Rare Earth Elements (REE)—Minerals in the Silius Fluorite Vein System

Ore Geology Reviews 74 (2016) 211–224

Contents lists available at ScienceDirect

Ore Geology Reviews

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

Rare earth elements (REE)—Minerals in the Silius ﬂuorite vein system (Sardinia, Italy)

N. Mondillo a,b,⁎,M.Bonia, G. Balassone a, S. Spoleto a, F. Stellato a,A.Marinoc, L. Santoro b,J.Sprattb a Dipartimento di Scienze della Terra, dell'Ambiente e delle Risorse, Università degli Studi di Napoli Federico II, via Mezzocannone 8, 80134 Napoli, Italy b Department of Earth Sciences, Natural History Museum, Cromwell Road, SW7 5BD London, UK c Fluorite di Silius S.p.A. Viale Merello 14, 09023 Cagliari, Italy article info abstract

Article history: The Silius vein system, located in SE Sardinia (Italy) is analogous to other late- to post-Hercynian mineral systems Received 1 July 2015 of this type in Europe. The Silius system consists of two main veins, characterized by several generations of fluo- Received in revised form 10 November 2015 rite, calcite and quartz, with initial ribbon-like geometries, followed by breccias and cockade-like textures. In this Accepted 11 November 2015 study, aimed at investigating the REE concentrations in the Silius vein system, a REE average of ~800 ppm (locally Available online 12 November 2015 ΣREE N1500 ppm) has been observed in the carbonate gangue of the fluorite orebody. These amounts are related to the presence of the REE-bearing minerals synchysite-(Ce) and xenotime-(Y). The chemical composition of Keywords: SE Sardinia synchysite-(Ce) has been obtained by wavelength dispersive spectrometry (WDS). The average synchysite- Hydrothermal veins (Ce) formula, built on the basis of (CO3)2F and 5 negative charges, is Ca1.07(La0.19,Ce0.36,Pr0.04,Nd0.15, REE Sm0.03,Gd0.03,Y0.13)(CO3)2F. From their geochemical characteristics, and their textural relationships with other Fluorite gangue phases, it is likely that synchysite-(Ce) and xenotime-(Y) formed at the same P-T-X conditions as the REE-minerals other minerals of the Silius fluorite mineralization. Synchysite-(Ce) and xenotime-(Y) at Silius could be related Carbonates to a local circulation phenomenon, where the REE are derived from a REE-bearing source rock in the basement of southeastern Sardinia, which has been leached by the same fluids precipitating the fluorite/calcite mineralization. REE concentrations contained in the carbonate gangue of still unexploited parts of the Silius vein deposit, as well as in dumps and tailings accumulated during past fluorite processing, could possibly represent a sub-economic by-product of the fluorite exploitation. © 2015 Elsevier B.V. All rights reserved.

1. Introduction REE concentrations have been also detected in several fluorite vein systems, not directly related to magmatism, occurring in the Hercynian Rare Earth Element (REE)-bearing minerals occur in many igneous, basement of Europe. Hercynian fluorite vein systems were emplaced in sedimentary, and metamorphic rocks, where they may be concen- the basement rocks between late Paleozoic and early Mesozoic, as a con- trated in ore deposits both related to igneous and hydrothermal pro- sequence of far-field tectonics (e.g. continental extension) associated cesses, or associated with sedimentary environments and weathering with the initial opening of the Atlantic Ocean, and circulation of brines (Chakhmouradian and Wall, 2012). REE-rich ores in carbonatites and within the basement (Muchez et al., 2005; Muñoz et al., 1999, 2005). peralkaline igneous rocks can be directly derived from magmatic pro- To date, REE analyses of these Paleozoic-Mesozoic systems have largely cesses, but also from metasomatism and hydrothermal remobilization focused at defining their occurrence and content in fluorite, caused by of magmatic REE minerals (Chakhmouradian and Zaitsev, 2012; Gysi ion-substitution processes (Dill et al., 2011; Möller, 1991; Möller et al., and Williams-Jones, 2013). Other REE-bearing mineral deposits can be 1976, 1994; Möller and Giese, 1997; Schwinn and Markl, 2005). Several associated with hydrothermal quartz- and fluorite-bearing veins of studies, carried out on fluorite veins in south Germany (e.g. Harz Moun- orthomagmatic derivation (Samson et al., 2004; Williams-Jones et al., tains, Schwarzwald, Bohemian Massif), revealed that the measured REE 2000). Typical weathering-related REE-mineral concentrations occur concentrations are associated with the occurrence of specific REE min- in placers, and in laterite caps and ion-absorption clay deposits erals (e.g. Dill et al., 2011; Gieré, 1996, and references therein; Haack (Chakhmouradian and Wall, 2012). et al., 1987; von Gehlen et al., 1986). Comparable research was also carried out on Triassic-Jurassic fluorite deposits in the Hercynian terrains of northern Africa (Bouabdellah et al., 2010; Cheilletz et al., 2010). ⁎ Corresponding author at: Dipartimento di Scienze della Terra, dell'Ambiente e delle We present here the results of an investigation aimed at evaluating Risorse, Università degli Studi di Napoli Federico II, via Mezzocannone 8, 80134 Napoli, fl Italy. the REE behavior and fractionation in the Silius uorite vein system, E-mail address: [email protected] (N. Mondillo). located in SE Sardinia, Italy. The Silius system is analogous to other

http://dx.doi.org/10.1016/j.oregeorev.2015.11.016 0169-1368/© 2015 Elsevier B.V. All rights reserved. 212 N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224 late- to post-Hercynian mineralization of this type in Europe (Boni et al., by low- to medium-grade metamorphism, and iii) an “inner zone”,in 2009; Natale, 1969). Fluorite and galena reserves were evaluated at 2 northern Sardinia, characterized by medium- to high-grade meta- million tons of raw material in 2006 (Castorina et al., 2008), when min- morphic signatures (Carmignani et al., 2001). ing operations ceased, but a new exploration campaign has been recent- The “external zone” successions (SW Sardinia), spanning in age from ly carried out, to find new fluorite resources. The REE contents in the early Cambrian to Devonian and Carboniferous, are represented by clas- Silius fluorite ore had been already reported by Castorina et al. (2008), tic and carbonate, mainly shallow water sedimentary rocks (Bechstädt in relation to a more general study on the different types of fluorite de- and Boni, 1994; Carmignani et al., 2001). posits in Sardinia. These authors used the REE content in fluorite to dis- The “nappe-zone” area, where the Silius mineralization is located, is tinguish between the different hydrothermal F-ore deposits in Sardinia, characterized by various lithotypes (Fig. 1B). The upper Cambrian- and compared their data with the REE contents of other Hercynian fluo- Ordovician successions mostly consist of siliciclastics (metasandstones, rite mineralizations in Europe. phyllites and quartzites), collectively grouped in the “Arenarie di San As for other classical studies on European fluorite deposits (e.g. Dill Vito” (Fig. 1B) and “Solanas” Formations (not outcropping in the study et al., 2011), we have re-evaluated the REE amounts contained in both area). These Formations are overlain by the so-called “Ordovician mag- fluorite and host rock. Moreover, we have also carried out a thorough matic and volcano-sedimentary complex” (Carmignani et al., 2001), mineralogical research on the various gangue components of the vein which consists of both effusive and intrusive igneous products, dis- system. The results of this study, aside from obtaining new data on the playing an almost complete arc-related calc-alkaline suite of middle to nature of the REE minerals in the Silius veins, have shed new light on late Ordovician age (Gaggero et al., 2012; Oggiano et al., 2010). In the the characteristics of this kind of fluorite ore and its gangue. Silius area the Ordovician volcano-sedimentary complex, named “Porfiroidi” Formation (Fig. 1B), is represented by rare andesitic lavas 2. Geological setting and abundant metasediments derived from the reworking of the for- mer, overlain by metarhyolites and metarhyodacites (Carmignani 2.1. Regional geology et al., 1994). The Ordovician magmatic and volcano-sedimentary complex is unconformably overlain by siliciclastic continental to neritic The Sardinia-Corsica Paleozoic basement represents a fragment of sediments, which were deposited during a period of active intraplate the Hercynian orogen (Arthaud and Matte, 1977; Carmignani et al., magmatism, resulting in alkali basalt flows, sills and dykes (late 1994; Crowley et al., 2000) that can be subdivided into three tectono- Ordovician-Silurian) (Gaggero et al., 2012; Oggiano et al., 2010) metamorphic zones (Fig. 1A): i) a “foreland zone”,outcroppingin (Fig. 1B). In the Silurian to Devonian period, the magmatic activity southwestern Sardinia, with low-grade or no metamorphism, ii) a stopped and pelagic sediments (Graptolitic Shales, or “Formazione “nappe-zone”, in southeastern and central parts of Sardinia, affected degli Scisti a Graptoliti”) were deposited (Fig. 1B). These consist of

Fig. 1. A) Major tectonic and metamorphic zones of the Hercynian basement in Sardinia (Italy) (Carmignani et al., 2001, modified); yellow star = Silius area. B) Geological sketch map of the Silius area (Nuova Mineraria Silius s.p.a., modified). N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224 213 black shales and pelagic limestones, and are interrupted at the top by veins are almost parallel in the topmost levels, with a spacing of 70 m the Carboniferous (Culm-type) flysch (Carmignani et al., 2001). on average, but they coalesce at a depth of 350 m a.s.l. in the NE sector, The “inner zone” is represented by a high-grade metamorphic com- and at a depth of 450 m a.s.l. in the SW sector of the mineralized area. plex, mainly consisting of amphibolites, which are southward bounded The San Giorgio vein was the first to be formed, and was displaced by an eclogite-bearing belt, interpreted as the suture zone of the and brecciated by the San Giuseppe vein. The San Giorgio vein contains Hercynian orogen (Carmignani et al., 1994). bands of chalcedony, pink fluorite, barite and calcite. The San Giuseppe The Hercynian compressive deformation was concluded by the em- vein is also banded, and shows several generations of fluorite, calcite placement of granite batholiths that occurred during the extensional and galena (Fig. 2A, B, C). The width of the veins ranges between 7 period associated with the collapse of the orogen (310–280 Ma; and 8 m in outcrop and 15–20 m in depth, where they are interconnect- Carmignani et al., 2001; Di Vincenzo et al., 1994). The entire basement ed (Fig. 2B). The Silius vein system has been mined from the surface was later intruded by Permian to Triassic magmatic dykes of bimodal (around 600 m a.s.l.) down to 200 m a.s.l., and explored to 100 m a.s.l. composition (Atzori and Traversa, 1986; Carmignani et al., 2001; The upper part of the vein system contains fluorite, barite and galena Ghezzo and Orsini, 1982). During this period, several hydrothermal with grades of 35%, 10% and 3%, respectively, whereas in the lower vein systems containing fluorite, barite and sulfides were developed in zones the amount of barite decreases, and the fluorite and galena grades SE Sardinia. The Silius mineralization corresponds to one of these vein can also reach 50% and 6%, respectively. Quartz and carbonates are the systems (Natale, 1969). Other post-Hercynian occurrences in the area main constituents of the gangue. Two less economically important are set within cataclastic and mylonitic belts, as the so-called “Filone veins (called San Giovanni and Davide) are also parallel to the other Argentifero del Sarrabus”, a complex association of stockworks and veins, and can be traced to the 500 m level. veins of Pb-Cu-sulfides and silver minerals, as well as the stibnite All the veins are mainly hosted by the “Porfiroidi” Formation and (±scheelite) vein systems of Su Suergiu and Corti Rosas (Belkin rarely by Late Ordovician sedimentary rocks of the Graptolitic Shales et al., 1984; Carmignani et al., 1978; Valera, 1974). Formation. The fluorite-barite mineralized veins cut locally the late- to post-Hercynian (Permian to Triassic felsic and mafic-intermediate) 2.2. The Silius hydrothermal veins magmatic dykes: this relationship defines a relative minimum age for the mineralization process (Natale, 1969). The veins are characterized The Silius vein system (Fig. 1B) crops out discontinuously for 2–3km by successive generations of fluorite, calcite and quartz, with initial along a NE-SW strike, between the Acqua Frida area (39°30′46″N-9°15′ ribbon-shaped geometries, followed by breccias and cockade-like tex- 05″E), the Muscadroxiu shaft (39°31′00″N-9°15′18″E) and the Genna tures (Fig. 2D). The different generations are always accompanied by Tres Montis shaft (39°31′38″N-9°16′00″E), at an elevation ranging be- sulfides (marcasite, sphalerite and galena), and in the late stages by bar- tween 600 and 700 m a.s.l. At the 400 m level, the veins reach a maxi- ite (Natale, 1969). mum length of 4 km. In previous fluid inclusion studies on the Silius mineralization, it was The system consists of two main veins, called “San Giorgio” and “San reported that most ore minerals precipitated at temperatures in the Giuseppe” (Fig. 2A), which have an almost vertical dip (about 75°), and range of 120–180 °C from fluids with salinities reaching up to a strike of N045E along the main sector Acqua Frida - Genna Tres ~18wt.%NaClequiv.(Boni et al., 2009). Due to the very low first melting Montis, and N065E northeastward from Genna Tres Montis. These two temperatures (~ −50 °C) recorded in both fluorite and calcite, the

Fig. 2. A) Silius Mine — production gallery at 235 m a.s.l.: ribbon structure (fluorite, calcite, galena) of the San Giuseppe vein (Boni et al., 2009, modified). B) Green-purple banded fluorite of the San Giorgio vein. C) & D) San Giuseppe vein (zoned fluorite, calcite/dolomite, galena) cutting and partly replacing the San Giorgio vein: chalcedony-rich fragments of the San Giorgio vein occur as breccia clasts cemented by San Giuseppe calcite (Boni et al., 2009, reinterpreted). 214 N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224 authors suggested the involvement of Ca-rich fluids in the mineraliza- 1000 °C. Rare earth and refractory elements were determined by ICP- tion process. Lead isotopes of sulfides in the Silius mineralization have MS following a LiBO2/Li2B4O7 fusion and nitric acid digestion. In addi- been determined as intermediate between the compositions of Ordovi- tion, a separate volume split was digested in Aqua Regia and analyzed cian metavolcanic rocks (“Porfiroidi” Formation) and Hercynian gran- by ICP-MS to detect the precious and base metals. Lower detection ites (Boni et al., 2009). Sr and Nd isotopes of Silius fluorite, measured limits obtained in the two laboratories are reported in Tables 1, 2 and 4. by Castorina et al. (2008), have a signature coherent with a brine circulation within both Paleozoic metasedimentary rocks and Hercynian 4. Mineralogy and major element geochemistry granitoids. 4.1. Fluorite 3. Sampling and analytical methods The analyzed samples consist of medium-sized (b1 cm) white/violet The samples analyzed in this study were mainly collected in both the fluorite crystals, associated with calcite, quartz, barite and a few minute San Giuseppe and the San Giorgio veins, along the two vertical sections sulfide intergrowths. In thin section fluorite appears of different gener- of the Muscadroxiu and the Central shafts of the mine (the latter is lo- ations, each showing a strong blue luminescence under CL (Fig. 3). As cated between the Muscadroxiu and the Genna Tres Montis shafts) at reported in the literature (Natale, 1969), it is possible to distinguish at every 20 m, from 100 m a.s.l. up to 200 m a.s.l. At each sampling site, least three fluorite generations: a first and more extensive generation we collected fluorite, calcite, and fragments of the host rock. Additional consisting of fluorite alone (Fig. 3A, B), a second generation of fluorite sampling was carried out at the Acqua Frida shaft, at 300 m elevation, with a calcite gangue (Fig. 3C, D), and a third one associated with quartz where we collected fluorite and calcite from both veins. In Table A1 and barite (Fig. 3E, F). (Supplementary material) are listed all the analyzed samples. Minor amounts of SiO , or various values of LOI or tot-C, detected by Fluorite and calcite were cleaned of the impurities (sulfides and Fe- 2 bulk chemical analyses (Table 1), are related to intergrown quartz or oxides) by handpicking under binocular microscope. The purity of the carbonate minerals in fluorite, also revealed by XRD. The most contam- minerals was tested with X-ray diffraction analysis, by using a Seifert– inated samples have not been considered when discussing the relation- GE ID3003 diffractometer (Dipartimento di Scienze della Terra, ships between REE sorption processes and genetic conditions. dell'Ambiente e delle Risorse-DiSTAR, Università degli Studi di Napoli Federico II, Italy), with CuKα radiation, Ni-filtered at 40 kV and 30 mA, 3–80 °2θ range, step scan 0.02°, time 10 s/step, and the RayfleX (GE) 4.2. Calcite and dolomite software package; a silicon wafer was used to check the instrumental setting. Sample holder was a zero-background plate of quartz crystal Mineralogical analyses have been conducted on several samples of cut and polished 6° of the c-axis. the carbonate gangue, in order to verify if this consisted solely of calcite, For polished thin section preparation, the samples were impregnat- as reported in the literature. X-ray diffraction clearly confirm that the ed with Araldite D and Raku Hardener EH 2950 (OMT Laboratory, Aosta, carbonates of both the San Giorgio and San Giuseppe veins consist of a Italy). Cathodoluminescence microscopy (CL) was carried out with a mixture of calcite and dolomite/ferroan dolomite, not yet reported for Hot Cathode Instrument at the Heidelberg University. Secondary elec- this vein system. The carbonate gangue can locally contain also small ra- tron imagining by scanning electron microscopy (SEM), and energy dis- dial aggregates of newly formed kaolinite and quartz. persion spectrometry (EDS) investigations on thin sections were carried Calcite has yellow-orange colors under CL (Fig. 3D), and generally out with a Jeol JSM 5310, equipped with the INCA X-stream pulse pro- appears zoned. This indicates that variable amounts of Mn2+ have cessor and the 4.08 version Inca software (Oxford Instruments detector) substituted for Ca2+ in the calcite lattice (Götze, 2012). Under micro- (DiSTAR, Università di Napoli, Italy), operating at 15 kV primary beam scopic observation, dolomite replaces calcite in patches (Fig. 3D). voltage, 50–100 mA filament current, variable spot size and 50 s net ac- Replacive dolomite shows an extremely low luminescence under CL quisition time. Reference standards were: albite (Si, Al, Na), orthoclase and consequently, when associated with calcite, it appears generally (K), wollastonite (Ca), diopside (Mg), almandine (Fe), rutile (Ti), barite dark (Fig. 3D, F). Where the replacement is strongly pervasive, the dolo- (Ba), strontianite (Sr), eskolaite (Cr), rhodonite (Mn), pyrite (S), sphal- mite colors under CL are dark-reddish. According to Götze (2012),the erite (Zn), galena (Pb), fluorite (F), apatite (P), sylvite (Cl), smithsonian red CL color of dolomite is generally due to variable amounts of phosphates (La, Ce, Nd, Sm, Y), pure vanadium (V) and Cornig glass (Th, Mn2+ ⇆ Mg2+ substitution in its structure. The low luminescence in- U). Analytical errors are 1% rel. For major elements and 3% rel. For minor tensity of the Silius dolomite is related to the presence of discrete elements. Fe2+ amounts in the cationic sites, which quench the luminescence ef- Quantitative data sets of selected samples were obtained by wave- fects related to Mn2+ (Götze, 2012). length dispersion spectrometry (WDS), using a Cameca SX100 electron SEM analyses confirmed that dolomite is not pure, but rather con- microprobe operating at 15 kV, 15 nA, and 10 μm spot size (Natural His- sists of the ferroan variety and ankerite, with Fe contents up to tory Museum, London, UK). Compositions of standards used are report- 23 wt.% FeO, and Mn up to 4 wt.% MnO (Fig. 4). Fe-bearing dolomite ed in Table A2 (Supplementary material). Interference corrections were has a mottled texture (Fig. 4A). Two replacive generations have been carried out prior to matrix correction for the following elements: F detected: an initial almost pure dolomite phase (b5 wt.% FeO, ~1 wt.% (overlapping element, Ce), Nd (Ce), Sm (Ce), Ba (Pr) and La (Nd). A MnO), which is followed by a high-Fe dolomite phase (N6 wt.% FeO, monazite (Manangotry) standard (composition and determinations in N1 wt.% MnO) and eventually by ankerite (Fig. 4B, C).

Table A3 - Supplementary material) was used as a quality control sam- Plotted on a diagram deﬁnedbyCaMg(CO3)2-CaFe(CO3)2- ple. WDS detection limits for each element are reported in Table 3. CaMn(CO3)2 (Fig. 5), the analyzed phases have a composition covering Details on WDS analyses are provided in the Appendix. the ﬁeld between pure dolomite and ferroan dolomite, with an almost

Major, minor and trace (REE) elements in ﬂuorite, carbonate min- continuous substitution between Ca1.08Mg0.86Fe0.03Mn0.03(CO3)2 and erals and host rock samples have been measured at the ACME laborato- Ca1.03Mg0.42Fe0.46Mn0.09(CO3)2. Only two outlier data show a more def- ries (Vancouver, Canada) and at the Activation Laboratories (Ancaster, inite ankeritic composition: Ca1.01Mg0.19Fe0.68Mn0.12(CO3)2 and Ca1.01 Canada). In both the laboratories, the same protocol was adopted: the Mg0.14Fe0.73Mn0.13(CO3)2. samples were pulverized to 85% -200 mesh, to obtain about 20 g of Chemical analyses on the carbonate gangue corroborate the Fe-rich pulp. Major oxides and several minor elements were analyzed by ICP- composition of the dolomite. In fact, several samples from the gangue of

OES following a LiBO2/Li2B4O7 fusion and dilute nitric digestion. Loss the ﬂuorite mineralization (Table 2) contain variable Mg and Fe amounts, on ignition (LOI) was calculated by weight difference after ignition at which result from the mixture of different phases, as calcite, dolomite, Fe- N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224 215

Table 1 Major (wt.%) and minor to trace elements (ppm) compositions of Silius ﬂuorite (ICP-OES-MS on bulk sample).

FL-01 FL-02 FL-03 FL-04 FL-05 FL-06 lower detection FL-07 FL-08 FL-09 FL-10 FL-11 FL-12 Lower limita detection limitb

SiO2 2.39 2.94 1.67 32.15 0.42 14.51 0.01 1.60 4.37 3.48 0.27 6.29 0.20 0.01

Al2O3 0.05 0.01 b0.01 0.01 b0.01 b0.01 0.01 0.08 0.06 0.03 0.02 0.04 0.02 0.01

Fe2O3t 0.10 0.12 0.08 0.64 0.08 0.32 0.04 0.03 0.03 0.08 0.03 0.15 0.04 0.01 MgO b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 0.01 0.03 0.01 0.11 0.02 0.02 0.01 0.01

Na2O 0.01 b0.01 b0.01 0.01 b0.01 0.01 0.01 b0.01 0.01 0.01 0.01 0.03 b0.01 0.01

K2O 0.02 b0.01 b0.01 b0.01 b0.01 b0.01 0.01 b0.01 b0.01 b0.01 b0.01 0.02 b0.01 0.01

TiO2 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 0.01 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 0.001

P2O5 b0.01 b0.01 b0.01 0.03 0.02 b0.01 0.01 b0.01 0.01 b0.01 0.02 0.03 0.03 0.01 MnO b0.01 b0.01 b0.01 b0.01 b0.01 0.13 0.01 0.005 0.005 0.013 0.005 0.088 0.004 0.001 LOI 1.20 1.70 1.20 3.40 0.60 11.40 - 1.13 1.45 1.19 0.80 6.89 0.54 - F 40.00 46.54 43.92 31.67 29.31 27.88 0.01 Not measured Cac 50.22 49.80 51.24 34.18 51.47 41.40 0.01 52.69 51.34 51.30 53.64 47.61 53.52 0.01 Sum 93.99 101.11 98.11 102.09 81.90 95.65 55.56 57.29 56.21 54.81 61.17 54.36

Sc b1 b1 b1 b1 b1 b11 1 1 b 11221 Ba 7 5 2 15 3 564 1 9 15 86 8 14 10 3 Nb 0.3 b0.1 b0.1 0.2 0.2 0.3 0.1 183 101 71 59 37 40 1 Rb 0.8 0.2 0.2 0.3 b0.1 b0.1 0.1 b2 b2 b2 b2 b2 b22 Sr 78.3 56.6 64.5 300.0 12.7 44.3 0.5 57 76 18 23 37 12 2 V b817b8 b8 13 17 8 11 12 11 10 11 10 5 W b0.5 b0.5 0.8 4.3 0.8 b0.5 0.5 2 2 b12121 Zr2.20.40.20.61.10.40.1 9876754 Mo 0.5 0.4 0.3 1.1 0.4 0.5 0.1 b2 b2 b2 b2 b2 b22 Cu 2.1 1.6 1.8 6.4 1.8 26.6 0.1 b10 b10 10 b10 30 30 10 Pb 2.0 1.4 1.4 1.9 1.7 4.3 0.1 b5 b5 b5 b594b55 Zn 4 3 8 54 5 17 1 b30 b30 b30 b30 b30 b30 30 As 0.8 0.7 b0.5 0.9 1.0 1.6 0.5 b5 b5 b5 b5 b5 b55 Sb b0.1 b0.1 b0.1 0.3 b0.1 0.1 0.1 b0.5 b0.5 b0.5 b0.5 b0.5 0.8 0.5

Y 29.8 24.7 135.9 126.9 182.2 96.4 0.1 37 26 88 135 171 203 2 La 3.5 3.7 5.9 31.9 10.5 35.0 0.1 5.0 3.1 9.4 9.4 79.8 17.0 0.1 Ce 6.8 6.6 14.3 51.5 13.3 67.2 0.1 10.0 6.9 15.4 18.5 166.0 21.8 0.1 Pr 0.87 0.85 1.97 5.87 1.34 7.92 0.02 1.33 0.94 1.98 2.73 21.30 2.31 0.05 Nd 3.9 3.0 10.0 23.3 5.3 32.3 0.3 5.6 4.4 10.5 15.4 89.4 8.9 0.1 Sm 1.35 1.06 3.85 6.33 1.62 9.16 0.05 2.1 1.4 3.3 5.9 27.9 2.4 0.1 Eu 1.56 1.53 2.52 3.35 0.90 3.09 0.02 1.81 1.65 2.03 3.50 8.28 0.90 0.05 Gd 2.25 1.85 7.62 9.83 4.33 11.88 0.05 2.7 1.9 6.1 11.1 28.4 4.3 0.1 Tb 0.51 0.46 1.74 2.53 0.93 1.81 0.01 0.6 0.4 0.9 1.5 4.3 0.8 0.1 Dy 3.62 3.43 12.82 19.17 6.76 8.87 0.05 4.0 3.1 4.3 6.9 19.0 5.8 0.1 Ho 0.78 0.72 2.93 4.40 1.69 1.35 0.02 0.8 0.7 0.7 1.1 2.8 1.2 0.1 Er 2.39 1.96 8.69 13.13 4.00 3.05 0.03 2.3 2.0 1.5 2.0 5.5 2.8 0.1 Tm 0.30 0.32 1.08 1.76 0.45 0.29 0.01 0.37 0.33 0.12 0.17 0.49 0.28 0.05 Yb 2.00 2.10 6.07 10.25 1.83 1.42 0.05 2.3 2.2 0.4 0.6 2.1 1.0 0.1 Lu 0.27 0.30 0.74 1.28 0.19 0.22 0.01 0.32 0.35 0.05 0.06 0.25 0.11 0.04 Y/Ho 38.2 34.3 46.4 28.8 107.8 71.4 46.3 37.1 125.7 122.7 61.1 169.2 La/Ho 4.5 5.1 2.0 7.3 6.2 25.9 6.3 4.4 13.4 8.5 28.5 14.2 Eu/Eu* 4.2 5.1 2.2 2.0 1.6 1.4 3.6 4.8 2.1 2.0 1.4 1.3

(La/Yb)n 0.1 0.1 0.1 0.2 0.4 1.8 0.2 0.1 1.7 1.2 2.8 1.3 Ce/Ce* 0.9 0.8 0.9 0.8 0.8 0.9 0.9 0.9 0.8 0.8 0.9 0.8

Fe2O3t = total iron;

Eu/Eu* = [Eun/√(Smn ·Gdn)];

(La/Yb)n =Lan/Ybn;

Ce/Ce* = [Cen/√(Lan ·Prn)]. a ACME Laboratories. b Activation Laboratories. c Ca is reported as element (not oxide) because it mostly occurs in fluorite, but the element also occurs as CaO in intergrown carbonate minerals. rich dolomite and ankerite. These samples have a very similar composi- In the analyzed samples, synchysite-(Ce) occurs as acicular and tion, though they were collected in different zones of the mine. tabular crystals (Fig. 4C, D), with an average size below 100 μm (Fig. 4E, F). In detail, synchysite-(Ce) has been observed both as sin- 4.3. REE-minerals gle crystals and as composite aggregates in two main locations: first- ly within the dolomitized calcite and Fe-bearing dolomite (Fig. 4C, Abundant REE-minerals have been detected within the carbonate D), and secondly within quartz veinlets and in cavities occurring gangue of both Silius mineralized veins (Fig. 4C, D, E, F, G, H). among the calcite and dolomite patches (Fig. 4E). In the first The mineralogy of these phases is quite simple and is mostly repre- setting, synchysite-(Ce) is distributed throughout the irregularly sented by synchysite-(Ce) (Table 3), a Ca-REE-fluorcarbonate dolomitized carbonate gangue, while it is almost absent in unaltered

[CaREE(CO3)2F], and by rare xenotime-(Y) (YPO4). These minerals calcite (Fig. 4D). When synchysite-(Ce) occurs in the quartz veinlets, have been observed only under electronic microscope, because it is paragenetically followed by the third ﬂuorite generation (Fig. 4F, G). their amounts is below (or very near to) the XRD detection limits No REE minerals have been found as inclusions in the ﬂuorite crystals of adopted in this study (b1wt.%). the main generation that have been selected for ICP analyses. 216 N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224

Table 2 Major (wt.%) and minor to trace elements (ppm) compositions of carbonate gangue of Silius veins (ICP-OES-MS on bulk samples).

CA-01 CA-02 CA-03 CA-04 Lower detection CA-05 CA-06 CA-07 CA-08 Lower limita detection limitb

SiO2 0.37 1.79 0.57 2.70 0.01 3.74 2.75 1.50 0.02 0.01

Al2O3 0.06 0.44 0.07 b0.01 0.01 0.02 0.02 0.02 b0.01 0.01

Fe2O3t 0.46 0.95 2.14 9.61 0.04 8.50 13.26 5.45 0.66 0.01 MgO 0.04 2.16 5.51 10.67 0.01 13.60 11.51 10.12 0.07 0.01 CaO 55.09 51.17 46.45 32.94 0.01 30.86 28.38 36.57 54.66 0.01

Na2O b0.01 b0.01 0.03 0.02 0.01 b0.01 b0.01 b0.01 b0.01 0.01

K2O b0.01 b0.01 0.02 b0.01 0.01 b0.01 b0.01 b0.01 b0.01 0.01

TiO2 b0.01 b0.01 b0.01 b0.01 0.01 b0.001 b0.001 b0.001 b0.001 0.001

P2O5 b0.01 b0.01 b0.01 b0.01 0.01 b0.01 b0.01 b0.01 b0.01 0.01 MnO 0.51 0.54 1.05 1.37 0.01 1.52 1.81 1.37 0.48 0.001 LOI 43.3 42.9 43.9 42.6 - 40.07 40.69 43.47 43.43 - F 0.10 0.06 0.08 0.15 0.01 Not measured Sum 99.83 99.95 99.82 99.91 98.31 98.42 98.50 99.32

Sc 22 15 2 7 1 1 2 10 2 1 Ba 10 2 958 6 1 16 2936 920 3124 3 Nb b0.1 b0.1 b0.1 b0.1 0.1 b1 b1 b1 b11 Rb 0.1 b0.1 1.1 0.1 0.1 b2 b2 b2 b22 Sr 31.1 18.4 45.2 18.4 0.5 22 58 38 71 2 V b8 b8 b8 b88 b5 b5765 W b0.5 b0.5 b0.5 b0.5 0.5 4 4 1 b11 Zr 0.2 0.3 1.3 0.4 0.1 b4 b4 b4 b44 Mo 0.1 0.3 0.1 0.3 0.1 b2 b2 b2 b22 Cu 2.0 1.3 1.4 1.3 0.1 20 10 b10 10 10 Pb 1.5 0.8 0.9 4 0.1 6 b523b55 Zn 3 13 13 32 1 80 50 40 b30 30 As b0.5 0.9 1.4 1 0.5 b5 b5 b5 b55 Sb b0.1 b0.1 b0.1 b0.1 0.1 b0.5 b0.5 b0.5 b0.5 0.5

Y 116.6 68.1 148.8 69.6 0.1 58 178 247 45 2 La 185.5 126.2 233.3 120.4 0.1 80.8 395 313 22.9 0.1 Ce 330.5 230.5 400.5 218.9 0.1 132 685 628 55.1 0.1 Pr 32.24 24.91 43.28 23.75 0.02 14.90 76.40 75.40 7.45 0.05 Nd 113.4 94.9 165.0 90.2 0.3 57 289 298 31 0.1 Sm 22.56 22.03 33.90 19.68 0.05 12.1 62.7 71.8 9.9 0.1 Eu 6.50 7.53 21.51 8.97 0.02 6.73 31.90 30.30 3.61 0.05 Gd 24.64 20.93 35.46 21.32 0.05 12 60.2 73.2 9.9 0.1 Tb 4.24 3.17 4.71 2.76 0.01 1.7 7.1 9.3 1.8 0.1 Dy 23.92 15.62 25.12 12.40 0.05 10 34.3 47.4 11.2 0.1 Ho 4.58 2.71 4.45 1.94 0.02 1.9 5.4 7.5 2.2 0.1 Er 12.30 6.13 11.12 4.71 0.03 5.2 12.3 16.7 6.8 0.1 Tm 1.88 0.92 1.44 0.61 0.01 0.63 1.35 1.88 1.15 0.05 Yb 13.22 6.12 8.07 4.13 0.05 3.6 6.7 9.9 8.6 0.1 Lu 1.76 0.85 1.03 0.58 0.01 0.47 0.86 1.23 1.31 0.04 Eu/Eu* 1.3 1.7 2.9 2.1 2.6 2.4 2.0 1.7

(La/Yb)n 1.0 1.5 2.1 2.2 1.7 4.4 2.3 0.2 Ce/Ce* 1.0 0.9 0.9 0.9 0.9 0.9 0.9 0.9

Fe2O3t = total iron;

Eu/Eu* = [Eun/√(Smn ·Gdn)];

(La/Yb)n =Lan/Ybn;

Ce/Ce* = [Cen/√(Lan ·Prn)]. a ACME Laboratories. b Activation Laboratories.

Table 3 shows the chemical composition of synchysite-(Ce). REE 5. REE geochemistry concentration in synchysite varies to a large extent, i.e. La ranges between 6.58 and 13.03 wt.% La2O3, Ce occurs in an interval between 5.1. Fluorite 15.90 and 21.10 wt.% Ce2O3, and Y between 2.58 and 8.45 wt.% Y2O3.It should be noticed that Gd concentration can be higher than Sm concen- Excluding the fluorite samples contaminated by intergrown gangue tration, as also shown in the WDS spectrum by the peak intensity of the minerals, the REE amounts determined by ICP on bulk fluorite are solely Gd Lβ1 peak marginally greater than that of the Sm Lβ1 peak (Fig. 6). related to REE diffused in its crystal structure. In fact, no discrete REE The formula of synchysite-(Ce) has been built on the basis of minerals have been detected within the analyzed fluorite crystals in

(CO3)2F and 5 negative charges, and is given in Table 3.Theaverage contrast to the carbonate gangue. synchysite-(Ce) formula is Ca1.07(La0.19,Ce0.36,Pr0.04,Nd0.15,Sm0.03, The REE contents of the Silius fluorite (Table 1) have been nor- Gd0.03,Y0.13)(CO3)2F. malized to the post-Archean Australian Shale (PAAS; Taylor and Xenotime-(Y), similarly to synchysite-(Ce), occurs as subhedral McLennan, 1985), in order to compare their distribution with that ob- crystals filling cavities within calcite and dolomite (Fig. 4H). The few an- served in other deposits of the same type. REE in Silius fluorite are alyzed specimens (EDS) are characterized by Dy and Yb contents of generally lower than PAAS. It is possible to observe that the REE concen- ~3 wt.%. trations normalized to PAAS have a roof-shaped pattern, characterized N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224 217

Fig. 3. A) & B) First fluorite generation, luminescing blue under CL. C) & D) Cubic crystal of the second fluorite generation in carbonate gangue: calcite (yellow-orange under CL) is replaced by dolomite (dark under CL). E) & F) Third fluorite generation consisting of microcrystals associated with quartz and barite in a veinlet, cutting the carbonate gangue (dark-reddish under CL). by positive Eu and Y anomalies (Fig. 7A). The Eu anomaly is also The bulk carbonate gangue is characterized by REE contents notably reflected in the Eu/Eu* ratio, which is always greater than 2, except in higher than those in fluorite (Table 2); in two of the considered samples, samples FL-06 and FL-11 (contaminated by quartz and carbonates), the total REE amounts are over 1500 ppm. In the diagram of PAAS- and in FL-05 and FL-12, which have no contamination. The patterns normalized REE-contents (Fig. 7B), it is possible to notice that the aver- are asymmetric, because the majority of the samples have HREE con- age pattern is almost flat for the carbonate samples, with positive drops tents very near to PAAS, whereas the LREE are strongly depleted; in in the Eu and Y positions. In detail, the CA-08 sample is the only one fact, several samples have a La/Ybn ratio b1. The sample FL-3 is the where HREE are more enriched than LREE, and LREE are less abundant only not-contaminated specimen to have the MREE and HREE more than PAAS. The other samples are characterized by an enrichment of enriched than PAAS, and the LREE more depleted than PAAS (La/Ybn both LREE and HREE alike (La/Ybn ratios around 1), or by a LREE enrich- ratio ~ 0.03). Other samples (Fl-06, Fl-09, FL-10) have a La/Ybn ratio ment higher than HREE (La/Ybn ratios are around 2; only the sample CA- slightly N1, which is too small to be considered as proof for LREE/ 06 has a La/Ybn ratio slightly N4). The Eu positive anomaly is marked by HREE fractionation. Moreover, these samples are characterized by nor- Eu/Eu* ratios between 1.3 and 2.92. malized MREE values, which are higher than the LREE and HREE values. 5.3. Vein system host rocks 5.2. Carbonates Two measured samples from the “Porfiroidi” Formation, which is Significant amounts of REE-bearing mineral inclusions, trapped in the main host rock of the Silius deposit, have a REE content around calcite and dolomite, clearly influence the REE concentrations deter- 100 ppm ΣREE (Table 4). One of these samples, P6 is clearly contam- mined by ICP on the bulk samples of the carbonate gangue. However, inated with fluorite. Among the other types of host rocks, the highest the amounts of REE possibly hosted in the calcite and dolomite structure REE contents have been detected in the samples consisting of late- are lower than the WDS detection limits. Hercynian porphyritic dykes, whereas the metasedimentary black 218 N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224

Fig. 4. A) Fe-bearing dolomite crystal showing a mottled texture. B) Calcite replaced by dolomite, which locally reaches ankerite composition. C) & D) Acicular and tabular crystals of synchysite irregularly distributed throughout the dolomite, Fe-bearing dolomite and ankerite gangue. E) Acicular crystals of synchysite within quartz veinlets and in cavities between calcite and dolomite. F) Synchysite across dolomitized calcite, quartz and ﬂuorite of the third generation. G) Synchysite aggregate of crystals, clearly cemented by ﬂuorite. H) Subhedral crystal of xenotime cutting across calcite and dolomite. Backscattered images (BSE) from SEM-EDS. N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224 219

Fig. 6. Synchysite-(Ce) WDS spectrum showing intensities of Sm and Gd Lβ1peaks.Note no peak overlap correction or background subtraction has been applied.

Fig. 5. Dolomite and Fe-bearing dolomite chemical compositions plotted in the ﬂ CaMg(CO3)2-CaFe(CO3)2–CaMn(CO3)2 diagram (EDS analyses). PAAS. In general, two types of patterns can be observed: at patterns and LREE-slightly depleted patterns (La/Ybn b 1). All of them do not show spikes or anomalies, either positive or negative. This behavior shales have very low REE values. All the host rocks have PAAS- is different from that of ﬂuorite, which has convex patterns with Eu normalized REE patterns depleted in REE, in comparison to the and Y positive anomalies, and from the carbonate gangue, which is adopted reference (Fig. 7C). Only a sample of basic-intermediate generally characterized by La/Ybn ratios N1 and by Eu and Y positive porphyry has REE concentrations that are on average higher than anomalies (Fig. 7).

Table 3 Chemical compositions of synchysite-(Ce) in the Silius veins.

Sample CA-1 CA-1 CA-1 CA-1 CA-2 CA-2 CA-2 CA-2 CA-3 CA-3 CA-3 CA-3

Crystal agglomerate ID T_6 T_7 C_05 C_05 N_2 N_5 N_5 N_8 3_01 3_02 3_03 3_03

Crystal ID 1/1. 1/2. 1/5. 1/10. 1/2. 1/1. 1/3. 1/3. 1/3. 1/1. 1/1. 1/3.

Detection wt.% limit

CaO 18.14 18.09 18.08 17.92 18.70 19.37 18.12 18.68 19.51 20.67 20.42 19.15 0.03 MnO 0.10 0.04 b.d.l. b.d.l. 0.09 0.02 0.03 b.d.l. b.d.l. 0.05 0.05 0.04 0.03

FeOt 0.44 0.26 0.02 0.01 0.59 0.13 0.00 0.04 b.d.l. 0.14 0.01 0.04 0.03

Y2O3 2.58 4.18 4.81 5.48 6.08 8.45 7.45 4.37 5.03 4.01 3.58 3.52 0.06

La2O3 13.03 12.11 10.39 9.88 9.61 6.58 7.17 10.10 10.50 10.17 10.49 9.76 0.05

Ce2O3 21.10 20.47 18.27 18.42 19.01 15.90 17.08 19.75 19.76 19.21 19.83 19.76 0.07

Pr2O3 2.06 2.01 2.12 2.06 1.97 1.95 2.17 2.13 2.09 2.00 2.10 2.12 0.08

Nd2O3 8.00 8.03 9.37 8.40 7.73 8.34 9.14 8.47 8.19 7.81 8.40 8.77 0.07

Sm2O3 1.41 1.49 2.64 2.42 1.79 2.50 2.63 1.93 1.68 1.59 1.68 1.71 0.05

Gd2O3 1.17 1.40 2.68 2.65 1.80 2.85 2.99 1.89 1.53 1.48 1.50 1.51 0.03 F measured 7.12 5.98 4.71 6.56 6.07 5.01 5.63 5.89 6.01 5.48 5.65 5.60 0.46 F calculateda 6.04 6.06 6.10 6.01 6.13 6.11 6.05 6.04 6.16 6.16 6.19 6.04 a CO2 27.99 28.09 28.26 27.83 28.41 28.30 28.01 28.00 28.56 28.55 28.68 27.97 –O=F −2.54 −2.55 −2.57 −2.53 −2.58 −2.57 −2.55 −2.54 −2.60 −2.60 −2.61 −2.54 Total 99.51 99.66 100.53 98.78 99.33 97.92 98.27 98.85 100.41 99.25 100.33 97.84

apfu on the basis of (CO3)2F Ca 1.017 1.011 1.004 1.011 1.033 1.074 1.015 1.047 1.072 1.136 1.117 1.074 Mn 0.004 0.002 ––0.004 0.001 0.001 ––0.002 0.002 0.002 Fe 0.019 0.011 0.001 0.000 0.025 0.006 0.000 0.002 – 0.006 0.001 0.002 Σ 1.041 1.024 1.005 1.011 1.063 1.081 1.016 1.049 1.072 1.144 1.120 1.078

Y 0.072 0.116 0.133 0.153 0.167 0.233 0.207 0.122 0.137 0.109 0.097 0.098 La 0.252 0.233 0.199 0.192 0.183 0.126 0.138 0.195 0.199 0.192 0.198 0.188 Ce 0.404 0.391 0.347 0.355 0.359 0.301 0.327 0.378 0.371 0.361 0.371 0.379 Pr 0.039 0.038 0.040 0.039 0.037 0.037 0.041 0.041 0.039 0.037 0.039 0.041 Nd 0.150 0.150 0.173 0.158 0.142 0.154 0.171 0.158 0.150 0.143 0.153 0.164 Sm 0.025 0.027 0.047 0.044 0.032 0.045 0.047 0.035 0.030 0.028 0.030 0.031 Gd 0.020 0.024 0.046 0.046 0.031 0.049 0.052 0.033 0.026 0.025 0.025 0.026 Σ 0.962 0.978 0.985 0.988 0.951 0.944 0.983 0.961 0.952 0.896 0.913 0.927

Σcations 2.003 2.002 1.990 1.998 2.013 2.024 2.000 2.010 2.024 2.041 2.033 2.005 a Determined by stoichiometry. 220 N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224

in Germany (e.g. Harz Mountains, and Schwarzwald), where it is associated with bastnäsite (Gieré, 1996, and references therein; Haack et al., 1987; von Gehlen et al., 1986). Small occurrences of synchysite-(Ce) in Sardinia are only recorded in association with molybdenite within Hercynian leucogranites in the small prospects of Su Senargiu (Sarroch) and Perda Pibera (Gonnosfanadiga) (Boni et al., 2003; Brizzi et al., 1994; Orlandi et al., 2013; Stara et al., 1996). In hydrothermal environments REE are transported mainly as non- fluoride complexes (e.g. chloride species) in fluorine-bearing solutions (Migdisov and Williams-Jones, 2014; Williams-Jones et al., 2012). These authors proposed a model in which fluoride ions would work as a binding ligand that promotes REE mineral deposition, rather than as an agent for REE transport. With this background, the co-genetic association of REE minerals and fluorite would be explained by the low solubility of both fluorite and REE minerals. Then, the ore-forming hydrothermal fluids carry appreciable concentrations of REE (and fluorine in the form of HF) in relatively acid conditions. The REE deposition may occur when the neutralization of the acidity of the solution induces the precipitation of fluoride-bearing minerals that in this model are represented only by bastnäsite. This happens, for example, when fluids in- teract with carbonate host rocks (e.g. dolomite or limestone), which buffer the pH to higher values, leading to the precipitation of the REE fluorcarbonate bastnäsite, according to the following reaction (Williams-Jones et al., 2012):

2þ þ þ − ¼ þ þ þ − ð Þ REECl HF HCO3 REECO3F 2H Cl 1 − Any mechanism that produces an increase in pH and/or HCO3 activity, or a decrease in Cl− activity will lead to the deposition of this mineral. If the host rock of the mineralization consists of dolomite, the ore ﬂuid will react with the carbonate, providing an effective mechanism to precipitate the ore minerals (Williams-Jones et al., 2012):

þ þ ð Þ ¼ 2þ þ 2þ þ − ð Þ 2H CaMg CO3 2 Ca Mg 2HCO3 2 Fig. 7. A) REE concentrations in ﬂuorite bulk samples, normalized to PAAS (post-Archean − Australian Shale; Taylor and McLennan, 1985). B) REE concentrations in carbonate gangue This reaction produces a sharp increase in both pH and HCO3 activ- bulk samples, normalized to PAAS. C) REE concentrations in host rock samples, normalized ity. Concerning synchysite-(Ce), we may propose, by analogy with to PAAS. ICP-OES-MS analyses. bastnäsite, a possible reaction for its formation based on the chemical reaction (1):

REECl2þ þ Ca2þ þ HF þ 2HCO− ¼ CaðREEÞðCO Þ F þ 3Hþ þ Cl−; ð3Þ 6. Discussion 3 3 2 where the acid solution is buffered as in reaction (2). 6.1. New insights on the mineralogy of the Silius vein system: Xenotime is a common accessory phase in granites and pegmatites, are REE-bearing minerals compatible with this mineralization type? and is a common detrital mineral in siliciclastic sedimentary rocks; however, it may also form during hydrothermal processes (Brown et al., 2002; In the previously published studies on the Silius mineralization (Boni Cook et al., 2013; Kositcin et al., 2003; Schaltegger et al., 2005). In Sardin- et al., 2009; Castorina et al., 2008; Natale, 1969), the deposit was de- ia, hydrothermal xenotime was found in the same leucogranite-hosted, scribed as a simple association of fluorite, barite, and mixed sulfides in a molybdenite-related vein occurrences of Su Senargiu and Perda Pibera gangue consisting of calcite and quartz. The dolomite-ankerite phases de- (Orlandi et al., 2013, 2014). In the Silius paragenesis, xenotime-(Y) has tected in the present study are paragenetically younger than the other been detected as a phase strictly associated with synchysite-(Ce). Consid- minerals of the veins because they replace earlier calcite. This event was ering that synchysite preferentially hosts LREE, whereas xenotime is the then followed by, or is partly contemporaneous with a minor fluorite- common HREE carrier, this is probably a consequence of the different barite-quartz mineralization stage, and by the precipitation of a late cal- fractionation of LREE and HREE between the two minerals from the cite phase. Consequently, the formation of both dolomite and Fe- REE-bearing hydrothermal fluid. Gysi et al. (2015) investigated experi- dolomite can be considered as evidence for a short and abrupt change mentally the thermodynamic properties of xenotime-(Y) and other in the chemical composition of the mineralizing fluids, which were tem- HREE phosphates in a hydrothermal environment, and in aqueous porarily enriched in Mg and Fe, in order to dolomitize the previously de- HClO4-H3PO4 solutions, at temperatures up to 250 °C. From these exper- posited calcite in both San Giorgio and San Giuseppe veins. The iments, the authors showed that the solubility of HREE phosphate occurrence of synchysite-(Ce) and xenotime-(Y), that are paragenetically solids increases sharply with decreasing temperature. At Silius, associated with dolomitized calcite, also marks a sharp temporary change most of the ore minerals precipitated at temperatures in the range in the chemistry of the fluidsthatforawhilebecameREE-bearing. of 120–180 °C from a dominant fluid consisting of a NaCl ± CaCl2- Synchysite (the main REE-carrier in the Silius fluorite veins) is gen- rich brine (Boni et al., 2009). Hence, we may assume that the erally associated with orthomagmatic-hydrothermal or hydrothermal- mineralizing fluid that precipitated xenotime-(Y), together with metasomatic REE-mineralizations, but it has been also found in synchysite-(Ce), can besimilartothemodelofGysi et al. (2015)

Alpine-type veins (Augé et al., 2014; Chakhmouradian and Wall, for xenotime-(Y) precipitation from H3PO4–HCl–HF-ﬂuid at 150 °C 2012; Förster, 2001; Gieré, 1996, and references therein; Williams- temperature. In these experimental conditions, xenotime is stable Jones et al., 2000). This mineral was detected in post-Hercynian veins in the entire pH range that is higher than ~2. This wide stability N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224 221

Table 4 Major (wt.%) and minor to trace (ppm) elements compositions of Silius host rocks (ICP-OES-MS on bulk samples).

SN 1 P 6 P 7 PG 1 PG 3 PR 1 PR 2 M 2 M 3 Lower detection limita

SiO2 78.39 66.27 75.43 48.13 75.55 62.89 85.70 50.30 82.58 0.01

Al2O3 5.77 4.77 11.25 15.52 11.50 16.11 6.08 14.81 6.80 0.01

Fe2O3t 1.92 0.73 1.29 8.04 1.56 4.87 0.91 7.82 1.95 0.04 MgO 0.67 0.36 0.59 6.89 0.62 1.70 0.12 3.37 0.47 0.01 CaOb 2.73 19.78 3.08 6.14 1.33 2.12 1.82 15.58 1.97 0.01

Na2O 0.04 0.02 0.06 1.29 0.13 2.99 0.09 1.05 0.06 0.01

K2O 1.60 1.68 3.93 2.89 5.19 4.61 2.64 3.79 2.70 0.01

TiO2 0.19 0.07 0.35 0.95 0.08 0.75 0.04 0.64 0.23 0.01

P2O5 0.05 0.10 0.17 0.19 0.02 0.19 0.01 0.15 0.12 0.01 MnO 0.04 0.02 0.02 0.11 0.08 0.05 b0.01 0.24 0.03 0.01 LOI 8.30 5.90 3.70 9.60 3.70 3.40 2.20 1.80 2.90 - F 1.28 13.53 1.41 0.13 0.03 0.06 1.02 0.06 0.12 0.01 Sum 100.98 113.23 101.28 99.88 99.79 99.74 100.63 99.61 99.93

Cr 20 b20 20 180 b20 b20 b20 40 20 20 Ni 25.1 3.4 6.4 100.9 2.8 3.4 3.3 74.0 21.2 0.1 Sc 6 3 5 25 3 15 2 15 6 1 Ba 968 136 431 390 796 1064 526 2057 529 1 Be 2 b1 b124 6b1521 Co 4.3 2.8 5.0 30.6 1.1 11.7 1.8 15.2 3.3 0.2 Cs 6.7 4.0 13.1 14.7 6.3 7.3 5.9 11.6 7.8 0.1 Ga 10.8 6.2 15.8 14.6 14.8 19.1 8.6 19.1 10.3 0.5 Hf 0.8 7.5 3.8 3.3 6.6 4.8 2.6 2.8 1.2 0.1 Nb 4.0 1.7 7.8 7.2 14.7 11.6 7.7 13.1 4.8 0.1 Rb 78.3 80.6 174.7 111.1 174.2 164.2 92.3 121.8 114.5 0.1 Sn 1 2 3 b13 4 21541 Sr 59.2 151.0 38.8 348.0 25.0 297.4 26.9 555.1 57.6 0.5 Ta 0.2 0.3 0.6 0.5 0.9 1.0 0.6 1.0 0.4 0.1 Th 2.5 9.2 10.2 3.8 17.1 11.6 10.0 10.9 3.3 0.2 U 5.1 2.8 4.1 0.7 3.6 4.1 3.4 3.1 2.1 0.1 V 466 b8 34 167 b8 102 b8 175 98 8 W 1.1 1.3 2.9 1.3 0.9 1.9 0.6 2.0 2.3 0.5 Zr 30.5 276.6 133.4 134.5 207.4 172.6 71.0 104.8 44.3 0.1 Mo 8.3 1.2 2.7 0.8 2.2 1.4 2.2 6.0 4.2 0.1 Cu 46.2 13.2 8.3 30.7 6.1 10.1 7.6 13.9 47.6 0.1 Pb 141.8 274.7 15.2 6.8 62.6 26.8 12.6 7.6 11.1 0.1 Zn 648 857 191 71 523 107 2271 25 43 1 As 32.9 12.1 11.0 2.2 5.5 8.4 54.1 188.1 103.9 0.5 Cd 5.6 4.4 0.9 b0.1 2.5 0.3 29.8 0.2 0.2 0.1 Sb 5.4 0.4 0.6 0.2 0.5 0.5 1.6 0.9 1.4 0.1 Bi 0.1 0.4 b0.1 b0.1 0.2 0.1 b0.1 0.4 0.4 0.1 Ag 2.6 0.4 0.2 0.4 0.3 0.1 0.2 b0.1 0.9 0.1 Au§ b0.5 1.3 1.3 0.8 1.3 0.6 3.9 1.4 156.3 0.5

Y 12.7 35.1 24.8 21.1 43.0 23.8 18.3 30.3 5.9 0.1 La 10.8 13.0 21.6 17.2 40.5 31.9 12.2 40.6 7.4 0.1 Ce 17.4 26.2 44.9 37.2 87.2 62.7 25.5 75.4 14.8 0.1 Pr 2.11 2.99 5.03 4.68 10.02 7.19 3.09 8.63 1.75 0.02 Nd 7.1 11.6 18.9 19.9 38.6 29.1 12.4 31.5 6.6 0.3 Sm 1.33 2.66 4.28 4.15 8.28 5.53 2.40 6.26 1.46 0.05 Eu 0.32 0.69 0.71 1.01 1.45 1.24 0.36 1.09 0.22 0.02 Gd 1.39 3.28 4.24 4.18 7.66 5.40 2.39 5.86 1.51 0.05 Tb 0.24 0.59 0.69 0.68 1.18 0.78 0.42 0.94 0.22 0.01 Dy 1.59 3.68 4.25 4.15 7.37 4.40 2.66 5.10 1.20 0.05 Ho 0.34 0.83 0.78 0.85 1.52 0.91 0.54 1.00 0.20 0.02 Er 1.23 2.42 2.37 2.53 4.88 2.63 1.99 2.89 0.59 0.03 Tm 0.17 0.43 0.32 0.37 0.76 0.38 0.31 0.43 0.08 0.01 Yb 1.01 2.93 2.12 2.21 5.07 2.66 2.03 2.95 0.50 0.05 Lu 0.17 0.44 0.27 0.32 0.72 0.38 0.32 0.39 0.09 0.01 Eu/Eu* 1.1 1.1 0.8 1.1 0.9 1.1 0.7 0.8 0.7

(La/Yb)n 0.8 0.3 0.8 0.6 0.6 0.9 0.4 1.0 1.1 Ce/Ce* 0.8 0.9 1.0 0.9 1.0 0.9 0.9 0.9 0.9

Fe2O3t = total iron; Eu/Eu* = [Eun/√(Smn · Gdn)]; (La/Yb)n = Lan/Ybn; Ce/Ce* = [Cen/√(Lan · Prn)]. a ACME Laboratories. b Ca is reported as oxide because it mostly occurs in carbonates, but the element also occurs as Ca in intergrown ﬂuorite. § ppb.

range can be also compatible with the conditions of precipitation of The occurrence of both synchysite-(Ce) and xenotime-(Y) at Ca-bearing REE phases, which, as previously reported, are condi- Silius is compatible with the genetic environment and the tempera- tioned by the buffering reactions between the acid REE-bearing tures determined in previous studies by Castorina et al. (2008) and ﬂuids and the carbonate host rock (Williams-Jones et al., 2012). Boni et al. (2009). 222 N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224

6.2. Geochemistry of fluorite and gangue minerals: a comparison with discussion the fluorite samples FL-04, FL-06 and FL-11 (Table 1)clearly previous studies contaminated by other minerals (quartz, barite, calcite), it appears that the Silius fluorite is characterized by a positive correlation between the The REE occurrence in fluorite has been largely sought in the La/Ho and Y/Ho ratios (Fig. 8). This would exclude a prolonged migra- Paleozoic-Mesozoic hydrothermal veins of Europe, as a way to inves- tion through the host rock before fluorite precipitation, though PAAS- tigate their genesis (e.g., Dill et al., 2011; Möller, 1991; Möller et al., normalized patterns showing positive Y anomalies (as in this case) gen- 1976, 1994; Schwinn and Markl, 2005). It is well known that REE are erally form when the distance between the sites of REE mobilization and hosted in the fluorite structure due to ion-substitution processes deposition is big (Bau and Dulski, 1995). (Möller et al., 1998): REE geochemistry of carbonates in Hercynian fluorite veins has been widely investigated, for example in the deposits of the Harz mountains 2Ca2+ =REE3+ +Na+ in Germany (e.g. Möller et al., 1984), in the small occurrences of the Valle de Tena in Spain (Subías and Fernández-Nieto, 1995), and in the El Hammam orebodies in Morocco (Cheilletz et al., 2010). More specif- 3Ca2+ =2REE3+ +(□) ically, the REE occurrence in the structure of carbonate minerals, i.e. calcite, siderite, and magnesite, has been described by Bau and Möller (1992) for the German ores. However, in our case study, it is necessary Ca2+ =REE3+ +F− to consider that the REE amounts in the carbonates are mostly related to the occurrence of interspersed synchysite-(Ce) and xenotime- Möller et al. (1998) also suggested that Eu, Ce and Y anomalies in (Y) mineral inclusions, and that they do not depend on substitution fluorite REE-PAAS-normalized patterns depend on the environmental processes of REE within calcite or dolomite lattices. Consequently, conditions (mainly T) during the precipitation of this mineral. In partic- it is challenging, for example, to explain why all REE patterns of the ular, the presence of Eu anomalies is strongly temperature-dependent, bulk carbonates at Silius are characterized by Eu positive anomalies, fl because at temperatures N200 °C, Eu3+ is reduced to Eu2+, which is similarly to those of pure uorite. In our opinion, the different possi- very mobile and difficult to be incorporated into the fluorite structure, bilities accounting for the observed Eu positive anomaly are: and therefore tends to remain in the fluid. This produces negative Eu anomalies in fluorites precipitated from a fluid with T N 200 °C. Instead, i) the Eu anomaly of the carbonate gangue depends on the physico- a Eu anomaly is lacking or slightly positive in fluorite precipitated from a chemical conditions during precipitation of the minerals; after Bau fluid with a temperature b200 °C (Möller et al., 1998). Another possi- and Möller (1992), a positive Eu anomaly in calcite indicates precipitation temperatures of at least 200–250 °C, reducing conditions, and bility is that the Eu anomaly was inherited from the REE source rock 2+ (Schwinn and Markl, 2005). Silius fluorite shows a clear positive Eu reduction to Eu , which is easily incorporated into the calcite Eu anomaly, which probably indicates precipitation temperatures structure; b200 °C, in agreement with the homogenization temperatures of the ii) the Eu positive anomaly in the carbonate gangue is inherited fluid inclusions measured by Boni et al. (2009). from the REE source rock: considering the high contribution of The presence or absence of a cerium anomaly is controlled by the ox- synchysite-(Ce) to the total REE concentration in the gangue, it is ygen fugacity of the fluid, which controls the redox reaction between probable that the REE were not fractionated on the basis of their ox- Ce3+ and Ce4+.Ce4+ does not migrate in the fluid together with idation state when entering the mineral, but were incorporated in fl other REE, because it is typically retained in Fe and Mn-hydroxides toto from the hydrothermal uid, thus preserving the footprint of fl (Möller and Bau, 1993). The absence of Ce anomalies in the REE patterns their original source. Considering also that uorite and its carbonate fl of Silius fluorite confirms that the fluids precipitating the main fluorite gangue precipitated together, also the Eu positive anomaly in uo- generation were relatively oxygen-poor, a fact that is also confirmed rite could be partly inherited from the REE source, not being only by the presence of sulfides (galena and pyrite N sphalerite) associated an effect of a temperature-dependent process. with fluorite. Bau and Dulski (1995) suggested that the Y/Ho ratio varies during the migration of the hydrothermal fluid through the rocks: in particular, Considering the geological, textural, and paragenetic characteristics a progressive migration of the fluids produces a negative correlation be- of the Silius mineralization, option (i) is highly likely, but considering tween the Y/Ho ratio and the La/Ho ratio. If we exclude from the the high contribution of synchysite-(Ce) to the REE concentration in the bulk gangue, option (ii) can be also valid. Consequently, it is not possible to totally exclude that the Eu anomaly was inherited from the REE composition of an original source rocks. A comparison between the REE patterns of fluorite and carbonate gangue with those of several host rocks at Silius shows that both fluorite and carbonates have patterns completely different from those of their host rocks (Fig. 7). This point would be enough to exclude a provenance of the REE from the leaching of the same rocks hosting the vein system. The association of REE- fluorcarbonates with Fe-bearing dolomite and quartz may indicate that the hydrothermal fluids have leached Mg-Fe silicatic rocks from the basement, which should have contained discrete REE amounts from the start. Moreover, the presence of a positive Eu anomaly in the normalized REE pattern of both fluorite and carbonates (if partly inherited from the original REE source), may suggest that the REE source was characterized by high feldspar (specifically plagioclase) contents. On this point, for example, Castorina et al. (2008), referring to the experimental work of Schwinn and Markl (2005), suggested a possible contribution of a gneiss-derived hydrothermal fluid to the formation of Fig. 8. Y/Ho vs. La/Ho in bulk samples of fluorite. Two samples are clearly contaminated by the Silius vein system. On the other hand, the upper Ordovician alkali carbonates (see Table 2). basalts occurring in Southeastern Sardinia have many characteristics, N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224 223 which fit the above constraints and make these rocks a potential REE- Boni, M., Balassone, G., Fedele, L., Mondillo, N., 2009. Post-Variscan hydrothermal activity and ore deposits in southern Sardinia (Italy): selected examples from Gerrei (Silius source for the Silius mineralization. In fact, they are Mg-Fe rich, contain vein system) and Iglesiente district. Per. Mineral. 78, 19–35. discrete amounts of REE, and also show positive Eu anomalies in the REE Bouabdellah, M., Banks, D., Klügel, A., 2010. Comments on “A late Triassic 40Ar/39Ar age for chondrite-normalized pattern (Gaggero et al., 2012). The occurrence of the El hammam high-REE fluorite deposit (Morocco): mineralization related to the Central Atlantic magmatic province?” by Cheilletz et al. (Miner. Deposita 45, xenotime-(Y) at Silius suggests also the presence of phosphorus within 323–329, 2010). Mineral. Deposita 45, 729–731. the source rocks. Phosphorus could have been leached from the Silurian Brizzi, G., Olmi, F., Sabelli, C., 1994. Gli arseniati di pira inferida, gonnosfanadiga (CA). Riv. black shales and pelagic sediments of the Graptolitic Shales Formation, Mineral. Ital. 3 (1994), 193–206. which locally host the fluorite mineralization. Brown, S.M., Fletcher, I.R., Stein, H.J., Snee, L.W., Groves, D.I., 2002. Geochronological constraints on pre-, syn-, and post-mineralization events at the world-class Cleo gold deposit, Eastern Goldfields province, Western Australia. Econ. Geol. 97, 541–559. 7. Conclusions Carmignani, L., Cortecci, G., Dessau, G., Duchi, G., Oggiano, G., Pertusati, P., Saitta, M., 1978. The antimony and tungsten deposit of Villasalto in South-Eastern Sardinia and its relationship with Hercynian tectonics. Schweiz. Mineral. Petrogr. Mitt. 58, 163–188. This study has shown that REE concentrations in the carbonate Carmignani, L., Oggiano, G., Pertusati, P.C., 1994. Geological outlines of the hercynian gangue of the Silius fluorite deposit (SE Sardinia, Italy), are related basement of Sardinia. In: petrology, geology and ore deposits of the paleozoic basement of Sardinia, guidebook to the B3 field excursion. 16th General Meeting of the In- to the presence of the REE-bearing minerals synchysite-(Ce) and ternational Mineralogical Association, Pisa, pp. 9–20. xenotime-(Y). From the discussion of the data and the comparison Carmignani, L., Oggiano, G., Barca, S., Conti, P., Salvadori, I., Eltrudis, A., Funedda, A., Pasci, with the existing literature, it is likely that both synchysite-(Ce) S., 2001. Geologia della Sardegna, Note illustrative della Carta Geologica della Sarde- gna alla scala 1:200000. Mem. Descr. Carta Geologica d'Italia, Serv. Geol. It., 60, pp and xenotime-(Y) formed at the same P-T-X conditions considered 283, Ist. Poligr. Zecca dello Stato, Roma. for the other minerals of the Silius fluorite mineralization. Rare Castorina, F., Masi, U., Padalino, G., Palomba, M., 2008. Trace-element and Sr–Nd isotopic earth elements are likely derived from a REE-bearing source rock in evidence for the origin of the Sardinian fluorite mineralization (Italy). Appl. Geochem. 23, 2906–2921. the basement of southeastern Sardinia, which has been leached by Chakhmouradian, A.R., Wall, F., 2012. Rare earth elements: minerals, mines, magnets the same fluids precipitating the fluorite/calcite mineralization. (and more). Elements 8, 333–342. In our opinion, the discovery of REE-minerals in the Silius fluorite Chakhmouradian, A.R., Zaitsev, A.N., 2012. Rare earth mineralization in igneous rocks: sources and processes. Elements 8, 347–356. mine may open interesting perspectives for the exploration of sub- Cheilletz, A., Gasquet, D., Filali, F., Archibald, D.A., Nespolo, M., 2010. A late Triassic economic REE concentrations in this type of deposits, where REE could 40Ar/39Ar age for the El Hammam high-REE fluorite deposit (Morocco): mineraliza- be recovered as by-product of the fluorite exploitation. Moreover, REE tion related to the Central Atlantic magmatic province? Mineral. Deposita 45, – minerals could possibly be found also in dumps and tailings accumulat- 323 329. Cook, N.J., Ciobanu, C.L., O'Rielly, D., Wilson, R., Das, K., Wade, B., 2013. Mineral chemistry ed during the fluorite concentration process. of rare earth element (REE) mineralization, browns ranges, Western Australia. Lithos 172-173, 192–213. Crowley, Q.G., Floyd, P.A., Winchester, J.A., Franke, W., Holland, J.G., 2000. Early Paleozoic Acknowledgments rift related magmatism in Variscan Europe: fragmentation of the Armorican terrane assemblage. Terra Nova 12, 171–180. We would like to thank Ing. G. Mura, Director of the Silius mine, for Di Vincenzo, G., Elter, F.M., Ghezzo, C., Palmeri, R., Ricci, C.A., 1994. Petrological evolution of the Palaeozoic basement of Sardinia. In: Carmignani, L., Ghezzo, C., Marcello, A., having allowed us to sample in the underground levels, and R. de' Pertusati, P.C., Pretti, S., Ricci, C.A., Salvadori, I. (Eds.), Petrology, Geology and Ore De- Gennaro (DiSTAR, Napoli) for the SEM-EDS analyses. Thanks are also posits of the Palaeozoic Basement of Sardinia. Guide Book to the Field Excursion 145, – due to the anonymous reviewers, whose comments have highly en- pp. 643 658. Dill, H.G., Hansen, B.T., Weber, B., 2011. REE contents, REE minerals and Sm/Nd isotopes of hanced the quality of the manuscript, and F. Pirajno for editorial han- granite- and unconformity-related fluorite mineralization at the western edge of the dling. N. Mondillo would like to thank B. Lehmann, for the positive Bohemian masSif: with special reference to the Nabburg-Wölsendorf District, SE discussion on the Harz fluorite, and R. Herrington for suggestions and Germany. Ore Geol. Rev. 40, 132–148. fi Förster, H.J., 2001. Synchysite-(Y) ± synchysite-(Ce) solid solutions from Markersbach, help. This work was partially nanced by the Università degli Studi di Erzgebirge, Germany: REE and Th mobility during high-T alteration of highly frac- Napoli Federico II, grant RDIP2014 to M. Boni and N. Mondillo. tionated aluminous A-type granites. Mineral. Petrol. 72, 259–280. Gaggero, L., Oggiano, G., Funedda, A., Buzzi, L., 2012. Rifting and arc-related early Paleozo- ic volcanism along the North Gondwana margin: geochemical and geological evi- Appendix. Supplementary data dence from Sardinia (Italy). J. Geol. 120, 273–292. Ghezzo, C., Orsini, J.B., 1982. Lineamenti Strutturali e Composizionali del Batolite Ercinico Sardo-Corso in Sardegna. In: Carmignani, L., et al. (Eds.), Guida Alla Geologia del Supplementary data to this article can be found online at http://dx. Paleozoico Sardo. Guide Regionali Soc. Geol. It., pp. 165–181. doi.org/10.1016/j.oregeorev.2015.11.016. Gieré, R., 1996. Formation of rare earth minerals in hydrothermal systems. In: Jones, A.P., Wall, F., Williams, C.T. (Eds.), Rare Earth Minerals: Chemistry, Origin and ore DepositsMineralogical Society Series 7. Chapman and Hall, London, pp. 105–150. References Götze, J., 2012. Application of cathodoluminescence microscopy and spectroscopy in geosciences. Microsc. Microanal. 18, 1270–1284. Arthaud, F., Matte, P., 1977. Late-Palaeozoic strike-slip faulting in Southern Europe and Gysi, A.P., Williams-Jones, A.E., 2013. Hydrothermal mobilization of pegmatite-hosted Northern Africa: results of a right lateral shear zone between the Appalachians and REE and Zr at Strange Lake, Canada: a reaction path model. Geochim. Cosmochim. the urals. Geol. Soc. Am. Bull. 88, 305–1320. Acta 122, 324–352. Atzori, P., Traversa, G., 1986. Post-granitic Permo-Triassic dyke magmatism in eastern Sar- Gysi, A.P., Williams-Jones, A.E., Harlov, D., 2015. The solubility of xenotime-(Y) and other

dinia (Sarrabus p.p., Barbagia, Mandrolisai, Goceano, Baronie and Gallura). Per. Min- HREE phosphates (DyPO4,ErPO4 and YbPO4) in aqueous solutions from 100 to 250 °C eral. 55, 203–231. and psat.Chem.Geol.401,83–95. Augé, T., Bailly, L., Wille, G., 2014. An unusual occurrence of synchysite-(Ce) in amygdales Haack, U., Schnorrer-Köhler, G., Lüders, V., 1987. Seltene Erd-minerale aus from the Esterel volcanic rocks, France: implications for rare-earth element mobility. hydrothermalen gängen des harzes. Chem. Erde 47, 41–45. Can. Mineral. 52, 837–856. Kositcin, N., McNaughton, N.J., Griffin, B.J., Fletcher, I.R., Groves, D.I., Rasmussen, B., 2003. Bau, M., Dulski, P., 1995. Comparative study of yttrium and rare earth element behaviours Textural and geochemical discrimination between xenotime of different origin in the in fluorine-rich hydrothermal fluids. Contrib. Mineral. Petrol. 119, 213–223. ArchaeanWitwatersrandBasin,SouthAfrica.Geochim.Cosmochim.Acta67,709–731. Bau, M., Möller, P., 1992. Rare earth element fractionation in metamorphogenic hydro- Migdisov, A.A., Williams-Jones, A.E., 2014. Hydrothermal transport and deposition of the thermal calcite, magnesite and siderite. Mineral. Petrol. 45, 231–246. rare earth elements by fluorine-bearing aqueous liquids. Mineral. Deposita 49, Bechstädt, T., Boni, M., 1994. Sedimentological, Stratigraphical and ore Deposits Field 987–997. Guide of the Autochthonous Cambro-Ordovician of Southwestern Sardinia. Servizio Möller, P., 1991. REE fractionation in hydrothermal fluorite and calcite. In: Pagel, M., Geologico d'Italia, Roma (434 pp.). Leroy, J. (Eds.), Source, Transport and Deposition of Metals. Balkema, Rotterdam, Belkin, H.E., De Vivo, B., Valera, R., 1984. Fluid inclusion study of some Sarrabus fluorite pp. 91–94. deposits, Sardinia, Italy. Econ. Geol. 79, 409–414. Möller, P., Parekh, P.P., Schneider, H.-J., 1976. The application of Tb/Ca–Tb/La abundance Boni, M., Stein, H.J., Zimmerman, A., Villa, I.M., 2003. Re–Os age for molybdenite from SW ratios to problems of fluorspar genesis. Mineral. Deposita 11, 111–116. Sardinia (Italy): A comparison with 40Ar/39Ar dating of Variscan granitoids. Mineral Möller, P., Bau, M., 1993. Rare-earth patterns with positive cerium anomaly in alkaline Exploration and Sustainable Development. Proceedings of the Seventh Biennial SGA waters from Lake Van, Turkey. Earth Planet. Sci. Lett. 117, 671–676. Meeting on Mineral Exploration and Sustainable Development, Athens, Greece, Möller, P., Giese, U., 1997. Determination of easily accessible metal fractions in rocks by pp. 247–250. batch leaching with acid cation — exchange resin. Chem. Geol. 137, 41–55. 224 N. Mondillo et al. / Ore Geology Reviews 74 (2016) 211–224

Möller, P., Morteani, G., Dulski, P., 1984. The origin of the calcites from Pb-Zn veins in the Samson, I.M., Wood, S.A., Finucane, K., 2004. Fluid inclusion characteristics and genesis of Harz mountains, Federal Republic of Germany. Chem. Geol. 45, 91–112. the fluorite–parisite mineralization in the snowbird deposit, Montana. Econ. Geol. 99, Möller, P., Franzke, H.J., Lüders, V., 1994. Post-orogene intraformationale bildung der 1727–1744. hydrothermalen Pb–Zn, F und F–Ba Lagerstätten im Schwarzwald. Technical Report Schaltegger, U., Pettke, T., Audétat, A., Reusser, E., Heinrich, C.A., 2005. Magmatic-to-hy- Mo 443/3 and Fr 874/1. Intraformationale Lagerstättenbildung, GFZ-Postdam, Germany. drothermal crystallization in the W–Sn mineralized mole granite (NSW, Australia) Möller, P., Bau, M., Dulski, P., Lüders, V., 1998. REE and Y fractionation in fluorite and their part I: crystallization of zircon and REE-phosphates over three million years — ageo- bearing on fluorite formation. Proceedings of the 9th Quadrennial IAGOD Symp., chemical and U–Pb geochronological study. Chem. Geol. 220, 215–235. Schweizerbart, Stuttgart, pp. 575–592. Schwinn, G., Markl, G., 2005. REE systematics in hydrothermal fluorite. Chem. Geol. 216, Muchez, P., Heijlen, W., Banks, D., Blundell, D., Boni, M., Grandia, F., 2005. Extensional tec- 225–248. tonics and the timing and formation of basin-hosted deposits in Europe. Ore Geol. Stara, P., Rizzo, R., Tanca, G.A., 1996. Iglesiente e arburese. Miniere e MineraliEnte Rev. 27, 241–267. Minerario Sardo (192 pp.). Muñoz, M., Boyce, A.J., Courjault-Radé, P., Fallick, A.E., Tollon, F., 1999. Continental basinal Subías, I., Fernández-Nieto, C., 1995. Hydrothermal events in the Valle de Tena (Spanish origin of ore fluids from southwestern massif central fluorite veins (Albigeois, Western Pyrenees) as evidenced by fluid inclusions and trace-element distribution France): evidence from fluid inclusion and stable isotope analyses. Appl. Geochem. from fluorite deposits. Chem. Geol. 124, 267–282. 14, 447–458. Taylor, S.R., McLennan, S.M., 1985. The Continental Crust: Its Composition and Evolution. Muñoz, M., Premo, W.R., Courjault-Radé, P., 2005. Sm/Nd dating of fluorite from the Blackwell Scientific Publications Oxford, London, Edinburgh (312 pp.). worldclass montroc fluorite deposit, Southern Massif Central, France. Mineral. Valera, R., 1974. Appunti Sulla morfologia, termometria e composizione delle inclusioni Deposita 39, 970–975. fluide di fluoriti sarde. Rend. Soc. Ital. Mineral. Petrol. 30, 459–480. Natale, P., 1969. Il giacimento di Silius nel gerrei. Boll. Ass. Miner. Subalpina 6, 1–35. Von Gehlen, K., Grauert, B., Nielsen, H., 1986. REE minerals in southern Schwarzwald Oggiano, G., Gaggero, L., Funedda, A., Buzzi, L., Tiepolo, M., 2010. Multiple early Palaeozoic veins and isotope studies on gypsum from the central Schwarzwald, F.R. Germany. volcanic events at the Northern Gondwana margin: U–Pb age evidence from the N. Jb. Miner. Mh. 9, 393–399. Southern Variscan branch (Sardinia, Italy). Gondwana Res. 17, 44–58. Williams-Jones, A.E., Samson, I.M., Olivo, G.R., 2000. The genesis of hydrothermal fluorite- Orlandi, P., Gelosa, M., Bonacina, E., Caboni, F., Mamberti, M., Tanca, G., Vinci, A., 2013. REE deposits in the Galllinas Mountains, New Mexico. Econ. Geol. 95, 327–342. Sardignaite, gelosaite et tancaite-(Ce): trois nouveaux minéraux de Su seinargiu, Williams-Jones, A.E., Migdisov, A.A., Samson, I.M., 2012. Hydrothermal mobilization of the sarroch, sardaigne, italie. Le Règne Minér. 112, 39–52. rare earth elements — ataleof“ceria” and “yttria”.Elements8,355–360.

Orlandi, P., Biagioni, C., Bindi, L., Nestola, F., 2014. Ichnusaite, Th(MoO4)2.3H2O, the ﬁrst natural thorium molybdate: occurrence, description, and crystal structure. Am. Min- eral. 99, 2089–2094.