 N0 3 method, aeeording to Boh et al. (1978).The U sequentialleaching has been earried out according to Airey's et al. (1987) method.
The Ó 18 0 in c1ays and quartz has been determined using a Finnigan-MAT-251 mass spectrometer, following the Borthwick and Harmon's (1982) chemical method, modified by Vennemann and Smith (1990).

Results
Characterization of the < 60 f.IJ1l fraction Mineralogical composition.The semiquantitative mineralogical composition of the < 60 ¡.tm fraction (table 4) shows that the total phyIlosilicates vary between 43 and 74 %, the principal components being  sericite and illite and minerals belonging to the smectite subgroup.Traces of kaolinite are observed in only three samples (DAR-2, LAR-2 and LAR-3).The minerals associated with the phyllosilicates, in order of abundance, are: quartz, K-feldspar, albite and apatite.Chemical composition.The chemical composition of the < 60 f!m fraction has been subdivided into major (table 5a) and trace elements (table 5b).
a) Majar elements.From all these elements, the high P 2 0 S is remarkable in all the samples, particulady in sample DAR-3, which is also very anomalous in CaGo The highest total S02 value is observed in sample DAR-1, and the variations observed in the rest of the major elements are due, fundamentally, to the 205 phyllosilicate contents of the samples.The high K 2 0 values, in almost all the samples, indicate that sericite, illite and K-feldspar are the predominant minerals.
The H 2 0-has been determined at 220 0 C, because smectites loose their moisture in two different steps, within a temperature range from 80 0 C to 170 0 C. b) Trace elements.If the values in table 5b are compared with the average values corresponding to the hydrothermally-altered and fresh granite from the mine gallery (Pérez del Villar et al., 1990), the following remarks can be made: -AH the samples are very positively anomalous in Ba, compared to the altered and fresh granite, whose average concentration are 47 and 29 ppm, respectively.The samples located in the U-mineralized vein are exceptionally high in Ba, due to the existence of barite associated with the U mineralization.Furthermore, sample DAR-1, which is the most anomalous in Ba, is also anomalous in total S02.
-Similady, the Sr content of the clayey samples is higher than in the altered and fresh granite, whose average concentrations are 9 and 5 ppm, respectively.
-The Ce, La and Y contents of the samples are also remarkably higher than those in the altered and fresh granites, whose average concentrations, in ppm, are: 27 (Ce), 7 (La) and 7 (Y); and 22 (Ce), 6 (La) and 6 (Y), respectively.All of these elements are mainly held in monazite and xenotime in the granite.Sn follows the same trend, with contents higher than in both the altered (17 ppm) and fresh granite (17 ppm), in which Sn occurs, fundamentally, as cassiterite.
-The Zn values are also higher than in the altered and fresh granite (121 and 89 ppm, respectively).This element is found in the structures of the micas and, fundamentally, as sphalerite in the granite.The clayey samples from the mineralized quartz vein and adjacent fractures (LAR-1, LAR-2 and LAR-3) are  positively anomalous in Zn and Cu.It should be rein U and Th in relation to the altered (22 and 9 ppm, minded that the quartz vein is also mineralized by respectively) and fresh (17 and 8 ppm, respectively) pyrite, chalcopyrite and sphalerite.granite.In the former, U occurs mainly as U0 2 ++ -The As is, in general, below the detection limit phosphates, and in the fresh granite U is fundamenof the ICP spectroscopy, except in some samples tally present as uraninite and Th as thorite and held from the quartz vein and adjacent fractures, mainly in monazite.DAR-1, LAR-2 and LAR-3.
-The highest U contents are found in samples 10- -The < 60 !Lm fractions are, in general, enriched cated in the U mineralized quartz vein (DAR-3 and DAR-4), in adjacent fractures (LAR-3) and in farther ones (LAR-5 and LAR-8).The two last samples were obtained from fractures that intersect the hydrothermally-altered granite, which is also enriched in U.
silicates and K-AI-Si-Cl complexes are gels trapped by the cIay minerals.
Chlorine is present in almost aH the minerals and mineraloids, sometimes in important concentrations.This anion could be a tracer of the hydrothermal so- Characterization of the < 2 ¡.tm fraction Size Control.The granulometric distribution has been obtained for only three samples (fig.5), which present unimodal distribution plots, with average .particIesizes of :::::: 1,0.78 and 0.43 ¡.tm.
Transmission and Scanning Electron Microscopy.
The chemical analyses obtained by TEM + EDX and the structural formula of the cIay minerals (ta-0.4bIes 7a, b, c, d) are given.The data in table 7a correspond to dioctahedral cIay micas, sericite, 0.3 hydromuscovite and illite, and the most numerous 0.2 chemical data of the cIays (table 7b) correspond to 0.1 randomly-interstratified illite-smectite.However, the absence of any evidence of interstratification by XRD suggests that these last data correspond to physical mixtures of illite and smectite, and the calculation of their structural formula is, therefore, pointless.The data corresponding to the aluminian dioctahedral smectite, specificaHy beidellite, and the chemical composition of kaolinite (tabla 7c and d) 0.7 complete the analytical results of the cIay minerals.
Similarly, the semiquantitative chemical composition of the accessory minerals and mineraloids associated with the cIay minerals has been obtained by EDX + SEM and the data are given in table 8.The Fe compounds and minerals (Fe oxyhydroxides, Fe 3 + sulfates and jarosite) occur either as coatings on the cIay particIes or as single grains.Something similar occurs with the Al oxyhydroxide gels.Gibbsite, calcite, gypsum, barite, apatite, monazite, torbernite and complex aluminosilicates generaHy appear as single euhedral to anhedral grains, sorne of them reaching up to 20 ¡.tm.According to the electron images of these minerals and mineraloids (fig.6), the REE  Uranium has been identified in sample DAR•3 as a 20llm euhedral torbernite crystal; Th has been detected in monazite, and both elements in an unidentified Th-U-REE-Si-AI-P mineral.

Chemical Composition:
a) Major efements.The analytical results of the major elements, including free SiOz, Al z 0 3 and Fez03 (table 9) show that: -As in the < 60 Ilm fraction, high PzOs contents are also present in the clay fractions, principally in those corresponding to the samples from the mineralized dyke.Among them, sample DAR-3, which is rich in apatite, is very anomalous both in PzOs and CaO.
The variations observed in the KzO contents are related to the (illite + K-feldspar)/smectite ratio.
-All the samples show high free SiOz and Fez03 contents.The former is associated with Al gels, and the latter occurs as Fe oxyhydroxides, as confirmed by EDX + SEM analyses.The free Alz0 3 values are lower than expected, according to the data obtained by the abovementioned techniques.However, this can be explained if most of the free Al oxyhydroxides occured as gibbsite, which is insoluble by the partial dissolution method used.
-The high organic COz content is remarkable in all the samples, particularly in samples LAR-l, DAR-2, LAR-2 and LAR-S.b) Trace elements.As for the major elements, the trace elements in the < 2 Ilm fraction (   The first four oxides are fundamentally related to the ciay mineral enrichment.The enrichment in AIz0 3 and FeZ03 is also due to the oxyhydroxide concentration in the < 2 ¡.tm fraction, and the enrichment in CaO and mineral COz is related to the calcite enrichment in this fraction.The enrichment observed in the trace elements will be discussed in another section. Cation Exchange Capacity.The CEC of the < 2 ¡.tm fractions, expressed in mEq/100 gr (tabie 13), shows that: i) the smectite-rich samples have the highest CEC.However, a good correlation between the amount of smectite and these values is not observed.This may be due to the semiquantitative mineralogical composition of the < 2 ¡.tm fractions, and ii) except in sample LAR-1, the most important been determined in aH the < 2 ¡,tm fractions and in six quartz samples from the U mineralized vein (table 15).Given that the < 2 ¡,tm fractions are mixtures of sericite-illite, smectite, kaolinite, quartz, feldspars and apatite, a first approach to calculate the theoretical o 18 0 value for illite has been done correlating the illite/smectite ratio with the experimental O 18 0 (fig.7).Thus, a theoretical O 18 0 value = + 17.45 %0 for pure smectite has been estimated.This value is in agreement with the theoretical O 18 0 value of a smectite in e~uilibrium with the El Berrocal meteoric water (o 1 O = -8 %0) at a T = 20°C (Reyes et al., in prep. a).
From the theoretical o 18 0 value for smectite and the measured O 18 0 for quartz, the theoretical O 18 0 for pure illite has been determined in samples DAR-1, DAR-2 and LAR-1, in which quartz is the only accessory mineral (table 16).
: lnterlayer charge.% : Percentage oC tetrahedric charge.exchange cations in aH the samples are Ca and Mg, and Na is always dominant over K.
U Sequential Leaching.The data of the U sequential leaching, classified by the reagents used and by the mineral and mineraloid phases with which the U is associated are given in table 14.The percentage of readily leachable U in the samples varies from 5 to approximately 35 % of the total U. GeneraHy, this U is fundamentaHy associated with the amorphous and crystaHine Fe oxyhydroxides, and the carbonate phase also plays an important role in the retention of this radioelement.Furthermore, the extremely low amount of U « 0,1 ppm) associated with the clay fraction as exchange cation should be pointed out, except in sample LAR-5 (2.9 ppm).

Discussion
The mineralogical, geochemical and isotopic characteristics of the clayey materials studied suggest that these materials were formed by the interaction between the granitic fault gouges and the hydrothermal solutions originated during the tectonization process, foHowed by weathering.The principal clay minerals found in the samples (sericite-illite) and the presence of Cl-in aH the minerals analyzed, including the clays and sorne mineraloids, suggest that the hydrothermal solutions had an alkali chloride nature, with an undetermined Na/K ratio.However, this ratio was possibly controHed by the Ab/KFd = 1.7 illite-water fractionation, iBite is formed at a T range between 70-120°C (fig.8a).Similarly, according to Clayton's el al. (1972) equation for quartz-water fractionation, the quartz vein is formed at a T range between 80-120°C (fig.8b).Consequently, and as a preliminary conclusion, sericite-illite and the quartz vein are formed within the same epithermal alteration process, as had been mentioned by Pérez del Villar el al. (1993).
The presence of significant amounts of kaolinite is limited to sample DAR-2, while only traces of this mineral have been detected in those fractures adjacent to the mineralized quartz vein.This can be explained considering that kaolinite is also a weathering clay mineral, formed from illite and/or beidellite, under strongly leaching acid conditions, according to the following reactions:  In fact, this process has started when the quartz vein was under the oxidizing action of the meteoric water, which produced, and still produces, sulfide, pitchblende and carbonate oxidation and dissolution, acidification of the resulting solutions, and the degradation of the clay minerals to form kaolinite, gibbsite and Al gels.Jarosite has also been detected, indicating that the upper pH limit of the solutions are = 4.This last weathering process is in agreement with the present pH values observed in the percolating waters of the El Berrocal mine gallery, ranging from 6.8-5.8 along the gallery and -3 in the mineralized vein zone (Turrero and Gómez, 1990).
Concerning the geochemistry of the clay fraction, in relation to its retention capacity, several trace element groups have been established for discussion: a) U and Th.The V of the clay fraction has two sources: i) directly from the uraninite of the granite and, above all, ií) from the epithermal V mineralization, mainly hosted in the principal quartz vein.The supergenic oxidation and leaching of this mineralization release VOz ++ to the solution, which precipitates mainly as VOz phosphate minerals, autunite and torbernite.

Samples
Ií '80eVoSMOW) According to the sequential leaching data of the < 2 f.lm fraction, practically no V is adsorbed in the clay minerals, except in one sample.The readily leachable V, approximately 25 % as an average value, is associated with carbonates and amorphous and crystalline Fe oxyhydroxides.Consequently, the rest of the V (= 75 %) is held in the accessory, inherited and refractory minerals, such as apatite and monazite, in other unidentified secondary Th-V-REE minerals (see Table 11), and mainly in secondary V minerals insoluble under the sequentialleaching chemical procedure.
Thorite-auerlite and monazite from the granite are the only possible sources of the Th detected in the < 60 f.lm and < 2 f.lm fractions.The high Th content in the first fraction is due to an artificial concentration of the Th minerals during sieving or to a preexisting natural concentration in the clayey materials due to the resistant nature of these minerals under the hydrothermal argillitization process.
The fact that the < 2 f.lm fraction is enriched in Th relative to the < 60 f.lill fraction suggests that, during the argillitization processes of the granitic fault gouge, a partial destabilization of the Thminerals took place.Consequently, the Th thus released underwent a slight remobilization and precipitated later on as unidentified Th-V-REE complex minerals (see table 11).However, Th is mainly heId in fine-grained monazite, which is inherited from the granite.b) Zn, Cu, Ni and Sn.Zinc would be supplied to the < 2 f.lm fraction, in part, by the destabilization of sphalerite from the granite and, principally, by the sulfides present in the mineralized quartz vein.The second source accounts for the highly anomalous Zn content of the clayey samples from the quartz vein and adjacent fractures.The anomalies of Cu and Ni are explained in the same way.
Two sources account for the Sn in the < 2 f.lm fraction: i) cassiterite from the granite and ii) sulfides in the mineralized vein, which generally contain Sn 2 + substituting the major metallic elements.Since Sn is easily mobilized as Sn 2 +, the second source is the most important.Thus, the sulfide destabilization by weathering releases Sn 2 + which, upon oxidation to Sn 4 + , causes its fixation in the clayey materials.This explains why the main Sn anomalies have been found in the clays from the mineralized quartz vein.
According to the chemical composition of the accessory minerals and mineraloids associated with the clay minerals, Zn and, fundamentally, Cu are present in almost all these minerals.Thus, Cu is present in primary minerals inherited from the granite, like apatite, in minerals associated with the pitchblende mineralization, like barite, and in secondary VOz++ minerals, like torbernite.Nickel seems to be controlled by Fe oxyhydroxides.Though Sn has not been detected in any of the minerals and mineraloids of the clay, this element can be associated with Al oxyhydroxides as Sn(OH)4 and/or free fine-grained cassiterite, inherited from the granite.c) V. Vanadium has been detected in the fresh and altered granite with contents varying between 5 and 6 ppm, mainly concentrated in chloritized biotite and muscovite (Pérez del Villar et al., 1990).Consequently, after the degradation of those phyllosilicates, Vis concentrated in the beidellite and illite lattices, substituting octahedral Al.Although V has not been detected in the Fe oxyhydroxides, these and Mn oxyhydroxides are usually accompanied by V. d) Ba.Given that the «El Berrocal» granite is highly evolved, its Ba content is low, though a remarkable increase is observed in the hydrothermally altered granite (Pérez del Villar et al., 1990).The high Ba concentration detected in the < 2 Jlm fraction must, therefore, come from the alteration of the granite and, more specifically, from the destabilization of the K-feldspar.This element has been transported as BaCl z and retained in the clays, either as cryptocrystalline barite or as an exchange cation.It should be noted that the U mineralization in the quartz vein is rich in barite, which explains the extremely high Ba contents of the samples from this zone.Barite has been detected in sample DAR-2 as an accessory mineral and coating the clay particles.
The possitility that sorne Ba could be found as an exchange cation in beidellite cannot be disregarded, given the high affinity between this mineral and Ba.The fact that this element has not been found in the CEC can be due to the use of NH 4 N0 3 as exchange agent and to the stronger adsorption of Ba by beidellite than by NH 4 +, Ca++ and Mg++.The exchange NH 4 + for Ba++ could not, therefore, have taken place.e) As.The As concentration in the fresh and altered granite is < 25 ppm (Pérez del Villar et al., 1990).In the < 60 Jlm fraction and, particularly, in the < 2 Jlm fraction, the As contents are very high, specially in those samples from the mineralized quartz vein and the adjacent fractures to the North of the vein.The possible explanation for the As anomalies is that it also comes from the destabilization of the sulfides from the principal quartz vein, and is retained as As0 4 -3 by the Fe oxyhydroxides and other active surfaces, like the clay minerals (Sillen and Martel, 1964).This element has not been detected, however, in any of the minerals and mineraloids of the < 2 Jlm fraction.
It should be noted that, among the As anomalous samples, those from the sulfide mineralized vein (DAR-1, 2, 3 and 4) are the least anomalous.The clue to this fact can be found in the pH = 3 of the present percolating water.Thus, at these low pH values, As can form H z As0 4 -, which is more soluble in water and, therefore, could have been leached (Hem, 1970).f) Ca, Sr, Ce, Y, P and mineral C. Calcium and mineral C are the only elements that show a remarkable enrichment in all the < 2 Jlm fractions compared with the < 60 Jlm fractions.This fact is explained as follows: i) Ca is the main exchangeable cation of the < 2 Jlm fractions and ii) the presence of cryptocrystalline calcite in the same fraction.On the other hand, sorne of the Ca-enriched samples are also enriched in PzOs.Consequently, a large part of the Ca is accounted for by the constant presence of apatite, sometimes in several percentages.
The fact that calcite and apatite occur as stable phases, despite the more or less acid pH of the percolating waters, suggests that those minerals must be protected by the clay minerals themselves.
The < 2 Jlm fractions enriched in PzOs are also generally enriched in Sr, Ce and Y, elements which usually replace Ca in the apatite, as is the case of this mineral in the fresh granite (Pérez del Villar et al., 1990).At the same time, monazite and other undetermined Ca, Y, Ce, La and REE phosphates, similar to xenotime, have also been detected in the < 2 Jlm fractions.
The existence of Ca, Y, Ce, La and REE complex silicate gels, trapped by clay minerals, suggests that, during the illitic process, monazite was partially destabilized, leading to the mobilization of these elements, which precipitated as silicates, given the high [H 4 Si0 4 l in the environment.However, it must be noted that part of the Y and lanthanides could be leached from the mineralized quartz vein, since these elements are particularly mobile under acid conditions, under which kaolinite is, in turn, stable (Burkov and Podporina, 1967).g) Organic C. The organic C enrichment observed in all the clay fractions is due to any or both of the following reasons: i) existence of present algal and/or bacterial activity in the clayey zones, and ii) the possible existence of clay-organic matter complexes.In this last case, the humic and/or fulvic acids, formed at the topographic surface are transported by percolating waters and retained by the clays.

Concluding remarks
1.The clayey materials associated with Nlloo-10ü°E fractures in the «El Berrocal» mine were formed by the interaction between the granitic fault gouges and hydrothermal solutions originated during the fracturation phase.They should not, the-refore, be considered as fracture fillings «sensu stricto».
2. The argillitization process of the granite was probably produced by two or three phases: the first, of illitic nature, was originated by the interaction between acid KCI-rich hydrothermal solutions and the granitic fault gouges, at a T ranging from 120-70°C.The second phase, of smectitic nature, probably took place during ancient and/or present weathering processes.In this last phase, the interaction between the pre-existing illitic material and more alkaline solutions took place, causing a partial transformation of illite to beidellite.During these weathering phases, the illite and/or beidellite were transformed to kaolinite in those fractures where a strong K+ leaching was produced and enough sulfide content remained to acidify the percolating water, after the sulfide oxidation.Furthermore, in this phase, carbonates, pitchblende and apatite were totally or partially oxidized and dissolved.Consequently, Al gels, gibbsite, Fe oxyhydroxides, jarosite, Fe 3 + sulfates and VOz +-t phosphates, autunite and torbernite, were formed.The presence of residual carbonates and apatite is possible due to their protection by the clay minerals themselves.
3. The accessory minerals and mineraloids, and in far less proportion the clay minerals, play a leading role in the retention of the heavy elements released during the alteration of the granite, sulfides and uraniferous mineralizations.
In relation to the natural radionuclides, V is either precipitated as VOz++ phosphates, mainly autunite and torbernite, adsorbed by Fe oxyhydroxides or retained by calcite and Th-V-REE unidentified mineraloids.Monazite and apatite, which are inherited from the granite, also contain V. Thorium is principally present in the minerals inherited from the granite, mainly monazite, and retained by Th-V-REE unidentified and secondary mineraloids.
Among the non-radioactive heavy elements, V would be fundamentally localized in the clay mineral lattices.Ytrium, Ce and La are present in the inherited minerals of the granite and forming Th-V-REE unidentified minerals and mineraloids.Strontium substitutes Ca in apatite, and Ba is found as crytocrystalline barite and, probably, as an exchange cation in beidellite.Copper and Zn occur in almost all the secondary minerals and mineraloids, whereas Ni seems to be controlled by the Fe oxyhydroxides.Arsenic is possibly retained by Fe oxyhydroxides and clay minerals.Tin is associated with Al gels and/or as cassiterite inherited from the granite.
4. Concerning the light elements, organic C could be due to present biological activity in the clayey materials or to organic acids transported from the topo-graphic surface.In this last case, it is probable that organic-clay complexes exist alongthe fractures.

Fig
Fig. l.-Geological situation and schematic sections oí the El Berrocal uranium mine.
Fig.2.-Fracture scheme of the west wall of the El Berrocal uranium mine, afterCapote (1991), and location of the c1ayey samples.

Fig. 4 .
Fig. 4.-Flowchart corresponding to the separation and studies carried out in the < 60 and < 2 I-trn fractions.

<
60 ¡.tm fraction are also detected in the ciay fraction, though generally enhanced.The ratio between the chemical composition of the < 2 and < 60 ¡.tm fractions has been calculated to obtain the chemical enrichment or impoverishment factor reached during the separation of the ciay fraction from the < 60 ¡.tm fraction (tables 11 and 12).In general, this factor will be useful for a better interpretation of the retention capacity of the ciays for light and heavy eiements, either by adsorption or precipi-tation.Thus, AIz0 3 , FeZ03' MgO, CaO, mineral COz, organic COz, As, Ba, Be, Cu, Ni, Sn, V, Zn, U and Th are generally enriched in the < 2 ¡.tm fraction.However, Ce, Y, Sr and Pare only enriched in certain samples, usually belonging to the mineralized vein zone.
Fig. 8.-a) Plot ofSavin and Lee's (1988) equation for different /) 18 0 of water.The shaded area corresponds to the theoretical /) 18 0 values for r.ure illite.b) Plot ofClayton's el al. (1972)  equation for different /) 18 0 of water.The shaded area corresponds to the measured /) 1 O for quartz samples.Both plots consider that the /) 18 0 for El Berrocal meteoric water is ranging from -6 %o to -8 %O.

Table 7b .
----<:hemical composition oC the physical mixtures between Ulite and beidelliteThe highest level of Rb, an element not determined in the < 60 ¡.tm fraction, is detected in the most KzO-rich samples, which have dominant illite + Kfeldspar over smectite.The same U and Th anomalies observed in the < 60 ¡.tm fraction.Thus, As, Ba, Cu and Zn positive anomalies are detected in the same samples as for the < 60 ¡.tm fraction, though enhanced.Cerium, La and y are also present in important amounts, though generally in slightly lower concentrations than in the < 60 ¡.tm fraction.

Table 8 .
-8emiquantitative chemical composition of the accessory minerals and mineraloids associated with c1ay minerals

Table 13 .
---eation exchange complex oC the < 2 !J.IIl Craction In an acid environment and at the T isotopically es-