Geological setting, occurrence and associated minerals

The Ozernovskoe deposit is located within the North Kamchatka ore region of the Central Kamchatka volcanic belt, 140 km north of the town of Klyuchi and 80 km west of the Bering Sea coast (Okrugin et al., Reference Okrugin, Vymazalová, Kozlov, Laufek, Stanley and Shkilev2022).

This Au–Te–Se deposit of epithermal high sulfidation type is localised within a Miocene basaltic palaeovolcano, composed of sills and dykes of andesite–basalts, cut by through by volcanic domes of andesites, andesite–dacites, dacites, magmatic breccias and tuffisites (Petrenko, Reference Petrenko1999; Litvinov et al., Reference Litvinov, Patoka and Markovsky1999; Demin, Reference Demin2015). The ore-bearing quartz veins and stockworks are hosted by hydrothermal silification zones, consisting of quartz, kaolinite, dickite and alunite-group minerals.

This deposit is distinguished by a variety of selenide and telluride minerals, including some new and unnamed species (Kovalenker and Plotinskaya, Reference Kovalenker and Plotinskaya2005; Spiridonov et al., Reference Spiridonov, Ignatov and Shubina1990; Reference Spiridonov, Filimonov and Bryzgalov2009; Reference Spiridonov2014; Okrugin et al., Reference Okrugin, Vymazalová, Kozlov, Laufek, Stanley and Shkilev2022) as well as the recently described minerals ozernovskite (Pekov et al., Reference Pekov, Britvin, Pletnev, Chukanov, Belakovskiy and Yapaskurt2021) and rudolfhermannite (Pekov et al., Reference Pekov, Britvin, Pletnev, Yapaskurt, Belakovskiy, Chukanov, Vigasina and Ponomarev2022).

Ore deposition is related to multiple repeated cycles of extensive boiling and related hydrothermal brecciation. From one to five stages of ore minerals deposition can be recognised on every cycle. The earliest stage is represented by quartz + kaolinite + alunite-group minerals (woodhouseite, etc.) + anatase + rutile + zircon + pyrite followed by cassiterite with very rare scheelite, scandium-bearing titanite, and thortveitite. Cassiterite replaces rutile crystals epitaxially as revealed by electron back-scattering diffraction (EBSD) orientation mapping. Cassiterite from this locality is found to contain unusually high concentrations of W, Ga, V, Ti and Al substituting Sn. In rare cases it also contains Sc (up to 0.2–0.4 wt.%). Cassiterite–rutile crystals are usually overgrown and are partly replaced by Se-bearing mohite and okruginite with other Sn-, W-, Mo-bearing sulfides and sulfoselenides (Se-bearing herzenbergite and svetlanaite, Se-mawsonite, Se-stannite, kiddcreekite, hemusite) followed by other ore minerals (a variety of tetrahedrite-group minerals, chalcopyrite, Bi and Bi–Pb selenotellurides, Au, Ag, Cu tellurides, native tellurium and native Te–Se alloys). The latest tellurides and native tellurium were deposited in close paragenesis with quartz and Mn-bearing dolomite.

Appearance, physical and optical properties

Okruginite occurs as distinct Se-rich zones in Se-bearing mohite crystals (Fig. 1) or forms aggregates of small crystals in size 10–15 μm in quartz and poubaite (Figs 2, 3). It occurs in association with cassiterite, rutile, mohite, mawsonite, kiddcreekite, hemusite, native tellurium, kostovite, Se-bearing fahlerz (Se-goldfieldite – Se(Bi)-tetrahedrite – Se-tennantite), svetlanaite and quartz.

Figure 1. Okruginite (Okg) occurring as distinct Se-rich zones in Se-bearing mohite (Moh) crystal with pyrite (Py) in quartz (Qz). Back-scattered electrons image, holotype BM 2021, 02.

Figure 2. Reflected light (left) and back-scattered electrons (right) images of okruginite (Okg), brighter then mohite in BSE), Se-bearing mohite (Moh), cassiterite (Cst) and rutile (Rt) in quartz (Qz). Holotype BM 2021, 02.

Figure 3. Back-scattered electrons image of okruginite (Okg) with poubaite (Pub) and native tellurium (Te) in quartz (Qz). Holotype BM 2021, 02.

Okruginite is opaque with a metallic lustre. The mineral is brittle. The density calculated on the basis of the empirical formula is 5.126 g/cm³. In plane-polarised reflected light, okruginite appears brownish grey against rutile, cassiterite, chalcopyrite and other minerals in the natural assemblage. Okruginite has weak anisotropy with rotation tints pale blue-grey to pale grey-brown, bireflectance and pleochroism was not discernible. It exhibits no internal reflections.

Reflectance measurements were made in air relative to a WTiC standard using a J & M TIDAS diode array spectrometer attached to a Zeiss Axiotron microscope. The results in comparison with its synthetic analogue (Cu₂SnSe₃) are tabulated in (Table 1) and illustrated in Fig. 4.

Table 1. Reflectance data for okruginite and its synthetic analogue Cu₂SnSe₃.

Note: The values required by the Commission on Ore Mineralogy are given in bold.

Figure 4. Reflectance data for okruginite, measured on 3 sites and its synthetic analogue. The reflectance values (R%) are plotted versus wavelength λ in nm.

Chemical composition

Chemical analyses were performed with a TESCAN LYRA3 FEG scanning electron microscope (SEM) with Shottky cathode located in the Kurchatov Institute, Moscow, Russia, using Oxford Instruments Inca Wave and IncaEnergy+ WDS–EDS system with Wave-500 WDS spectrometer and X-Max80 EDS detector. The following X-ray lines were used for quantification: SnLα, SeLα, SKα, CuKα, FeKα, MnKα and VKα. The following conditions were applied for the data collection: wavelength dispersive spectroscopy; acceleration voltage of 20 kV; beam current of 12 nA, beam diameter of 1 μm, number of analyses = 27. Acquisition time was 30 s on peak and 15 × 2 s on background for major elements (Cu, Sn, Se and S) and 50 s on peak and 25 × 2 s on background for minor elements (V, Mn and Fe). The presence of Sb and Te was also tested for, but they were not detected. The MAC set of pure elements (for Sn, Se, Mn and V) and natural stoichiometric CuFeS₂ for Cu, Fe and S were used as standards. The matrix correction procedure of Pouchou and Pichoir (XPP), integrated in Inca software, was used.

Chemical analyses of the synthetic analogue were performed with a CAMECA SX-100 electron probe microanalyser in wavelength-dispersive mode using an electron beam focussed to 1–2 μm, located in the Institute of Geology, Academy of Sciences of the Czech Republic. Cuprite (Cu₂O) and pure Sn and Se were used as standards. Concentrations were quantified on the CuLα, SnLα and SeLα (with an overlap correction on SnL _c3) lines with an accelerating voltage of 15 keV, and a beam current of 10 nA on the Faraday cup.

The results are given in Table 2. The empirical formula of okruginite calculated on the basis of 6 apfu is Cu_1.99Sn_1.02(Se_1.82S_1.17)_Σ2.99 and Cu_1.86Sn_1.03Se_3.11 for its synthetic analogue. Alternative recalculation of the empirical formula of okruginite based on 3 Se (Se+S) atoms yields Cu_1.99Sn_1.01(Se_1.83S_1.17)_Σ3.00 and Cu_1.79Sn_0.99Se_3.00 for its synthetic equivalent. The ideal formula Cu₂SnSe₃ requires Cu 26.33, Sn 24.59, Se 49.08, total 100 wt.%.

Table 2. Electron-microprobe analyses (wt.%) of studied grains of okruginite and its synthetic analogue.

S.D. – Standard deviation

Synthetic analogue

The size (10–15 μm) of inclusions of okruginite, intimately intergrown with mohite or embedded in quartz, prevented their extraction in an amount required for relevant crystallographic and structural investigations. Therefore, these investigations were performed on the synthetic Cu₂SnSe₃.

The synthetic phase Cu₂SnSe₃ was prepared in an evacuated and sealed silica-glass tube in a horizontal furnace in the Laboratory of Experimental Mineralogy of the Czech Geological Survey in Prague. A charge of ~400 mg was carefully weighed out from the native elements. We used, as starting chemicals copper powder (99.9999% purity), tin ingot (99.9999% purity) and selenium pebbles (99.9999% purity). To prevent loss of material to the vapour phase during the experiment, the free space in the tube was reduced by placing a closely fitting silica glass rod against the charge. The evacuated tube with its charge was sealed and then annealed at 600°C for 1 week. After cooling by cold-water bath, the charge was ground into powder in acetone using an agate mortar, and thoroughly mixed to homogenise. The pulverised charge was sealed in an evacuated silica-glass tube again, and reheated at 200°C for 6 months. Afterwards, after homogenisation, the sample was heated at 500°C for 3 weeks and cooled in a slow regime (5°C per hour).

X-ray crystallography

The powder X-ray diffraction pattern of synthetic Cu₂SnSe₃, used for structure refinement, was collected in Bragg-Brentano geometry on a Bruker D8 Advance diffractometer, equipped with a LynxEye XE detector using CuKα radiation and a 10 mm automatic divergence slit. The data were collected in the angular range from 10 to 140° of 2θ with a 0.015° step and with a counting time of 1.4 sec per step. The details of data collection and basic crystallographic data are given in Table 3.

Table 3. Data collection and Rietveld analysis of synthetic of Cu₂SnSe₃, the synthetic analogue of okruginite.

The crystal structure of Cu₂SnSe₃, the synthetic equivalent of okruginite was refined using the Rietveld method by the Topas 5 program (Bruker AXS, 2014). Preliminary inspection of the powder data revealed that the main reflections can be indexed by a cubic sphalerite-based structure (F $\bar{4}$3m) with a = 5.69 Å, as was suggested by Sharma et al. (Reference Sharma, Ayyar and Singh1977) for the synthetic Cu₂SnSe₃ phase. However, this cubic model could not fit weak superstructure diffractions visible in our powder X-ray diffraction pattern and could not describe visible peak splitting indicating lowering of the symmetry (Fig. 5). It is worth noting that even Sharma et al. (Reference Sharma, Ayyar and Singh1977) observed peak splitting in the diffraction pattern of Cu₂SnSe₃ and mentioned that the true symmetry of the Cu₂SnSe₃ phase might be lower than the cubic one. All their attempts to index the collected Debye–Scherrer diffraction patterns of Cu₂SnSe₃ in lower symmetry cells failed. Based on the indexing of collected diffraction patterns, Hahn et al. (Reference Hahn, Klingen, Ness and Schulze1966) proposed a tetragonal chalcopyrite-structure type (I $\bar{4}$2d symmetry) for synthetic Cu₂SnSe₃. However, no proper crystal structure analysis (including the refinement of structure) was carried out in their original work; the atomic coordinates were taken from the Inorganic Crystal Structure Database (ICSD, 2023, #629091) and were estimated by an editor of the database. This tetragonal chalcopyrite-type structure model also does not account for observed peak splitting in the diffraction pattern. Subsequently, the monoclinic structure model (Cc symmetry) suggested by Delgado et al. (Reference Delgado, Mora, Marcano and Rincón2003) and also later by Nomura et al. (Reference Nomura, Maeda, Takei, Morihama and Wada2013) for the synthetic Cu₂SnSe₃ phase provides not only an accurate description of peak splitting, but also fits well to weak, though discernible, superstructure reflections (Fig. 6). The refinement of this monoclinic structure model resulted in a significant drop of R _wp profile agreement factor from 9.87% (F $\bar{4}$3m structure) to 3.42% (Cc structure).

Figure 5. Detail of Rietveld fit of synthetic Cu₂SnSe₃, an analogue of okruginite, using sphalerite, chalcopyrite and Cc structure models. Only the model with Cc symmetry (Delgado et al., Reference Delgado, Mora, Marcano and Rincón2003) can fit all the observed reflections in the diffraction pattern. Note the significant difference in the R _wp factor.

Figure 6. The final Rietveld fit for Cu₂SnSe₃, the synthetic analogue of okruginite.

The Rietveld refinement involved refinement of unit-cell parameters, background Chebychev polynomial of the 5^th order, atomic coordinates, isotropic size and strain and occupancy parameters for Cu1, Cu2 and Sn positions. Our refinement revealed mixed sites at Cu1, Cu2 and Sn positions (see Table 4), indicating a partial Cu/Sn disorder at these positions. No mixed sites were revealed at Se positions.

Table 4. Atomic positions (space group Cc) and isotropic displacement parameters for the synthetic analogue of okruginite, Cu₂SnSe₃. The isotropic displacement parameter (B _iso) was constrained for Cu/Sn and Se positions.

No significant deviations of crystal coordinates from their starting positions (structure model of Delgado et al., Reference Delgado, Mora, Marcano and Rincón2003) were observed; as the refined fractional coordinates do not deviate by more than 0.01 from the starting values. However, a partial disorder at Cu1, Cu2 and Sn positions was detected in our refinement (Table 4), which was not reported by Delgado et al. (Reference Delgado, Mora, Marcano and Rincón2003). Crystal structure and powder diffraction data are presented in Tables 4 and 5, respectively. Okruginite shows monoclinic symmetry, however the mineral is strongly pseudocubic. Because of strong partial reflection overlapping in its powder diffraction pattern, it is not possible to fit individual positions and intensities of reflections and perform subsequent traditional unit-cell refinement based on d _calc and d _obs. Hence, output from Rietveld refinement is presented in Table 5 (only I _calc ≥ 1 are presented). The final Rietveld fit is shown in Fig. 6. The crystallographic information files have been deposited with the Principal Editor of Mineralogical Magazine and are available as Supplementary material (see below).

Table 5. Powder X-ray diffraction data (output from Rietveld refinement) of Cu₂SnSe₃, the synthetic analogue of okruginite (CuKα radiation, Bruker D8 Advance, Bragg-Brentano geometry).

Structure description

The okruginite crystal structure is shown in Fig. 7. It can be described as a derivative of the sphalerite structure. As is typical for this group of minerals, the cations (Cu1, Cu2 and Sn) show tetrahedral coordination by Se atoms. The tetrahedra share Se-corners and form a three-dimensional network. The Cu–Se distances are within the range 2.34–2.49 Å (mean value 2.43 Å), Sn–Se distances vary from 2.47 to 2.56 Å (mean value 2.53 Å). These mean values are comparable with bond distances observed in similar adamantine structures, e.g. eskebornite CuFeSe₂ (Cu–Se: 2.418–2.424 Å, Delgado et al., Reference Delgado, de Delgado G., Quintero and Woolley1992), synthetic phases CuInSe₂ (CuSe: 2.432 Å, Knight, Reference Knight1992) and Cu₂FeSnSe₄ (Cu–Se: 2.417 Å, Sn–Se: 2.56 Å, Roque Infante et al., Reference Infante E., Delgado and Lopez Rivera1997). From a chemical point of view, okruginite is a selenide analogue of mohite, Cu₂SnS₃. Mohite was described as a monoclinic mineral by Kovalenker et al. (Reference Kovalenker1983), however its crystal structure has not been determined. A synthetic phase, Cu₂SnS₃ shows (at least three) modifications. The high-temperature modification (above 780°C) was reported by Palatnik et al. (Reference Palatnik, Komnik, Belova and Adroschenko1961) and by Moh (Reference Moh1963) with the sphalerite structure (i.e. disordered structure). Chen et al. (Reference Chen, Wada, Sato and Mieno1998) described a tetragonal stannite-type structure for synthetic Cu₂SnS₃ prepared at 700°C. A monoclinic modification (Cc symmetry) was prepared by Onoda et al. (Reference Onoda, Chen, Sato and Wada2000). Mohite is probably isostructural with this monoclinic modification of Cu₂SnS₃; however its direct detailed crystallographic study has not yet been performed. Okruginite is isostructural with this monoclinic Cu₂SnS₃ phase (Onoda et al., Reference Onoda, Chen, Sato and Wada2000). Other chemically related minerals petříčekite CuSe₂ (Bindi et al. Reference Bindi, Förster, Grundmann, Keutsch and Stanley2016) and svetlanaite SnSe (Okrugin et al., Reference Okrugin, Vymazalová, Kozlov, Laufek, Stanley and Shkilev2022) adopt marcasite and GeS-type structures, respectively, and hence are structurally very different from okruginite.

Figure 7. Crystal structure of Cu₂SnSe₃, the synthetic analogue of okruginite, showing tetrahedral coordination by selenium around Cu and Sn cations. Drawn using Diamond (Diamond - Crystal and Molecular Structure Visualization, Crystal Impact - Dr. H. Putz & Dr. K. Brandenburg GbR, Kreuzherrenstr. 102, 53227 Bonn, Germany, https://www.crystalimpact.de/diamond).

Our refinement of the synthetic analogue of okruginite revealed a partial disorder at the Cu1, Cu2 and Sn sites (i.e. Cu/Sn mixed sites) indicating that the structure is not fully ordered. By analogy to phase transitions observed for Cu₂SnS₃ phase, it is very likely that at higher temperatures (>600°C) okruginite transforms to a completely disordered sphalerite-based structure. The substitution of Se by S observed in natural material is in accordance with the work of Nomura et al. (Reference Nomura, Maeda, Takei, Morihama and Wada2013), who described the Cu₂SnSe₃–Cu₂SnS₃ solid solution.

Proof of identity of natural and synthetic okruginite

The structural identity of natural okruginite and its synthetic analogue was supported by EBSD measurements on the natural sample and comparison of the structural model for Cu₂SnSe₃ from our Rietveld refinement. Rietveld refinement of synthetic Cu₂SnSe₃ confirmed the monoclinic structure model of Delgado et al. (Reference Delgado, Mora, Marcano and Rincón2003), later also described by Nomura et al. (Reference Nomura, Maeda, Takei, Morihama and Wada2013). Hence, this structural model for Cu₂SnSe₃ was used in the EBSD measurement of the natural material.

A TESCAN LYRA3 FEG SEM combined with an EBSD system (Oxford Instruments AztecHKL with NordlysNano EBSD detector) was used for the measurements. The sample surface was prepared for EBSD by polishing with colloidal silica followed by broad beam ion milling (Gatan PECS II argon ion beam system operated at 2 kV for 60 min). Acquisition conditions were: accelerating voltage = 30 kV, beam current = 0.3 nA and no binning (full EBSP resolution is 1344 × 1024). The measured EBSD patterns obtained from a natural sample (several measurements from different grains of okruginite with different crystallographic orientations) were compared with patterns generated from the cubic, tetragonal and monoclinic structural models described above. Indexing conditions were: refined accuracy mode, 12 bands, 65 reflectors (the monoclinic model). All three models gave an acceptable match with reasonable mean angular deviation (MAD, i.e. goodness of fit of the solution) ranging between 0.26°and 0.36° (Fig. 8). It should be noted that in general the positive match obtained with EBSD is not a conclusive proof of structural identity, as EBSD cannot necessarily distinguish between closely similar structures (i.e. different derivatives of the sphalerite structure) that differ in a few reflectors observable in specific orientations only. Nevertheless, our synthesis and the relatively recent works of Delgado et al. (Reference Delgado, Mora, Marcano and Rincón2003) and Nomura et al. (Reference Nomura, Maeda, Takei, Morihama and Wada2013) favour the monoclinic Cc structure for Cu₂SnSe₃. It is also very likely that the cubic and tetragonal modifications of Cu₂SnSe₃ exist at higher temperatures (i.e. above 600°C) and the monoclinic phase is stable at lower temperatures. Considering the paragenetic sequence (conditions of formation in Nature of ~200–300°C) of okruginite, this mineral shows very likely monoclinic structure with Cc symmetry observed for synthetic Cu₂SnSe₃ by this work. Furthermore, the optical properties of okruginite indicate the monoclinic symmetry.

Figure 8. Electron back-scattered diffraction patterns from okruginite obtained from grains with different orientations; in the right pane the Kikuchi bands are solved (indexed). The cross marks the position of the pattern centre. Acquisition conditions were: accelerating voltage = 30 kV, beam current = 0.3 nA, no binning (full EBSD resolution is 1344 × 1024). Indexing conditions: 12 bands, 65 reflectors.

Raman spectroscopy

The Raman spectroscopic investigation was carried out using a Renishaw inVia Reflex Raman system coupled with a Leica microscope. The samples were measured with a 100× objective lens with excitation provided by a 785 nm diode laser and the signal was recorded by a thermoelectrically cooled CCD detector (spectral range of 100–4000 cm^–1, spectral resolution of 2 cm^–1). To enhance the signal-to-noise ratio, 20 scans were accumulated, each 20 s exposure time, with laser power at the source kept at a maximum of 15 mW to avoid thermal degradation. A polystyrene standard was used to check the wavenumber calibration and the spectra obtained were analysed using the GRAMS/AI 9.1 software package (Thermo Fisher Scientific).

We were able obtain the Raman spectrum of the synthetic phase which showed seven Raman bands around 75, 118, 209, 235, 251, and 366 cm^–1 with the dominant band at 180 cm^–1 (Fig. 9). These positions are generally in agreement with the Raman spectrum reported for monoclinic Cu₂SnSe₃ by Marcano et al. (Reference Marcano, Rincón, López, Pérez, Herrera-Perez, Mendoza-Alvarez and Rodríguez2011) who also provided tentative band assignments. Based on this, the bands at 75, 180 and 251 cm^–1 probably represent A′ modes, and bands at 209 and 235 cm^–1 A′′ modes. The weak band around 366 cm^–1 is probably an overtone of the strong 180 cm^–1 band, but the other the weak band around 118 cm^–1 has no counterpart in the spectra reported previously. We were unable to measure a Raman spectrum of the natural phase due to very small size of grains and spectral interferences from other phases (a S-rich analogue or Ti-oxides).

Figure 9. Raman spectrum of synthetic Cu₂SnSe₃ in the 50–500 cm^–1 range.