Introduction

Tungsten deposits can broadly be classified into granitoid-related and non-granitoid. The first type of deposits includes tungsten-bearing granitoids, greisens, hydrothermal and tungsten skarn deposits (e.g. Hosking 1982; Raith and Prochaska 1995). In contrast, non-granitoid deposits include volcanogenic or exhalative tungsten deposits that may be remobilised by subsequent metamorphism. Examples include carbonate-hosted scheelite and wolframite of the Nock Mountains, Austria (Neinavaie et al. 1989), Western Hohe Tauern, Eastern Alps (Niedermayr and Schroll 1983) and Mittersill and Bohemian Massif, Austria (Thalhammer et al. 1989; Beran et al. 1985). Scheelite in the Bohemian Massif and Eastern Alps is associated with calc-silicate rocks, pelitic metasediments, quartzite and tourmalinite of higher amphibolite facies.

The tungsten mineralisation under discussion occurs around Miniki Gol, Chitral District, northern Pakistan. Miniki Gol lies within the NE Hindu Kush Range, some 8 km NW of Tirich Boundary Zone that is considered as suture zone between the Karakorum and Eastern Hindu Kush by Zanchi et al. (2000). The mineralisation is mainly represented by scheelite that is particularly concentrated in calc-silicate rocks and subordinate tourmalinites. The best grade tungsten mineralisation zone is confined to a 2-km wide belt in the vicinity of Miniki Gol. Potential ore-forming two-mica leucogranite (Garam Chashma pluton) is emplaced 400 m to the southeast of the mineralised zone (Figs. 1 and 2). The Garam Chashma pluton has been dated at 22 Ma (Faisal et al. 2015; Hilderband et al. 1998). Pegmatite pockets occur within the leucogranite and associated metasediments. Moreover, marble-hosted stratiform zinc-lead mineralisation of Besti Gol is located 10 km to the NNE of Miniki Gol scheelite deposits (Fig. 1).

Fig. 1
figure 1

Simplified geological map of the Chitral area (after Zanchi and Gaetani 2011). The box in the Arkari Gol shows the location of the investigated area at Miniki Gol

Full size image
Fig. 2
figure 2

Simplified geological map of the Miniki Gol and adjoining area

Full size image

The rocks of the investigated area belong mostly to the Arkari Formation of probable Jurassic age (Leake et al. 1989). These, dominantly metasedimentary, rocks include garnet-mica schist, graphitic schist, carbonaceous material, phyllite, micaceous quartzite, calc-silicate rocks and carbonates (Fig. 2). The garnet-mica schist locally contains patches of tourmalinites. These rocks are metamorphosed up to amphibolite facies and have undergone at least two phases of deformation related to continental collision during the Late Cretaceous and Eocene, (Leake et al. 1989; Heuberger et al. 2007; Zanchi and Gaetani 2011).

The leucogranite and tungsten mineralization do not appear to be spatially associated at least on the surface; however, a possible genetic linkage between the two cannot be ruled out (Zahid et al. 2013). Based on mineralogical and geochemical investigation of the scheelite-bearing calc-silicate rocks and tourmalinites along with barren metasediments, the current study explains the genesis of the tungsten mineralization.

Elsewhere in Pakistan, scheelite has been reported in Saindak area in Balochistan and Oghi in Hazara, NW Himalaya. In Saindak, scheelite is associated with a pyroclastic sequence, quartz porphyry dykes and tourmaline intrusive breccia pipes. Siddiqui et al. (1986, 2012) have considered it as a hypogene mineralisation associated with the Creto-Tertiary Chagai calc-alkaline magmatic belt. Scheelite in the Oghi area has been noted in granite gneiss (Asrarullah 1982) and seems to be related to the Mansehra granite which has a Rb-Sr age of 516 ± 16 Ma (Le Fort et al. 1980).

Geology of the area

The Eastern Hindu Kush terrain, hosting the tungsten mineralization in Chitral, runs from northeastern Afghanistan to north-western Pakistan (Fig. 1) and joins the Karakoram Range to the east. Two main blocks crop out in Chitral: the western part of the Karakoram block between the Tirich Mir Fault Zone and Reshun Fault, and the East Hindu Kush block northwest of the Tirich Mir Fault (Fig. 1). Lying 20 km to the east of the mineralised zone, the Tirich Boundary Zone (TBZ) separates the Karakoram block from East Hindu Kush-Wakhan (Zanchi and Gaetani 2011). The TBZ comprises amphibolites, metagabbros, peridotites, serpentinites, gneisses and quartzites, and is considered as Jurassic-Early Cretaceous in age.

The rock units between TBZ and Reshun Fault comprise metasediments (Fig. 1) and are mainly composed of highly strained phyllite and graphitic schists adjacent to Reshun Fault. Carbonate-rich lenses within these phyllites are common (Hildebrand et al. 2000). Along the Arkari valley, Leake et al. (1989) outline the Sewakht Formation north of the Cretaceous Krinji Limestone (Desio 1966). The formation comprises greenschists, limestones, dolomitic carbonates and sandstones which may correlate with the western continuation of the Devonian Owir Series. Leake et al. (1989) also introduced the term Lutkho Formation for the monotonous greenish phyllites extending between the Sewakt Formation and the Tirich Mir pluton. These rocks have been previously considered by Pudsey et al. (1985) as Lun Shales.

The rocks between the Tirich Boundary Zone and the Afghanistan border are metapelites which are commonly highly graphitic. They are interlayered with occasional quartz-rich lenses, calc-silicates and amphibolites. These lithologies are probably the southwestward continuation of the Wakhan slates, which have yielded Palaeozoic fossils (Buchroithner 1980; Gaetani and Leven 1993). In this area, the Wakhan slates pass into composite metamorphic succession reaching up to amphibolite grade, named as Arkari Formation by Leake et al. (1989). The formation includes mica schists, phyllites, marbles, quartzites and feldspathic gneisses. The Arkari Formation hosts the studied scheelite-bearing calc-silicate rocks (Fig. 2). Based on the Belemnite remains found 75 years ago near Besti, Tipper in Pascoe (1924) had suggested a Jurassic age for at least part of the protolith of the metamorphic complex.

Metamorphism and deformation history

The Eastern Hindu Kush has experienced at least two episodes of deformation along with two distinct metamorphic phases (Leake et al. 1989; Hildebrand et al. 2001). The earlier phase of metamorphism took place prior to the D2 deformation. Typical pre-D2 metapelite assemblages contain garnet, staurolite, biotite, quartz, muscovite, plagioclase, graphite, andalusite, tourmaline, apatite and monazite. Deformation phases have been recorded in the scheelite-bearing tourmalinites; at Miniki Gol, the muscovite flakes in these define S1 foliation and crenulation displays S2 foliation (Fig. 3a).

Fig. 3
figure 3

a Sketch of tourmalinite showing overprint foliation (S1), crenulation cleavage (S2), quartz veins and scheelite patches. Note that some scheelite patches are emplaced within the quartz veins and S2 foliation. b sketch of schist and calc-silicate rocks containing concordant and discordant scheelite patches at Miniki Gol

Full size image

Metamorphic mineral assemblages have been studied in only two areas including the Tirich Mir Boundary Zone and the upper Lutkho valley (Hildebrand et al. 2001). In both places, this later metamorphism is characterised by the appearance of sillimanite at the expense of staurolite and partial annealing of S1 by late mica growth (Hildebrand et al. 2001). Staurolite has been replaced by fibrolite and muscovite laths along the traverse from Gharam Chashma to the Tirich Mir Boundary Zone. The upper Lutkho valley is characterised by sillimanite K-feldspar schists, gneisses and migmatites. The abundant intrusions of leucogranite signify crustal melting and the generation of leucogranite seems to be contemporaneous with the second deformation (D2) (Leake et al. 1989).

Tungsten mineralisation

Scheelite in the study area occurs mainly in the calc-silicate rocks and subordinately in tourmalinites. It forms discontinuous patches, stringers and conformable small veins within the calc-silicate rocks and can be considered as stratabound type (Figs. 3b and 4a). In the tourmalinites, it occurs as finely disseminated grains as well as lenses (Fig. 3a). These tourmalinites are located at three sites within 2 km strike length and comprise tourmaline (up to 80 % by volume), garnet and quartz. The occurrence of scheelite in the drag folded sections and along the axial planes of S2 suggests that it developed during the second phase of deformation when the Miniki Gol leucogranite was also intruded (Fig. 3a). Some scheelite lenses occur within the crosscutting quartz veins (Fig. 3a) indicating post-magmatic development. The study area has undergone, at least, two major deformational events (D1, D2), with associated foliations S1 and S2 (Fig. 3a).

Fig. 4
figure 4

a Florescent scheelite veins within the calc-silicate rocks along the bedding. b Photomicrograph showing small scheelite vein cross-cutting clinozoisite (blue) grains

Full size image

The calc-silicate rocks are mainly composed of clinozoisite, quartz, calcic-amphibole and plagioclase. Clinozoisite grains are secondary in character and are intergrown with scheelite (Fig. 5a). Some scheelite veins also cut across the early formed main clinozoisite phase (Fig. 4b). Other ore minerals associated with scheelite include pyrite (partly altered to goethite), chalcopyrite, pyrrhotite and melnikovite (Leake et al. 1989).

Fig. 5
figure 5

Photomicrographs showing a scheelite grains (Sch) intergrown with clinozoisite (Czo) in calc-silicate rock, and b almandine (Alm) garnet porphyroblasts enclosed by biotite and muscovite (Mca) within garnet mica schist at Miniki Gol

Full size image

The minor occurrences and showings of scheelite mineralisation extend along strike length for 7 km up to Garam Chashma. However, the best grade tungsten mineralisation zone is confined to a 2-km belt in the vicinity of Miniki Gol leucogranite

The length of the scheelite lenses reaches up to 55 cm whereas the width and thickness of these patches extends up to 14 and 12 cm, respectively (Figs. 3 and 4). The average grade of mineralisation is around 1 %, with high concentration (up to 3 %) at Miniki Gol according to our study (Table 2). The mineralisation does not seem to be economical at least on the surface because of the limited concentration of the ore even in the more promising 2-km belt near the exposed leucogranite. An exploratory drilling is required to estimate the actual ore reserve tonnage.

Sampling methods and analytical techniques

Forty three samples, representing both the scheelite-bearing calc-silicate rocks and barren metasediments, were made into polished thin sections for detailed investigation with EPMA at the University of Leicester (UK). The analyses were performed through Jeol Superprobe model JXA-8600, with an on-line computer for ZAF corrections, using wavelength dispersive system. Scheelite was analysed by a program consisting of CaO, WO3, MoO2, SnO2,

FeO (total), MnO, MgO, Y2O3, TiO2, ZrO2, Nb2O5 and Ta2O5. Pure synthetic metals were used for each of W, Mo, Sn, Nb, Ta and magnetite (synthetic) for Fe. The accuracy of the ZAF correction is generally better than 2 %.

For XRF Spectrometry, 15 g powdered sample was mixed in a glass beaker with mount (10 to 20 drops) of Mowiol 88 (a solution of polyvinyl alcohol in a 1:5 mix of methanol and distilled deionised water). This mixture was then placed in a small dye and pressed into the shape of pellet. The powder pellets were run through a Philips Pw 1400 XRF spectrometer equipped with either a 3 kW rhodium anode tube or a tungsten anode tube. Each sample was analysed for Ga, Sc, Co, Ni, Cr, V, Cu, Mo, Sn, W, As, Pb, Zn, Th, Ba, Rb, Sr, Nb, Y, Zr, La, Ce, Nd, F, Cl and S. A set of international and internal standards was run with samples to monitor accuracy. The analytical results demonstrate a high degree of machine accuracy and precision (i.e., greater than 2 % at the 98 % confidence level) for all the major oxides as well as most of the minor and trace elements.

The powdered samples were also analysed for Be and Li by (ICPES). For these, 0.1 g of each sample was weighed into a labelled PTFE test tube to which 2 ml of HNO3 (70 %) was added. All test tubes were put in a heating block at 50 °C and left overnight. Then 1 ml of HClO3 (60 %) and 5 ml of HF (40 %) were added to each tube and mixed gently. This was followed by heating of the tubes in the block successively at 100 °C for 3 h, 140 °C for 3 h and 190 °C overnight. After this, 1 ml of conc. HCl was added and the tubes were heated at 50 °C. The samples were mixed thoroughly with deionised water. All tubes were decanted into 10-ml-labelled vials. The solutions were run through a Philips PV 8050 spectrometer linked to a PV 8490 source. The degree of precision and accuracy of the method was quantified by repeating 10 % of the analysed samples and 5 % of an internal standard. These estimates were later used to correct the data with the help of an “Excel Macro”.

Samples from calc-silicate rocks were analysed by instrumental neutron activation analysis for REE, B, Ag, Cs, Hf, Hg, Sb, Ta, Th, U and W. The analyses were performed at the Activation Laboratories in Ontario, Canada by the procedure of Lindstrom and Anderson (1985). In this method, 0.5 to 30 g of sample, encapsulated in a polythene vial, is placed in a beam of thermalised neutrons. Samples were counted with a high purity Ge detector and loss-free counting system for the Doppler prompt gamma at 477.6 KeV. Net peak areas were adjusted for Na interference.

Fluid inclusion studies were made at Royal Holloway and Bedford New College, University of London. For the thermometric studies, doubly polished wafers were analysed using a Linkam TH 600 programmable heating-freezing stage attached to a Zeiss transmitted-light microscope. The stage has a wide range of temperature between 180 and 600 °C with the following accuracy: (1) −180 to −20 °C = ±0.1 °C, (2) −20 to +50 °C = ±0.2 °C (3) 50 to 500 °C = ±0.5 °C. The stage was calibrated using synthetic fluid inclusions in quartz crystals.

Mineralogy and petrography

Calc-silicate rocks and metasediments

The calc-silicate rocks, hosting the bulk of the scheelite mineralization at Miniki Gol, are composed of clinozoisite, quartz, calcic-amphibole, plagioclase, and small quantities of chlorite, biotite, calcite, sphene, garnet and scheelite. The clinozoisite, reaching up to 60 modal % in some samples, coexists with scheelite (Fig. 5a). The plagioclase seems to be primary and mostly changed into clinozoisite. Most calcite grains are primary; however, some are re-precipitated. Some calc-silicate rocks are banded, with darker layers containing higher proportions of amphibole, chlorite and clinozoisite. The non-banded or granular variety largely consists essentially of quartz and clinozoisite. Scheelite grains occur mostly in the non-banded variety.

The calc-silicate rocks are mineralogically quite distinct from the associated garnet-mica schist and micaceous quartzite (see Fig. 5 for comparison). The schist is mainly composed of almandine garnet, muscovite and biotite (Fig. 5b), whereas the micaceous quartzite consists of more than 90 % quartz. In contrast, the calc-silicate rocks comprise dominantly clinozoisite, quartz, calcic-amphibole and scheelite (Fig. 5a).

Other metasediments include phyllite, quartz-mica schist and metapsammite. The schist is composed of variable amounts of quartz, plagioclase, muscovite, biotite, graphite and tourmaline, with porphyroblasts of garnet and staurolite. Kyanite has been reported by Leake et al. (1989). The micaceous quartzite is composed of abundant granular quartz (>90 %); the remainder being biotite, muscovite, plagioclase and chlorite. In contrast, metapsammite contains less than 90 % quartz and additional plagioclase grains. Micaceous quartzite is abundant as compared to metapsammite. Some calc-silicate bands also cross-cut the micaceous quartzite and hence postdate the micaceous quartzite.

Scheelite

Scheelite from calc-silicate rocks consists mainly of CaO (20.28 to 20.80 wt%) and WO3 (77.33 to 79.39 wt%); it is almost free of Mo, Y, Ti, Sn, Fe and Mn (Table 1). The noticeable feature of the analysed scheelite is the occurrence of substantial amounts of ZrO2 (up to 0.46 wt%, average 0.28 %) and Ta2O5 (up to 0.35 wt%, average 0.15 %). Zr and Ta are usually associated with scheelite in pegmatite and similar values of Ta2O5 have also been reported in pegmatite-bearing scheelite from Thailand (Linnen and Williams-Jones 1993). The presence of significant amounts of ZrO2 and Ta2O5 in scheelite is very rare (Zahid 1996; Zahid et al. 2013). Very high levels of Zr and Ta occur in scheelite from alkaline igneous rocks and from plutonic rocks relatively rich in soda (Deer et al. 1982; Tukiainen 1988). Tantalum concentration in scheelite seems to rise with increasing degree of differentiation of granitoid magmas (Boissavy-Vinau and Roger 1980). Taylor (1965) argued that Ta is concentrated in Zr-containing minerals from post-magmatic environments and that Zr4+ and Ta5+ replace W4+ and W6+, respectively.

Table 1 Composition (wt%) of scheelite from the investigated calc-silicate rocks
Full size table

Geochemistry of calc-silicate rocks and micaceous quartzite

The geochemical data for the Miniki Gol scheelite-bearing calc-silicate rocks and barren micaceous quartzite are summarised in Table 2. The chemistry of the Miniki Gol calc-silicate rocks differs significantly from the post-Archean shale of Krauskopf (1967), by containing high levels of CaO (mean 9.1 wt%) and SiO2 (mean 73.5 wt%) and low levels of Al2O3 (mean 11.0 wt%), FeO (mean 3.53 wt%), K2O (mean 0.3 wt%) and Na2O (mean 0.4 wt%).

Table 2 Summary of the geochemical data of the Miniki Gol micaceous quartzite and calc-silicate rocks (major elements wt%)
Full size table

Compared with the schist from the study area, they are also enriched in CaO and Al2O3, and depleted in K2O and Na2O (Tables 2 and 3). A positive correlation exists between CaO and Al2O3 in the calc-silicate rocks and it is likely that both these oxides are mainly contained in the clinozoisite (Fig. 6a). The CaO content of the calc-silicate rocks reaches up to 15.27 wt%, which is very high compared to the quartz-rich Phanerozoic greywacke (sandstone), Australia, or even quartz-poor greywacke where the maximum CaO is recorded as 5.52 wt% (see Taylor and McLennan 1985). The additional Ca might have been introduced into these rocks by hydrothermal fluids. A positive correlation also exists between Fe and Mg in the calc-silicate rocks (Fig. 6b).

Table 3 Geochemistry of the schist from Miniki Gol and surrounding area (major elements wt%)
Full size table
Fig. 6
figure 6

a–b Relationship of oxides, CaO vs. Al2O3 and FeO vs. MgO, in the Miniki Gol calc-silicate rocks. c-d Correlation of Ga vs. Y and Nb in the studied calc-silicate rocks

Full size image

The calc-silicate rocks are represented by high values of W, Sn, Zr and Hf, and low levels of B, F, Li, Ba, Rb, Cs, V, Cr and Zn (Table 2). The scheelite-bearing calc-silicate rocks are characterised by higher average concentration of Zr (455 ppm), Hf (16 ppm), Be (8.4 ppm), Th (24 ppm), U (6.1 ppm), Ga (31 ppm), Nb (14.7 ppm) and Y (35.3 ppm) as compared to the micaceous quartzite and schist (Tables 2 and 3). Tungsten levels are highly variable in these rocks and range from 0.4 ppm to 3 wt%. Gallium has a linear positive relationship with Y and Nb in the calc-silicate rocks (Fig. 6c–d). Very low concentrations of B (mean 4.9 ppm), Hg (<1 ppm), Sb (mean 0.7 ppm) and MnO (mean 0.1 wt%) and relatively low F (mean 770 ppm) are recorded in the scheelite-bearing calc-silicate rocks. This may be considered as an indication of the negligible role of exhalative activity (see also Sonnet et al. 1985; Plimer 1984; Maucher 1976). The depletion of Ba (Plimer 1987), coupled with low level of the P2O5 (mean 0.3 %) in the scheelite-bearing calc-silicate rocks, also rule out exhalative activity.

The Miniki Gol calc-silicate rocks are enriched in REE contents (Table 2) compared with the North American shale (Taylor and McLennan 1985). The total REE abundances in the scheelite-bearing calc-silicate rocks are relatively high as compared to the barren micaceous quartzite. Maximum REE contents (ppm) in the scheelite-bearing rocks shown in Table 2 are as follows: La (130), Ce (280), Nd (100), Sm (19), Eu (2.9), Gd (13), Tb (3.9), Yb (9) and Lu (1.39). Almost half of the scheelite-bearing calc-silicate rocks as well as the unmineralised micaceous quartzite exhibit positive Eu anomalies (Fig. 7a). The REE pattern of the calc-silicate rocks is more or less similar to that of Miniki Gol leucogranite (Fig. 7a, c). The calc-silicate rocks are represented by low chondrite-normalised (La/Lu) N ratios (<10) whereas in most of the analysed micaceous quartzite, (La / Lu) N ratios are greater than 10 (Table 2). The majority of the scheelite-bearing calc-silicate rocks (13 out of 21) are enriched in HREE. This enrichment is apparent from the LA/Yb ratio of less than 10. In comparison, the LA/Yb ratio in the micaceous quartzite is very high, reaching up to 19 (Table 2). A strong positive correlation exists among the REE and Zr, Hf, Nb, Y, Ta, together with a weak linear relationship with F (Fig. 8a–f). However, no correlation has been found between REE and ore elements such as W and Sn.

Fig. 7
figure 7

Chondrite-normalised REE pattern of a calc-silicate rocks (zc 65a, zc 70 and zm 66 and micaceous quartzite (sample zc 6), b garnet mica schist, and c leucogranite at Miniki Gol

Full size image
Fig. 8
figure 8

Correlations of REE in calc-silicate rocks with other trace elements (ppm): a Yttrium, b Fluorine, c Tantalum, d Niobium, e Zirconium and f Hafnium. Identical relationship holds for other REE with the corresponding elements

Full size image

Schist

Summary of the major and trace elements chemistry of the schist from the Miniki Gol and surrounding area is given in Table 3. Comparison of the major element chemistry of the studied metasediments with the average post-Archean shale of Krauskopf (1967), shows that the Miniki Gol schistose rocks are richer in Na and poorer in Ca (Table 3). The CaO content of the calcite-rich schist reaches up to 9.15 wt%, whereas the garnet-rich schist contains up to 12.2 wt% FeO and 2.1 wt% MnO. Comparing the trace element data of the rocks with the average shale of Turekian and Wedepohl (1961), the Miniki Gol schist is significantly depleted in S and to a lesser extent, Cl, Ni, Cu and Mo (Table 3). The studied schist is enriched in F, Li and Cs and to a lesser extent, Be, W, Pb, Th, Zr, Y, Hf and Nb (Table 3). Leake et al. (1989) have also reported high level of B (minimum 201 ppm), W (mean 11 ppm) and As (up to 26 ppm) in the studied schist. The enrichment of these elements is a clear expression of metasomatic activity (see also Christensen et al. 1983; Leake et al. 1989; Van de Haar et al. 1993), in the Miniki Gol schist.

The REE contents of the Miniki Gol metasediments are significantly higher (Table 3, Fig. 7b) than the North American shale (see Taylor and McLennan 1985). The maximum Eu/Eu* ratios in the studied schist is observed as 0.9 (Table 3). The mobility of REE in metamorphic rocks is poorly understood (see Grauch 1989) and many investigators (e.g. Cullers et al. (1974) and Lottermoser (1989a)) consider that REE content is unaffected during progressive metamorphism. However, partial melting and retrogressive metamorphism do change (loss or gain) the REE content (cf. Lottermoser 1989a). The enrichment of REE abundances in the proposed metasediments relative to the North American shale might be indicative of metasomatic activity.

Microthermometry

Aqueous fluid inclusions were found in the Miniki Gol scheelite-bearing calc-silicate rocks (Fig. 9); however, no CO2-bearing fluid inclusions were observed. The fluid inclusions in these rocks are smaller (15 to 20 μm in diameter) than those observed in the pegmatite and leucogranite of the area (Zahid 1996). These inclusions occur in quartz with well-defined grain boundaries and are intergrown with clinozoisite and scheelite crystals (Fig. 5a). The aqueous fluid inclusions have very low salinity, with the maximum value as 2 wt% NaCl equivalent. The salinity of CO2-bearing fluid inclusions in the muscovite pegmatite at Miniki Gol ranges from 8 to 10 wt% NaCl equivalent, whereas the primary fluid inclusions in leucogranite are characterised by relatively low salinity (between 2 and 8 wt% NaCl equivalent). The homogenisation of the two phases in these inclusions in the calc-silicate rocks occurred at temperature between 160 °C and 500 °C (Fig. 10). The fluid inclusions in the pegmatite homogenised at maximum temperature 495 °C and in leucogranite at 440 °C. It is also worth mentioning that the salinity progressively decreases from pegmatite through leucogranite to scheelite-bearing calc-silicate rocks at Miniki Gol (Zahid 1996).

Fig. 9
figure 9

Aqueous fluid inclusions in the scheelite-bearing calc-silicate rocks at Miniki Gol. a Fluid inclusions in the granular quartz intergrown with scheelite. b Aqueous fluid inclusions along with trail of mono-phase fluid inclusions in the quartz grain

Full size image
Fig. 10
figure 10

Total homogenisation temperature (T H) of Miniki Gol scheelite-bearing calc-silicate rocks

Full size image

Discussion

The genesis of the Miniki Gol scheelite mineralisation and its protolith is difficult to assess as the study area has undergone multiple episodes of deformation and metamorphism, followed by the intrusion of leucogranite and post-magmatic hydrothermal activity. On the basis of high levels of Mn and Zn and low levels of K and Rb in the tourmaline-quartz gneiss compared with associated schist, Leake et al. (1989) have considered these tourmaline-rich rocks as siliceous chemical precipitates. These authors have ascribed a sedimentary/exhalative or diagenetic origin to the Miniki Gol scheelite mineralisation and argued that subsequent metamorphism recrystallised the tungsten as scheelite prior to the emplacement of leucogranite. The present study, however, suggests that hydrothermal activity has played a significant role in the transformation of micaceous quartzite into calc-silicate assemblages at Miniki Gol. The mineral paragenesis such as clinozoisite,

quartz, calcic-amphibole, plagioclase, chlorite, biotite, calcite, sphene and garnet, along with scheelite in the calc-silicate rocks, undoubtedly specifies a skarn type environment. The occurrence of clinozoisite (up to 60 % by volume), the coexistence of scheelite and clinozoisite grains (Fig. 5a), and the presence of the calc-silicate minerals indicate that hydrothermal solutions have converted the micaceous quartzites into calc-silicate rocks. The enrichment of Zr, Hf, Be, Sn, W, Th, U, Ga, Nb, F and Y in the scheelite-bearing calc-silicate rocks compared with micaceous quartzite is compatible with hydrothermal alteration. This metasomatic activity is more prominent in the scheelite-bearing calc-silicate rocks than in the unmineralised micaceous quartzite. The CaO level in the calc-silicate rocks is high (mean 9.1 wt%) as compared to the micaceous quartzite (mean 4 wt%). This shows that hydrothermal fluids have leached CaO from associated marble and reprecipitated in the calc-silicate rocks, altering micaceous quartzite into calc-silicate assemblages. The positive correlation among Ga, Nb and Y (Fig. 6c, d) in the scheelite-bearing calc-silicate rocks indicates similar geochemical behaviour of these elements during the hydrothermal regime. Moreover, the high concentrations of F, Li and Cs, and to a lesser extent Be, W, Sn, Pb, Th, Zr, Y, Hf and Nb in Miniki Gol schist (Table 3), compared with the average shale (Turekian and Wedepohl 1961), suggest that these elements have been introduced by metasomatic fluids.

Hydrothermal activity has also affected the REE pattern of the Miniki Gol calc-silicate rocks. Some of the scheelite-bearing rocks, such as ZC 65 a (Fig. 7a), are rich in REE whereas some, such as ZC 70, are depleted in total REE (Fig. 7a). This enrichment and depletion of the REE in the scheelite-bearing calc-silicate rocks, enrichment of HREE over LREE and positive Eu anomalies are suggestive of a pronounced metasomatic activity in these rocks. HREE enrichment and low (La / Lu) N ratios (Table 2) demonstrate hydrothermal alteration and could be related to scheelite deposition with the decreasing temperature of the hydrothermal fluid (Lottermoser 1989b). Positive Eu anomalies have also been reported by Sonnet et al. (1985) in scheelite-bearing calc-silicate rocks of peri-granitic origin. The compatibility of the REE pattern between leucogranite and calc-silicate rocks at Miniki Gol (Fig. 7a, c) also establishes a genetic association between these two rocks. The strong correlation between the elements such as Zr, Hf, Nb, Y, Ta, F and REE (see Fig. 8) indicates an identical geochemical behaviour of these elements in the calc-silicate rocks and also suggests that REE were mobile in the aqueous fluids. The mobility of the REE appears to increases with the change from early-magmatic to late-stage hydrothermal solutions (cf. Lottermoser 1992). The fluoride and carbonate complexes can significantly facilitate the transportation of the REE along with Y in hydrothermal regime (Wood 1990). pt?>The transportation of Zr, Ti, U, Th, Y along with REE in the presence of alkali-rich- fluorine and phosphorus complexes have also been reported by Giere (1990) and Rubin et al. (1993). Moreover, both acidic and alkaline environments facilitate the transportation of H2WO4 and WO4 2− respectively (Robb 2005). The most interesting feature of the Miniki Gol scheelite is the relatively higher levels of Zr and, to a lesser extent, Ta (up to 0.46 and 0.35 wt%, respectively). This enrichment, particularly of Ta, has been related to late-stage granitic and pegmatitic intrusions (Linnen et al. 2012; Stilling et al. 2006; Deer et al. 2013).

The occurrence of tourmalinites at three locations within 2 km strike length, the absence of Sb and Hg in the studied rocks and the lack of cinnabar-stibnite mineralisation at Miniki Gol (Zahid et al. 2013) are inconsistent with the exhalative or volcano-sedimentary origin. The lack of evidence for the presence of evaporites such as secondary halides in the study area rules out the possibility of lacustrine environment. Boron-rich clay is also an inadequate precursor as clay minerals are not known to contain more than 2000 ppm of boron (Slack et al. 1984). This amount of boron would only be sufficient for 9 % tourmaline by volume (Plimer 1986), whereas Miniki Gol tourmalinites contain up to 80 % tourmaline by volume. Leucogranite can be considered as a potential source for the excessive boron of these tourmalinites. The mineral chemistry of tourmaline, although not conclusive, can provide some constraints on any genetic relationship between the leucogranite and tourmalinisation. The average FeO# of the tourmaline (0.6) from the Miniki Gol tourmalinites corresponds to magmatic hydrothermal (greisen-related) Sn-W deposits (cf. Zahid et al. 2013). This ratio of the tourmaline from studied tourmalinites is higher than that of the associated leucogranite, thereby reflecting an Fe-enrichment trend away from the granite as a result of late-stage evolved magmatic differentitaion (see also Benard et al. 1985; Taylor et al. 1992). Moreover, the reverse trend of chemical zoning, i.e. higher MnO in rims than cores, in the tourmaline grains from the tourmalinites at Miniki Gol corresponds to a metasomatic origin (cf. Zahid et al. 2013).

Aqueous fluid inclusions in the scheelite-bearing calc-silicate rocks are of very low salinity. A generally decreasing trend of the homogenisation temperatures from CO2-bearing fluid inclusions to aqueous fluid inclusions within the Miniki Gol pegmatite correlates with decrease in salinity (Zahid 1996). This suggests that magmatic fluids were probably mixed with meteoric water. Such a phenomenon is a characteristic feature of many hydrothermal and greisen-related deposits, such as the tungsten deposits of Xihuashan, China (Giuliani et al. 1988) and Cligga Head, SW England (Spooner 1981). The occurrence of calcite in the Miniki Gol scheelite-bearing rocks demonstrates that some CO2 was possibly present at the early stage of the hydrothermal fluids. Tungsten was introduced into the site of deposition (calc-silicate rocks) probably during the last phase of the hydrothermal activity, when pH was relatively high. As pointed out by Absar (1991), clinozoisite usually forms from high pH and CO2-deficient fluid whereas calcite precipitates from CO2-rich fluids. The continuous separation and consumption of CO2 by calcite from the hydrothermal fluid may have increased the pH of the fluid and facilitated the deposition of clinozoisite together with scheelite from the CO2-deficient fluid. The scarcity of the CO2-bearing fluid inclusions in the scheelite-bearing calc-silicate rocks and the association of clinozoisite with the scheelite imply that the fluid phase at the time of growth of these minerals was predominantly aqueous. This is consistent with the laboratory experiments of Keppler and Wyllie (1991) that solubility of W is very high in the presence of water as the sole volatile.

Conclusions

The main conclusions derived from this study are as follows:

  1. 1.

    Hydrothermal fluids played a key role in the transformation of micaceous quartzite into scheelite-bearing calc-silicate mineralogy.

  2. 2.

    The association of clinozoisite, calcic-amphibole, plagioclase, chlorite, biotite, calcite, sphene and garnet indicates skarn style mineral assemblages.

  3. 3.

    The enrichment of Zr, Hf, Be, Sn, W, Th, U, Ga, Nb, F and Y, along with total REE in the scheelite-bearing calc-silicate rocks and metasediments, suggests a significant metasomatic activity at Miniki Gol.

  4. 4.

    It is concluded that the intergrown scheelite and clinozoisite were formed from CO2-deficient and high-pH hydrothermal fluids.