Bosiite, NaFe³⁺₃(Al₄Mg₂)(Si₆O₁₈)(BO₃)₃(OH)₃O, a new ferric member of the tourmaline supergroup from the Darasun gold deposit, Transbaikalia, Russia

Tourmaline, rich in Fe³⁺, has been recorded from many different localities: from the type locality of povondraite (Žáček et al., 1998), from the cap rock of a salt dome in the Gulf of Mexico (Henry et al., 1999), and from evaporite formations in Namibia (Henry et al., 2008). Tourmalines from the Larderello geothermal field, Italy (Cavaretta & Puxeddu, 1990), from northern Scandinavia, and from porphyry-copper deposits formed during hydrothermal processes (Baksheev et al., 2011). The origin of Al-poor tourmaline from the Western Carpathians is still not clear (Bačík et al., 2008). Tourmaline from the Darasun gold deposit is, except for the Mg content, similar to tourmaline formed in evaporites and similar to tourmaline from the Western Carpathians, which are also enriched in Ca, but have a different compositional evolution: from Fe³⁺-rich tourmaline (bosiite) to Fe-bearing dravite in Darasun, whereas in the Western Carpathians it evolved from schorl–dravite to povondraite. Samples from Darasun are very similar in composition to those from the Vetka porphyry-copper deposit, Chukotka, Russia, and also in Mg and Ca contents (Baksheev et al., 2010, 2011). The maximum Fe_tot content of tourmaline from Darasun (~5 apfu) is higher than in tourmalines from deposits in porphyry-copper deposits (~3 apfu) (Lynch & Ortega, 1997). In addition, the Fe³⁺/Fe_tot value in tourmalines from the Darasun gold deposit is significantly higher. Tourmaline from porphyry tin deposits also displays a homovalent AlFe₋₁ substitution, but is characterized by a lower Fe³⁺/Fe_tot value (Baksheev et al., 2010, 2011, 2012). The latter authors concluded that the homovalent Fe³⁺→Al substitution in tourmaline (also observed in tourmalines from Darasun gold deposit) may be used as prospecting guide for porphyry-style deposits and that tourmalines from Cu, Sn and Au porphyry-type deposits can be distinguished from each other on the basis of Li concentration and the ratio Fe³⁺/Fe_tot.

The Darasun gold deposit (Darasun mine), Vershino-Darasunskiy, Transbaikal Krai, Eastern-Siberian Region, Russia (52°20′24″N, 115°29′23″E), was the second mine in terms of gold reserves of the Transbaikal Krai. This deposit is hosted in basic and intermediate igneous rocks. The most ancient rock is metamorphosed Early Palaeozoic gabbro, which was intruded by Middle Palaeozoic and Middle Mesozoic granodiorite, diorite, granite, granosyenite and syenite, and also by granitic rocks of the Amanansky complex of Early Jurassic age. The formation of gold deposits in Transbaikalia seems to be controlled by small intermediate and felsic subvolcanic intrusions enriched in alkalis, which belong to the Middle to Late Jurassic Amudzhikan–Sretensky complex and comprise granite and granodiorite porphyry, fracture-related intrusions and stocks, quartz diorite porphyry, granophyre, felsite and pitchstone dykes (Baksheev et al., 2011). The Darasun deposit comprises extensive steeply dipping gold-bearing quartz veins and mineralized zones spatially related to the Amudzhikan–Sretensky subvolcanic K-rich granodiorite porphyry intrusion. Ores of this deposit are rich in sulfides (up to 60%). Gangue minerals are quartz, tourmaline, calcite and dolomite.

Bosiite was found in a sulfide- and gold-bearing quartz vein. Native gold, calcite, dolomite, arsenopyrite, chalcopyrite, pyrrhotite, tetrahedrite, sphalerite and galena occur in the same vein. Bosiite coexists with quartz and pyrite (Fig. 1 and Fig. S1 in Supplementary Material, linked to this article and freely available online at the GSW website of the journal: http://eurjmin.geoscienceworld.org/) and is intimately associated with other tourmalines (Figs 2 and 3). The first tourmaline generation is bosiite, which is followed by a second generation (Fig. 4) of a new tourmaline (recently described as oxy-dravite by Bosi & Skogby, 2013), and a third generation, which is represented by dravite (~20 μm thick veins). Depletion of Fe in the second and third generations is caused by deposition of early sulfides.

Baksheev et al. (2011) have studied 73 primary fluid inclusions in quartz associated with tourmaline from the bosiite type locality. At room temperature, they found vapour-dominant and aqueous inclusions, indicating boiling in the fluid. Microthermometry indicated that vapour-dominant inclusions homogenize at ~360–380°C and contain a fluid with a salinity of about 1.7 wt% equiv. NaCl. Aqueous inclusions homogenize to liquid at ~350–390°C and contain fluid with a salinity ranging from ~7 to 10 wt% equiv. NaCl. The calculated pressure from vapour-dominated inclusions is 180–230 bars, which corresponds to a depth of 600–800 m. These authors concluded that the fluid inclusions in the associated quartz indicate that first-stage tourmaline (bosiite) precipitated from boiling fluid at low pressure. Boiling favoured the increasing oxygen activity (causing the loss of H₂ into the vapour phase), and resulted in the increase of Fe³⁺ in that tourmaline. Based on these data, bosiite crystallized probably during a hydrothermal event at PT conditions of ~0.02 GPa/380°C.

The holotype specimen of bosiite (crystal used for the structural study and coarse-grained sample) is deposited in the collection of the Naturhistorisches Museum Wien, Vienna, Austria, under the catalogue number N 9793. Parts of the cotype material (coarse-grained sample) have also been deposited in the collection of the Fersman Mineralogical Museum, Moscow, Russia (catalogue no. 94999).

A fragment of bosiite was separated from the same hand specimen from which the chemistry had been determined; it was subsequently studied on a Bruker AXS Kappa APEXII single-crystal diffractometer equipped with a charge-coupled device (CCD) area detector and an Incoatec Microfocus Source IμS (30 W, multilayer mirror, MoKα). Single-crystal X-ray diffraction data were collected at room temperature (up to 80° 2θ), integrated and corrected for Lorentz and polarization factors with an absorption correction by evaluation of partial multiscans. The structure was refined with SHELXL97 (Sheldrick, 1998) using scattering factors for neutral atoms and a tourmaline starting model from Ertl et al. (2010). The H atom bonded to the O3 atom was located from a difference-Fourier map and subsequently refined. Refinement was done with anisotropic displacement parameters for all non-H atoms. Table 1 provides crystal data and details of the structure refinement. Site occupancies were refined according to well-known characteristics of the tourmaline structure (Na and Ca were refined at the X site, Fe and Al were refined at the Y and Z sites; for other details, see Table 2). The refinement converged at a R1(F) value of ~3.8% (Table 1). Atom parameters and equivalent isotropic-displacement parameters are given in Table 2, selected interatomic distances in Table 3.

A tourmaline fragment was prepared as a section (polished on one side of the sample) for chemical analysis. Concentrations of all elements except B, Li, Be and H were determined with a JEOL JSM–6480LV electron microscope equipped with an Inca Energy–350 energy-dispersion system (EDS) (Laboratory of Analytical Techniques of High Spatial Resolution, Department of Petrology, Lomonosov Moscow State University, Moscow, Russia). Chemical analyses were done by a combination of EDS Inca Energy–350 and WD (wavelength-dispersion) Inca Wave–500 (accelerating potential 15 kV, 64 ± 0.1nA, ~5 μm beam diameter). The concentration of F (MgF₂ was used as a standard) was measured with a WD TAP crystal and the content of the other elements with EDS. For the EDS analyses, line profiles of characteristic X-ray radiation were optimized and normalized using natural silicates as standards (Jarosewich et al., 1980). The error for elements with a concentration of ≥5 wt% does not exceed 1.5 relative %.

The Li content was determined by ICP–MS (Institute of Geology of Ore Deposits, Petrography, Mineralogy and Geochemistry (IGEM), Russian Academy of Sciences, Moscow, Russia). Initially, a sample was decomposed with H₂SO₄ and HF in a MILESTONE microwave device according to a four-step program. After the solution was evaporated, treatment by water and dissolution in HNO₃ followed. The measurements were made by using a PLASMA QUAD mass spectrometer (VG Instruments Company).

The crystal-chemical formula was calculated on the basis of 31 (OH, O, F) atoms. The B₂O₃ content was calculated to fill the B site completely. Because of the small amount of available sample, H₂O was not analysed, but was calculated for a charge-balanced formula assuming that there are no vacancies at the Y, Z, and T sites. Analytical data are given in Table 4. Back-scattered electron images were obtained at an accelerating potential of 15 kV and a current intensity of ~15 nA.

Approximately 5 mg of tourmaline, consisting mainly of bosiite, were gently crushed under acetone, then mixed with a sugar–acetone solution designed to form sugar coatings around each grain and prevent preferred orientation. The ⁵⁷Fe spectrum was acquired at 295 K using a SM–2201 Mössbauer spectrometer operating in constant acceleration mode with a ⁵⁷Co (in Cr) source (Institute of Mineralogy, Ural Division, Russian Academy of Sciences, Miass, Russia). Calibration was done with a standard sample of sodium nitroprusside. Isomer shift refers to an α-Fe absorber at 293 K. Results of the measurements were processed by a least-squares procedure using the SPECTR program in fitting a thin absorber (using a Lorentzian line-shape).

Peak areas were not corrected for differential recoil-free fractions for Fe²⁺ and Fe³⁺ because the appropriate correction factors do not exist.

In plane-polarized transmitted light, bosiite is pleochroic, O = yellow-brown, E = red-brown. Bosiite is uniaxial negative, ω = 1.760(5), ϵ = 1.687(5). An infrared spectrum of a first-generation tourmaline single crystal (bosiite) was recorded by Baksheev et al. (2011, their Fig. 3B). These authors concluded that the spectrum indicates that a part of Fe³⁺ occurs at the Z site. The interactions of three OH groups at the V sites with one Y-site cation and two Z-site cations are indicated by a band in the range 3580–3480 cm⁻¹, with a complex asymmetric profile. In the single-crystal absorption spectrum of bosiite, this band is located at a lower frequency with a maximum at 3552 cm⁻¹. The interaction of the hydroxyl groups at the W site with three Y-site cations is represented by a single weak band at 3733 cm⁻¹ (Baksheev et al., 2011).

Bosiite is brittle and has a Mohs hardness of 7; it is non-fluorescent, has no observable parting and cleavage, and has a vitreous lustre. The megascopic colour of bosiite is black and the streak is dark brown. It has a measured density of 3.23(3) g/cm³ (by pycnometry) and a calculated density of 3.26(1) g/cm³ from the empirical formula. The fracture is irregular to uneven and sub-conchoidal. It has a prismatic habit and only unspecified prismatic forms are developed. Twinning was not observed. The c:a ratio calculated from the refined unit-cell parameters is 0.455.

The Mössbauer spectrum of bosiite (Fig. 5 in Baksheev et al., 2011) was fitted to three doublets (Table 5) that were assigned according to Fuchs et al. (1998); Béziat et al. (1999) and Dyar et al. (1998). The first two doublets (^[Y1]Fe²⁺, ^[Y2]Fe²⁺) are subcomponents of a distribution corresponding to Fe²⁺ at the Y site and cover a total area of about 15%. This is also in agreement with the findings of Andreozzi et al. (2008), who modelled Fe²⁺ at the Y site in tourmalines of the elbaite–schorl–dravite series by three doublets (ΔE_Q = 2.45, 2.19, and 1.72 mm/s). The third strongest doublet, with an isomer shift of 0.57 mm/s, is about 85%, implying that most of the iron in the tourmaline sample examined occurs as Fe³⁺ in octahedral coordination (Baksheev et al., 2011). However, the real amount of Fe³⁺ can be higher or lower because of the chemical heterogeneity of the studied sample, of the significant overlap of the peaks in the spectrum and of the additional uncertainty introduced in assuming the same recoil-free fraction for Fe²⁺ and Fe³⁺. As was shown in previous studies (e.g., Dyar et al., 1998; Andreozzi et al., 2008), the assignment of Fe³⁺ to Y or Z sites on the basis of Mössbauer data alone is problematic and sometimes not possible. Because of the limited resolution of our Mössbauer data (Table 5), we are not able to use it to establish whether Fe³⁺ occupies both the Y and Z sites. It is possible that the two expected Fe³⁺ doublets are hidden in a single doublet.

The X site in bosiite is occupied mainly by Na (0.73 apfu), but contains also significant amounts of Ca (0.23 apfu) (Table 4). The content of K is at the detection limit (Table 4). The refinement is in excellent agreement with the chemistry (Na_0.75(2)Ca_0.25(2); Table 2).

The Y site is occupied by Fe³⁺ (~1.5 apfu), Mg (~0.8 apfu), Fe²⁺ (~0.6 apfu) and Al (~0.1 apfu). Small amounts of Ti⁴⁺ (0.01 apfu) might also occupy this site. Only traces of Li occupy the Y site (Li content is ~3 ppm; Table 4). The Y-site occupancy is consistent with the refinement results (Table 7) and is also in good agreement with the relatively large <Y–O> distance of 2.054 Å (Table 3).

The Z site is mainly occupied by Al (~3.2 apfu), but contains also significant amounts of Fe³⁺ (~1.9 apfu) and Mg (~0.9 apfu). The Z-site occupancy is consistent with the refinement results (Table 7) and is also in good agreement with the enlarged <Z–O> distance of 1.969 Å (Table 3). Hawthorne (1996) noted that the occurrence of Al³⁺ at the Y site and Mg²⁺ at the Z site is due to disorder rather than differences in chemical composition, and that the order–disorder reaction can be expressed as ^YMg + ^ZAl↔ ^YAl + ^ZMg [^YMg^ZAl(^YAl^ZMg)₋₁]. This disorder is driven by the short-range requirements of O²⁻ at the W site (=O1 site) that is associated with 2Al + Mg configurations at the coordinating Y sites so as to satisfy its bond-valence requirements (Hawthorne, 1996, 2002). Taylor et al. (1995) showed that disorder of Al and Mg over the Y and Z sites in a calcic tourmaline occurs in tandem with significant O²⁻ at the W site. Also in bosiite, the W site is mainly occupied by O²⁻ (Table 4). All tourmalines from the type locality of bosiite have an Mg content of ~2 apfu, the W site is dominated by O²⁻, and chemical variations are in accord with the FeAl₋₁ or AlOFe₋₁OH₋₁ exchange vector (or both). In the suite, the composition of bosiite is substantially deficient in Al (<6 apfu), consistent with a predominance of the FeAl₋₁ exchange (see also Baksheev et al., 2011).

The T site in bosiite is mainly occupied by Si and contains relatively small amounts of Al (~0.1 apfu), consistent with the slightly enlarged <T–O> distance of 1.621 Å (Table 3) and data given by MacDonald & Hawthorne (1995) on the Si ↔ Al substitution in tourmaline. There are no indications for significant amounts of ^[4]B (the <T–O> distance is not reduced). Also the amount of ^YAl is too small (~0.1 apfu) for significant amounts of ^[4]B to occur through short-range order (Fig. 1 in Ertl et al., 2008).

An H atom (H3) at the site associated with O3 was easily located in the refinement. Ertl et al. (2002) showed that the bond-angle distortion (σ_oct²) of the ZO₆ octahedron in a tourmaline is largely a function of the <Y–O> distance in that tourmaline, although the occupant of the O(3) site (V site in the general formula) also affects that distortion. The correlation coefficient, r, of <Y–O> and σ_oct² of the ZO₆ octahedron is −0.99 (Fig. 2 in Ertl et al., 2005) for all investigated tourmalines whose V site is occupied by 3 (OH) groups. Bosiite (with ^Zσ_oct² = 42.47 and <Y–O> 2.054 Å; Table 3) lies exactly on the V site = 3 (OH) line. Hence, the V site of bosiite is completely occupied by (OH)₃. No H associated with the O(1) site (W site) was found by refinement. Hence, this site is clearly dominated by oxygen, as indicated by the chemical composition. This is also consistent with the infrared spectrum, which shows only a weak band for the hydroxyl groups at the W site (see above).

According to the IMA-CNMNC guidelines (Henry et al., 2011), the X site is dominated by Na, the Y site is dominated by Fe³⁺, the Z site is dominated by Al along with Mg (valency-imposed double-site occupancy), the W site by O²⁻. Accordingly, this tourmaline is a new species.

Bosiite is a Na-, Fe³⁺- and Mg-bearing tourmaline, and is isotypic with povondraite and other trigonal (R3m) members of the tourmaline group such as oxy-dravite (Bosi & Skogby, 2013). Their ideal chemical compositions are given in Table 8. For comparison, chemical analyses and selected physical properties of the valid minerals of the tourmaline supergroup (Henry et al., 2011) are given in Table 9. A triangular plot (Fe–Al–Mg) of tourmalines from the bosiite type-locality is shown in Fig. 5 and highlights the main compositional changes.

The refined formula is ^X(Na_0.75Ca_0.25)_Σ1.00^Y(Fe_2.01Al_0.99)_Σ3.00^Z(Al_4.15Fe_1.85)_Σ6.00^T(Si₆O₁₈)(BO₃)₃^V(OH)₃^WO, in reasonable agreement with the optimized formula (1). Slight differences may reflect minor chemical zonation.

The name follows the tourmaline nomenclature recommendations approved by the IMA-CNMNC (Henry et al., 2011), which necessitate a new root name for a mineral of the alkali-subgroup 3 with the formula NaFe³⁺₃(Al₄Mg₂)(Si₆O₁₈)(BO₃)₃(OH)₃O. The new mineral is named after Dr. Ferdinando Bosi (b. 1967), researcher at the University of Rome La Sapienza, Italy, and expert on the crystallography and mineralogy of the tourmaline-supergroup minerals and the spinels.

Bosiite, ideally NaFe³⁺₃(Al₄Mg₂)(Si₆O₁₈)(BO₃)₃(OH)₃O, is related to end-member povondraite, ideally NaFe³⁺₃(Fe³⁺₄Mg₂)(Si₆O₁₈)(BO₃)₃(OH)₃O, by the substitution ^ZAl₄ → ^ZFe³⁺₄. Further, bosiite is related to oxy-dravite, ideally Na(Al₂Mg)(Al₅Mg)(Si₆O₁₈)(BO₃)₃(OH)₃O, by the substitutions ^[6]Fe³⁺₃ → ^[6]Al₃. It can be distinguished from schorl, povondraite, fluor-buergerite, dravite, oxy-dravite or other dark tourmalines only by accurate determination of the chemical composition, preferably by a combination of single-crystal structure refinement and chemical analysis including Mössbauer spectroscopy. It may also be confused with luinaite-(OH) (IMA 2009–046) and its currently unnamed F-analogue, both representing monoclinically distorted variants (Cm) of the tourmaline structure. Relatively large lattice parameters (a ≥ 16.1, c ≥ 7.3 Å) and the occurrence in a hydrothermal environment could be preliminary indications for the occurrence of bosiite.

We thank V.O. Yapaskurt (Department of Petrology, Lomonosov Moscow State University, Moscow, Russia), S.A. Gorbacheva (IGEM, Russian Academy of Sciences, Moscow, Russia) and A. B. Mironov (Institute of Mineralogy, Ural Division, Russian Academy of Sciences, Miass, Russia) for analytical work. We are grateful to Andreas Wagner (Institut für Mineralogie und Kristallographie, Vienna, Austria) for taking photographs of bosiite. We sincerely thank Frank C. Hawthorne and an anonymous reviewer for constructive comments on this manuscript. This work was supported in part by Austrian Science Fund (FWF) project no. P23012-N19 and no. P26903-N19 to AE, by IGCP 540 and the Russian Foundation for Basic Research (project nos. 09–05–00697a and 09–05–12037ofi–m). The purchase of the electron microscope was financially supported by the Program for the Moscow State University Development.