Mn–Fe-rich genthelvite from pegmatites associated with the Madeira Sn–Nb–Ta deposit, Pitinga, Brazil: new constraints on the magmatic-hydrothermal transition in the albite-enriched granite system

Abstract

Genthelvite from pegmatite veins hosted by the albite-enriched granite (ca.1.8 Ga) corresponding to the Sn–Nb–Ta (F, REE, Li, Zr, U, Th) Madeira deposit, Amazonas, Brazil was studied. Genthelvite, the exclusive Be-bearing mineral within the deposit, occurs as massive crystals of up to 4.7 cm in size. Compositions are homogeneous within individual crystals, although there is moderate variation in the overall composition reflecting relatively limited substitutions within the helvine–genthelvite–danalite solid-solution series, with relatively high Zn contents (36.96 to 49.45 wt.% ZnO), lower Mn contents (0.61 to 3.03 wt.% MnO), and variable Fe contents (2.10 to 10.94 wt.% FeO), completing an existing compositional gap in this system. Genthelvite formed in an alkaline and subaluminous environment, under stable conditions within the late-evolved fluids, at relatively high temperature (>400°C), in a reducing environment. The extremely high concentration of fluorine in the magma and the crystallisation of magmatic galena resulted in an effective reduction of H₂S fugacity. This resulted in the stabilisation of genthelvite during the transition from the late magmatic to early hydrothermal stages of the albite-enriched granite evolution. The variability in Fe content within genthelvite is associated primarily with localised variations in the mineral assemblage (e.g. the presence of riebeckite and polylithionite). Genthelvite was altered by low-temperature aqueous fluids rich in F which resulted in the incorporation of Fe, Mn, Mg, Pb, Ba, Na, K, U and REE into the Zn²⁺ structural site and the allocation of excess Si, Al, Ti and P in the ^IVSi and ^IVBe structural sites. The substantial content of U and REE substituting for Zn, together with Si substituting for Be, is charge balanced by the presence of vacancies at the A site.

Introduction

Helvine-group minerals are anhydrous sulfosilicates, cubic and isostructural with the space group P $\bar{4}$3n and are sodalite structural types. They have the general formula A₈Be₆(SiO₄)₆S₂, in which the species are defined by the cation in the A crystallographic site. The species helvine (Mn₄Be₃Si₃O₁₂S), danalite (Fe₄Be₃Si₃O₁₂S) and genthelvite (Zn₄Be₃Si₃O₁₂S) form a solid solution, whose proportions are defined by the states of reduction, sulfidation and alkalinity of the system (Burt, 1980). Complete miscibility should exist between the three end-members (Hassan and Grundy, 1985), however, compositional gaps exist between the end-members Zn–Fe, Zn–Mn and Mn–Fe and pure danalite has not been observed in Nature (Oftedal and Saebø, 1963; Clark and Fejer, 1976; Dunn, 1976; Larsen, 1988; Perez et al., 1990; Langhof et al., 2000; Bilal, 2013).

The helvine-group minerals are only found in peralkaline and alkaline granites, syenites, rare metal pegmatites, albitites, greisens, skarns and contact zones (Deer et al., 2004). In this work, we have investigated the occurrence of genthelvite in pegmatite veins hosted by the albite-enriched granite facies of the Madeira granite. This facies corresponds to the Madeira deposit, which is characterised by an association of Sn with cryolite, Nb, Ta, Y, rare earth elements (REE), Li, Zr, U and Th in the same albite-enriched granite that hosts a massive cryolite deposit. The genthelvite crystals occur in pegmatite veins found in the most differentiated portion in the centre of the pluton. These pegmatites fit in the CMS (Dill, 2016) classification as the 24dE type because they are hosted in peralkaline igneous rocks and are carriers of REE–Y ores. According to the classification in Černý and Ercit (2005), they belong to the Rare Elements class and the NYF family, as they are rich in REE, Nb, Y and F, and are associated with A-type granites in environments with low pressures and temperatures. The study of genthelvite brings new constraints on the conditions of the magmatic-hydrothermal transition in the albite-enriched granite system.

Previous work

Geological setting

The Pitinga Province is located (Fig. 1) in the southern portion of the Guyana Shield (Almeida et al., 1981), in the Tapajos-Parima Tectonic Province (Santos et al., 2000). The volcanic rocks of the Iricoume Group (Veiga et al., 1979) predominate in the Pitinga Province (Fig. 1) and have ²⁰⁷Pb/²⁰⁶Pb zircon ages between 1881 ± 2 and 1890 ± 2 Ma (Ferron et al., 2006). They consist mostly of effusive and hypabyssal rhyolites, highly welded ignimbrites, ignimbritic tuffs, and surge deposits formed in a subaerial environment with cyclic effusive and explosive activities (Pierosan et al., 2011; Simões et al., 2014). The Iricoume Group host the Madeira Granite (Fig. 1).

Figure 1. (a) Location map and (b) geological map of the Madeira Granite, Pitinga, Brazil (modified from Costi, 2000).

The Madeira granite (Figs 1, 2) contains four facies (Horbe et al., 1991; Lenharo et al., 2003; Costi et al., 2005, 2009; Bastos Neto et al., 2009). The oldest, mostly metaluminous porphyritic amphibole–biotite granite (1824 ± 2 Ma, Costi et al., 2000), contains plagioclase-mantled K-feldspar megacrysts, sometimes have reverse-zoned K-feldspar-mantled plagioclase ovoids, and is usually referred to as the ‘rapakivi’ subfacies. The amphibole–biotite granite was followed by the metaluminous biotite granite (1822 ± 2 Ma, Costi et al., 2000). The younger facies are the hypersolvus porphyritic alkali feldspar granite (1818 ± 2 Ma, Costi et al., 2000) and the albite-enriched granite (Fig. 2). The latter is the host to the pegmatites investigated in this work. The age of the albite-enriched granite is only very roughly constrained at 1822 ± 22 Ma (Bastos Neto et al., 2014) due to the metasomatic alteration of zircons. According to Costi (2000), these younger facies were emplaced simultaneously. The hypersolvus granite has K-feldspar phenocrysts in a fine- to medium-grained matrix composed dominantly of K-feldspar and quartz.

Figure 2. Geological map of the albite-enriched granite (modified from Minuzzi, 2005).

The albite-enriched granite

The albite-enriched granite is an oval-shaped body with a surface outcrop of ~2×1.3 km. It is divided into subfacies of a core albite-enriched granite and a border albite-enriched granite. The core albite-enriched granite is a peralkaline subsolvus granite, porphyriticto-seriate in texture, fine-to-medium grained, and composed of quartz, albite and K-feldspar in approximately equal proportions (25–30%). The accessory minerals are cryolite (5%), polylithionite (4%), green–brown mica (3%), zircon (2%) and riebeckite (2%). Pyrochlore, cassiterite, xenotime, columbite, thorite, magnetite and galena occur in minor proportions. The border albite-enriched granite is peraluminous and exhibits types of texture and essential mineralogy similar to that of the core albite-enriched granite, except for being richer in zircon, with fluorite instead of cryolite. Iron-rich silicate minerals are absent, having been completely removed by autometasomatic processes (Costi et al., 2000, 2009).

The Madeira rare-metal deposit

The Pitinga Province is the largest Sn producer in Brazil. The alluvial ore deposits were discovered in 1979 (Veiga et al., 1979) and are almost exhausted. The primary ores are associated mainly with the Madeira deposit, which has been exploited since 1989. The Madeira deposit corresponds to the albite-enriched granite facies of the Madeira Granite and presents disseminated ore with grades of 0.17 wt.% Sn (cassiterite), 0.20 wt.% Nb₂O₅) and 0.024 wt.% Ta₂O₅ (both in pyrochlore and columbite). The potential by-products of the disseminated ore are F (4.2 wt.% cryolite), Y and heavy rare earth elements (HREE) (xenotime), Zr and Hf (zircon), Th (0.07 wt.% ThO₂, thorite), and U (pyrochlore). Regardless of the disseminated character of the albite-enriched granite mineralisation, there are small zones of enrichment associated with the granite in which particular minerals are considerably abundant. These are: (1) thick pods and bands of the pegmatitic albite-enriched granite (~50 cm, rarely up to 10 m thick; Stolnik, 2015) that show gradational contacts with the albite-enriched granite; These contain the same minerals as the core albite-enriched granite, although polylithionite, riebeckite, xenotime and thorite are more abundant and of large size; (2) border pegmatites at the contact between the border albite-enriched granite and the older facies (Fig. 2) characterised by increased size and amounts of quartz and zircon, advanced alterations of K-feldspar and biotite and by local enrichments in fluorite, polylithionite, thorite and secondary hematite (Lengler, 2016); (3) several bodies of massive cryolite intercalated with core albite-enriched granite and hypersolvus granite, which are sub-horizontal, up to 300 m long and 30 m thick and composed of cryolite (~87 vol.%), quartz, zircon and feldspar (Minuzzi et al., 2006); and (4) pegmatite veins (Paludo et al., 2018; Ronchi et al., 2019) which are not mappable, occur more commonly in the central, northern and northwest parts of the core albite-enriched granite, and have thicknesses ranging from a few centimetres up to 2 m. Genthelvite occurs only in these pegmatite veins and was first identified by Ronchi et al. (2011). There are two types of pegmatite veins visible in the whole mine front (140 to 210 meters above sea level) hosted by the core albite-enriched granite. The prevalent type is that of tabular bodies emplaced in horizontal extension fractures. The other group is formed by tabular bodies emplaced in subvertical reverse fault planes. These fractures and faults served as a conduit for the fluids, with transportation from southwest to northeast, in a compressive system, with horizontal tension and at low solidus temperature (Ronchi et al., 2019). These have been separated into three groups in Paludo et al. (2018) on the basis of composition and modal values as: (1) those rich in amphibole (riebeckite, fluoro-arfvedsonite and fluoro-eckermannite), with intermediate contents of K and Na; (2) rich in polylithionite, with high K contents; and (3) cryolite-rich, with high Na contents.

Evolution of the albite-enriched granite system

Bastos Neto et al. (2009, 2014) consider that the A-type magmatism in Pitinga evolved in a post-collisional extensional setting, probably in a within-plate scenario in which extensional and trans-tensional tectonic regimes dominated. In this context, the albite-enriched granite magma would have been related to the isotherm rise, which occurred when the mantle fluid ascended further into the crust promoting fenitisation-type reactions (Martin, 2006) in rocks previously enriched in Sn, and introduced elements such as F, Nb, Y, REE and Th in anomalous concentrations. The input of a F-rich fluid generated metasomatism causing the rock to become fusible.

Lenharo (1998) and Costi (2000) considered that the albite-enriched granite magma evolved towards an extremely Na-, F-enriched residual melt. Costi (2000) interpreted that, at the point of H₂O saturation, the extremely F-enriched residual fluid separated into an aqueous, relatively F-poor portion and a low-H₂O, Na–Al–F-rich portion. The latter fraction resulted in the formation of massive cryolite bodies, whereas the H₂O-rich fraction formed the associated quartz-, feldspar- and mica-bearing pegmatitic rocks. In accordance with Bastos Neto et al. (2009), extreme fluorine enrichment in the residual melt is unlikely to have been attained because the F content was buffered by crystallisation of magmatic cryolite (Dolejs and Baker, 2007). Furthermore, fluid-inclusion data (Bastos Neto et al., 2009; Ronchi et al., 2011) show that the massive cryolite deposit was formed from an aqueous, saline hydrothermal fluid. The higher homogenisation temperature of 400°C, measured in massive cryolite, determines the minimum starting temperature for the hydrothermal process.

Methods

Genthelvite crystals from several pegmatitic veins were described and identified by combining optical properties, using compositional analyses and powder X-ray diffraction data. The pegmatite veins are spread throughout the core albite-enriched granite and are not mappable due to their small size (up to 2 m thick). Sampling was carried out mainly in the central area of the core albite-enriched granite, on the surface of the open pit. Over 50 thin sections of the pegmatites were analysed and, among those which contained genthelvite, 10 were examined by back-scattered electron microscopy (BSE), with qualitative analysis using an energy-dispersive X-ray detector (Zeiss, model EVO MA10) at the Centre for Microscopy and Microanalysis in Universidade Federal do Rio Grande do Sul (UFRGS).

Mineral compositions of genthelvite were obtained using electron probe micro-analysis (EPMA) (JEOL JXA-8230) at the EPMA Laboratory of the Universidade de Brasília (UnB). The operation conditions were: 15 kV accelerating voltage and 10 nA beam current (F, Mg, Zn, Al, Si, Hf, Nb, P, Cl, S, Bi, Ti, Mn, Y, Ta, Sn, Ca, Zr, Fe, V and Rb), and 20 kV and 50 nA (Na, Er, Tm, Yb, Ho, Lu, K, Pb, Dy, Tb, Sm, Gd, Eu, Sr, Th, Pr, Nd, Ce, La, Ba and U), 1 μm beam diameter, and interference corrections were applied in all cases of peak overlap. The wavelength dispersive X-ray spectrometer (WDS) crystals used were: TAP (Si, Zn, Na, Al); PETJ (Nb, P, Hf, Cl, S, K, Bi, Sr, Y, Ta, Sn, Th, Pb); PETH (Rb, Zr, U); LIF (Ti, Mn, Sm, Eu, Gd, Dy, Er, Ho, Tb, Tm, Yb, Lu), LIFH (Ca, Fe, Ba, V, La, Ce, Pr, Nd); and LDE1 (F). The counting times on the peaks were 10 s for all elements, and half that time for background counts on both sides of the peaks. The following natural and synthetic standards were used: microcline (Si, K, Al); albite (Na); apatite (P; Ca); andradite (Fe); topaz (F); forsterite (Mg); vanadinite (V, Pb, Cl); pyrite (S); MnTiO₃ (Mn); YFe₂O₁₂ (Y); LiNbO₃ (Nb); LiTaO₃ (Ta); MnTiO₃ (Ti, Mn); ZnS (Zn); Bi₂O₃ (Bi); RbSi (Rb); BaSO₄ (Ba); baddeleyite (Zr); PbS (Pb); HfO₂; SrSO₄ (Sr); SnO₂; ThO₂; UO₂; and synthetic REE-bearing glasses.

Crystallographic investigations were performed in the X-ray diffraction laboratory at UFRGS using a Siemens D5000 X-ray Diffractometer (XRD) with a scanning step of 0.05°2θ, a time of 1 s, between 5 and 100°2θ, CuKα radiation (1.5418 Å) and a Ni filter. Crystallographic parameters were determined using the UnitCell program (Holland and Redfern, 1997), and the diffractions of 19 (reflections) faces were processed. The error in the processed values was 0.00017, with 95% reliability. The crystallographic information files have been deposited with the Principal Editor of Mineralogical Magazine and are available as Supplementary material (see below).

Results

Mineralogy and petrography

The contact of the pegmatite veins with the host rock is abrupt, marked by a thin, well-defined border, typically on the scale of centimetres. From the border towards the centre of the bodies, there is a systematic increase in the size of the minerals, though without the establishment of a clear zoning pattern. The interior of the bodies is homogeneous, characterised by anhedral to subhedral minerals. In most veins, the pegmatitic texture is marked by crystals (up to 10 cm) of polylithionite, quartz, cryolite, microcline and albite, by crystals (up to 7 cm) of riebeckite, xenotime and genthelvite, and crystals (up to 3 cm) of thorite, galena and, more rarely, zircon, cassiterite and gagarinite.

Both the horizontal and the subvertical pegmatite veins exhibit the same mineralogical composition. The paragenetic succession is illustrated in Fig. 3. The major mineral assemblage consists of quartz, microcline, cryolite, polylithionite and riebeckite. The minor minerals are xenotime, thorite, gagarinite-(Y), and genthelvite. Accessory minerals are zircon, cassiterite, pyrochlore, columbite, galena, sphalerite, native lead, native bismuth, pyrite, hematite and chlorite. The matrix primarily consists of albite, quartz and grains are commonly anhedral, colourless in natural 0.3–1 cm. Additionally, polylithionite, cryolite and riebeckite are present. The entire magmatic paragenesis has undergone significant alteration due to the influence of F-rich hydrothermal fluids, ultimately leading to the formation of the secondary phases depicted in Fig. 3.

Figure 3. Paragenetic evolution of the genthelvite-bearing pegmatite veins. The line thickness is indicative of the abundance of the mineral. The precursor minerals for the most important replacement reactions are indicated by arrows.

Genthelvite crystals from the pegmatite veins hosted by the core albite-enriched granite have sizes from 1.0 mm to 4.7 cm and present a light pink colour in macroscopic samples (Fig. 4). Optically, the grains are commonly anhedral, colourless in natural light and isotropic in polarised light (Fig. 5a,b). In the pegmatites genthelvite occurs predominantly as massive aggregates of crystals surrounding polylithionite and magmatic quartz and includes the accessory minerals pyrochlore, thorite and zircon (Fig. 5a–d). Subordinately, genthelvite occurs filling voids and microfractures in magmatic quartz (Fig. 5e) and polylithionite phenocrysts or arranged interstitially in the matrix with magmatic quartz and orthoclase (Fig. 5f).

Figure 4. Macroscopic sample of genthelvite from the pegmatite veins hosted by the core albite-enriched granite. Genthelvite (Ghv) occurs surrounding polylithionite (Pln) crystals.

Figure 5. Microphotographs and BSE image of genthelvite from the pegmatite veins hosted by the core albite-enriched granite: (a) typical genthelvite from the pegmatites, with triangular cleavage, associated with polylithionite, plane polarised light (PPL); (b) same as in (a), crossed polarised light (XPL); (c) genthelvite filling the space between polylithionite crystals, PPL; (d) genthelvite associated with magmatic quartz, pyrochlore, thorite and zircon, PPL; (e) genthelvite filling voids in magmatic quartz, XPL; (f) genthelvite in the matrix with magmatic quartz and microcline, XPL; (g) BSE image of genthelvite with microfractures filled by hydrothermal cryolite; (h) hydrothermal quartz along genthelvite growth lines, XPL. Abbreviations: Crl II = hydrothermal cryolite, Ghv = genthelvite, Mcc = microcline, Pcl = pyrochlore, Pln = polylithionite, Qz I = magmatic quartz, Qz II = hydrothermal quartz, Thr = thorite and Zrn = zircon. Mineral symbols after Warr (2021).

The contact with polylithionite is rectilinear and slightly reactive. The contact with magmatic quartz is undulated and shows reactive rims. The contact with pyrochlore, zircon and thorite is undulated and reactive, and these minerals have a partially dissolved aspect. The samples show evidence of hydrothermal alteration. Genthelvite is characterised by corrosion features such as cavities and microfractures which are commonly filled by hydrothermal cryolite (Fig. 5g). Hydrothermal quartz also occurs associated with genthelvite, specifically filling the channels opened along genthelvite growth lines (Fig. 5h). On the basis of these characteristics, genthelvite is considered a mineral of late crystallisation, preceded by the crystallisation of polylithionite and early magmatic quartz and formed before the hydrothermal cryolite.

Genthelvite composition

The helvine-group minerals can be represented by the general formula A²⁺₈Be₆Si₆O₂₄S₂. Structural formulae calculated on the basis of 26 O and S atoms and Be = 6 atoms per formula unit (apfu) (Zito and Hanson, 2017) show cations systematically in deficit on the A site and in excess on the Si site. Therefore, in this study calculations were performed with the following assumptions: the crystal structure is charge balanced; the anion site is fully occupied (i.e. O + S = 26); the Si site is fully occupied by P⁵⁺, Si⁴⁺, Ti⁴⁺ and Al³⁺ (preferential order of occupation is P > Si > Ti > Al according to ionic potential); the excess in the Si site (i.e. ^IVSi > 6 apfu) is allocated in the ^IVBe site with the preferential order of occupation Al > Ti > Si > P with Be²⁺ calculated to yield a total of Be + Si = 12 apfu (i.e. site ^IVBe = 6) (Dunn, 1976; Finch, 1990); and the A site has a vacancy (□) and is occupied by U⁴⁺, A³⁺ (Ce, Pr, Nd, Sm, Eu, Gd, Ho and Er), A²⁺ (Zn, Fe, Mn, Mg, Pb and Ba) and A⁺ (Na and K) (i.e. site ^IVA = 8 – □).

Representative compositions for genthelvite are given in Table 1. The low analytical totals are probably due to the existence of vacancies within the genthelvite structure, hydration as a result of hydrothermal alteration, and the inherent errors associated with Be calculation. The elements P, Ti, Al, REE, Ba, Na and K have concentrations near the detection limits of the EPMA. Consequently, discussions regarding the incorporation of these elements into the crystal structure of genthelvite must be approached with caution. Additionally, low concentrations (hundreds to thousands of ppm) of V, Th, Cr, La, Dy, Tm, Yb, Lu, Bi, Ca, Ni, Sr and Cl were detected. These elements were not considered in the totals of the analyses and in structural calculations.

Table 1. Representative compositions from EPMA (wt.%) for Mn–Fe-rich genthelvite in the pegmatite veins hosted by the albite-enriched granite.

*BeO calculated considering a total of Be + Si = 12 apfu; n.d. = not detected.

Genthelvite has a homogeneous composition within individual grains and is characterised as a Mn–Fe-rich genthelvite. It represents a solid solution within the genthelvite–danalite–helvine system, with relatively limited substitutions occurring between Zn²⁺, Fe²⁺ and Mn²⁺. However, genthelvite does exhibit moderate overall compositional variability, falling within the range of 36.96 to 49.45 wt.% ZnO, 2.10 to 10.94 wt.% FeO, and 0.61 to 3.03 wt.% MnO. The composition of genthelvite plotted in terms of the relative proportions of Zn, Fe and Mn (expressed as percentages of [Zn + Fe + Mn] atoms) (Fig. 6) reflect the predominance of compositions along the upper part of the Zn–Fe join, although invariably the presence of a small component of helvine occurs.

Figure 6. Compositions of genthelvite, danalite and helvine expressed as percentages of Zn + Fe + Mn atoms.

In genthelvite, other cations that occupy the structural site of Zn, in addition to Mn and Fe, are uncommon, and trace concentrations of K, Ca and Mg are the most commonly reported. In genthelvite from Pitinga the maximum values were 0.062 wt.% MgO, 0.061 wt.% PbO, 0.14 wt.% BaO, 0.33 wt.% Na₂O and 0.05 wt.% K₂O. Uranium concentrations have not been reported in genthelvite from other localities, but in Pitinga it occurs in all genthelvite samples with 0.13 to 0.25 wt.% UO₂. Additionally, the genthelvite samples have high contents of REE (maximum 0.40 wt.% REE₂O₃) relative to that from Cheyenne Canyon (USA, 4.1 ppm REE₂O₃; Zito and Hanson, 2017) and other helvine-group minerals such as the Mn–Zn-rich danalite from Sucuri (Brazil, maximum 363 ppm REE₂O₃; Raimbault and Bilal, 1993) and the Zn–Fe-rich helvine from Dajishan (China, maximum 13 ppm REE₂O₃; Raimbault and Bilal, 1993). The average concentration of light rare earth elements (LREE) (723 ppm) is slightly higher than that of HREE (565 ppm) in genthelvite from Pitinga.

Genthelvite from this work exhibits a REE normalised distribution pattern (Fig. 7) that closely resembles that of the host pegmatite, except for a positive anomaly in Pr and Eu, and the absence of La, Tb, Dy, Tm, Yb and Lu. The mineral incorporated LREE preferentially, although helvine-group minerals have high affinity with HREE (Raimbault and Bilal, 1993; Deer et al., 2004).

Figure 7. Chondrite-normalised REE patterns (chondrite of Anders and Grevesse, 1989).

The strong negative correlation (R ² = 0.963) between Zn and the sum of the cations Fe, Mn, Mg, Pb, Ba, Na, K, U and REE attest to their location at the A site (Fig. 8a). The positive correlation of Mn and Fe (R ² = 0.672, Fig. 8b) is evidence of their concomitant entrance when substituting for Zn. Concerning the elements in the ^IVBe and ^IVSi structural sites, the values are similar in all the samples, with little variations in the concentrations of Si (29.30 to 32.19 wt.% SiO₂), S (5.07 to 5.52 wt.% S), and calculated Be (11.73 to 12.49 wt.% BeO). The BeO concentration is similar to those found by inductively coupled plasma atomic emission spectroscopy (ICP–AES) analyses in danalite, which had an average content of 13.1% BeO (Raimbault and Bilal, 1993). In addition, maximum values of 0.38 wt.% P₂O₅, 0.30 wt.% TiO₂ and 0.25 wt.% Al₂O₃ were observed. In contrast to all other minerals found in the pegmatites, genthelvite does not contain fluorine. This absence is probably due to the fact that F has an ionic radius that is too small for sodalite-type structures. The ΣA site gave a negative correlation with Si + Al + P + Ti (R ² = 0.900, Fig. 8c), meaning that the Si substituting for Be is charge balanced by the presence of vacancies at the A site.

Figure 8. Binary diagrams for genthelvite from the pegmatites associated with the albite-enriched granite: (a) ΣA site (– Zn) versus Zn; (b) Mn versus Fe; and (c) ΣA site versus Be and ΣA site versus Si + Al + P + Ti. Concentrations are expressed in apfu.

Lattice parameters of genthelvite

Genthelvite from the pegmatites associated with the albite-enriched granite have an average a parameter of 8.127 Å, varying between 8.117 Å and 8.134 Å (error of 0.00017 Å), which is in accordance with the expected values for this mineral (Table 2). In the helvine-group minerals the constancy of the structural dimensions of the BeO₄ and SiO₄ tetrahedra (Hassan and Grundy, 1985) imply that variations in the unit cell parameter are related to the Mn–Zn–Fe proportions (Oftedal and Saebø, 1963), as well as the proportions of other elements in the A site. For genthelvite from Pitinga and other localities (Table 2), the correlation between the average a parameter and the average Mn content is strongly positive (R ² = 0.84, Fig. 9a), but with the average Zn + Fe content this parameter presents a strong negative trend (R ² = 0.89, Fig. 9b). Though there is a moderate overlap of unit cells, as indicated by the error bars, the correlation can be associated tentatively with the ionic radii of the elements. The ionic radius is larger for Mn (0.66 Å) and smaller for Zn (0.60 Å) and Fe (0.63 Å) (Shannon, 1976).

Table 2. Crystallographic parameters of genthelvite from the pegmatites (Madeira deposit, Pitinga Mine) and other localities, in descending order of the average of parameter a.

Abbreviations: n.a. = not available.

Figure 9. Correlation of the unit cell parameter a (Å) versus the Mn (a) and Zn + Fe (b) concentrations of genthelvite from Pitinga (this work, filled circle) and other localities (open circles, Table 2). Mn, Zn and Fe are expressed as wt.%. The bars indicate the error for the parameter a.

Discussion

Geological environment of genthelvite occurrences

Beryllium is a rare element both in meteorites and on Earth, but it is a crustal element par excellence, with an average of 2.1 ppm BeO in rocks of the upper continental crust, in contrast to 1.4 ppm BeO in the lower crust and 0.07 ppm BeO in the mantle (Rudnick and Gao, 2005). The first paragenesis of magmas are formed by minerals whose structure inhibits the capture of Be in melting. Therefore, Be enrichment occurs in the final stages of magmatic crystallisation, mainly in granitic pegmatites and alkaline rocks (Grew, 2002). In the later stages of differentiation, there is a supersaturation of SiO₂ and accumulation of alkalis and volatiles, allowing the formation of Be minerals and quartz (Pulz et al., 1998).

A significant majority of occurrences of genthelvite are in pegmatites and in late-formation rocks as hydrothermal veins, greisens and skarns, associated with alkaline to peralkaline granites and syenites (Table 3). These occurrences have in common a highly evolved magma enriched in high-field-strength elements and associated hydrothermal events and/or metasomatism. The genthelvite investigated occurs in pegmatite veins that represent the most evolved fluid of a peralkaline magmatic system, but contrast with other genthelvite occurrences of the world showing an unusual mineralogical association and geochemical trend.

Table 3. The genthelvite investigated from Madeira deposit, Pitinga Mine, Brazil and comparison with other occurrences.

References: (1) This work; (2) Glass et al. (1944); (3) Zito and Hanson (2017); (4) Langhof et al. (2000); (5) Vasil'ev (1961); (6) Von Knorring and Dyson (1959); (7) Raimbault and Bilal (1993); (8) Bilal (2013); (9) Miranda (2018); (10) Dunn (1976); (11) Haapala and Ojanperã (1972); (12) Haapala and Lukkari (2005); (13) Botelho (1992); (14) Freitas (2000); (15) Lunts and Saldau (1963); (16) Morgan (1967); (17) Clark and Fejer (1976); (18) Es'kova (1957); (19) Oftedal and Saebø (1963); (20) Larsen (1988); (21) Bollinberg and Petersen (1967); (22) Dushin et al. (2018); (23) Kingsbury (1961); (24) Bulakh and Frank-Kamenetsky (1961); (25) Cianciulli and Verbeek (2003); (26) Leavens et al. (2009)

*Findings by Kingsbury are considered dubious see e.g. Ryback et al. (1998)

Mineral assemblages of genthelvite worldwide

Considering the stability fields of minerals, genthelvite can be formed from willemite and phenakite; willemite can be altered to sphalerite; genthelvite can be altered to sphalerite and phenakite or bertrandite; and all of them together with quartz (Burt, 1988). Therefore, commonly associated minerals with genthelvite are quartz, feldspar, micas and other Zn-bearing phases such as sphalerite, willemite and gahnite, as well as other Be-containing phases such as phenakite and bertrandite (Burt, 1988). There is also willemite in these associations, especially in peralkaline rocks. Genthelvite with willemite, phenakite, Na-fluorides as gagarinite, weberite and pachnolite have been reported in metasomatic peralkaline rocks from Russia (Kudrin, 1978). In the Ilimaussaq Complex, Greenland, willemite with chkalovite was reported as the only Be-bearing mineral, whereas genthelvite was found at a different location within the Complex (Metcalf-Johnson, 1977). In Mont St. Hilaire, Canada, willemite and genthelvite were reported, together with sphalerite and galena (Bank, 1975; Dunn, 1976).

In this investigation, genthelvite is the only Be-bearing phase, and it is associated with polylithionite, riebeckite, quartz, microcline, albite, xenotime, cryolite and the accessory phases pyrochlore, columbite, zircon, gagarinite, galena, sphalerite and hematite. Beryl, willemite and phenakite or bertrandite were not observed. This occurrence is notably unusual because it lacks the helvine-group minerals danalite and helvine, which are commonly associated with genthelvite in other deposits. This absence is probably linked to the physical-chemical conditions discussed below. In peralkaline conditions, genthelvite associated with aluminous minerals such as beryl and topaz are restricted, but genthelvite with Na-fluorides are typical (Burt, 1988), as observed in the genthelvite–cryolite association in the Pitinga pegmatites.

Controls on genthelvite composition

Crystallographic and structural data, the ionic radius of the A-site cations and the structural geometric model, indicate that complete miscibility should exist between the three end-members of the helvine–genthelvite–danalite solid solution (Hassan and Grundy, 1985), however, there are apparent compositional gaps in the Zn–Fe, Zn–Mn and Mn–Fe trends, and pure danalite has not been observed in Nature (Oftedal and Saebø, 1963; Clark and Fejer, 1976; Dunn, 1976; Larsen, 1988; Perez et al., 1990; Langhof et al., 2000; Bilal, 2013). According to Antao and Hassan (2010) the absence of pure danalite in Nature might simply indicate that another phase is more stable than danalite.

Genthelvite in this investigation has high concentrations of Zn and small concentrations of Fe in comparison with genthelvite from the Rockport Deposit, Massachusetts, USA (Dunn, 1976) and from the Cairngorm Mountains, UK (Clark and Fejer, 1976). It also has a higher Zn and lower Fe and Mn content than the more Mn–Fe-enriched genthelvite reported from Cumberland, Rhode Island, USA (Dunn, 1976) and from the Sucuri Granite in Brazil (Miranda, 2018). It is also different from the typical Fe–Mn-rich genthelvite reported in Norway (Oftedal and Saebø, 1963; Larsen, 1988), the Air Mountains of Niger (Perez et al., 1990) and Uto, Sweden (Langhof et al., 2000). Finally, with this work we demonstrate the existence of natural genthelvite in the upper Zn–Fe trend of the Zn–Fe–Mn ternary diagram (Fig. 6), filling a compositional gap.

Genthelvite is one of only two known silicates in which Zn and Be occur together. Regardless that Zn (ionic radius 0.60 Å; Shannon, 1976) and the much smaller Be (ionic radius 0.27 Å; Shannon, 1976) represent different chemical affinities, they both have the tendency to be concentrated by fractionated crystallisation and to adopt four-fold coordination (Burt, 1988). Iron and Mn (^IVMn = 0.66 Å, ^IVFe = 0.63 Å; Shannon, 1976) also adopt this coordination. Regardless of these coordination controls compositional variations in helvine-group minerals have been shown to have a direct association with temperature and S and O fugacity, rather than with the availability of Zn, Fe and Mn in the fluid. An investigation of several genthelvite crystals from a granitic massif (Antao and Hassan, 2010) showed that Mn-rich genthelvite formed at lower temperatures than Mn-poor genthelvite. In the Taghouaji Alkaline Complex, Niger, Mn-poor genthelvite occurs with sphalerite and galena in a low $f_{{\rm O}_2}$ and high $f_{{\rm S}_2}$ environment and a crystallisation temperature higher than 375°C, whereas the Mn-rich genthelvite occurs with hematite at temperatures of ~288°C (Perez et al., 1990).

The Mn-poor composition of the genthelvite investigated here is probably influenced by its formation at relatively high temperatures, in a low $f_{{\rm O}_2}$ and high $f_{{\rm S}_2}$ environment. However, the significant variation in Zn–Fe content between individual grains (36.96 to 49.45 wt.% ZnO, 2.10 to 10.94 wt.% FeO) is probably controlled by the local mineral assemblage. In the magmatic stage of the pegmatite veins, Fe and Zn were buffered by the presence of riebeckite (~30 wt.% FeO + Fe₂O₃; ~2 wt.% ZnO) and polylithionite (~6 wt.% FeO; ~1 wt.% ZnO). In these minerals, the Mn content is typically <1 wt.% MnO (Paludo et al., 2018). Additionally, the available Pb and S combined to form magmatic galena. During the early hydrothermal stage, Fe was incorporated into altered thorite and altered pyrochlore, whereas both Fe and Mn were widely incorporated into secondary Mn–Fe-rich columbite. Zinc and S were incorporated into secondary sphalerite, and the remaining available S formed galena and pyrite. In the late hydrothermal stage, Fe contributed to the formation of hematite, surrounding all the previous formed minerals (Minuzzi et al., 2005; Bastos Neto et al., 2009; Hadlich et al., 2019). Genthelvite crystallised during the transition between the magmatic and hydrothermal stages of pegmatite evolution, following the crystallisation of riebeckite and polylithionite. Thus, the compositional variation in the genthelvite is closely related to the prior formation of these minerals, which played a significant role in buffering the Fe content relative to Zn in their immediate vicinity.

These variations in composition are also verified by the REE contents. Genthelvite from Pitinga has a REE pattern enriched in LREE relative to its host pegmatite. The average LREE content is slightly higher than that of HREE content, which differs from those of other localities, which have higher HREE contents. The HREE enrichment in relation to the LREE in the helvine-group minerals from Sucuri, Brazil and Dajishan, China (Raimbault and Bilal, 1993) was attributed by these authors only to crystallographic controls. The higher concentration of HREE in danalite from Cheyenne, USA (Zito and Hanson, 2017), was attributed to the presence of late F-enriched fluids. At Pitinga, the REE are primarily concentrated in xenotime (with a predominance of HREE as Dy, Yb and Lu, Bastos Neto et al., 2012) and polylithionite. As these minerals crystallise before genthelvite, the REE content in genthelvite is lower and skewed towards the LREE, even though the late-stage fluids are rich in fluorine. The Eu positive anomaly probably reflects a reducing magma in which the Eu²⁺ could preferentially substitute for Zn²⁺ and because of the crystallisation of albite instead of anorthite in the pegmatites.

The possibility that the presence of REE and U in genthelvite might result from hydrothermal alteration cannot be ruled out. In such a scenario, the incorporated U might have originated from the alteration of U–Pb–LREE-bearing pyrochlore into columbite during the early hydrothermal stage of the pegmatite veins (Hadlich et al., unpublished work). The released Pb was incorporated predominantly into secondary galena, whereas the U found partial accommodation within secondary U-rich columbite, and solid solutions of thorite–coffinite–xenotime and zircon.

Origin of genthelvite

Few Be-bearing minerals form in the magmatic stage of pegmatite consolidation, with beryl being dominant (Černý, 2002). Paragenetically late (supercritical-to-hydrothermal) beryllium minerals are divided in two broad categories: (1) alteration products of the early phases of Be; (2) minerals occuring in miarolitic cavities and fissures (Černý, 2002). In the pegmatites investigated, genthelvite is observed as massive aggregates of crystals surrounding pegmatitic crystals and including accessory minerals. This observation can be interpreted as the crystallisation of a late-evolved fluid, wherein Be was preserved though complexation (e.g. with F) as no other Be mineral is present.

Compositional zoning or intergrowths between helvine-group minerals is relatively common (Haapala and Ojanperä, 1972; Clark and Fejer, 1976; Perez et al., 1990; Antao and Hassan, 2010), and is attributed to changes in the physical-chemical conditions during crystallisation, such as temperature and S fugacity. The preservation of this zoning or intergrowths would require low-temperature crystallisation and a rapid crystallisation process (Antao and Hassan, 2010). On the basis of these observations, it is plausible to infer that the lack of zoning in the genthelvite grains signifies diffusion between Zn, Fe and Mn. This diffusion process is probably facilitated by a relatively slow crystallisation within the late-evolved fluids of pegmatite formation, which occurred under relatively high temperatures.

The mineralogical and petrographic variations of the albite-enriched granite were mapped and described in detail in Bastos Neto et al. (2009), and the fluid-inclusion assemblies or associations in Ronchi et al. (2011). In these investigations, it was concluded that the large hydrothermal massive cryolite deposit in the centre of albite-enriched granite is part of an evolutionary process of a magma originally rich in volatiles, which during its polyphase crystallisation process resulted in the exsolution of hydrothermal saline deuteric fluids (salinity between 0 and 25% eq. NaCl and homogenisation temperatures from 100 to 400°C). Furthermore, the authors concluded that these fluids lowered the solidus of the system, within the albite-enriched granite dykes of the pegmatitic albite-enriched granite and pegmatite veins. In this process, phases rich in microcline, polylithionite, cryolite and genthelvite were formed and hydrothermal alteration was promoted, resultiing in albitisation, silicification, clay-formation, fluoritisation and oxidation of iron-rich minerals.

In the pegmatite veins, hosted by the albite-enriched granite, the sequence of genthelvite formation following riebeckite and polylithionite suggests that genthelvite is one of the last minerals to crystallise. However, it does form prior to hydrothermal cryolite, which is observed corroding genthelvite crystals. This observation emphasises the strong association of genthelvite with the transitional period between the late magmatic and early hydrothermal stages of pegmatite evolution.

Genthelvite formation conditions

Genthelvite is a rare mineral compared to other Be-bearing minerals or even to other members of the helvine group, as a result of its small stability field (Burt, 1988). The elements that constitute genthelvite (Zn, Mn, Fe, Be, Si and S) are commonly found as trace elements in highly fractionated granitic systems, therefore, this mineral is typical of systems at a late stage of differentiation, in which stability is due to local and transient conditions, generally atypical in the consolidation of granitic pegmatites, including low alumina activity and relatively reducing conditions that facilitate the coexistence of sulfides and silicates (Burt, 1988; Bilal and Fonteilles, 1988).

Genthelvite stability in a paragenesis is restricted to systems with low S activity (Burt, 1988). Because of the differing chalcophile behaviour of Zn>>Fe>Mn, in systems with high SO_–1 (the acid anhydride of H₂S), the Zn₂SiO₄ component would have been destabilised to form an assemblage with sphalerite and quartz (Burt, 1988). In contrast, under low SO_–1 conditions danalite and helvine are not stable, and the instability of the FeS and MnS components will lead to the formation of silicates or oxides (Burt, 1988). The low content of Fe in genthelvite also indicates high O₂ fugacity during crystallisation, with the crystallisation of hematite (Burt, 1980).

The wide compositional variation in helvine-group minerals, genthelvite in the core and danalite in the border, in the albitites associated with the Sucuri Granite (Brazil) suggest that the increased alkalinity (albitisation) in the system favoured genthelvite growth and the subsequent increase in S fugacity favoured danalite crystallisation by the reaction: Zn₈Be₆Si₆O₂₄S₂ + 8FeS₂ ↔ Fe₈Be₆Si₆O₂₄S₂ + 8ZnS + 4S₂ (Miranda, 2018).

Within the albite-enriched granite, the pegmatites originated through a process of continuous magma fractionation, resulting in a peralkaline composition highly enriched in high-field-strength elements. The extensive crystallisation of microcline, albite and polylithionite effectively buffered the Al content in the later fluids, reducing the alumina activity within the system. Simultaneously, the late-stage magmatic processes involved in the genesis of the pegmatites were characterised by a substantial increase in fluorine content (Paludo et al., 2018). The pegmatites that are rich in amphibole have ~3.35 wt.% F, whereas the ones rich in polylithionite have ~4.80 wt.% F, and those rich in cryolite have an average of 37.32 wt.% F. The remarkably high concentration of fluorine, coupled with the crystallisation of magmatic galena, effectively lowered the fugacity of H₂S, thus stabilising genthelvite during the transition from the late magmatic to the early hydrothermal stages of pegmatite evolution. The low Fe content in genthelvite can be attributed to the prior incorporation of Fe into riebeckite and polylithionite and indicates an O activity that is too elevated to favour danalite formation. The widespread formation of Mn and Fe oxides (columbite and hematite) during the hydrothermal stage further underscores the high O activity within the system.

Conclusion

This investigation of genthelvite in pegmatite veins hosted by the Madeira albite-enriched granite has yielded the following conclusions.

Among the various pegmatites hosted by the albite-enriched granite, the genthelvite-bearing bodies are rich in riebeckite, polylithionite and xenotime, confirming the observations by Paludo et al. (2018).

Genthelvite was formed during the transition between the late magmatic and early hydrothermal stages of pegmatite evolution. It stands out as the sole Be-bearing mineral in this context. Its formation can be attributed to Be being preserved in the late-evolved fluid, probably by complexation with fluorine. Genthelvite crystallised surrounding polylithionite, riebeckite, quartz, xenotime, pyrochlore, thorite and zircon, and it was subsequently corroded by hydrothermal cryolite.

Genthelvite exhibits a homogeneous composition within individual grains, but the overall composition displays moderate variation. It is characterised as a Mn–Fe-rich genthelvite, forming a solid solution within the genthelvite–danalite–helvine system. The mineral features relatively high levels of Zn (36.96 to 49.45 wt.% ZnO), lower Mn content (0.61 to 3.03 wt.% MnO), and variable Fe content (2.10 to 10.94 wt.% FeO), in comparison to the compositions found in other locations, and fills a compositional gap along the upper part of the Zn–Fe join in the Zn–Fe–Mn ternary diagram. Genthelvite exhibits elevated U levels (up to 0.25 wt.% UO₂), and significant REE contents (up to 0.40 wt.% REE₂O₃). Notably, the average LREE content is higher than that of HREE, a deviation from compositions found in other locations, where HREE content tends to be the higher one. Additionally, genthelvite from this site lacks fluorine. Regardless of these compositional disparities, the crystallographic parameters (a = 8.127 Å) of genthelvite from the pegmatite veins closely resemble those determined for crystals from other localities.

The unusual composition of genthelvite is attributed to the buffering of Fe, F and HREE by minerals that crystallised before genthelvite. These include riebeckite, polylithionite, xenotime and magmatic cryolite.

Genthelvite was affected during the late hydrothermal stage, which was associated with F-rich aqueous fluids responsible for the formation of the massive cryolite deposit, and the disseminated hydrothermal cryolite within the albite-enriched granite. The highest observed fluid-inclusion homogenisation temperature of 400°C, measured in hydrothermal cryolite (Bastos Neto et al., 2009), establishes the minimum initial temperature for the hydrothermal process. The structure of genthelvite facilitated the incorporation of Fe, Mn, Mg, Pb, Ba, Na, K, U and REE into the Zn²⁺ structural site. Additionally, it allowed the allocation of excess Si, Al, Ti and P to the ^IVSi and ^IVBe structural sites. The high levels of U and REE substituting for Zn, as well as Si substituting for Be, are charge balanced by the presence of vacancies at the A site.

This crystallochemical study of genthelvite has revealed that its formation occurred in an alkaline and subaluminous environment, under stable conditions within the late-evolved fluids, and at relatively high temperatures (>400°C), in a reducing environment. The substantial presence of fluorine and the crystallisation of magmatic galena effectively reduced the fugacity of H₂S, ensuring the stability of genthelvite during the transition from the late magmatic to early hydrothermal stages in the evolution of the albite-enriched granite. The moderate variation in Fe content within genthelvite can be attributed primarily to the presence of riebeckite and polylithionite, which acted as buffers for Fe content in their immediate surroundings. In addition, the extensive formation of Mn and Fe oxides (columbite and hematite) during the hydrothermal stage, underscores an O activity that is too high to favour the formation of danalite.