The composition of sulfides was determined using two CAMECA SX100 electron probe microanalyzers (EPMA) at Masaryk University, Brno, and at the National Museum, Prague, in wavelength dispersion mode. For EMPA at Masaryk University, Brno, an accelerating voltage of 25 kV, a beam current of 20 nA, and a beam spot size of 2 µm were used. Standards used for the analysis were Cu on metallic Cu, Ni and As on pararammelsbergite, Co on metallic Co, Zn on ZnS₂, Fe on FeS₂, Se on PbSe, Ag on metallic Ag, S on chalcopyrite, Hg on HgTe, Sb on metallic Sb, Bi on Bi, and In on InAs. For all minerals, the correction procedure (X-PHI) described by Merlet (1994) was applied. For EMPA at the National Museum, Prague, an accelerating voltage of 25 kV, a beam current of 20 nA, and a beam spot size ca. 1 µm were used. Standards used were Sb on Sb₂S₃, Ag on metallic Ag, Pb on PbS, Te on PbTe, Cl on halite, Sn on Sn, Fe on pyrite, Ni on Ni, Co on metallic Co, Zn on ZnS, Cu and S on chalcopyrite, Hg on HgTe, Tl on Tl(Br,I), Bi on Bi₂Se₃, Cd on CdTe, Au on metallic Au, Cr on metallic Cr, As on NiAs, Se on PbSe, Ge on metallic Ge, In on InAs, Mn on metallic Mn, Ga on GaAs, K on sanidine, Ba on baryte, and P on fluorapatite. The PAP matrix correction was applied (Pouchou and Pichoir, 1985). Overlap corrections and peak calibrations were systematically incorporated. Compositional data from EMPA are provided in the Supplement (Tables S2–13).

The whole-rock abundances of both main and trace elements were acquired during the deposit study done by Geomet s.r.o. Parts (typically halves) of drill cores in intervals ranging from 20 to 200 cm were, after crushing and homogenization, analyzed in ALS Global employing the ME-4ACD81 (base metals by ICP-AES) and ME-MS81 (lithium borate fusion followed by ICP-MS) packages from the ALS Global Geochemistry Fee Schedules (https://www.alsglobal.com/en/Resources-and-downloads, last access: 7 May 2026).

The Tescan integrated mineral analyzer (TIMA) solution in the R&D laboratory of TESCAN GROUP a.s., Brno, Czech Republic, was used to characterize selected thin sections and epoxy blocks in order to obtain the modal composition of the rocks, grain size distribution, and associations of the relevant ore minerals, bulk chemical composition, and elemental deportment and to understand the textures on a thin-section level. The machine combines signals from a BSE detector and up to four EDS detectors to distinguish individual phases and create mineral images, which are interpreted in dedicated software (see Hrstka et al., 2018). The following analytical conditions were used for acquisition: an accelerating voltage of 25 kV and probe current of 10 nA. Samples C1/205 and P1/250 were analyzed using the “dot mapping” mode (see details in Hrstka et al., 2018), with a 1 µm BSE and 5 µm EDS grid; samples C73, C74, and C75 were analyzed with dot mapping, with a 2 µm BSE and 6 µm EDS grid; and samples C62–64, C65–67, C68, C69, C70, C71, C72, C76, and C77–78 were analyzed using Bright-phase search sections (Hrstka et al., 2018), with a 2 µm BSE grid and 2 µm EDS grid for phases with brightness above the 25 % BSE threshold and their immediate surrounding and a 6 µm EDS grid for phases with brightness below the 25 % BSE threshold. For all samples, SE and CL images were acquired simultaneously with BSE and EDS. The data were measured and processed in TIMA 2.12 software.

Identification of some mineral species (pyrite, marcasite, enargite) was done using the Raman spectrometer Horiba Labram HR Evolution at the Department of Geological Sciences, Masaryk University, Brno. This dispersive, edge-filter-based system is equipped with an Olympus BX 41 optical microscope, a diffraction grating with 600 grooves per millimeter, and a Peltier-cooled, Si-based charge-coupled device (CCD) detector. The Raman signal was excited by a 633 nm laser. The nominal laser beam energy of 50 mW was attenuated to 5 or 12.5 mW using a neutral density filter to avoid the thermal damage of the analyzed area. The Raman signal was collected in the 100–900 cm⁻¹ range using a 600 gr mm⁻¹ grating and a 50× objective. The system was operated in confocal mode with a beam diameter of ∼ 1 µm. No visual damage to the analyzed surface was observed after the excitation. Raman-shift calibration was done using a silicon wafer. The wavenumber accuracy was ∼ 0.5 cm⁻¹, and the spectral resolution was ∼ 2 cm⁻¹.

Flat-lying quartz–zinnwaldite veins are identical in their mineralogical composition to steep veins (Štemprok, 1987) and most likely formed during the same genetic event; besides field observations (Štemprok, 1987), this is supported by the geochemical similarity of their quartz (Müller et al., 2018).

Sulfides occur in fractures (Fig. 3a, b) or in pockets (Fig. 3c), typically in centers of quartz–zinnwaldite veins; they form accumulations and irregular lenses of arsenopyrite, molybdenite, sphalerite, tennantite, galena, stannite, Cu sulfides, Bi minerals, pyrite, and opal (David, 1991). They either accompany the Li–Sn–W mineralization or the veins contain sulfide minerals only. The veins are characterized by the presence of several generations of coarsely crystalline quartz from the hydrothermal greisen stage preceding the sulfide mineralization and by the presence of fine-grained quartz on fractures filled with sulfides. In quartz–zinnwaldite veins, sulfide mineralization can form massive accumulations more than 10 cm large (Fig. 3).

Disseminated mineralization is volumetrically the most common type of sulfide mineralization at the deposit. It occurs in greisen and greisenized granites in the form of disseminated typically fine-grained subhedral grains to smaller sulfide aggregates. The main sulfide minerals encountered during past mining were galena and sphalerite, which accounted for 80 wt %–85 wt % (galena) and 5 wt %–10 wt % (sphalerite) of sulfide concentrates (David, 1991). In contrast to the veins, copper minerals (tennantite, chalcopyrite) are rare. Interaction of the fluids producing disseminated sulfides with Sn,W mineralization was locally encountered as texturally distinct assemblages. Thick-tabular crystals of wolframite I are recrystallized to Mn-enriched rims and/or needle-like crystals of wolframite II that overgrow sphalerite (see Hreus et al., 2021) and can be covered by later galena and replaced by scheelite. Muscovitization of zinnwaldite and formation of secondary cavities filled with clay minerals are associated with the process.

Baryte–fluorite veins seem to be the youngest hydrothermal assemblage at the deposit, cutting granite, greisens, and the Teplice rhyolite. They are usually devoid of sulfides, but they frequently interfered with earlier mineralization, especially those with galena, and tennantite. The baryte veins therefore locally contain corroded/recrystallized tennantite, galena, and secondary Cu sulfides (covellite, digenite/roxbyite, anilite, djurleite), as well as malachite (Fig. 3d, e, f), minor Pb-rich baryte (“hokutolite”), and anglesite.

Classification of Fe-rich members of the kësterite–stannite solid solution is rather complex. Kissin and Owens (1989) defined ferrokësterite as a low-temperature polymorph of stannite and Fe analogue of kësterite. Kissin (1989) noted a compositional Fe / Zn boundary between stannite and ferrokësterite at ca. Fe_0.75Zn_0.25 at 800 °C and showed the presence of minerals with a kësterite structure (either single or with stannite) in Fe-dominant synthetic systems down to 500 °C. From 700 to 600 °C, a single kësterite-structure phase was present from the Fe_0.5Zn_0.5 boundary to the Zn₁₀₀ composition; at 500 °C, a single phase was observed already from Fe_0.6Zn_0.4 (Kissin, 1989). On the other hand, Bonazzi et al. (2003) and Schorr et al. (2007) demonstrated the presence of a stannite structure ordering up to Fe_0.3Zn_0.7 on materials synthesized at 750 °C. The identification of ferrokësterite is rather problematic even from structural data (see, e.g., Bonnazzi et al., 2003), and there is a lack of experimental and especially structural works on transitional kësterite–stannite compositions that formed at temperatures lower than 600 °C; therefore, it is very problematic to classify them as either stannite or ferrokësterite based on their composition only. In this paper, we chose the conservative classification as stannite; however, further structural work would be necessary to clarify the mineral structural order and classification. Similar stannite and kësterite occurrences in the Erzgebirge include Vernéřov (Breiter et al., 2009b), where stannite (0.65–0.85 apfu Fe, 0.2–0.4 apfu Zn) replaces cassiterite and locally contains exsolution lamellae of kësterite (ca. 0.65 apfu Zn).

The stable Fe / Zn ratio in stannoidite from different mineralization stages indicates that it might be constrained by its crystal structure stability. Yamanaka and Kato (1976) showed that two out of three Fe atoms in the stannoidite structure are oxidized to Fe³⁺, and the crystal–chemical formula of stannoidite is therefore Cu¹⁺Fe23+Fe²⁺Sn24+S122-; as Zn substitutes for Fe²⁺ only, the ratio Fe / (Fe + Zn) = 0.66 represents the structurally constrained limit for Fe–Zn exchange. The Fe / (Fe + Zn) range of 0.65–0.74 for the Cínovec stannoidite agrees with its formation in Zn-saturated local systems involving sphalerite or tennantite-(Zn). The texture where stannoidite I forms rims around corroded Fe-bearing kësterite (Fig. 5) indicates kësterite replacement; the process can be represented by the following reactions: 4Cu2+(Fe2+0.3Zn0.7)SnS4+0.8FeS+2O2=Cu8+(Fe23+Zn)Sn2S12+1.8ZnS+3S+2SnO24Cu2+(Fe0.32+Zn0.7)SnS4+0.8FeS+4H2O=Cu8+(Fe23+Zn)Sn2S12+1.8ZnS+3H2S+H2+2SnO2. A change in redox conditions took place with an influx of hydrothermal solutions producing sphalerite. FeS can be sourced from FeS ↔ ZnS exchange that depletes kësterite I from iron, resulting in kësterite II (Fe-depleted rims in Fig. 5a). The same texture was also described as a reaction rim by Ramdohr (1980; see his Fig. 107).

Tetrahedrite-group minerals (Biagioni et al., 2020) from the Cínovec deposit are the most important carriers of As within the sulfide paragenesis. In most of the samples analyzed, tennantite-(Zn) is a dominant phase of this group. Tetrahedrite-group minerals from the deposit frequently show elevated to high contents of Bi (up to 1.52 apfu Bi). Much higher Bi contents (locally up to 2.65 apfu) were reported by Gołębiowska et al. (2012) from Rędziny, Poland, and by Sejkora et al. (2025), who defined the first Bi-dominant member of the tetrahedrite group, annivite-(Zn). Other occurrences of the Bi-dominant analogue of tennantite and tetrahedrite are reported, e.g., by Bortnikov et al. (1979), Kieft and Eriksson (1984), Spiridonov et al. (1986), Velebil and Sejkora (2018), and Dolníček et al. (2024). The tennantite–tetrahedrite from Cínovec is locally enriched in Ag (e.g., David, 1991), which correlates positively with Bi in Bi-rich tennantite II (at Bi > 1 apfu only) and Sb in tennantite III (Fig. 10c, d). Similar Bi-rich compositions are known from elsewhere in the Erzgebirge (e.g., Niederschlema-Alberoda and Jáchymov; Förster et al., 2004; Velebil and Sejkora, 2018) and Variscan orogene (e.g., central Schwarzwald, Staude et al., 2010) and are distinctly different from the Freiberg district (Burisch et al., 2019).

The chemistry of berryite from the Cínovec deposit shows significant variability (Fig. 12). Compared to the ideal formula Cu₃Ag₂Pb₃Bi₇S₁₆, it has elevated contents of Cu and in two samples also Zn. The excess of Cu has previously been described by several authors (e.g., Lowry et al., 1994; Cook, 1998; Topa et al., 2006; Sejkora et al., 2021). Variability of berryite chemistry within the deposit covers the variability of previously published berryite analyses from other localities around the world. A second occurrence from the Erzgebirge at Hřebečná (Sejkora et al., 2021) shows smaller ranges in Cu, Pb, and Bi concentrations and minor contents of Se.

The chemical composition of aikinite–bismuthinite series minerals shows wide variation. At the deposit, bismuthinite (naik 2.55–9.08), gladite (naik 36.72–38.16), salzburgite (naik 39.73–40.04), lindströmite (naik 61.49–65.18), friedrichite (78.27), and aikinite (86.22–100.82) have been distinguished. In several cases, aikinite with naik over 100 was identified; it could be caused by other substitutions or possibly by the presence of a vacancy. Such a wide compositional variation within the aikinite–bismuthinite series within one locality is not common; it was known from the metamorphosed scheelite deposit in Felbertal, Austria (Topa et al., 2002), similarly also in the Gemerská Poloma tin deposit and in the associated hydrothermal veins (Števko and Sejkora, 2021; Kondela et al., 2025). The common feature of all the localities is a multi-stage hydrothermal overprint of the previous lithology.

Besides early magmatic/metasomatic assemblages (Breiter et al., 2017b, c), the paragenetic sequence at the Cínovec/Zinnwald deposit (Fig. 15) reveals five distinct late metasomatic and hydrothermal mineralization stages, demonstrating the complex evolution of the deposit; their distribution generally corresponds to variance of Zn, Pb, Cu, and Ba within the deposit (Fig. 2). The earliest metasomatic (greisenization) stage is characterized by oxide–silicate assemblages (Johan and Johan, 1994, 2005; Breiter et al., 2017c; Hreus et al., 2021), with quartz and zinnwaldite formation in veins and granites, associated with, e.g., topaz, wolframite, and fluorite. It is important to note the positive correlation of Zn with Li (Fig. 2), which indicates that a large portion of Zn is contained in zinnwaldite due to ZnFe₋₁ exchange (up to ∼ 1100 ppm Zn; Breiter et al., 2023, unpublished data of authors). The first sulfides are related to the subsequent potassic stage. It is marked by feldspathization, muscovitization (Breiter et al., 2017a), and the formation of cassiterite I, Mo-rich scheelite I, and likely also arsenopyrite and molybdenite, although the formation of the latter during alteration of scheelite I seems more likely. The early sulfide stage begins by the formation of early kësterite I (Zn-rich stannite or Fe-bearing kësterite), followed by its reaction with sphalerite, resulting in the stannoidite, kësterite II, and cassiterite II assemblage. The intermediate sulfide stage with galena introduced significant Ag and Bi contents into the system. The late sulfide stage is characterized by the introduction of As, Sb, and further Cu and likely Fe, resulting in common tennantite–tetrahedrite minerals, chalcopyrite, and remobilization of elements such as Fe + In from sphalerite and Bi+Ag from galena (producing aikinite–bismuthinite series, gustavite, berryite, emplectite). Native bismuth found locally in the studied samples but widespread at the deposit must have formed below 271 °C (melting temperature of Bi). Its position towards other minerals is unclear due to its scarcity.

The latest fluorite–baryte stage represents the waning hydrothermal system, with dominant baryte, minor fluorite, and (likely remobilized?) galena or tennantite. The occurrence of enargite in tennantite cut by the fluorite–baryte vein (sample C73) indicates temperatures above ∼ 280–300 °C (Pósfai and Buseck, 1998); however, enargite likely represents the tennantite (or galena) stage rather than the baryte–fluorite overprint. The low-temperature alteration of tennantite produced both sulfides (Cu sulfides ranging from covellite to djurleite, cupropearceite, etc.) and oxygen-bearing (e.g., malachite, anglesite, mimetite, agardite-(Y), chrysocolla) phases. Stromeyerite represents the lowest-temperature assemblages below 93 °C (Skinner, 1966). The alteration is clearly (at least in the studied samples) related to the fluorite–baryte stage – Cu sulfides in tennantite voids are sometimes overgrown by fluorite and baryte, which are later followed by Cu arsenates (sample C73). Alternatively, Cu sulfides (covellite and djurleite) associated with mimetite occur in baryte voids, followed by calcite, muscovite, and kaolinite (sample C69). The textures suggest that the fluorite–baryte stage either followed after or, more likely, caused partial low-temperature overprint of the earlier sulfide assemblages.

Rare grains of cryolite matching the ideal formula Na₃AlF₆ were found as inclusions in low-Fe,Cu sphalerite and in galena, in both cases in the vicinity of tennantite and fluorite veinlets (samples C71 and C74). We assume that cryolite represents a rare product of albite and/or topaz reaction with remobilized fluorine-rich fluids.

David (1991) and Štemprok (1987) placed the tennantite stage after the fluorite–baryte stage. However, the textural-paragenetic features observed in our samples clearly show that tennantite precedes baryte formation. Fluorite occurs in several stages, including the end of the tennantite crystallization, which locally forms voids filled with fluorite. We assume that the earlier studies had to work with a limited number of samples where the two mineralizations intersect.

This sequence documents the evolving conditions from high-temperature (> 400 °C) magmatic–hydrothermal fluids to cooler (< 200 °C) late-stage fluids. A similar paragenetic sequence was encountered in the Krupka district located 6.5 km SE, also sourced from the Cínovec granite (e.g., Žák, 1966; Sejkora and Breiter, 1999), and is very similar for the whole eastern part of the Erzgebirge (e.g., Sadisdorf; Leopardi et al., 2024). The mineralization can be divided into two basic types – metasomatic bodies (greisens and “zwitters”) and ore veins, where sulfide and fluorite–carbonate stages can be distinguished (Žák, 1966; Sejkora and Breiter, 1999; Krejčí Kotlánová et al., 2024).

The NW–SE structural trend in the Teplice caldera related to intrusion of A-type granites (Breiter et al., 2017b) plays a fundamental role in controlling the spatial distribution of Sn–W–Li mineralization, as evidenced by numerous deposits on both the Czech and the German sides of the metallogenic province. Key examples include the Sadisdorf vein-type mineralization, the Krupka vein-type and greisen deposits, and the Cínovec/Zinnwald greisen system.

The relationship between Pb–Zn mineralization (galena, sphalerite) and greisen–vein systems can sometimes be characterized by spatial and temporal zonation linked to magmatic–hydrothermal fluid evolution. Greisen alteration driven by high-temperature (> 400 °C) fluids exsolved during late-stage granite crystallization typically hosts proximal Sn–W mineralization (e.g., cassiterite, wolframite) within or near granitic intrusions (Leopardi et al., 2024; Korges et al., 2018). As later fluids migrate through the intrusion fracture system outward, cooling (250–400 °C) and interaction with host rocks promote sulfide precipitation, which may lead to distal Cu + Pb + Zn ± Ag assemblages in vein and stockwork systems (e.g., Moore and Jackson, 1977; Mlynarczyk et al., 2003; Liu et al., 2016; Jiang et al., 2022; Leopardi et al., 2024).

At Sadisdorf (Germany), Pb–Zn sulfides (galena, sphalerite) occur in peripheral quartz–arsenopyrite–chalcopyrite veins, contrasting with proximal greisen-hosted Sn–W oxides; their mode of occurrence reflects fluid oxidation state and sulfur activity changes. Fluid inclusion data suggest mixing with meteoric water and decreasing salinity (0 wt %–10 wt % NaCl equiv.) in distal zones, facilitating base metal transport as chloride complexes and subsequent deposition via pH buffering (Leopardi et al., 2024). This zonation mirrors global greisen systems (e.g., San Rafael, Peru; Cligga Head, UK), where Pb–Zn mineralization marks the waning stages of hydrothermal activity, often overprinted by late fluorite–carbonate veins (Mlynarczyk et al., 2003; Moore and Jackson, 1977).

In Krupka, sulfide mineralization starts in the greisen stage, as evidenced by quartz–molybdenite veins with Sn–W mineralization (Knöttel deposit; Pauliš et al., 2022). The paragenetic sequence of sulfides (Žák, 1966) is very similar to that observed in Cínovec, starting with a sphalerite–chalcopyrite–pyrite assemblage followed by galena, Bi sulfides, and tennantite. The sulfide stage is followed by a final fluorite–carbonate stage with rare baryte. Krejčí Kotlánová et al. (2024) showed that the greisenization process took place at a temperature of 370–490 °C and a pressure of 155–371 bar, whereas the subsequent stages took place in a wide temperature range of < 120–370 °C. The source of fluids was not only magmatogenic but also meteoric water or fluids derived from sedimentary rocks.

A widely developed structural feature in the eastern Erzgebirge is the presence of flat-lying (subhorizontal) fractures, which are intensely mineralized with Sn–W–Li ores and associated sulfides. These structures are particularly well expressed in the Cínovec/Zinnwald cupola (e.g., Breiter et al., 2017b; Müller et al., 2018) and partly also in the Krupka district (Eisenreich and Breiter, 1993; Sejkora and Breiter, 1999; Pauliš et al., 2022). Their primary tectonic control is linked to the active exposure of granite intrusions, which promotes the formation of contractional fractures that broadly follow the morphology of the intrusion. Many of these subhorizontal fissures have been extensively exploited in the past and are notably enriched in Sn, W, Li, Cu, and Bi mineralization. These features collectively highlight the significance of contractional, intrusion-parallel structures for ore deposition in this segment of the Erzgebirge (David, 1991); the southern part of the Cínovec deposit, strongly dominated by massive greisens, represents an important anomaly.

Small vein-type mineral deposits combining cassiterite with common sulfides were historically mined in the central Erzgebirge; unfortunately, only basic mineralogical information is available. The W–E trending quartz veins at Hora svatého Šebestiána are located about 400 m above hidden peraluminous granite pluton and evolved in two stages: (1) quartz–cassiterite–chlorite and (2) quartz–cassiterite–sphalerite (with pyrrhotite, galena, chalcopyrite, tetrahedrite) (Breiter, 1981). About 20 NE- to SW-oriented veins were mined at Hora svaté Kateřiny. Mineralization evolved from the quartz–cassiterite stage and quartz–chlorite–sulfide stage (mainly sphalerite, chalcopyrite, and tetrahedrite) to, finally, the fluorite–carbonate stage (Breiter, 1982). The nearest outcrop of a rare-metal granite of type A was found ca. 1.5 km to the SW from the deposit (Breiter, 2008).