Origen y naturaleza de los minerales fosfato-sulfato alumínicos (APS) asociados a la mineralización de uranio en los red-beds triásicos (Cordillera Ibérica, España)

This study focuses on the mineralogical and chemical study of an Aluminium–phosphate–sulphate (APS) mineralization that occurs in a clastic sequence from the Triassic (Buntsandstein) of the Iberian Range. The deposit is constituted by sandstones, mudstones, and conglomerates with arenaceous matrix, which were deposited in fluvial to shallow-marine environments. In addition to APS minerals, the following diagenetic minerals are present in the clastic sequence: quartz, K-feldspar, kaolinite group minerals, illite, Fe-oxides-hidroxides, carbonate-sulphate cement-replacements and secondary uraniferous minerals. APS minerals were identified and characterized by optical microscopy, X-ray diffraction, scanning electron microscopy, and electron microprobe. Microcrystalline APS crystals occur replacing uraniferous minerals, associated with kaolinite, mica and filling pores, in distal fluvial-to-tidal arkoses-subarkoses. Given their Ca, Sr, and Ba contents, the APS minerals can be defined as a solid solution of crandallite-goyacite-gorceixite (0.53 Ca, 0.46 Sr and 0.01 Ba). The chemical composition, low LREE concentration and Sr > S suggest that the APS mineral were originated during the supergene alteration of the Buntsandstein sandstones due to the presence of the mineralizing fluids which causes the development of U-bearing sandstones in a distal alteration area precipitating from partially dissolved and altered detrital minerals. Besides, the occurrence of dickite associated with APS minerals indicates they were precipitated at diagenetic temperatures (higher than 80oC), related to the uplifting occurred during the late Cretaceous post-rift thermal stage.


Introduction
Aluminium phosphate-sulphate (APS) minerals occur in a wide range of geological environments, ranging from superficial weathering through sedimentary, diagenetic, hydrothermal and metamorphic environments, to post-magmatic systems (Dill, 2001).But, APS minerals have been neglected by geologists and geochemists due to their extremely small sizes (< 0.1-10 µm) and their low concentrations, generally less than 0.05 wt % (Gaboreau et al., 2007).Nevertheless, APS minerals are abundant in the clay mineral host-rock alteration assemblages associated with some uranium areas in the middle Proterozoic basins (Wilson 1985;Beaufort et al., 2005;Gaboreau et al., 2005;Gaboreau et al., 2007) and complex APS minerals were identified by Spötl (1990) as an early diagenetic precipitation in the Late Permian sandstones of the Northern Calcareous Alps (Austria).The model proposed for their origin suggests the precipitation of APS minerals as a consequence of the dissolution of detrital apatite in low pH environment.In an extensive review, Dill (2001) remarked that the precipitation of these minerals indicates extreme acid and oxidising conditions at shallow depths, and describes that these are common occurrences in different geological settings.This origin was also proposed by Benito et al., (2005) and Galán Abellán et al., (2008) who studied the APS minerals group in palaeosols from the southeastern continental Late Permian of the Iberian Range.According to these authors, APS mineral formation occurred shortly after sedimentation, due to the circulation of acid meteoric groundwaters, and concluded that the lack of carbonates and the presence of APS minerals at the base of the Triassic are a clear indication of extremely acid conditions during the Permian-Triassic transition in the studied area.
In Australian sedimentary formations of marine sandstones, Rasmussen (1996) described the presence of authigenic APS minerals (florencite, crandallite, gorceixite) as volumetrically minor but widespread components of Archaean to Cretaceous sandstones.In these sediments trace quantities of crandallite and gorceixite, as well as florencite, occur within pockets and linings of detrital clay and quartz surfaces.According to this author, the authigenic minerals were precipitated shortly after burial within the zone of sulphate reduction and methanogenesis.In the Permo-Triassic sandstones of Australia, the above author described early-dia-genetic REE-phosphate minerals with crandallite and gorceixite crystals lining cavities formed in the sandstones after grain dissolution.Phosphates were probably formed by the release of P and REE to the pore waters of fluvio-marine sediments, but they could also be formed by the partial dissolution and alteration of detrital minerals (monazite, clay matrix, feldspars and mica).The above considerations confirm that APS minerals are key minerals, and as Gaboreau et al., (2005) asserted, there is a need of more systematic analysis in all types of environments and particularly in siliciclastic sedimentary basins.Similarly, few data is available on the APS compositional variations in relations to the uranium deposits (e.g.Quirt et al., 1991;Gaboreau et al., (2005).The presence of APS minerals in the Buntsandstein sandstones where the uranium mineralization is widespread (Castañon et al., 1981;De la Cruz et al., 1987), is important because the timing of APS precipitation is clearly related to the composition and circulation of the mineralizing fluids.The study of the clays mineralogy of the Buntsandstein in this area of the Iberian Range by SEM, XRD, DTG and DTA has recently been accounted by Marfil et al., (2012) and reveals the illite-dickite-kaolinite mineral assemblage.The purpose of the present study is to contribute to a better understanding of the origin of these APS minerals and their relationship to the uranium mineral alteration.

Geological setting
The Iberian Range is an intracratonic, folded segment of the Alpine chain that developed as a rift basin between Permian and late Cretaceous times (Salas & Casas, 1993;Salas et al., 2001).The former authors indicate that the Iberian Basin was filled with thick clastic sequences during two rifting cycles, the first spanning from Late Permian to Late Triassic, and the second from Late Jurassic to Early Albian.During these active periods, sedimentation took place mainly in alluvial to lacustrine environments in a complex system of extensional basins.Both cycles gave way to periods of post-rift thermal subsidence during which shallow-marine carbonates were deposited.The sedimentation of the Permian and Triassic red-bed sandstones and mudstones took place during the first stage of rifting in half-graben basins.During Early Permian (Autunian) times, a series of half-graben, intermontane basins become unfilled by alluvial fans, slope breccias and lacustrine deposits associated in their lower part with volcaniclastic rocks of cal-alkaline affinities, such as andesites with subordinate rhyolites and basalts (Muñoz et al., 1983).
The Buntsandstein sediments can be divided into two major successions that are not always both present in the Iberian area: a lower one, comprising laterally restricted conglomerates and more widespread sandstones, and an upper one of irregular distribution, comprising sandstones, sandstones and siltstones and siltstones.The lower succession shows a marked fining -upward sequence interpreted as being of alluvial fan origin and forming an apron derived from an elevated SW margin of the riff basin.The overlying formation consists, of pink to red bedded sandstones of sandy braided river origin.These sandstones consist of amalgamate channel deposits totally devoid of fine grained flood plain sediments and are interpreted as having been deposited during a slow subsidence rate (Lopez-Gomez et al, 2011).
This study is based on core samples from borehole "Sigüenza 44-3" (drilled by the former Spanish Junta de Energía Nuclear (today CIEMAT) and by the Shell Company in 1978), with a depth of 999 m.This borehole is located on the Sigüenza High some 110 km Northeast of Madrid (Fig. 1).Two large Buntsandstein units were sampled, a lower unit (538 to 369 m) composed of conglomerates R. Marfil, A. La Iglesia, J. Estupiñan Estudios Geológicos, 69(1), 21-34, enero-junio 2013 4, analysis 67617-5).E) and F) SEM images of secondary uranium mineral grains (U).F) U grain is replacing clay matrix composed by kaolinite.The analyses of these grains are in Table 4 (analysis 20492-1 and 67616-3).. and sandstones with sporadic interbedded mudstones at the top, and an upper unit (369 to 300 m) constituted by sandstones and mudstones and occasional conglomerates with arenaceous matrix.Both units are red beds, separated by a 10 m-thick conglomerate, and were sedimentologically distinguished in the original survey logs (Fig. 1).The Gamma ray spectrometry shows a radioactive anomaly at a depth between 340 and 345 m, which is related to the presence of secondary uranium minerals in grey mineralized zones.On the contrary, in this area, exploration of Buntsandstein targets has not been successful, despite the fact that surface oil seeps are present in the Sigüenza area (Marfil et al., 1996b).

Samples and methods
The mineralogical composition of the almost 50 sandstones and mudstones selected from the two Buntsandstein units was determined by modal analysis (400 points per thin section) using a polarizing microscope.X-ray powder diffraction (XRD) analyses were performed on the 2-20 µm and <2µm fractions.A Bruker D8 Advance diffractometer, equipped with a Sol-X detector was used.The experimental conditions were: measurement interval 2-70º 2θ, time per step1 s, and a step size of 0.02º 2θ.The mineneralogical composition of the crystalline phases was estimated following Chung´s (1975) method and using Bruker software (EVA).Selected by XRD 5 sandstones and mud-stones samples were examined using a JEOL JSM-6400 scanning electron microscope equipped with a Link Systems energy-dispersive X-ray microanalyser (accelerating voltage 15 KV, beam current 3nA) by secondary electron (SE) and backscattered electron (BSE) modes.Finally, the chemical composition of the APS minerals was determined using a JEOL JXA-8900 electron microprobe (accelerating voltage 15 kV, current 21.5 nA, beam diameter 5 µm and the counting time 100s).The detection limits in ppm were 250 for Al, 370 for Sr; 245 for Fe, 205 for P, 135 for Ca, 240 for S, and 285 for Ba.For LREE the detection limits were 222 for Nd, 90 for La, 60 for Ce, 53 for Th and 158 for Pr.The following standard were used: For P apatite; for Ca augite; for Sr CO 3 Sr; for Al sillimanite; for Fe almandine; for Ba CO 3 Ba, for F apatite; for S chalcopyrite, and for LREE synthetic glass from PH developments.

Petrography, mineralogy, and chemical characteristic of the APS minerals
Significant amounts of APS minerals were detected in thin sections and SEM.APS minerals occur as very small idiomorphic crystals filling secondary pores, stylolites and microfractures (Fig. 4 B).These minerals replace uraniferous grains or they are enclosed in the distorted and broken mica crystals (Fig. 4 C).APS crystals (sizes ranging from 5 to 10 µm) also appear filling cavities as cement that surrounds corroded quartz grains (Fig. 4 D).APS minerals are preferentially found among expanded muscovite flakes which are totally twisted by mechanical compaction and altered to kaolinite (Fig. 4E, F).The quartz present in these sandstones has suturated and stylolitic contacts marked by iron oxides, indicative of pressure-solution processes.
The chemical compositions of the APS minerals and the structural formulae calculated on the basis of 2 moles of (XO 4 ) and 6 moles of (OH) per formula unit, following Scott´s method (1987), are given in Table 3.According to the criteria of this author, the analysed minerals belong to the plumbogummite group since trivalent anions are >1.5 moles, divalent cations in A-sites are dominant and Al>Fe in B-sites.The results of the analyses are plotted on Scott's diagram (Fig. 6).The calculated average value of (XO4) 3-is 1.51, which fits within the plumbogummite field.Considering the contents of Ca, Sr, and Ba, these APS minerals can be defined as a solid solution of crandallite-goyacite-gorceixite (Ca 0.53, Sr 0.46, Ba 0.01).The LREE (La, Ce, Pr, Nd and Th) contents are very low; in the case of La and Ce the average value is 282 ppm and 434 ppm, respectively.The content of Pr, Nd and Th in all the analyzed APS minerals was usually below the detection limit.Table 4 shows the chemical composition of the selected sand grains containing high values in UO 3 .Their values in Cu, As and P suggest an uranyl phosphate arseniate, of the autunite group, specifically torbernite-zeunerite serie: Cu(UO 2 )2(PO 4 , AsO 4 ) 2 .10H 2 O.

Discussion
The geochemical conditions in the Triassic basin during the APS mineral precipitation were oxidising and slightly acid, as indicated by the presence of the paragenesis formed by kaolinite-illite-quartz and hematite.Furthermore, the red sandstones beds are bleached and mottled corresponding to hematite dissolution caused by repeated process of acidification as described by Hoeve & Quirt (1984) and Kister et al., (2006).According to Rasmussen (1996), the critical factor controlling the APS mineral precipitation is the availability of Al.In the Buntsandstein sandstones, enough Al for the precipitation of APS minerals was supplied by the alteration of feldspars, mica or kaolinite.The Sr  content of the APS minerals could be related to the presence of celestite and strontianite in the interbedded mudstones and sandstones (Morad et al. 1992).Consequently, the compaction of the interbedded mudstones could have released the Sr and S needed for the APS mineral precipitation in the most porous sandstones.Another S source may be the sulphates present in fluids expelled from overlying evaporitic formations (Keuper facies).Such conditions are more likely to be associated with the introduction of oxidising meteoric waters during an uplift event in intra-formational unconformities in the Iberian Basin.
Idiomorphic APS crystals grow between the collapsed mica, which is partially altered to kaolinite fans and dickite (Marfil et al., 2012).This implies that APS minerals postdate the mechanical and chemical compaction, indicating that APS were formed during diagenesis.Constraints placed by vitrinite reflectance deduced temperatures from the associated organic matter in the underlying Permian shales of 135 to 159ºC around 900 m depth (Marfil et al., 1996 a, b).In addition, K/Ar age determinations of the associated fibrous illite, indicate that the episode of illitization in the Buntsand-stein occurred around 199.6 ± 4.3 M.a.during Lower and Middle Jurassic (Marfil et al., 1996 a).These ages coincide with the period of the most intensive extensional tectonics in this basin coeval with volcanic events (Salas & Casas, 1993).Thus, the above mentioned facts support the hypothesis of a diagenetic origin and high temperature for the APS minerals, similar origin as for these minerals in the Athabasca basin in Canada (Gaboreau et al., 2007).The assemblage of the APS minerals and dickite implies as well high diagenetic temperature rates.According to Ehrenberg et al., (1993) and Mcaulay et al. (1994) the transformation kaolinitedickite occurs at temperatures between 80 and 130ºC.Thus, the mineralizing fluids, related to tectonic events that occurred during the late Cretaceous post-rift thermal stage, were meteoric waters that transported the uranium into the buried red Buntsandstein sandstones.The fluids were enriched in uranium by circulating through granitic and gnessic rocks of the Spanish Central System (Castañon et al., 1981;De la Cruz et al., 1987), or even through Permian volcanic rhyolitic rocks.In this area of the Iberian Range significant Permian deposits composed by rhyolitic rocks (Peña et al., 1977  ZrO 2 1,81 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 Total 100,00 100,00 100,00 100,00 100,00 100,00 100,00 100,00 100,00 100,00 100,00 100,00 100,00  Muñoz et al., 1985) crop out.However, the time of the uranium mineralization and the origin and timing of migrating fluids in the Buntsandstein are not yet fully understood.The low content in LREE of the APS crystals is not indicative of a clear relationship with the uranium mineralization detected in the studied arkosic sandstones.However, the Gamma ray spectrometry (Fig. 1) shows a radioactive anomaly located at a depth of around 340 m, due to the presence of uranium minerals (Marfil et al., 1996b) Nevertheless, in agreement with the origin proposed by Gaboreau et al. (2005), APS minerals in the Triassic Buntsandstein facies could be related to the U-bearing arkosic sandstones in a distal alteration area, as shown in figure 7. Plotting the chemical analyses of the studied APS on the Gaboreau et al.( 2005) diagram, they fit in the area of distal alteration or barren areas.The above authors explain that the chemical variation in APS minerals is influenced by the host-rock lithology.For an equivalent alteration zone, the APS minerals formed in basement rocks are systematically enriched in P and LREE compared with those formed in the sandstones of the barren areas.Monazite (Ce) in the gneissic weathering profiles of the Leucogia area (Papoulis et al., 2004), is partially altered to a Th-silicate with very low concentration of LREE.Likewise, the above authors found that monazite in hydrothermaly altered rhyolitic rocks of Kos Island, were altered to Pbearing crandallite-goyazite and during this alteration process LREE are depleted.The alteration processes that occurred in the studied U-bearing sandstones could be similar to those described by the above mentioned authors.

Conclusions
In the studied Triassic sequence, several levels of the Buntsandstein sandstones contain important concentrations of APS minerals deposited in secondary pores formed by dissolution of the framework grains, mainly feldspars and uraniferous minerals.They appear as well replacing kaoliniteillite matrix, inside microfractures of the quartz grains and between the exfoliation planes of the mica.
APS minerals were precipitated under diagenetic conditions related to the uplifting occurred during the late Cretaceous post-rift thermal stage at high diagenetic temperatures, suggested by the presence of dickite in the studied samples APS minerals occurred in barren sandstones in zones of distal alteration with low LREE contents, and related to some tectonic event responsible for the introduction of meteoric and mineralizing waters into the buried sediments.
The intergranular clay matrix composed of kaolinite, illite and the feldspars constituted the main Al source, whereas secondary uranium minerals and apatite were the most probable P sources.Fluids migrating from the Permian evaporitic formations, or, more probably, from the overlying Keuper evaporitic facies can account for the Ca, S and Sr contents.
Based on their chemical composition, the APS minerals have been defined as a solid solution of crandallite-goyacite-gorceixite.

Fig. 3 .
Fig. 3.-A) Photomicrograph (crossed nicols) of a subarkosic sandstone showing frequent quartz overgrowths and K-feldspars replaced by kaolin minerals.(Small rests of feldspar are yellow stained).B) SEM image of abundant matrix of fibrous illite and a more recent phase of coarse dickite crystals (D).Some of these crystals are intergrowth with quartz cement, (Qc).C) Photomicrograph (crossed nicols) showing secondary uranium minerals (U) replacing frame-work detrital grains (possibly feldspars and clay matrix), as well as, quartz overgrowths (Qc).D) SEM image of the secondary uranium mineral (U) partially replaced by APS (Table4, analysis 67617-5).E) and F) SEM images of secondary uranium mineral grains (U).F) U grain is replacing clay matrix composed by kaolinite.The analyses of these grains are in Table 4 (analysis 20492-1 and 67616-3)..
and kaolinite minerals.Quartz cement occurs as syntaxial overgrowths frequently corroded by clays mainly kaolinite-dickite and fibrous illite (Fig.3A, B).Ductile rock fragments consist of micaschists, argillites, and early eroded and transported (rip-up) clay clasts.Mica includes muscovite and subordinate amounts of biotite.Turmaline, apatite, rutile, anatase, sphene, malachite, ilmenite, monazite, torbernite and xenotime are characteristic accessory components.Framework grains and the matrix are replaced by secondary uraniferous minerals in green and grey sandstones beds.These secondary uraniferous minerals appear in different associations: as cement and replacing feldspars, illite epimatrix and quartz cement (Fig.3C), filling microfractures of the quartz grains and the exfoliation planes of the mica.They also become visible as individual grains (Fig.3D, E, F).Several of these uraniferous mineral grains are replaced by APS as the figures 3D, F and 4A show.Pyrite appears to be associated with the development of anhydrite nodules.. Other cements include scarce K-feldspar overgrowths, Fe and Ti oxides or oxyhydroxides (as coatings and patches), dolomite and Fe-dolomite/ankerite. Occasionally, abundant calcite, anhydrite and barite are common cement replacements.

Fig. 4 .
Fig. 4.-A) SEM image of several secondary uranium mineral phases (U) replaced by APS microcrystals.In the centre, remains of Kfeldspar partially replaced by kaolinite (F-K to K) (Analyses 67616-4 of Table4).B) Abundant APS crystals filling fractured quartz grains and replacing kaolinite matrix.C) SEM image of aggregates of small APS euhedral crystals lining the planes of distorted muscovite and corroding quartz, replacing feldspars, and partially filling the pore space.D) SEM image under BSE mode of APS zoned microcrystals filling the porosity and corroding the quartz grains.E) SEM images of sandstone showing kaolinite matrix and very rich in APS crystals.Illite is rarely observed in this area where mica is altered to kaolinite-dickite.F) SEM image showing APS crystals between twisted kaolinite booklets and extremely broken quartz grains.