Applied Geochemistry 58 (2015) 123–135

Contents lists available at ScienceDirect

Applied Geochemistry

journal homepage: www.elsevier .com/ locate/apgeochem

Uranium migration and retention during weathering of a granitic wasterock pile

http://dx.doi.org/10.1016/j.apgeochem.2015.02.0120883-2927/� 2015 Published by Elsevier Ltd.

⇑ Corresponding author at: Institut für Geologie und Paläontologie, WestfälischeWilhelms-Universität, 48149 Münster, Germany. Tel.: +33 1 44 27 50 84.

E-mail address: floraboekhout@hotmail.com (F. Boekhout).

F. Boekhout a,⇑, M. Gérard a, A. Kanzari a, A. Michel b, A. Déjeant a, L. Galoisy a, G. Calas a, M. Descostes c

a Institut de Minéralogie, de Physique des Matériaux, et de Cosmochimie (IMPMC), Sorbonne Universités, UPMC Univ Paris 06, CNRS, IRD UMR 206, 4 place Jussieu,F-75005 Paris, Franceb Institut du Physique du Globe, 75005 Paris, Francec AREVA, BG Mines, R&D, Paris la Défense, France

a r t i c l e i n f o

Article history:Available online 25 February 2015Editorial handling by M. Kersten

a b s t r a c t

This study investigates the post-mining evolution of S-type granitic waste rocks around a former uraniummine, Vieilles Sagnes (Haute Vienne, NW Massif Central, France). This mine was operated between 1957and 1965 in the La Crouzille former world-class uranium mining district and is representative of intra-granitic vein-type deposits. 50 years after mine closure and the construction and subsequent re-vegeta-tion of the granitic waste rock pile, we evaluate the environmental evolution of the rock pile, includingrock alteration, neo-formation of U-bearing phases during weathering, and U migration. Vertical trencheshave been excavated through the rock pile down to an underlying paleo-soil, allowing the investigation ofthe vertical differentiation of the rock pile and its influence on water pathways, weathering processes andU migration and retention. Arenization dominantly drives liberation of U, by dissolution of uraniniteinclusions in the most alterable granitic minerals (i.e. K-feldspar and biotite). Retention of U in the matrixat the base of the waste rock pile, and in the underlying paleo-soil most likely occurs by precipitation of(nano-) uranyl phosphates or a combination of co-precipitation and adsorption reactions of U onto Fe(oxy)hydroxides and/or clay minerals. Even though U-migration was observed, U is retained in stable sec-ondary mineral phases, provided the current conditions will not be modified.

� 2015 Published by Elsevier Ltd.

1. Introduction

In the near future, the combination of an intensifying demandfor uranium resources, declining ore grades, and the exploitationof lower commodity grade ores will lead to an exponential increaseof the annual volume of waste rock (Kahouli, 2011). The largestvolume of waste products are produced in the nuclear fuel-cycleduring mining and milling (Abdelouas, 2006), triggering a growinginterest in waste rock management (e.g., Kipp et al., 2009; Miaoet al., 2013; Schindler et al., 2013; Neiva et al., 2014). The term‘waste rock’ is defined as untreated rocks that do not containenough U to be economically processed. Waste rock piles arehighly heterogeneous in their nature, comprising barren rockremobilized from the mine surroundings (access roads, miningworks), overburden from overlying soils and rock covering theore deposit and un-reclaimed, sub-economic ore extracted fromthe mine. A major characteristic of waste rock piles is theirincreased erosional surface compared to natural granitic outcrops,

inducing an accelerated weathering rate. Apart from being poten-tially harmful for the environment this enhanced weathering pro-vides an end-member example of accelerated continentalalteration processes. Rock pile weathering overprints the latehydrothermal alteration of the rock and the natural supergenepre-mining alteration of the site.

For a long period, U vein-type deposits yielded the bulk of glo-bal U production, whereas less than 10 percent of the U was pro-duced from deposits of this type at the end of the 1980s. Thismining activity left behind a significant amount of granitic wasterock that form the oldest U-bearing rock piles in many inhabitedareas, hence their importance for the evaluation of the weatheringevolution of mine wastes for remediation strategies. One of themost representative regions with granitic vein-type deposits isthe ‘La Crouzille’ district in the French Massif Central. S-type gran-ites of the Massif Central region have an elevated U concentration,often larger than 20 ppm (Barbier, 1970; Cuney, 2014). Such a highbackground concentration dominantly arises from resistate U-bearing phases, located in accessory phases (zircon, sphene, allan-ite, monazite, apatite, or magmatic uraninite, ect.) (Bajo et al.,1983; Berzina et al., 1975; Cuney, 2009), as opposed to secondaryU phases within primary minerals (Speer et al., 1981; Tieh et al.,

124 F. Boekhout et al. / Applied Geochemistry 58 (2015) 123–135

1980), concentrated at textural heterogeneities (grain boundaries,mineral cleavages and fractures) in late-stage deuteric and earlyweathering materials (Dill et al., 2010; Guthrie and Kleeman,1986; Kikuchi et al., 2007; Leroy and Turpin, 1988; Waber et al.,1992).

A question of major environmental relevance with respect to Uwaste rocks concerns the extent of U mobilization since the pile’sconstruction. A mine-impacted wetland in the La Crouzille miningarea has U values of up to 14,000 ppm. Wetlands often act as sinksfor U and other trace metals, with U accumulating up to severalthousand ppm (Owen and Otton, 1995; Regenspurt et al., 2010;Wang et al., 2013). Wang et al. (2013) recently showed that forthe La Crouzille wetland the presence of U(IV) in soil, as a non-crystalline species is bound to amorphous Al–P–Fe–Si aggregates,and in pore water, occurs as a distinct species associated with Feand organic matter colloids. Also, natural soils in this same regiondemonstrate that uranium exerts a high pressure on soil bacterialcommunities and suggest the existence of a uranium redox cyclemediated by bacteria in the soil (Mondani et al., 2011).Furthermore, naturally occurring high U concentrations of up to3000–4000 ppm are common in alpine soils even though the sur-rounding granitic bedrock only contains trace amounts of U(Owen and Otton, 1995; Regenspurt et al., 2010). Experimentalwork revealed that U, liberated by granitic weathering in thisregion, was concentrated in soil organic matter (rather than tomineral phases) and was present primarily in the hexavalent state(Regenspurt et al., 2010).

This study refers to the influence of weathering on U migrationin a re-vegetated granitic waste rock pile constructed 50 years agoon the site of the Vieilles Sagnes uranium mine, with the greateraim of improving mine waste management in this context. TheVieilles Sagnes site is located in the La Crouzille former mining dis-trict, Massif Central, France (Fig. 1a) and bears one of the oldest andmost geologically representative waste rock piles in the region,constructed between 1957 and 1965. The weathering-driven dif-ferentiation of the waste rock pile was directly investigated byexcavating trenches down to the underlying paleo-soil (Fig. 1band c). Geochemical and mineralogical data on the different hori-zons within the waste rock pile and the paleo-soil enable an eval-uation of the impact of weathering processes over 50 years. Theappearance of neo-formation of uranyl-phosphate minerals resultsfrom weathering and more generally from the complex interplaybetween rock weathering, U migration and its fate. A major roleis played by the paleo-soil underlying the pile and its highly Usorbing phases, resulting from weathering, that limit the extentof U mobility to the immediate surroundings.

Fig. 1. Vieilles Sagnes site. (a) Map of France indicating the location of the VieillesSagnes site with a star. The shaded area represents the Limousin region. (b) The re-vegetated Vieilles Sagnes waste rock pile in La Crouzille mining district, FrenchMassif Central, before trench excavation. (c) VSA trench through (upper part of) theVieilles Sagnes waste rock pile exposing vertical cross sections down to theunderlying paleo-soil.

2. Geological background

Uranium deposits related to granites are best exemplified bythe mid-European Variscan uranium province, which extends overmore than 2000 km from Spain to the Bohemian Massif (Cuney,2014). They are located in late Carboniferous peraluminousleucogranites (French Massif Central) and in their metamorphichost rocks. Uranium deposition occurred 30–50 Ma after theemplacement of the granites, at 270 ± 15 Ma (Holliger et al.,1986; Kribek et al., 2009) during a regional extensional event.The ore forming fluids are low-salinity and low-temperature fluids(Dubessy et al., 1987). Uranium deposition results from the mixingof oxidized meteoric fluids leaching magmatic uraninite from gran-ites with fluids derived from an overlying basin (Turpin et al.,1990). This implies that waste rocks derived from French MassifCentral have experienced a multiphase geological history, respon-sible for an important mineralogical heterogeneity. At least threeperiods of U-mineral formation, alteration or neo-formation can

be distinguished in this region (Fig. 2). The first period is theemplacement of S-type granites (Fig. 2a) (324 ± 4 Ma), in whichmost U is located in accessory phases such as zircon, monazite,apatite and magmatic uraninite (hereafter referred to as uraninite

Fig. 2. Cartoon illustrating the different periods of mineral growth, alteration and weathering of the rocks found at the Vieilles Sagnes site based on literature (see Section 2for references). (a) Granite emplacement (�324 ± 4 Ma) forming the primary magmatic minerals quartz, biotite, muscovite, plagioclase and K-feldspar and accessory mineralsas zircon, monazite and uraninite I (Holliger et al., 1986). (b) The ore forming phase when the main ore minerals coffinite (U(SiO4) � nH2O) and uraninite II (UO2) precipitatedduring (late) hydrothermal alteration. (c) Mining of the Vieilles Sagnes site (1957–1965) and waste rock pile construction after which the granitic blocks is re-vegetated andexposed to weathering over the course of about 50 years.

F. Boekhout et al. / Applied Geochemistry 58 (2015) 123–135 125

I) (Alderton et al., 1980; Förster et al., 1999). A late emplacement oflamprophyric dykes crosscutting these intrusive bodies does notappear to be linked to an ore-forming event (Scaillet et al., 1996).The mineralization event (Fig. 2b) is related to the two-stageexhumation of the Western French Massif Central (Le Carlier deVeslud et al., 2013), with precipitation of coffinite(U(SiO4) � nH2O) and uraninite (UO2; hereafter referred to as ura-ninite II, also known as pitchblend). Fluid circulation and subse-quent hydrothermal alteration continued until at least 280 Maduring post-collisional Permian rifting, as extensive tectonicsexhumed these magmatic bodies (Bouchot et al., 2005; Cuneyet al., 1990; Dill et al., 2011; Le Carlier de Veslud et al., 2013).The maximum supergene alteration depth in the French MassifCentral is up to about 100 m (Cathelineau, 1983). During miningof the Vieilles Sagnes site (1957–1965) rocks were taken from adepth of around 225 m and the non-processed rocks were placedon an on-site rock pile located next to the underground mine(Fig. 2c). Recent environmental studies on La Crouzille district haveconcerned the control of U-speciation on U migration into rivers(Allard et al., 1999) and wetlands (Wang et al., 2013).

3. Site description and sampling

The Vieilles Sagnes mine is located approximately 20 km north ofLimoges (Fig. 1a) and experiences a mild oceanic climate with aver-age temperatures of 18 �C in summer to 4.5 �C in winter and averagerainfall of 1022 mm/yr. This site produced 214,000 t of ore with anaverage grade of 1.14‰, and 262,000 t of waste rocks with an U-con-tent lower than a cut-off grade of 200 ppm U (Leroy, 1978) The pho-ton flux on site is between 400 and 550 cps (gamma radiation) andsampling was conducted with a Geiger counter. The rock pile isheterogeneous, with a wide range of size from meter-size bouldersto a sandy and clay-size matrix. The present day pH varies between4 and 6. This site is particularly suitable for studying U migration ona waste rock pile, as no remodeling of the site took place since thepile construction and re-vegetation 50 years ago.

Petrographical and geochemical characteristics of the LaCrouzille district have been extensively studied (Barbier, 1970;Barbier and Ranchin, 1969; Brocandel, 1987; Friedrich, 1984;Girard, 1990). The so called ‘fertile’ S-type granites of this areashow a U background of 20 ppm due to the presence of uraniniteI (Barbier, 1970; Cuney, 2014). Ore mineralization resulted fromthe hydrothermal alteration of the surrounding host rocks, withthe formation of secondary monazite, rutile, iron oxides, pyriteand clay minerals (Cathelineau, 1986; Dudoignon et al., 1988;Leroy and Turpin, 1988; Meunier and Velde, 1982; Nishimotoand Yoshida, 2010). Additionally, fluorite, barite and calcite are

observed in the zones of disseminated mineralization (Barbier,1970; Barbier and Ranchin, 1969; Leroy and Sonet, 1976). The pre-sent-day rock pile material can be described as variably hydrother-mally altered rocks within a fine-grained sandy matrix exposed topost-mining weathering.

Isolated since 50 years, this site has been overgrown by mossand meter-high ferns and partly re-vegetated by small birch trees(Fig. 1b). In order to study the vertical differentiation in the wasterock pile, two trenches were excavated with the help of a digger(Fig. 1c): one at the top of the waste rock pile (VSA – 46.00241.3649) and one at the bottom (VSB – 46.0023 1.3640). Fig. 3shows the schematic profile of the VSA trench at the upper partof the rock pile. The trenches are 25 and 20 meters long respec-tively and a maximum vertical depth of up to around 2.5 meters.The shallowest blocks and the upper 40–50 cm of the rock pileare intermingled with roots from the anthropogenic soil cover(Fig. 3a and b). The matrix of the waste rock pile is generally greyto yellowish brown. Some zones in the middle of the trench showyellow discoloration (Fig. 3c and d). A slight increase in the matrixto waste rock ratio can be observed towards the deepest parts ofthe waste rock pile (Fig. 3e), which is located in the second trench(VSB – sketch not shown here). In some levels of the rock pile thefragile rock chips are more yellowish to red and seem more coarse-grained than the surrounding matrix (Fig. 3f). Different types ofsamples were taken according to their gamma radiation or locationin the vertical profile as indicated by the insets in Fig. 3:

� Waste rocks: waste rocks (up to several decimeters in diameter)that were removed from the trenches during excavation (not in-situ) and in-situ blocks in the wall of the trenches. The graniticblocks vary in color from yellow to pinkish to brown and have agrain size variation between 2 and 5 mm. Two blocks of up to20 cm had a non-granitic mineralogy and contain predomi-nantly dominantly quartz and mica (lamprophyre, Fig. 3g).� Matrix of the waste rock pile: greyish brown to yellow fine-

grained matrix of waste rock material including fragile graniticrock chips and arenization products (Fig. 3h).� Paleo-soil/anthropogenic soil: Anthropogenic re-vegetated soil

above and paleo-soil underneath the rock pile. The paleo-soilis only encountered underneath the lower trench (VSB). Theupper part of the rock pile might have been partly emplacedon outcropping granite.

4. Methods and sample preparation

Elemental analyses were performed to determine the bulk com-position of the different parts of the waste rock pile and underlying

Fig. 3. Schematic profile of the VSA trench through the upper part of the Vieilles Sagnes waste rock pile. Sample location of figures a–d, and h are indicated in the schematiccross section of trench VSA. Figures e, f and g are located in the second trench (cross section not shown). (a) + (b) The most superficial rocks consist of larger blocksintermingled with roots and moss from the soil cover that developed in 50 years. (c) + (d) Zones in the middle of the trench show yellow discoloration from pyrite oxidation tojarosite. (e) A slight increase of matrix to waste rock ratio can be observed towards the deepest parts of the waste rock pile and the underlying paleo-soil with a U content>1000 ppm. (f) The matrix of the waste rock pile is generally grey to yellowish brown. In some parts of the rock pile the arenized rock chips are more yellowish to red andmore coarse-grained than the surrounding matrix. (g) Lamprophyre block VSB8. (h) Sample location of a high U matrix sample VSA22 (904 ppm) at the bottom of the VSAtrench. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

126 F. Boekhout et al. / Applied Geochemistry 58 (2015) 123–135

F. Boekhout et al. / Applied Geochemistry 58 (2015) 123–135 127

paleo-soil. Samples were ground with an agate mortar and pestleand using an agate-ring mill. About 50 mg bulk sample were dis-solved using a mixture solution of HNO3 + HF on a hot plate at90�C for 1 h. The precipitate is eluted with a solution containingboric acid in order to remove fluoride silicate and evaporated againat 60 �C. The eluted sample was diluted with 2% HNO3 for themajor-element measurement. This preparation procedure doesnot allow measuring Si directly using ICP-MS since the main partof Si is lost during evaporation. Replicate analyses (every 3–5 mea-surements) of GSN and MA-N granitic reference samples gave anaccuracy of 3–8% for all major and trace elements. Analyses werecarried out using an inductively coupled plasma mass spectrome-ter (ICP-MS; Agilent 7500 CX�) at the IRD, Bondy (France).

X-ray diffraction analyses were performed with a X’Pert ProPanalytical diffractometer (IMPMC-UPMC) with Ni-filtered CuKaradiation at 40 kV and 40 mA. For bulk samples, disoriented pow-ders were scanned from 2� to 90� (2h) with 3 s counting every0.0167�, in an aluminum rotating sample-holder.

The morphology of weathering phases was studied by using aZeiss supra 55VP and Zeiss EVO LS15 scanning electron microscope(SEM) with an acceleration voltage of 15 keV (IMPMC-UPMC).Micro-morphological observations were coupled with EDS micro-probe analyses (Energy Dispersive Spectrometry; microprobeOxford Instrument INCA Energy 350XMax), conducted at 15 keVwith a 450 pA probe current and a working distance of 7.5 mm.Semi-quantification of the spectra was achieved using theESPRIT� software (Bruker). Some images were acquired with asmaller working distance (down to 2.5 mm) with the microscopeoperating at lower voltage (down to 3 keV). SEM observations werealso carried out on thin sections from granitic blocks and rock chipsat different levels of the vertical profile. Matrix samples wereimpregnated before thin section preparation but also observed aspowder. Additionally, to study the intermediate phase of alterationbetween solid blocks and matrix, several fragile rock chips wereimpregnated and then made into thin sections.

5. Results

5.1. SEM-EDS analysis

Based on the mineralogy and geochemical analyses of theVieilles Sagnes rock pile, samples were selected for a detailedSEM study on thin sections and rock fragments. SEM observationsshow the presence of accessory phases not previously identified byXRD and give insight on the alteration processes and the modifica-tion of rock texture. U-bearing minerals present in the waste rockpile are either inherited from the pre-mining history of the graniticrocks or neo-formed during weathering. SEM-EDS analyses weremainly focused on the detection of neo-formed U-bearing mineralphases during weathering.

5.1.1. Pre-mining mineralogyThe heterogeneity of the rock pile at the time of mining is a

result of magmatic, (late) hydrothermal and possible supergenemineral formation and alteration (see Section 2). An overview ofthe pre-mining mineralogy is given in Fig. 4. All waste rocks con-tain major magmatic minerals quartz, K-feldspar, plagioclase, bio-tite and muscovite, and varying abundances of clay minerals(smectite, kaolinite and traces of chlorite), with the exception oftwo small blocks of lamprophyre that contain dominantly quartzand mica (VSA5 and VSB8). Accessory minerals as zircon, monaziteand apatite are common in all samples. Other (pre-mining) acces-sory minerals as Ti-oxide, pyrite, Fe-Ti oxides, Fe phosphates,hydrothermal monazite, fluorite, uranyl phosphates and U-oxideinclusions in primary magmatic minerals, are observed in varying

abundances. Pyrite is a common accessory mineral throughoutthe entire rock pile.

U-bearing pre-mining minerals can be identified in both thewaste rocks and interstitial matrix samples. Apart from zircon,monazite and apatite, U oxides are found as inclusions in primarymagmatic minerals, and uranyl phosphates associated with (late)hydrothermal accessory minerals. U oxide is present as sub-micro-metric inclusions with undefined morphologies (Fig. 4a), or ascubic or spheroidal mineral inclusions with a diameter of up to30 lm. The latter occur as inclusion in apatite, K-feldspar andquartz (Fig. 4b and c). Uranyl phosphates are found associated witha Fe (oxy)hydroxide minerals (Fig. 4d), as inclusions in quartz, or asnodules in a Fe phosphate coating in micro-fractures (Fig. 4e) andmicro-voids. This specific occurrence of Fe phosphate was found inthe block of lamprophyre (VSB8), with uranyl phosphates of up to40 lm in diameter. U inclusions in primary magmatic mineralsshow a varying resistance to alteration and weathering dependingon their host mineral. U oxide inclusions in quartz (Fig. 4b) showno signs of alteration, whereas uraninite inclusions in K-feldsparand biotite show the possible liberation of U into the surroundingprimary magmatic mineral (Fig. 4f).

Smectite alteromorph after plagioclase is associated with Ti-minerals intergrown with monazite and zircon (Fig. 4g). Thesemineral assemblages are found in most of rock samples. Fe- andTi-oxides also occur in between mica sheets (Fig. 4h). Ti oxide isoften found in a matrix of kaolinite alteromorph after biotite(Fig. 4i). Additionally, zoned aluminum phosphate sulphate (APS)minerals rich in As, with traces of P, Sr and REE, are found in fewsamples as VSA10 (32 ppm U) and VSB3.

5.1.2. Post-mining mineralogyThroughout the rock pile, the waste rocks show different

degrees of weathering from incipient weathering of granitic blocksto complete arenization. Neo-formation of minerals can beobserved on a field scale by yellow to reddish weathering aureoles(Fig. 3c and d). On a micro-scale, primary silicate minerals trans-form into clay minerals together with oxidation and alteration ofpyrite and APS, as well as neo-formation of uranyl phosphatesand Fe oxides.

The oxidation of pyrite occurs in the granitic waste rocks andthe lamprophyre (Fig. 5a). Pyrite shows varying degrees of oxida-tion within a sample: while some pyrite is un-oxidized, protectedby the surrounding primary or clay minerals, pyrite adjacent tomicro-voids shows signs of weathering. When associated with aclay mineral (kaolinite or smectite), oxidizing pyrite is associatedwith the precipitation of sub-micrometric crystals of (Ca) uranylphosphates (Fig. 5b and c). A final-stage coating of Fe phosphatecovers some micro-voids (Fig. 5c). In the most weathered samplespyrite is completely transformed into agglomerates of jarosite(Fig. 5d). APS shows oxidation rims in a matrix of smectite (Fig. 5e).

In the arenized matrix (e.g. VSB11), there is evidence of uranylphosphates up to 30 lm. SEM-EDS analyses show formation of asecondary uranyl phosphate surrounding this primary uranylphosphate, with a slightly different composition by the additionalCu and minor As (Fig. 5f). This secondary uranyl phosphate precip-itates dominantly along grain boundaries and within pores. Micro-morphological observation on impregnated thin sections and frag-ile rock chips show the occurrence of these (Cu, As) uranyl phos-phates throughout the matrix (e.g. VSA18; Fig. 5g and h).

With increasing degree of weathering, biotite expansion alongcleavage planes intensifies as well as the abundance of K-feldsparor biotite transformed into smectite. Advanced weathering to com-plete arenization of the rocks (e.g. VSB16) implies the mobilizationof Fe as sub-lm Fe oxide crystals (Fig. 5i). The mobility of Fethrough the rock pile is illustrated by the enrichment of Fe with

ea VSA18 VSA18

VSB11

U oxideU oxideqz

Fe oxide

c

d

b

10 µm2 µm

20 µm

U oxide

VSA21R

3 µm

qz

qz

VSA18

Kfs2 µm

U oxide

fVSB8e

20 µm

qz

Fe pht

U pht

VSA9

Ti oxide

kln

bt

10 µm

VSB3

20 µm

mnz

zrnsme

plTi oxide

g

Fe oxide

VSA16

whitemica

i

10 µm

h

U pht

Kfs

Fig. 4. SEM-EDS analyses of the pre-mining mineralogy of the Vieilles Sagnes waste rock pile. (a) sub- micrometric U oxide inclusions in K-felspar (Kfs). (b) Cubic U oxide inquartz (qz). (c) Spheroidial U oxide in quartz. (d) Uranyl phosphate (U pht) associated with Fe oxide included in quartz (qz); (e) alteration of U oxide in K-feldspar (Kfs). (f)Uranyl phosphate (U pht) nodules growing in a fracture coating of Fe phosphate (Fe pht) g) typical pre-mining mineralogy with zircon (zrn), monazite (mnz) and Ti oxideminerals associated with smectite (sme) after plagioclase (plg). (h) Formation of Fe (oxy)hydroxide phase during hydrothermal alteration, associated with mica. (i) Pyrite (py)in a matrix of kaolinite (kln) alteromorph after biotite (bt).

Fig. 5. SEM-EDS analyses of the post-mining mineralogy of the Vieilles Sagnes waste rock pile. (a) Oxidation: jarosite (jrs) coating alteromorph after pyrite (py). (b)Precipitation of sub-micrometric crystals of uranyl phosphate (U pht) around pyrite (py). (c) Precipitation of (Ca) uranyl phosphate (U pht) around pyrite (py), coated by Fephosphate (Fe pht). (d) Jarosite (jrs) neo-formation. (e) Alteration of APS minerals in smectite. (f) Weathering of uranyl phosphate (U pht) into (Cu, As) uranyl phosphate (Upht). (g) (Cu, As) uranyl phosphate (U pht) coating a micro-void. h) (Cu, As) uranyl phosphate (U pht). (i) Fe leaching: sub-micrometric Fe oxide neo-formation on biotite (bt)sheets.

128 F. Boekhout et al. / Applied Geochemistry 58 (2015) 123–135

F. Boekhout et al. / Applied Geochemistry 58 (2015) 123–135 129

depth, characterizing the matrix samples closest to the base of thewaste rock pile and the underlying paleo-soil.

The paleo-soil underlying the waste rock pile (VSB12, VSB13and VSB14; Fig. 3e) contains high U concentrations, but the onlyoccurrence of a U-bearing phase has been found as scarce (<10–20 fragments per SEM sample) small (<1 lm) uranyl phosphatecoatings on fluorapatite. A (Cu) uranyl phosphate occasionally alsogrows within agglomerates of sub-micrometric grains of differentprimary minerals in the paleo-soil horizons.

5.2. Geochemistry and mineralogy

ICP-MS and XRD bulk rock analyses enable the evaluation of themajor mineralogical components in the waste rock pile and pro-vide an overview of the overall degree of weathering. Also, U-richhorizons in the trenches are localised. Waste rocks and the are-nized matrix have a signature of differentiated granite (Fig. 6):all samples (waste rock and fine-grained matrix) show low Sr val-ues and plot parallel to the Ba–Rb side of the Ba–Rb–Sr ternary

Fig. 6. Ternary Rb–Ba–Sr granite discrimination plot of El Bouseily and El Sokkary(1975) of granitic waste rocks and matrix samples.

Fig. 7. (a) Primitive mantle normalized REE plot of granitic waste rocks and matrix samplnormalized trace element plot of granitic waste rocks and matrix samples (Sun and Minterpretation of the references to color in this figure legend, the reader is referred to th

discrimination diagram used for tracing differentiation trends ingranitic rocks (El Bouseily and El Sokkary, 1975).

The primitive mantle-normalized (Sun and McDonough, 1989)REE diagram of waste rocks and matrix samples reveals enrich-ment in LREE with La/Yb ratios of �15–132 for waste rock andmatrix samples, the line color representing log U of the individualsamples (Fig. 7a). There is no correlation between La/Yb and U con-tent. Negative Eu anomalies are present in all samples. Primitivemantle-normalized (Sun and McDonough, 1989) trace elementpatterns for granitic waste rock and matrix samples of theVieilles Sagnes site have a similar signature with negative P, andTi anomalies, and a positive, largely varying anomaly in U(Fig. 7b). The REE and trace element patterns are very similar, withfew exceptions: sample VSA15 is significantly depleted withrespect to the rest of the samples, whereas samples VSA19 andVSB10r have a slightly more enriched signature when comparedto rest of the rock pile.

The chemical index of alteration (CIA) (Nesbitt and Young,1982) is used as a proxy for the alteration state of the differentsamples and is defined as Al2O3/(Al2O3 + CaO + Na2O + K2O). Asalteration of the bulk rock proceeds, primary minerals are replacedby clay minerals resulting in an increase of the CIA values (Priceand Velbel, 2003). Fig. 8a shows the CIA vs log U content for the dif-ferent sample groups. Even though the mineralogy of the wasterocks and the matrix is similar, they display a varying U content.Waste rocks (Fig. 8a) fall below the 200 ppm of U cut-off gradewith values between 9 and 191 ppm U, with the exception ofVSB10r with a value of 333 ppm U. The matrix (Fig. 8a) is moreenriched in U, with values ranging between 114 and 1530 ppm.Fifteen of the nineteen matrix samples have a U content abovethe cut-off grade of 200 ppm. An increase in U content is specifi-cally seen in the matrix samples closest to the base of the rock pile,for example, VSB11, VSA12, VSA22 and VSA23 (Figs. 3e and 8a).Overall, it can be observed that the matrix is more homogeneousin comparison to the waste rocks, indicated by the smaller rangein CIA values, and relatively more altered (richer in clay minerals).XRD analyses also illustrate this increase in clay mineral content inthe arenized matrix with respect to the waste rocks. The main par-agenesis of the granitic waste rocks, analysed by XRD bulk rockanalyses, consists of varying quantities of quartz, K-feldspar, pla-gioclase, biotite, muscovite and clay minerals smectite, kaoliniteand traces of chlorite (Fig. 8b). Samples VSA5 and VSB8, two lam-prophyric samples of up to 20 cm in diameter, have a non-graniticmineralogy of dominantly quartz and mica. The matrix of the

es (Sun and McDonough, 1989). Color gradient represents log U. (b) Primitive mantlecDonough, 1989). Lamprophyre samples were not included in these figures. (Fore web version of this article.)

Fig. 8. Degree of weathering of the Vieilles Sagnes waste rock pile. (a) The CIA vs log U diagram for the different sample groups, indicating the 200 ppm cut-off grade for theVieilles Sagnes site. The top horizon of the paleo-soil underneath the waste rock pile is most enriched in uranium and functions as an attenuation zone (indicated by thedashed line). Deeper horizons (VSB14) are not enriched in U. (b) XRD pattern comparison of waste rock and matrix XRD diffraction patterns of matrix sample VSA21m⁄ androck sample VSA21r⁄ (these samples with the highest and lowest CIA value respectively indicated with a thicker contour line in Fig. 8a). Minerals were identified according tothe position of the (001) series of peak areas of the basal reflections for the main clay mineral groups of smectite (1: 1.5 nm), mica (2: 1.0 nm), K-Feldspar (3: 3.24) and quartz(4: 3.33 nm). Enrichment in smectite (1) can be observed in the matrix at the expense of K-feldspar (3) and mica (2) that show a reduction in peak intensity from the rock tothe matrix sample. Wrp = waste rock pile.

130 F. Boekhout et al. / Applied Geochemistry 58 (2015) 123–135

waste rock pile consists of fragile rock chips and arenization prod-ucts forming a sandy matrix in between the waste rocks. The min-eralogy of this matrix closely resembles the granitic mineralogy ofthe waste rocks with the same major rock forming minerals asmentioned above. When comparing waste rock and matrix XRDdiffraction patterns (Fig. 8b), enrichment in clay mineral contentcan be observed in the matrix (dominantly smectite). Matrix sam-ple VSA21m and rock sample VSA21r (the samples with the high-est and lowest CIA value respectively; Fig. 8a), illustrate a claymineral increase in the matrix at the expense of feldspar and micathat show a reduction in peak intensity (Fig. 8b). On a field scalesome zones in the middle of the trench show yellow discoloration(Fig. 3c and d), and reddish zones of Fe oxide formation. Someminor mineral phases were isolated and analysed by XRD. Thealteration mineral that coats some waste rocks and is responsiblefor yellow color of the weathering aureoles is identified by XRDas jarosite ((KFe3+

3 (OH)5(SO4)2).Five soil samples, above and below the waste rock pile, were

also analysed (Fig. 8a). The re-vegetated top-soil (20–30 cm thick)is an anthropogenic soil covering the waste rock piles that containsU concentrations of 12 (VSB20) and 38 ppm (VSA24). Abundantroots of trees and small bushes can be found in this anthropogenicsoil cover. Sample VSA24 consist of silty sand and is similar in min-eralogy to the matrix of the waste rock pile described above,whereas VSB20 is a dominantly quartz-rich sandy layer. The claymineral content of these soils is low and consists mostly of kaolin-ite and chlorite.

The paleo-soil underneath the waste rock pile is enriched in Uwith values between 1285 and 275 ppm, decreasing with depth(Figs. 3e and 8a). The outcrop underneath the waste rock pile ispossibly discontinuous and has only been found underneath partof trench VSB. This paleo-soil consists, from top to bottom, of a yel-low white 10 cm thick horizon (VSB12 – 1286 ppm U) with abun-dant roots and organic fibers. Clay minerals are predominantlysmectite with minor kaolinite and traces of chlorite. This layer is

underlain by a non-continuous 2 cm thick black horizon (VSB13– 1249 ppm U) consisting of quartz with traces of vermiculite.The deepest horizon (VSB14 – 276 ppm U) is a brownish >15 cmthick layer with small chunks of granite of several centimeters indiameter. The mineralogy is similar to VSB13 with dominantlyquartz and traces of vermiculite.

6. Discussion

6.1. Weathering and arenization of the Vieilles Sagnes waste rock pile

The greater surface area of the rock pile with respect to a natu-rally occurring granitic outcrop, will accelerate weathering rates,giving rise to measurable modifications of the granitic rock geo-chemistry, petrology and mineralogy in a timespan of a few dec-ades. However, it is necessary to decipher this recent weatheringfrom the various alteration stages, from incipient weathering ofgranitic blocks to complete arenization into the sandy matrix inbetween granite blocks.

The waste rocks retain a signature typical of differentiated gran-ites (Fig. 6). Leaving aside the two rare occurrences of small lam-prophyric blocks, the granitic rocks and matrix in between theblocks have very similar REE and trace element content and behav-ior (Fig. 7a and b). The influence of post-mining weathering can beobserved on both field- and micro-scale, overprinting the hetero-geneous pre-mining mineralogy. At the outcrop scale, weatheringfeatures in the waste rock pile consist of arenization and discoloredyellow weathering aureoles, as well as reddish Fe rich oxidationzones. On a micro-scale, clay mineral formation, pyrite oxidation,and the neo-formation of uranyl phosphate and jarosite are indica-tive of post-mining weathering. During weathering, Ti and Zr areconsidered the least mobile due to the stability of resistate miner-als such as zircon, rutile or ilmenite (Brown and Calas, 2012). In aternary Al–Ti–Zr plot (Garcia et al., 1994), the compositional vari-ation observed at Vieilles Sagnes (Fig. 9) is based on the premise

Fig. 9. Ternary plot of Al2O3 � 15 – Zr – TiO2 � 300 for waste rocks and matrixsamples (after Garcia et al., 1994). Arrow indicating a slight weathering trendtowards immature sandstone (after Nesbitt and Young, 1982). SPG = ‘stronglyperaluminous granite’ field.

F. Boekhout et al. / Applied Geochemistry 58 (2015) 123–135 131

that the matrix developed from arenization and mechanical trans-port from peraluminous granitic waste rocks (Sawyer, 1986). Thematrix samples plot on an arenization trend towards Zr (sandstoneend-member), indicating the development of an immaturesandstone.

Mineralogical investigations indicate that kaolinite, smectiteand minor chlorite are present in varying abundance, but their for-mation conditions are unclear. At the Vieilles Sagnes site, thehydrothermal alteration leads to albitization and desilification(episyenitization) in the vicinity of the mineralization (Barbier,1970; Barbier and Ranchin, 1969; El Jarray, 1993; Girard, 1990;Leroy and Sonet, 1976). The fact that the lowest CIA value foundfor the waste rock samples is 59, indicates that the least alteredsampled rock already contains more clay minerals than a freshgranite (average CIA value of 45–55 (Bahlburg and Dobrzinski,2011). This leaves the possibility that all rocks on the pile wereto a certain affected degree by hydrothermal alteration. However,a comparison between XRD diffraction patterns of the waste rockand the directly surrounding matrix (Fig. 8b) reflects the decreas-ing content of primary minerals at the expense of smectite phasesduring arenization. This shows that at least part of the smectite for-mation occurred during weathering.

Not only does the amount of smectite increase during weather-ing, also the biotite expansion along cleavage planes intensifies.Advanced weathering of the granitic blocks shows almost com-plete arenization and the mobilization of Fe by oxidation of pyriteand re-precipitation of sub-micrometric crystals of Fe (oxy)hy-droxides in between the remaining biotite and clay mineral sheets(Fig. 5i). In the field, Fe re-mobilization is shown by the presence ofred oxidized zones in the matrix samples close to the base of thewaste rock pile and in the underlying paleo-soil. Pyrite is presentthroughout the rock pile, as a remnant of the hydrothermal pre-mining phase (Cathelineau, 1983; Leroy, 1978). Chemical leachingfrom pyrite oxidation causes an acid environment in the absence ofacid neutralizing minerals as carbonate. This leads to arenization ofthe most weathered zones (e.g. Fig. 3d) and might trigger the for-mation of preferential water pathways through the waste rock pileas observed on Australian waste dumps (Tran et al., 2003). The for-mation of matrix materials and larger void spaces creates path-ways for fluid flow around the granitic cobbles and boulderswaste rocks. This jarosite-rich zone correlates with a depletion ofbulk REE (and U) as in VSA15 (Figs. 3d and 7). This is the only

horizon that shows significant depletion of REE and trace elementsand formation of jarosite during weathering. However, VSA19,sampled from another yellow horizon (Fig. 3c) shows enrichmentin LREE. The REE and trace element patterns of the rest of the sam-ples are relatively similar and show no signs of mobilization duringweathering, implying that REE are dominantly controlled by acces-sory mineral phases that are not susceptible to weathering.

6.2. Assessment of U speciation in granitic waste rock piles

Different pre- and post-mining U-bearing minerals were identi-fied by SEM-EDS analyses. The majority of the waste rocks is bar-ren, with a U content around the geological background of20 ppm U, or overburden rock with U values up to 190 ppm(Fig. 8a). Overburden rocks usually contain at least trace amountsof the ore. Pagel (1982), Cuney et al. (1990) and Leroy (1978) notethat uraninite is always present in granites surrounding economicU mineralization, associated with primary magmatic minerals,which is in line with the findings at the Vieilles Sagnes site (Fig. 4).

The pre-mining U-bearing minerals are identified as magmaticand hydrothermal or supergene accessory phases. U is either incor-porated in zircon or monazite or occurs as U-minerals (U oxides oruranyl phosphates). Lamprophyre waste rocks VSA5 (4637 ppm)and VSB8 (4307 ppm) can be classified as sub-economic ore.These two blocks of up to 20 cm are the only rocks encounteredin the rock pile with a U content far above the cut-off grade ofthe site. The uranyl phosphates of up to 40 lm in diameter, inthese samples are most likely coeval with late hydrothermal orsupergene minerals (Leroy and Turpin, 1988). The presence ofabundant accessory phases as zircon and monazite can in someanomalous cases be responsible for an elevated U content in bulkrock geochemistry, e.g. sample VSB10r with a U content of333 ppm. In this case SEM observations show the presence of largequantities of sub-micrometric hydrothermal monazite. This is alsoreflected by its elevated and flatter REE pattern (Fig. 7a). This sam-ple further contains abundant (Al, Ba, Ca) phosphates and Fe- andTi-oxides. The post-mining (sub-) micrometric (Ca) and (Cu) uranylphosphates observed by SEM are formed during weathering andprecipitated in the vicinity of oxidizing pyrite (Fig. 5b) or alonggrain boundaries and in pores and micro-voids (Fig. 5c, f and g).These neo-formed uranyl phosphate minerals are most likely autu-nite and torbenite. Some secondary uranyl phosphates are associ-ation with apatite. This association is also experimentallyobserved (Ohnuki et al., 2004) and has been proposed and usedfor U stabilization in mine- and mill tailings (Abdelouas, 2006).

Even though the matrix samples closest to the base of the pile,and the underlying paleo-soil were the most enriched in U, there isno correlation with the presence of U-bearing phases, besides a fewsub-micrometric uranyl phosphates (10–20 per sample). Eitherthese uranyl phases are too small to be visible by SEM, or anotherU retention process is responsible for these high U values. Manymaterials, including several metal hydroxides (Fe, Al, Mn) as wellas clays are known to efficiently adsorb U (Turner et al., 1996;Chisholm-Brause et al., 2001; Brown and Calas, 2012; Maheret al., 2012 and references therein) and control U release and trans-port under oxic conditions (Latta et al., 2012; Noubactep et al.,2006). Sato et al. (1997) reported sorption of U(VI) and P by Fe-ox-ides as a precursor to the formation of uranyl phosphates, also ura-nyl-phosphate ternary complexes were reported on the internalsurface area of low-T opals (Calas et al., 2009; Massey et al.,2014). Ulrich et al. (2006) later showed that the major U precipi-tate forming in mine water is rapidly agglomerating colloidal ferri-hydrite, though this depends on the precipitation rates of thevarious nano-crystalline phases (e.g. silica) in mine waters(Allard et al., 1999). Murakami et al. (2005) found that during crys-tallization of goethite and hematite from ferrihydrite, U, P, and Cu

132 F. Boekhout et al. / Applied Geochemistry 58 (2015) 123–135

or Mg adsorbed onto ferrihydrite are released, and lead to the for-mation of uranyl phosphate nano-crystals accumulated on the sur-face of the newly formed minerals.

Jarosite, which has an extended maximum stability field frompH 1 to 3.5, is present in some specific horizons in the rock pile.Hence, the pH close to pyrite and jarosite must be locally lower(Fig. 3c and d). Adsorption of U on a positively charged Fe-hydrox-ide surface would be limited under acidic conditions. However,considering the co-sorption of U with arsenate and phosphateshows that pre-adsorption of the latter oxy-anions provides thenecessary negative surface charge under acidic conditions for theenhanced adsorption of the positive charged uranyl ions (Bighamand Nordstrom, 2000).

U(VI) phosphates are probably the most common uranyl phasesat sites where natural weathering processes and/or anthropogenicactivity result in the release of U to the environment (Fanizza et al.,2013). For example, at the unmined Coles Hill uranium deposit(Virginia) hosted in quartzo-feldspathic gneisses, Ba uranyl phos-phates precipitate in the presence of secondary barite veins andphosphorous influences the speciation and mobility of uraniumin soil/saprolite systems developed under oxidizing conditions(Jerden Jr. and Sinha, 2006). In the vicinity of the U ore deposit ofKoongarra, Australia, saléeite and nano-meta-torbernite wereidentified (Murakami et al., 2005). The host rock, a chlorite-quartzschist, has been weathered for more than one million years (Aireyand Ivanovich, 1986). Downstream of this U ore deposit, U ismainly associated with goethite and hematite. Nano-crystals(20–100 nm in size) of uranyl phosphates such as meta-torbernitewere scattered between and firmly attached to nano-crystals (2–50 nm) of goethite and hematite that formed nodules and veins.These variations in U speciation due to weathering mostly dependon different pH conditions and availability of elements in solution.

Fig. 10. Schematic cartoon of U dissolution/oxidation and precipitation and/or sorptiodissolution and oxidation of uraninite inclusions in main magmatic minerals most vulnethe matrix and underlying soil, occurs by precipitation of (nano-) phases, as possibly ur

An additional example is provided by the impact of the mine dumpat the Pinhal do Souto mine (exploited between 1978 and 1989),Portugal, mineralized in autunite and torbernite. In the sedimentsand soils surrounding this dump, Fe-oxide precipitates retain thehighest concentration of metals, including U with concentrationsof up to 485 ppm (Neiva et al., 2014).

Overall, when it comes to U speciation at and around graniticvein-type U deposits, the precipitation of (nano-) uranyl phos-phates, and U sorption onto metal-(hydr)oxides or amorphousaggregates, can be expected. Sorbtion of U on clay minerals hasso far not been identified as a dominant process of U retention atmining sites, but more likely as an initiation mechanism of nano-precipitation of uranyl phosphates, as described by e.g.Murakami et al. (2005). This implies that, apart from the visibleneo-formed (Ca) and (Cu) uranyl phosphates observed at theVieilles Sagnes site by SEM, U concentrates as U-bearing nano-phases and/or uranyl phosphate complexes sorbed on Fe (oxy)hy-droxides or clay minerals (Fig. 10).

6.3. Environmental impact of U migration in a granitic environment

Processes of U mobilization depend on U-speciation (Brown andCalas, 2012; Maher et al., 2012) (Fig. 10). Under oxidizing and pHneutral conditions, U is often mobilized as uranyl phosphate com-plexes (Brugger et al., 2003). Organic ligands and bacterial activitymay also lead to an important mobility of U(IV) as found in somewetlands of the La Crouzille district (Wang et al., 2013). There isevidence that U and Fe have been mobilized in the VieillesSagnes waste rock pile over 50 years, with the presence of red oxi-dized zones and an increasing U content (up to 1285 ppm) down tothe underlying paleo-soil. Below the paleo-soil, the U contentdecreases with depth (Fig. 8a). It appears that the surface on which

n during weathering. (a) Arenization drives liberation of U (UO22+), dominantly by

rable to weathering (biotite and K-feldspar). (b) Retention of U in the deeper part ofanyl phosphates, or U sorption onto clay minerals and Fe (oxy)hydroxides.

F. Boekhout et al. / Applied Geochemistry 58 (2015) 123–135 133

the rock pile was deposited acts as an attenuation zone, as hasbeen observed at old mining sites (Pittauerová and Goliáš, 2002;Tran et al., 2003). However, there is no evidence of the integrityof this paleo-soil underneath the whole waste rock pile, and partof the rock pile may occur directly on outcropping granite.

The source of U is most likely the weathering of uraninite inclu-sions in K-feldspar and biotite (Fig. 10). During arenization U oxideinclusions are exposed to oxidative alteration after the dissolutionof primary mineral phases that are the least resistant to weather-ing (Donahue et al., 2009). The other U-bearing minerals presentin the Vieilles Sagnes waste rock pile seem unlikely candidatesfor liberation of U under these conditions on a human timescale:at circum-neutral pH, zircon and monazite will have a minimalcontribution to U liberation with negligible dissolution rates. Uoxide inclusions in quartz can also be excluded as a U source onthis time scale as we see no evidence of weathering of quartzgrains.

The formation of uranyl phosphates in the Vieilles Sagnes wasterock may arise from presence of dissolved phosphate in the perco-lating water. Even minor phosphate concentrations (10–8 M) pro-mote the formation of (nano-) uranyl phosphates (e.g. autunite),thereby limiting the mobility of uranyl (UO2

2+) in subsurface envi-ronments (Wellman et al., 2006). At a re-vegetated site, as theVieilles Sagnes site, P should be widely available due to the pres-ence of the soil cover with a P2O5 content of up to 0.40 wt%.Moreover, a neutral (pH 4–6 for the Vieilles Sagnes site), or slightlylower pH as seen in the vicinity of oxidizing pyrite, facilitates autu-nite formation by precipitation of nano-(Ca) uranyl phosphates(Cretaz et al., 2013; Mehta et al., 2014; Ohnuki et al., 2004). Themaximum adsorption of uranyl on Fe-oxides and oxyhydroxidesoccurs between pH 6 and 7 (Hsi and Langmuir, 1985). U retentionon different types of smectite is also most effective at circum neu-tral pH conditions (Catalano and Brown Jr., 2005; Schlegel andDescostes, 2009). Further speciation studies will shed light on themolecular-scale processes governing the geochemical and miner-alogical transformations revealed by the present study.

7. Conclusions and outlook

The influence of weathering on the stability of granitic wasterock piles has important implications for modeling their environ-mental impact and developing strategies for waste rock site man-agement. Weathering affects rock texture, and hence thepercolating of water through the pile, as well as U migrationthrough dissolution and neo-formation of U-bearing phases. Thefollowing main conclusions can be drawn from this study:

� Besides magmatic minerals (e.g., quartz, K-feldspar, plagioclase,biotite, muscovite) the Vieilles Sagnes granitic waste rocksshows varying quantities of clay minerals, smectite, kaoliniteand traces of chlorite. The matrix of the waste rock pile consistsof arenization products forming a sandy matrix, with a mineral-ogy that closely resembles that of the granitic blocks, thoughwith a slightly elevated clay mineral content, illustratingweathering of the primary aluminosilicate minerals.� The waste rock pile is currently under oxic conditions, which

result in the weathering and oxidation of biotite and reducedU-bearing phases. Arenization liberates U from uraninite inclu-sions in primary magmatic minerals, and mobilizes Fe. The oxi-dation of pyrite can locally drive these processes. Other U-bearing accessory phases, as zircon and monazite, are noteffected by weathering on a human time scale.� The majority of the waste rocks is barren- or overburden rock

and their U content falls within the cut-off grade of theVieilles Sagnes site with values <200 ppm U. The matrix of the

waste rock samples is enriched in U up to 1300 ppm. The tophorizons of the underlying paleo-soil are also enriched in Uand serve as an attenuation zone, as the U concentration dropsdown below this paleo-soil.� The main neo-formed U-bearing phases are sub-micrometric

(Ca) and (Cu) uranyl phosphates.� As these uranyl phosphates are resistant to oxidative dissolu-

tion under circum-neutral pH conditions, their presence doesnot pose an elevated environment impact on a human timescale.� For evaluating the environmental impact of granitic waste rock

piles, the key factor is not the bulk U content of the rock, butrather U speciation and solubility, including the resistance toweathering of the U-bearing phases. This study shows that Umigration is spatially limited at the Vieilles Sagnes site, by theformation of secondary phases that are stable under current siteconditions.� As in other granitic rock piles and naturally outcropping granitic

bodies, the formation of stable uranyl phosphates may play adominant role in limiting U-mobility under circum-neutral pHconditions. Under these conditions, U migration may also behindered by adsorption onto Fe (oxy)hydroxides and/or clayminerals.

Acknowledgements

The Scanning Electron Microscope (SEM) facility at IMPMC issupported by IdF SESAME Grant 2006 N�I-07-593/R, INSU-CNRS,INP-CNRS, University Pierre et Marie Curie – Paris 6, and by ANRGrant No. ANR-07-BLAN-0124-01. This manuscript benefited fromdiscussions with, and comments and lab assistance from T. Allard,F. Le Cornec, I. Esteve, S. Caquineau, T. Pilorge, L. Delbes, S. Biassand V. Phrommavanh. We also thank the two anonymous review-ers, whos comments greatly improved the manuscript.

References

Abdelouas, A., 2006. Uranium mill tailings: geochemistry, mineralogy, andenvironmental impact. Elements 2, 335–341.

Airey, P.L., Ivanovich, M., 1986. Geochemical analogues of highlevel radioactivewaste repositories. Chem. Geol., 203–213

Alderton, D.H.M., Pearce, J.A., Potts, P.J., 1980. Rare earth element mobility duringgranite alteration: evidence from southwest England. Earth Planet. Sci. Lett. 49,149–165.

Allard, T., Ildefonse, P., Beaucaire, C., Calas, G., 1999. Structural chemistry ofuranium associated with Si, Al, Fe gels in a granitic uranium mine. Chem. Geol.158, 81–103.

Bahlburg, H., Dobrzinski, N., 2011. Chapter 6 A review of the chemical index ofalteration (CIA) and its application of the study of Neoproterozoic glacialdeposits and climate transitions. Geol. Soc., Lond., Memoirs 36, 81–92.

Bajo, C., Rybach, L., Weibel, M., 1983. Extraction of uranium and thorium from Swissgranites and their microdistribution. 2. Microdistribution of uranium andthorium. Chem. Geol. 39, 299–318.

Barbier, J., 1970. Zonalités géochimiques et métallogèniques dans le massif deSaint-Sylvestre (Limousin-France). Mineral Deposita 5, 145–156.

Barbier, J., Ranchin, G., 1969. Influence de l’alteration météorique sur l’uranium àl’etat de traces dans le granite à dex micas de St-Sylvestre. Geochim.Cosmochim. Acta 33, 33–47.

Berzina, I.G., Yeliseyeva, O.P., Popenko, D.P., 1975. Distribution relationships ofuranium in intrusive rocks of northern Kazakhstan. Int. Geol. Rev. 16 (11),1191–1204.

Bigham, J.M., Nordstrom, D.K., 2000. Iron and aluminum hydroxysulfates from acidsulfate waters. In: Alpers, C.N., Jambor, J.L., Nordstrom, D.K. (Eds.), Sulfateminerals: Crystallography, geochemistry and environmental significance. Rev.Mineral Geochem., vol. 40, pp. 351–403.

Bouchot, V., Ledru, P., Lerouge, C., Lescuyer, J.-L., Milesi, J.-P., 2005. 5: Late Variscanmineralizing systems related to orogenic processes: the French Massif Central.Ore Geol. Rev. 27, 169–197.

Brocandel, M., 1987. Apport de la métallogénie de l’uranium de la rechercheminière, étude du secteur de Bellezane (Massif granitique de St Sylvestre, NOMassif Central). COGEMA, GREGU p. 88.

Brown, G.E., Calas, G., 2012. Mineral-aqueous solution interfaces and their impacton the environment. Chem. Perspect. 1, 483–742.

134 F. Boekhout et al. / Applied Geochemistry 58 (2015) 123–135

Brugger, J., Burns, P., Meisser, N., 2003. Contribution to the mineralogy of aciddrainage of uranium minerals: marécottite and the zippeite group. Am. Mineral.88, 676–685.

Calas, G., Galoisy, L., Allard, T., 2009. Uranium trapping on opals from the Nopalnatural analogue: evidence for complexation on internal surface of opal.Geochim. Cosmochim. Acta 73, A185.

Catalano, J.G., Brown Jr., G.E., 2005. Uranyl adsorption onto montmorillonite:evolution of binding sites and carbonate complexation. Geochim. Cosmochim.Acta 69, 2995–3005.

Cathelineau, M., 1983. Potassic alteration in French hydrothermal uraniumdeposits. Miner. Deposita 18, 89–97.

Cathelineau, M., 1986. The hydrothermal alkali metasomatism effect on graniticrocks: quartz dissolution and related subsolidus changes. J. Petrol. 27, 945–965.

Chisholm-Brause, C.J., Berg, J.M., Matzner, R.A., Morris, D.E., 2001. Uranium(VI)sorption complexes on montmorillonite as a function of solution chemistry. J.Colloid Interface Sci. 233, 38–49.

Cretaz, F., Szenknect, S., Clavier, N., Vitorge, P., Mesbah, A., Descostes, M., Poinssot,C., Dacheux, N., 2013. Solubility properties of synthetic and natural meta-torbernite. J. Nucl. Mater. 442, 195–207.

Cuney, M., 2009. The extreme diversity of uranium deposits. Miner. Deposita 44, 3–9.

Cuney, M., 2014. Felsic magmatism and uranium deposits Bulletin de la SociétéGéologique de France 185, in press.

Cuney, M., Friedrich, M., Blumenfeld, P., Bourguignon, A., Boiron, C., Vigneresse, J.L.,Poty, B., 1990. Metallogenesis in the French part of the Variscan orogen. Part 1:U preconcentrations in pre-Variscan and Variscan formations – a comparisonwith Sn, W and Au. Tectonophysics 177, 39–57.

Dill, H.G., Gerdes, A., Weber, B., 2010. Age and mineralogy of supergene uraniumminerals – tools to unravel geomorphological and palaeohydrological processesin granitic terrains (Bohemian Massif, SE Germany). Geomorphology 117, 44–65.

Dill, H.G., Gerdes, A., Weber, B., 2011. Dating of Pleistocene uranyl phosphates in thesupergene alteration zone of Late Variscan granites by Laser-Ablation-Inductive-Coupled-Plasma Mass Spectrometry with a review of U minerals ofgeochronological relevance to Quaternary geology. Chem. Erde 71, 201–206.

Donahue, K., Dunbar, N., Heizler, L., McLemore, V., 2009. Clay mineralogy of theGoathill Nort rock pile, Questa mine, Taos country, NM: Origins and Indicationsof In-situ Weathering. SME Annual Meeting.

Dubessy, J., Ramboz, C., N’guyen-Trung, C., Cathelineau, M., Charoy, B., Cuney, M.,Leroy, J., Poty, B., Weisbrod, A., 1987. Physical and chemical controls (fO2, T, pH)of the opposite behaviour of U and Sn-W as exemplified by hydrothermaldeposits in France and Great Britain, and solubility data. Bull. Minér. 110, 261–281.

Dudoignon, P., Beaufort, D., Meunier, A., 1988. Hydrothermal and supergenealterations in the granitic cupola of montebras, Creuse, France. Clays ClayMiner. 36, 505–520.

El Bouseily, A.M., El Sokkary, A.A., 1975. The relation between Rb, Ba and Sr ingranitic rocks. Chem. Geol. 16, 207–219.

El Jarray, A., 1993. Circulations fluides et altérations hydrothermales associées auxminéralisations à U (As, F) dans le Massif de St Sylvestre (NW du Massif Centralfrançais). CREGU, p. 324.

Fanizza, M.F., Yoon, H., Zhang, C., Oostrom, M., Wietsma, T.W., Hess, N.J., Bowden,M.E., Strathmann, T.J., Finneran, K.T., Werth, C.J., 2013. Pore-scale evaluation ofuranyl phosphate precipitation in a model groundwater system. Water Resour.Res. 49, 874–890.

Förster, H.-J., Tischendorf, G., Trumbull, R.B., Gottesmann, B., 1999. Late-collisionalgranites in the Variscan Erzgebirge, Germany. J. Petrol. 40, 1613–1645.

Friedrich, M., 1984. Le complex granitique hyper-alumineux de St Sylvestre nord-ouest du Massif central français. CREGU Géologie et Géochimie de l’Uranium, p.361.

Garcia, D., Fonteilles, M., Moutte, J., 1994. Sedimentary fractionations between Al, Tiand Zr and the genesis of strongly peraluminous granites. J. Geol. 102, 441–442.

Girard, C., 1990. Contrôle structural, pétrologique et géochimique desminéralisations uranifères dans le Massif de St Sylvestre. COGEMA, GREGU, p.246.

Guthrie, V.A., Kleeman, J.D., 1986. Changing uranium distributions duringweathering of granite. Chem. Geol. 54, 113–126.

Holliger, P., Cuney, M., Friedrich, M., Turpin, L., 1986. Age Carbonifère de l’unité deBrâme du complex granitique peralumineux de Saint Sylvestre (NO MassifCentral) défini par les données isotopiques sur zircon et monazite. ComptesRendu de l’Acedémie des Sciences, Paris 14, 1309–1314.

Hsi, C.-K.D., Langmuir, D., 1985. Adsorption of uranyl onto ferric oxyhydroxides:application of the surface complexation site-binding model. Geochim.Cosmochim. Acta 49, 1931–1941.

Jerden Jr., J.L., Sinha, A.K., 2006. Geochemical coupling of uranium and phosphorousin soils overlying and unmined uranium deposit: Coles Hill, Virginia. J.Geochem. Explor. 91, 56–70.

Kahouli, S., 2011. Re-examining uranium supply and demand: new insights. EnergyPolicy 39, 358–376.

Kikuchi, M., Hidaka, H., Horie, K., Gauthier-Lafaye, F., 2007. Redistribution of REE, Pband U by supergene weathering studied from in-situ isotopic analyses of theBangombé natural reactor, Gabon. Geochim. Cosmochim. Acta 71, 4716–4726.

Kipp, G.G., Stone, J.J., Stetler, L.D., 2009. Arsenic and uranium transport in sedimentsnear abandoned uranium mines in Harding County, South Dakota. Appl.Geochem. 24, 2246–2255.

Kribek, B., Zak, K., Dobes, P., Leichmann, J., Pudilova, M., Rene, M., Scharm, B.,Scharmova, M., Hajek, A., Holeczy, D., Hein, U., Lehmann, B., 2009. The Roznauranium deposit (Bohemian Massif, Czech Republic): shear zone-hosted, lateVariscan and post-Variscan hydrothermal mineralization. Miner. Deposita 44,99–128.

Latta, D.E., Boyanov, M.I., Kemner, K.M., O’Loughlin, E.J., Scherer, M.M., 2012. Abioticreduction of uranium by Fe(II) in soil. Appl. Geochem. 27, 1512–1524.

Le Carlier de Veslud, C., Alexandre, P., Ruffet, G., Cuney, M., Gheilletz, A., 2013. Atwo-stage exhumation in Western French Massif Central: new geochronologicalevidences of syn-collisional extension. Lithos 175–176, 1–15.

Leroy, J., 1978. The Margnac and Fanay uranium deposits of the La Crouzille District(Western Massif central, France): geologic and fluid inclusion studies. Econ.Geol. 73, 1611–1634.

Leroy, J., Sonet, J., 1976. Contribution à l’etude géo-chronologique des filons delamprophyres recoupant le granite à deux micas de Saint-Sylvestre (Limousin,Massif Central français). Comptes Rendu de l’Acedémie des Sciences, Paris, SérieII 283, 1477–1480.

Leroy, J.L., Turpin, L., 1988. REE, Th and U behaviour during hydrothermal andsupergene process in a granitic environment. Chem. Geol. 68, 239–251.

Maher, K., Bargar, J.R., Brown Jr., G.E., 2012. Environmental speciation of actinides.Inorg. Chem. 52, 3510–3530.

Massey, M.S., Lezama-Pacheco, J.S., Nelson, J.M., Fendorf, S., Maher, K., 2014.Uranium incorporation into amorphous silica. Environ. Sci. Technol. 48, 8636–8644.

Mehta, V.S., Maillot, F., Wang, Z., Catalano, J.G., Giammar, D.E., 2014. Effect of co-solutes on the products and solubility of uranium(VI) precipitated withphosphate. Chem. Geol. 364, 66–75.

Meunier, A., Velde, B., 1982. Phengitization, seriticitization and potassium–beidellite in a hydrothermally-altered granite. Clay Miner. 17, 285–299.

Miao, Z., Akyol, H.N., McMillan, A.L., Brusseau, M.L., 2013. Transport and fate ofammonium and its impact on uranium and other trace elements at a formeruranium mill tailings site. Appl. Geochem. 38, 24–32.

Mondani, L., Benzerara, K., Carrière, M., Christen, R., Mamindy-Pajany, Y., Février, L.,Marmier, N., Achouak, W., Nardoux, P., Berthomieu, C., Chapon, V., 2011.Influence of uranium on bacterial communities: a comparison of naturaluranium-rich soils with controls. PLoS ONE 6.

Murakami, T., Sato, T., Ohnuki, T., Isobe, H., 2005. Field evidence for uraniumnanocrystallization and its implications for uranium transport. Chem. Geol. 221,117–126.

Neiva, A.M.R., Carvalho, P.C.S., Antunes, I.M.H.R., Silva, M.M.V.G., Santos, A.C.T.,Cabral Pinto, M.M.S., Cunha, P.P., 2014. Contaminated water, stream sedimentsand soils close to the abandoned Pinhal do Souto uranium mine, centralPortugal. J. Geochem. Explor. 136, 102–117.

Nesbitt, H.W., Young, G.M., 1982. Early Proterozoic climate and plate motionsinferred from major element chemistry of lutites. Nature 299, 715–717.

Nishimoto, S., Yoshida, H., 2010. Hydrothermal alteration of deep fractured granite:effects of dissolution and precipitation. Lithos, 153–162.

Noubactep, C., Sonnefeld, J., Sauter, M., 2006. Uranium release from a natural rockunder near-natural oxidizing conditions. J. Radioanal. Nucl. Chem. 267, 591–602.

Ohnuki, T., Kozai, N., Samadfam, M., Yasuda, R., Yamamoto, S., Narumi, K.,Naramoto, H., Murakami, T., 2004. The formation of autunite(Ca(UO2)2(PO4)2�nH2O) within the leached layer of dissolving apatite:incorporation mechanism of uranium by apatite. Chem. Geol. 211.

Owen, D.E., Otton, J.K., 1995. Mountain wetlands: efficient uranium filters –potential impacts. Ecol. Eng. 5, 77–93.

Pagel, M., 1982. The mineralogy and geochemistry of uranium thorium, and rare-earth elements in the two radioactive granites of the Vosges, France. Mineral.Mag. 46.

Pittauerová, D., Goliáš, V., 2002. Weathering of mine wastes after historical silvermining in the Jachymov ore district (Czech Republic) and migration of uranium.Proc. Int. Workshop, 157–160.

Price, J.R., Velbel, M.A., 2003. Chemical weathering indices applied to weatheringprofiles developed on heterogeneous felsic metamorphic parent rock. Chem.Geol. 202, 397–416.

Regenspurt, S., Margot-Roquier, C., Harfouche, M., Froidevaux, P., Steinmann, P.,Junier, P., Bernier-Latmani, R., 2010. Speciation of naturally-accumulateduranium in an organic-rich soil of an alpine region (Switzerland). Geochim.Cosmochim. Acta 74, 2082–2098.

Sato, T., Murakami, T., Yanase, N., Isobe, H., Payne, T.E., Airey, P.L., 1997. Iron nodulesscavenging uranium from groundwater. Environ. Sci. Technol. 31, 2854–2858.

Sawyer, E.W., 1986. The influence of source rock type, chemical weathering andsorting on the geochemistry of clastic sediments from the QueticoMetasedimentary Belt, Superior Province. Chem. Geol. 55, 77–95.

Scaillet, S., Cuney, M., le Carlier de Veslud, C., Cheilletz, A., Royer, J.J., 1996. Coolingpattern and mineralization history of the Saint Sylvestre and western Marcheleucogranite pluton, French Massif Central: II. Thermal modelling andimplications of the mechanisms of uranium mineralization. Geochim.Cosmochim. Acta 60, 4673–4688.

Schindler, M., Durocher, J.L., Kotzer, T.G., Hawthorne, F.C., 2013. Uranium-bearingphases in a U-mill disposal site in Northern Canada: products of the interactionbetween leachate/raffinate and tailings material. Appl. Geochem. 29, 151–161.

Schlegel, M.L., Descostes, M., 2009. Uranium uptake by hectorite andmontmorillonite: a solution chemistry and polarized EXAFS study. Environ.Sci. Technol. 43, 8593–8598.

F. Boekhout et al. / Applied Geochemistry 58 (2015) 123–135 135

Speer, J.A., Solberg, T.N., Becker, S., 1981. Petrography of the uranium-bearingminerals of the Liberty Hill Pluton, South Carolina; phase assemblages andmigration of uranium in granitoid rocks. Econ. Geol. Bull. Soc. Econ. Geol. 76 (8),2162–2175.

Sun, S.s., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanicbasalts: implications for mantle composition and processes. Geol. Soc., Lond. 42,313–345 (special publications).

Tieh, T.T., Ledger, E.B., Rowe, M.W., 1980. Release of uranium from granitic rocksduring in situ weathering and initial erosion (central Texas). Chem. Geol. 29,227–248.

Tran, A.B., Miller, S., Williams, D.J., Fines, P., Wilson, G.W., 2003. Geochemical andmineralogical characterisation of two contrasting waste rock dumps—the INAPwaste rock dump characterisation project. In: Proceedings from the 5thinternational conference on acid rock drainage (ICARD): The AustralasianInstitute of Mining and Metallurgy, Melbourne, pp. 939–947.

Turner, G.D., Zachara, J.M., McKinley, J.P., Smith, S.C., 1996. Surface-chargeproperties and UO2

2+ adsorption of a subsurface smectite. Geochim.Cosmochim. Acta 60, 339–3414.

Turpin, L., Leroy, J.L., Sheppard, S.M.F., 1990. Isotopic systematics (O, H, C, Sr, Nd) ofsuperimposed barren and U-bearing hydrothermal systems in a Hercyniangranite, Massif Central, France. Chem. Geol. 88, 85–98.

Ulrich, K.-U., Rossberg, A., Foestendorf, H., Zänker, H., Scheinost, A., 2006. Molecularcharacterization of uranium(VI) sorption complexes on iron(III)-rich acid minewater colloids. Geochim. Cosmochim. Acta 70, 5469–5487.

Waber, N., Schorscher, H.D., Peters, T., 1992. Hydrothermal and supergene uraniummineralization at the Osama Utsumi mine, Poços de Caldas.

Wang, Y., Frutchi, M., Suvorova, E., Phrommavanh, V., Descostes, M., Osman, A.A.A.,Geipel, G., Bernier-Latmani, R., 2013. Mobile uranium(IV)-bearing colloids in amining-impacted wetland. Nature Commun. 4, 2942.

Wellman, D.W., Icenhower, J.P., Gamerdinger, A.P., Forrester, S.W., 2006. Effects ofpH, temperature, and aqueous organic material on the dissolution kinetics ofmeta-autunite minerals, (Na, Ca)2–1[(UO2)(PO4)]2�3H2O. Am. Mineral. 91, 143–158.