Mineral chemistry of the Geyer SW tin skarn deposit: understanding variable fluid/rock ratios and metal fluxes

Abstract

The Geyer tin skarn in the Erzgebirge, Germany, comprises an early skarnoid stage (stage I, ~ 320 Ma) and a younger metasomatic stage (stage II, ~ 305 Ma), but yet, the source and distribution of Sn and the physicochemical conditions of skarn alteration were not constrained. Our results illustrate that contact metamorphic skarnoids of stage I contain only little Sn. REE patterns and elevated concentrations of HFSE indicate that garnet, titanite and vesuvianite of stage I formed under rock-buffered conditions (low fluid/rock ratios). Prograde assemblages of stage II, in contrast, contain two generations of stanniferous garnet, titanite-malayaite and vesuvianite. Oscillation between rock-buffered and fluid-buffered conditions are marked by variable concentrations of HFSE, W, In, and Sn in metasomatic garnet. Trace and REE element signatures of minerals formed under high fluid/rock ratios appear to mimic the signature of the magmatic-hydrothermal fluid which gave rise to metasomatic skarn alteration. Concomitantly with lower fluid-rock ratio, tin was remobilized from Sn-rich silicates and re-precipitated as malayaite. Ingress of meteoric water and decreasing temperatures towards the end of stage II led to the formation of cassiterite, low-Sn amphibole, chlorite, and sulfide minerals. Minor and trace element compositions of cassiterite do not show much variation, even if host rock and gangue minerals vary significantly, suggesting a predominance of a magmatic-hydrothermal fluid and high fluid/rock ratios. The mineral chemistry of major skarn-forming minerals, hence, records the change in the fluid/rock ratio, and the arrival, distribution, and remobilization of tin by magmatic fluids in polyphase tin skarn systems.

Introduction

The skarns of the Erzgebirge are host to a significant share of Sn resources in Europe with a known endowment exceeding 360 kt of Sn (Elsner 2014). The regions with the most prolific Sn skarns are the Schwarzenberg District in the western part of the Erzgebirge (Korges et al. 2020; Lefebvre et al. 2019; Reinhardt et al. 2021), and the Geyer SW Sn-skarn deposit in the central part of the Erzgebirge. Indicated resources of the Geyer SW skarn bodies have been estimated to 8.3 Mt of ore at 0.53 wt% Sn and up to 1.12 wt% Zn (Elsner 2014; Hösel et al. 1996). However, some of the Sn is hosted by calc-silicates and thus cannot be economically extracted. The skarns of the Schwarzenberg District and Geyer SW are exceptionally well exposed by existing underground mine workings (Schwarzenberg District) and exploration drilling (Geyer SW). Yet, published data on the petrography and mineral chemistry of the skarns are very limited (Hösel et al. 1996; Schuppan and Hiller 2012).

A resurgence of Sn exploration has spurred scientific interest with recent studies providing a robust geological, geochronological, and mineralogical context (e.g., Bauer et al. 2019; Burisch et al. 2019; Kern et al. 2019; Korges et al. 2020; Lefebvre et al. 2019; Meyer et al. 2024; Reinhardt et al. 2021). Some of these studies have documented that most Sn in the skarn deposits of the Schwarzenberg District and at Geyer SW is hosted by cassiterite accompanied by actinolite, chlorite and quartz related to late-stage skarn alteration (Kern et al. 2019; Meyer et al. 2024). However, Kern et al. (2019) illustrated for the Hämmerlein skarn in the Tellerhäuser deposit that up to 0.1 wt% of total Sn may be hosted by calc-silicate minerals, including stanniferous garnet, malayaite (CaSnSiO₅), amphibole and epidote. As a result of detailed petrographic descriptions of the Geyer SW skarns (Meyer et al. 2024), a similarly complex distribution of Sn between various silicates and cassiterite has been rather tentatively proposed. Hence, the present follow-up study was conducted to provide a comprehensive set of mineral chemical data for the Geyer SW skarns. Results allow an assessment of the partitioning of Sn amongst various skarn minerals, and thus provides valuable information on the distribution of Sn across different paragenetic stages which is important for possible metal recovery. The results can also be used to provide important constraints on the origin of Sn in the skarns of the Erzgebirge. The recognition of a pre-metamorphic background enrichment of Sn in the metasediments that also host skarn mineralization (Romer and Kroner 2015; Romer et al. 2022; Weber et al. 2023) has led to a resurgence of the hypothesis that a significant share of the Sn in the Erzgebirge skarns is related to proximal redistribution of Sn of sedimentary origin (Lefebvre et al. 2019). This contrasts with the hypothesis that Sn was introduced from magmatic sources (Lehmann 2021).

Despite the recent advances in understanding skarn-hosted Sn mineralization of the Erzgebirge, systematic investigations of trace elements are scarce up to date. Here, we employ LA-ICP-MS and EPMA to determine the major, minor and trace element mineral chemistry of calc-silicates and cassiterite from a suite of drill core samples from the Geyer SW skarn bodies. We integrate the mineral chemical data with detailed petrographic observations and available petrogenetic constraints (Meyer et al. 2024) in order to better understand the mode of skarn formation in different environments (rock-buffered vs. fluid-buffered) and related fluctuations in fluid/rock ratios. Furthermore, we aim to constrain Sn fluxes (magmatic-hydrothermal vs. sedimentary) and the redistribution of Sn amongst various silicates and oxides during the evolution of a polyphase skarn mineral system. We include previously unpublished EPMA data from Hämmerlein (Erzgebirge) and Crown Mine (Cornwall) into our discussion to illustrate that our findings are relevant to other Sn skarn systems as well.

Skarn formation and genetic implications of calc-silicate compositions

Skarn assemblages form during regional or contact metamorphism, due to metasomatic replacement reactions of any kind of lithology, but especially carbonate-bearing rocks, with most economic skarns being related to magmatic-hydrothermal fluids (Meinert et al. 2005). Commonly, skarns are divided into endo- and exoskarn indicating an igneous or sedimentary protolith, respectively. Exoskarns can further be subdivided into skarns sensu stricto replacing calcite marble (calcic) or dolomite marble (magnesian) and skarn sensu lato, which relates to skarn alteration of marly schists with higher content of silica and aluminum (Meinert 1992). Most skarns show a transition from small scale diffusion-controlled reactions to infiltration metasomatism within their spatiotemporal evolution (Einaudi et al. 1981; Harlov and Austrheim 2013; Jamtveit and Hervig 1994; Kwak 1987; Meinert 1992; Park et al. 2017a). The interaction between fluid and host rock is mainly controlled by diffusion rates, composition of the fluid and the protolith, dictating textural characteristics and chemical compositions of skarn minerals (Einaudi et al. 1981; Pertsev 1974; Watanabe 1960; Zharikov 1970). Within reaction skarns, diffusion relies on chemical gradients in pore solutions with low fluid/rock ratio and is inefficient in transporting components over a larger distance, which typically results in the formation of skarnoid textures (Einaudi et al. 1981; Jamtveit and Hervig 1994). In contrast, advective metasomatic skarn formation is characterized by fluid flow driven by pressure gradients along percolation pathways, forming anhydrous (early) and hydrous (late) skarn assemblages (Harlov and Austrheim 2013; Meinert et al. 2005; Norton 1987). In these skarns, textures and compositions of minerals are a direct result of interaction with a (magmatic-) hydrothermal fluid that can efficiently transport chemical constituents. The mineral chemistry, zoning textures and cross-cutting relationships of skarn minerals, hence, record physicochemical changes during the spatiotemporal evolution of a skarn system.

An increasing number of trace element studies on skarn-related garnet have recently become available (Chen et al. 2022; Gaspar et al. 2008; Hong et al. 2022; Huang et al. 2022; Jamtveit et al. 1993; Jamtveit and Hervig 1994; Park et al. 2017b). The vast majority of these studies are dedicated to W, Cu-Au or polymetallic skarns; no trace element data have yet been published for calc-silicate minerals in tin-dominated skarn systems. These studies illustrate that the composition of skarn minerals formed under low fluid/rock ratios is primarily controlled by the bulk composition of the host rock. In such an environment, HFSE (high field strength elements) and Al can be transported over short distances to then be incorporated in newly formed skarn minerals (Jamtveit and Hervig 1994). In contrast, the chemical composition of skarn minerals formed under high fluid/rock ratios will more likely reflect the characteristics of the magmatic-hydrothermal fluids involved. Typically, these fluids transport an abundance of Fe and Si, and can introduce elements such as B, F, Li, W, Sn (and many more) into the skarn mineral assemblage (Einaudi and Burt 1982).

The rare earth elements (REE) are typical examples of HFSE – and the study of their distribution may provide interesting insight into the origin of skarn minerals. The overall concentration of REE in magmatic-hydrothermal fluids is typically low and marked by relative enrichment in LREEs and relative depletion in HREEs (Ayers and Eggler 1995; Bai and Koster van Groos 1999; Bauer et al. 2019; Reed et al. 2000). The extent to which the signature of the fluid is incorporated or whether the incorporation of the REE into minerals follows crystallographic restrictions depends largely on the kinetics of the growing minerals (Gaspar et al. 2008; Jamtveit and Hervig 1994). The more the growth rate of the minerals exceeds the diffusion rate of the fluid, the more the signature of the fluid is incorporated. Additionally, high (but variable) fluid flow will lead to strong oscillatory chemical zonation, euhedral mineral shapes, and complex twinning, which is best described as an isometric texture (Gaspar et al. 2008; Jamtveit and Hervig 1994; Meinert 1992).

Regional geology

The Erzgebirge (in German) or Krušné Hory (in Czech) region straddles the border between the Free State of Saxony in Germany and the Czech Republic (Fig. 1A). It comprises a sequence of metamorphic nappes (Fig. 1B) that formed during the Variscan orogeny in central Europe (Kroner et al. 2007). Protolith to these nappes was a variable volcano-sedimentary succession comprising greywackes, conglomerates, sandstones, clay stones, marls, and carbonate rocks as well as minor volcanic rocks of mafic to felsic composition (Ondrus et al. 2003; Tichomirowa et al. 2012), ranging from early Paleozoic to Neoproterozoic in age. These were deposited along the northern edge of the Gondwana Supercontinent (Mingram 1998); they underwent variable degrees of deformation and metamorphism during the Variscan Orogeny, owing to the collision of Gondwana with Laurussia during the Carboniferous. Peak metamorphism occurred at approximately 340 Ma, followed by rapid exhumation of the nappes to middle to upper crustal level (Kröner and Willner 1998) and the intrusion of voluminous granitic intrusions (Fig. 1B), facilitated by extensional tectonics (Wenzel et al. 1997). This late-Variscan phase (Breiter et al. 1999) was characterized by post-kinematic peraluminous magmatism and took place between 327 and 310 million years (Romer and Meixner 2014; Tichomirowa et al. 2019, 2022). Intrusions of this period comprise of crustal and minor mantle-derived melts, and their geochemical heterogeneity is attributed to the lithological variability of the stacked continental crust (Förster et al. 1999). Granitic batholiths of A-, I- and S-type composition (such as syeno- and monzogranites), evolved porphyries, lamprophyres, and extensive rhyolitic and dacitic volcanic rocks were emplaced after the main collisional stage (Breiter 2012; Förster et al. 1999; Tischendorf and Förster 1990). Later on, during the post-orogenic extensional tectonic phase from 305 to 286 million years, mainly (sub-)volcanic felsic magmatism occurred (Hoffmann et al. 2013; Luthardt et al. 2018; Zieger et al. 2019).

Fig. 1

(A) Geographical position of the Erzgebirge region in Central Europe. (B) Geological overview of the Erzgebirge with the location of the Geyer system (modified after Meyer et al. 2024). (C) Some relevant geological features in the Geyer-Ehrenfriedersdorf district, focusing on important localities where magmatic-hydrothermal mineralization styles (greisen, skarn) are known to occur (after Hösel 1994)

Following the collapse of the Variscan Orogen, subsidence and basin formation led to the deposition of a substantial (2–3 km, Wolff et al. 2015) sedimentary sequence that covered the Variscan basement (Pälchen and Walter 2011). During the Cretaceous Period, reactivation of the Elbe rift led to marine transgression and subsequent sedimentation. Rifting of the Eger Graben caused the Erzgebirge block to tilt towards the north in the Cenozoic. Rapid uplift and erosion were locally accompanied by local mafic (sub)volcanic magmatism – processes that resulted in the current exposure level as a Variscan basement window in Central Europe (Pälchen and Walter 2011).

Skarns in the Erzgebirge

The Variscan basement of the Erzgebirge is well known to host various types of magmatic-hydrothermal mineral systems, including greisen (Baumann et al. 2000; Korges et al. 2018), skarn (Burisch et al. 2019; Reinhardt et al. 2021), and epithermal vein-type mineralization (Swinkels et al. 2021). Recent geochronological studies have documented that the bulk of magmatic-hydrothermal mineralization in the Erzgebirge is associated with post-kinematic, late-orogenic (330–310 Ma) and post-orogenic granitic magmatism (305–285 Burisch et al. 2019; Meyer et al. 2024; Reinhardt et al. 2022; Romer et al. 2007; Tichomirowa et al. 2019). This includes the known skarn deposits (Reinhardt et al. 2022; Meyer et al. 2024).

Skarns across the Erzgebirge are mostly associated with carbonate-bearing metasediments of the Cambro-Ordovician succession of the Schwarzenberg Anticlinorium, which forms a SW-NE trending belt across the Erzgebirge. The metasediments consist of alternating sequences of metapelites, mica schists, marbles, and mafic metavolcanic rocks. The skarn bodies vary significantly in size, mineralogy, and the metal tenor: They may host subeconomic to economic grades of Sn, W, Fe, Zn, and/or In (Burisch et al. 2019; Hösel 2002; Reinhardt et al. 2021; Schuppan and Hiller 2012).

The most common skarn lithotypes in the Erzgebirge comprise garnet-pyroxene ± magnetite ± amphibole ± chlorite skarns marked by a polyphase origin (Burisch et al. 2019; Reinhardt et al. 2021). Oldest skarn occurrences are typically small and linked to the peak of regional metamorphism around 340 Ma (Burisch et al. 2019; Reinhardt et al. 2022) without significant metal endowment. Skarns of considerable size and, in some cases, well-endowed in tin formed later in two distinct stages that have been distinguished by garnet U-Pb dating (Burisch et al. 2019; Reinhardt et al. 2022; Meyer et al. 2024). Garnets of the first stage have ages of 325 to 313 Ma, coinciding with the age of many greisen deposits that are associated with the emplacement of late-collisional granites (Burisch et al. 2019; Romer et al. 2007). Garnets of the second stage of skarn formation range between and 308 to 295 Ma in age. These garnets are associated with abundant cassiterite mineralization in greisen and veins, and they are contemporaneous with post-collisional magmatism and the earliest onset of Permian rifting (Burisch et al. 2019; Hoffmann et al. 2013; Meyer et al. 2024).

Geyer SW skarn

The skarns at Geyer SW (Fig. 1C) are bound to three units of interbedded calcite marble, dolomitic marble, and muscovite-biotite schist (Fig. 2A). The regional metamorphic, amphibolite-facies metasedimentary host rocks strike northeast, dip 30–40° NW, and are underlain at 200–400 m depth by granite of the Ehrenfriedersdorf batholith (Fig. 2B; Hösel et al. 1996). The skarns are 1–10 m thick and can be followed along strike for up to 300 m; their Sn content decreases gradually with distance to the granite contact – from 0.67% Sn at 100 m to 0.30% Sn at 400 m from the granite contact (Bolduan 1963; Hösel et al. 1996; Meyer et al. 2024).

Fig. 2

(A) Geological map of the Geyer system with the exposed Geyersberg stock and profile line of B-B’. Collar positions of sampled drill cores are indicated with yellow stars (after Hösel et al. 1996). (B) Cross-section of the Geyer SW area including skarn- and vein-hosted Sn mineralization underlain by the Ehrenfriedersdorf composite batholith (after Hösel et al. 1996)

Based on careful petrographic studies, Meyer et al. (2024) distinguished in all three skarn units two temporally distinct events of skarn formation at Geyer SW. The older stage I assemblage is devoid of any Sn mineralization and exhibits characteristics typical of a proximal contact metamorphic skarnoid (Meinert et al. 2005), e.g., the texture of the host rock is still recognizable, fractures are absent, and anhydrous skarn minerals show small, anhedral crystal sizes with intense intergrowths. U–Pb ages of ~ 322 Ma for garnets of stage I relate the origin of these skarnoids to the emplacement of the large Ehrenfriedersdorf batholith at 324 to 315 Ma (Meyer et al. 2024; Romer et al. 2007). These skarnoid assemblages formed under lithostatic conditions, experiencing a minimum pressure of 200 MPa at depths between 2.5 and 4 km, with limited fluid flow at a temperature of 450 °C and 470 °C (Breiter et al. 1999; Meyer et al. 2024; Romer et al. 2007). Extensive retrograde alteration as well as cassiterite mineralization are conspicuously absent from stage I assemblages (Meyer et al. 2024).

Tin mineralization, specifically the formation of malayaite and cassiterite, is linked to a second, distinctly younger event of skarn formation (stage II; ~ 305 Ma from overlapping garnet and cassiterite U–Pb ages; Meyer et al. 2024). While no granitic intrusions of this age are known to occur in the Geyer district, abundant rhyolitic volcanic rocks of this age occur across the Erzgebirge, providing ample evidence for province-scale felsic magmatism (Förster et al. 1999; Hoffmann et al. 2013; Löcse et al. 2023; Luthardt et al. 2018).

According to fluid inclusion analyses, the onset of stage II skarn formation occurred at > 330 °C under hydrostatic conditions (10–15 MPa). The presence of high salinity fluid inclusion assemblages with heterogeneous entrapment hosted by prograde pyroxene grains of stage II has been used to suggest boiling (Meyer et al. 2024). Low fluid inclusion salinities hosted by amphibole, chlorite, sphalerite and cassiterite suggest dilution of the ore-fluid by meteoric fluids during the retrograde phase of stage II (Meyer et al. 2024). The bulk of cassiterite mineralization in the Geyer magmatic-hydrothermal system is related to late retrograde veinlets of stage II comprising fluorite or topaz, chlorite, white mica and/or quartz (Bolduan 1963; Hösel et al. 1996; Meyer et al. 2024).

Yet unpublished mineral chemical data from two other tin skarn localities are used in this study for comparison with the data from Geyer SW. These additional data are from the Hämmerlein skarn in the Schwarzenberg District (Kern et al. 2019) and from the Crown Mine of the Cornwall District, UK (van Marcke de Lummen 1986). More detailed information on the geological setting of these two Sn skarns is provided in electronic supplementary ESM 1.

Samples and methods

Access to six drill cores from the Wismut SDAG and Tin International exploration campaigns of the Geyer SW prospect was provided by the Geological Survey of the Free State of Saxony. These were sampled, resulting in a total of 60 individual samples of half drill cores, ranging in length from 5 to 20 cm. The main target of sampling was the selection of a wide range of macroscopically different unmineralized and mineralized portions of the three skarn units. 84 polished thin sections for petrographic investigations were produced of the 60 samples, 48 of which were analyzed by electron microprobe and LA-ICP-MS. Sample location, analytical condition, and representative analyses as well as the mineral calculation procedures are provided in ESM-1.

All sections were first examined by transmitted and reflected light microscopy on a Leica DM750P microscope equipped with an Olympus single-lens reflex camera. In addition, a Phenom XL scanning electron microscope (SEM) with an energy dispersive (EDS) X-ray spectrometer at the University of Tübingen was used for mineral identification and texture documentation in thin sections. Backscattered electron (BSE) images were acquired at an acceleration voltage of 15 kV and a beam current of 15 nA. Element distribution maps of polished hand specimens and thin sections were generated by microbeam X-ray fluorescence element mapping using a Bruker Tornado M4 instrument at the University of Tübingen. This instrument is equipped with a rhodium target X-ray tube operating at 50 kV and 500 nA and two EDS X-ray detectors, which produce unquantified false-color maps corresponding to the height of the Kα peaks of the respective elements.

This study includes three electron microprobe datasets that were obtained under slightly different measurement conditions and corrections. The data sets are collated in ESM-2 (Geyer SW), ESM-3 (Hämmerlein) and ESM-4 (Crown Mine). Analytical conditions used to obtain this data are summarized below and reported in detail in ESM 1.

Major and minor element chemistry of skarn minerals of Geyer SW was determined using a JEOL JXA-8230 electron microprobe at the Department of Geosciences, University of Tübingen. The analyses were conducted in wavelength-dispersive (WD) mode, employing an acceleration voltage ranging from 15 to 20 kV, a probe current of 20 nA, and a focused beam for all minerals except for amphibole and chlorite, for which a spot size of 5 μm was used. Counting times of 16/8 sec (peak/background) were applied for major elements, while minor elements were counted for 30/15 sec. To process the data, the PRZ correction was applied to all minerals (garnet, clinopyroxene, vesuvianite, epidote, amphibole, chlorite, titanite, malayaite and cassiterite).

Quantitative analyses of silicates and oxides were performed on samples from the Hämmerlein Deposit, Erzgebirge, and Crown Mine, Cornwall, with a JEOL JXA 8530 F Hyperprobe at the Helmholtz Institute Freiberg for Resource Technology, Germany, using WDS. An acceleration voltage of 20 kV, a probe current of 40 nA and spot sizes between 0.1 and 4 μm were chosen. Analyses were recalculated by the JEOL software with reference materials used for calibration matching the measured phases as closely as possible. Analyzed concentrations were corrected offline for instrument drift using natural mineral reference materials (ASTIMEX MIN25 and MAC Cassiterite). Mutual interferences were corrected following the procedure described in Osbahr et al. (2015).

Forty trace elements, including rare earth elements (REE), of gangue and ore minerals were analyzed by LA-ICP-MS at the Laboratory of Environment and Raw Materials Analysis (LERA), Karlsruhe Institute of Technology (KIT, Germany), using a Teledyne 193 nm Excimer Laser coupled to a sector-field ICP-MS (Element XR ThermoFisher). Ablation was done under a helium atmosphere and mixed with argon and nitrogen before the plasma torch. The ICP-MS was tuned for maximum sensitivity while keeping the oxide formation (UO/U < 0.1%) and element fractionation low (i.e., Th/U = ~ 1). Static ablation used a spot size of 35 μm in diameter and an energy flux of ca. 5 J/cm² at 10 Hz for 30 s. Before the analysis, each spot was pre-ablated (ca. 1 μm) to remove surface contamination. Calibration and data quality checking was done using the National Institute of Standards (NIST) 612 glass, BIR, BHVO-1 and BCR-2G. Silicon concentrations, previously determined by electron microprobe, served as an internal standard for garnet, clinopyroxene, epidote, amphibole, chlorite, titanite and malayaite. For cassiterite, the concentration of Sn was used as an internal standard. Data were reduced and evaluated using the iolite software package v4. The complete dataset of LA-ICP-MS analyses of Geyer SW may be found in ESM-5. All REE data were normalized to C1 chondrite after McDonough and Sun (1995), an approach adopted from recent studies dealing with the distribution of trace elements in skarn systems (e.g. Chen et al. 2022; Gaspar et al. 2008; Huang et al. 2022; Liu et al. 2022; Ordosch et al. 2019; Park et al. 2017a, b; Smith et al. 2004) in order to improve comparability. Eu anomalies were calculated based on the formula Eu/Eu* = (Eu)_cn/[(Sm)_cn x (Gd)_cn]^0.5 of McLennan (1989), where values below 1 show a negative Eu anomaly and vice versa.

Results

In the following, a brief summary of the macroscopic and microscopic characteristics of skarns from Geyer SW is provided. Please refer to Meyer et al. (2024) for a more detailed description. Following that, mineral chemistry data are presented for the localities Geyer SW, Hämmerlein and Crown Mine.

Petrography Geyer SW

Based on observations from drill cores (Fig. 3) and detailed petrography (Figs. 4 and 5), three characteristic mineral assemblages are recognized in samples from the Geyer SW skarn. All three of these assemblages may occur in close spatial proximity to one another, often rendering the recognition of paragenetic relations difficult on hand specimen scale. However, the identification of discrete generations of, e.g., garnet can be based on cross-cutting relationships, textures, crystal shape and color. Different garnet generations also include core to rim associations, which are regarded as generations because of strikingly different appearances. Similar features apply to other skarn minerals, which have been assigned to distinct assemblages following Meyer et al. (2024).

Fig. 3

Representative samples from drill cores (core and depth in brackets). (A) Mica schist replaced by pyroxene skarn (667-7-01; 391.1–391.2 m). (B) Fine-crystalline pyroxene (Cpx) skarn with remnants of mica shist and areas with higher proportion of red garnet (Grt; 667-5-01; 382.5–382.6 m). (C) Garnet-rich sample with cream-colored garnet and red vesuvianite (Ves; 667-11-01; 409.3–409.4 m). (D) Large pocket of red garnet with wollastonite (Wol) in a pyroxene-dominated area of stage I skarn (626-07; 85.0–85.2 m). (E) Densely intergrown large crystals of dark brown garnet and dark green pyroxene (667-15-01; 429.6–429.7 m). (F) Green clinopyroxene skarn cut by a vein containing vesuvianite at the rim and green garnet in the central part (626-02; 45.6–45.7 m). (G) Garnet-rich skarn with dark green, euhedral garnet crystals which are crosscut by veinlets containing hematite (Hem) and cassiterite (Cst; GM-08-1; 22.0–22.1 m). (H) Malayaite (Mly)-dominated parts in a garnet skarn under short-wave ultraviolet light (GM-08-2; 22.3–22.4 m). (I) Skarn mainly composed of amphibole with abundant sphalerite (Sph) (617-02; 167.2–167.4 m). (J) Vein with epidote (Ep) and euhedral, red sphalerite crosscutting skarn containing disseminated chalcopyrite mineralization (606-02; 81.4–81.5 m). (K) Chlorite- and cassiterite-filled veins intersecting epidote- and amphibole-rich skarn (657-03; 114.9-115.1). (L) Gneiss crosscut by a cassiterite (Cst)-quartz (Qz)-arsenopyrite (Apy)-bearing vein. (606-5; 242.6–242.7 m)

Fig. 4

Typical textures and intergrowths of garnet illustrated in SEM-BSE (B, C, E, H, I, J) and transmitted light microscopic images (A, D, F, J). (A) Euhedral, sector twinned garnet Grt I (plane and cross polarized image; 667-1-01a). (B) SEM-BSE image of Grt Ia overgrown by Grt Ib with wollastonite (Wol; 657-01). (C) Replacement of Grt Ia and Ib by Grt IIa (SEM-BSE; 667-4-05). (D) Poikilitic cores of Grt IIa overgrown by Grt IIb (plane polarized image; 626-08). (E) Grt IIa intergrown with fluorite and pyroxene, overgrown by Grt IIb (SEM-BSE; 667-10-02). (F) Brown Grt IIb crosscut by green Grt IIIa in veins (626-05; plane polarized image). (G) Stanniferous Fe-rich Grt IIb replaced by Al-rich Grt IIc, starting from veins. Al-mapping done with microprobe; black and blue represent low Al (and high Fe) areas while green to yellow represents Al-rich areas (667-10-04). (H) Anhedral Grt Ia and Ib intersected by Grt IIIa (SEM-BSE; 626-03). (I) Euhedral, lath shaped malayaite (Mly) in the vicinity of a replaced stanniferous garnet (SEM-BSE; 626-04). (J) Poikilitic Grt IIc intergrown with malayaite, fluorite and clinopyroxene overgrown by Grt IId with strong oscillatory zonation (SEM-BSE combined with plane polarized image; GM-8-02)

Fig. 5

Other minerals. (A) Green clinopyroxene intergrown with anhedral, poikilitic Grt Ia in transmitted light (667-8-04). (B) Lath-shaped, euhedral vesuvianite (Ves I) with anhedral Grt Ia (667-11-01). Circles in Vesuvianite are laser ablation marks (transmitted light). (C) Small euhedral epidote I (Ep I) replacing biotite (Bt) and feldspar (Kfs; SEM image; 667-14-02). (D) Titanite enclosed by Grt I. Where titanite is no longer shielded by the garnet, it is replaced by malayaite (SEM image; 667-1-01)). (E) Titanite containing remnants of rutile (Rt) and ilmenite (Ilm) is replaced in the upper left corner by malayaite (Mly) (SEM image; 606-01). (F) Fibrous, densely intergrown vesuvianite (Ves II) with clinopyroxene (transmitted light, crossed polarized; 606-01). (G) Large, lath shaped clinopyroxene (Cpx II) intergrown with magnetite (Mag). Both are subsequently replaced by sphalerite (Sph; reflected light; 626-06). (H) Euhedral vesuvianite intergrown with early euhedral amphibole without replacement (667-16-04). I) Green epidote (Ep II) in a vein together with deep red sphalerite and subhedral fluorite (Fl; 606-02). J) Clinopyroxene replaced by euhedral, internally zoned amphibole (Amp IIb; 606-04)). K) Chlorite and sphalerite replace clinopyroxene and amphibole (638-04). L) Partly replaced amphibole (Amp IIc) enclosed by chalcopyrite (reflected light; 657-06)

Stage I

Macroscopically, stage I skarn shows remnants of host rock and metamorphic banding is preserved to variable degrees (Fig. 3A and B). Predominantly calcite marble and marly mica schist are replaced by skarn. Lenses of skarn exhibiting stage I alteration do not commonly exceed 1 m in thickness; the transition to surrounding altered mica schist is gradational. Skarn I mineral associations are characterized by variable ratios of garnet to clinopyroxene with a significant proportion of wollastonite, vesuvianite, quartz, feldspar, and epidote (Fig. 3C and D).

Garnet-rich areas display a fine crystalline texture, albeit with multiple larger garnet crystals reaching sizes of up to 1 cm (Fig. 3D). Two garnet generations (Grt Ia and Grt Ib) were identified. Garnet Ia is usually some millimeters up to 1.5 cm large, has an orange to beige color and irregular rims. These garnets are strongly anisotropic, very weakly zoned and commonly show sector or dodecahedral twins (Fig. 4A). They are intergrown with green clinopyroxene (Fig. 5A), euhedral, red vesuvianite (Ves I; Fig. 5B) and epidote (Ep I; Fig. 5C). Titanite (Ttn I) is commonly enclosed in garnet (Fig. 5D). In contact between former marble and mica schist, titanite occurs as aggregates with inclusions of relict ilmenite and rutile (Fig. 5C). The outer zones of garnet Ia are overgrown by garnet Ib. This alteration generally does not exceed 100 μm and is best visible with the SEM (Fig. 4B and C). Garnet Ib is accompanied by increasing amounts of wollastonite, fluorite, and quartz.

Stage II

Early skarn stage IIA

Drill cores illustrating substage IIA show a highly variable clinopyroxene to garnet ratio, where minerals are generally coarse-grained and show isometric textures Fig. 3e, F, G). This association is most widespread in the studied sample set, and the thickness of skarn lenses differ, ranging from 30 cm up to 3 m. Yellow to brown garnet crystals reach sizes of up to 6 cm (Fig. 3E, F) and, more rarely, a mineral association which is dominated by green garnet can be encountered (Fig. 3F and G), occurring in lenses and small veins with sizes up to 50 cm. Fluorite and malayaite constitute additional phases that mark the IIa mineral assemblage (Fig. 3H).

Microscopically, the first garnets of the stage II skarn (Grt IIa) are of anhedral, porous appearance with numerous inclusions of other minerals, primarily clinopyroxene (Cpx II), remnants of earlier formed garnet, and fluorite. This garnet generation has cream to light brown colors, is slightly anisotropic without recognizable twins, but notably shows distinct dodecahedral crystal shapes (Fig. 4E). Vesuvianite (Ves II) has deep red to pale green colors and forms lath-shaped to acicular crystals (Fig. 5F). Occasionally, Ves II is zoned, with a darker core passing to a brighter rim. Clinopyroxene is only euhedral and of considerable size when it is intergrown with magnetite, forming up to 1.5 cm large crystals. (Fig. 5G). Malayaite replaces titanite to variable degrees, forming titanite-malayaite solid solutions (Ttn-Mly_ss), resulting in patchy zonation with malayaite-rich and -poor domains. In some samples, titanite and the surrounding host minerals are crosscut by fractures, with only the portion no longer shielded by the host mineral being more Sn-rich (Fig. 5D). Garnet IIa is subsequently overgrown by garnet IIb (Fig. 4D and E). The most distinctive feature of these garnets is the scarcity of inclusions, their strongly oscillatory zoning, and their euhedral shape (Fig. 4F). Their coloration ranges from yellow to dark brown and they are nearly isotropic. Towards the edge of single grains, the color intensity gradually diminishes, and the spacing between individual bands increase. Very locally, garnet IIb and vesuvianite II are intergrown with green, euhedral amphibole (Amp IIa; Fig. 5H).

As skarn alteration progresses, mineral formation starts from veins and replaces previously formed minerals. The first generation of garnet in these veins (Grt IIc) is of bright green color with weak oscillatory zoning (Fig. 4F). This generation of garnet crosscuts garnet I and garnet IIa and b, but also clinopyroxene (Fig. 5G, H, I). Malayaite (Mly) is the most common inclusion in garnet IIc and is usually present as euhedral, lath-shaped crystals or acicular aggregates up to 1 mm (Fig. 4I). Garnet IId, which overgrows garnet IIc, forms euhedral, dodecahedral, mostly isotropic crystals up to 2 cm in diameter (Fig. 4J). Compared to other garnet generations, growth zones are of considerable size, up to 400 μm wide, and they may be bent or tilted (Fig. 4J). Inclusions are scarce; fluorite is observed as a minor mineral only.

Late-stage skarn IIB

The late-stage alteration of stage II skarn (substage IIB) comprises epidote, amphibole, chlorite, and other hydrous minerals in varying proportions (Fig. 3I, J and K). These minerals occur as massive aggregates, but also in veins and their alteration aureoles (Fig. 3J and K). The late-stage skarn alteration is in some samples associated with base metal sulfides (Fig. 3J and K) or with cassiterite mineralization (Fig. 3L).

Green epidote (Ep II) and dark green to black amphibole (Amp IIb and c) occur in veins (Fig. 5I) as dense aggregates or as replacement products (Fig. 5J), but they are rarely intergrown with each other. Epidote and amphibole crystals can reach sizes of up to 1 cm, featuring well-defined growth zones and only few inclusions. In certain areas, amphibole may undergo local replacement by chlorite (Fig. 5K) or become enclosed by chalcopyrite, sphalerite, and other sulfides (Fig. 5L). The formation of cassiterite is consistently associated with veins, veinlets, and their alteration aureoles, typically affecting the host rock (skarn as well as metasedimentary units) within a limited zone, rarely extending more than a few tens of centimeters. Cassiterite crystals are predominantly euhedral and exhibit well-defined growth zones. In veins, they can attain sizes of up to 1.5 cm, although they are generally smaller in impregnated areas. The highest quantities of cassiterite are typically encountered in fluorite- and chlorite-filled veinlets. These veinlets often exhibit the most substantial alteration halos. However, cassiterite was also found overgrowing magnetite and sphalerite in clinopyroxene-dominated skarn, replacing garnet and malayaite in garnet-dominated skarn and appearing as interstitial phase between amphibole, hematite, and chlorite (Fig. 6A). Manganiferous calcite, prehnite, feldspar and fluorite together with “wood tin” and stokesite (CaSn(Si₃O₉)*2H₂O) occur in different proportions and are the latest observed minerals associated with the Sn system; they replace most of the previously formed skarn minerals.

Fig. 6

(A) Occurrences of cassiterite (Cst), commonly in veins with fluorite (Flr) or chlorite (Chl) crosscutting garnet (Grt) or pyroxene (Px) skarn or in impregnated areas with hematite (Hem) or chalcopyrite (Ccp; transmitted light; only sulfide skarn image in reflected light). (B) Trace elements of cassiterite normalized to the high- fluorine S-type Eibenstock granite. The elements of cassiterite are remarkably consistent irrespective of different host rocks and different gangue minerals.¹Förster et al. 2021; *divided by 1000. (C) REE pattern of cassiterite from veins with different gangue minerals and host rocks. Please note that colors of frames around pictures in (A) correspond to line colors in graphs B and C

Mineral chemistry

In the following, the results obtained for the major, minor and trace element chemistry of all major groups of silicate minerals present in the Geyer SW skarn units will be reported, following their position in the different stages (Fig. 7). Results are presented for each mineral group – and within these groups compositional differences between the three different stages (I, IIA, IIB) will be highlighted. In addition, we will report the mineral chemistry of cassiterite. The procedures for mineral formula calculation and representative chemical analyses of the mineral groups may be found in the ESM-1.

Fig. 7

(modified from Meyer et al. 2024) analyzed by EPMA and LA-ICP-MS. The sketch on the right side schematically illustrates the different generations of garnet and their typical intergrowths

Simplified paragenetic sequence of minerals.

Garnet Group

The first garnet generation in stage I (Grt Ia) occupies a large compositional range of And_4–55Grs_43–95 with up to 0.16 apfu F. Sn remains invariably below or near the detection limit of the electron microprobe (EPMA, LOD is 0.05 wt%), but LA-ICP-MS data show concentrations between 11 and 600 µg/g. Grt Ia is overgrown and replaced by garnet Ib, which still belongs to stage I. It is identified only by its Fe concentration that is always higher than that of the immediately adjacent Grt Ia. Grt Ib also covers a wide range of compositions from And_19–82Grs_15–77 with F concentrations up to 0.11 wt%. Sn contents in Grt Ib only rarely exceed the LOD of the EPMA (Fig. 8A), but Sn is well detectable with LA-ICP-MS (12–980 µg/g).

Fig. 8

Electron microprobe analyses of skarn minerals of the different skarn stages. (A) Garnet with Sn per formula unit against Al divided by the sum of total Fe and Al, reflecting the compositional range between grossular and andradite. (B) Sn versus Ti in atoms per formular unit of titanite group minerals showing the solid solution between titanite and malayaite with a prominent miscibility gap. (C) Sn per formula unit against the clinozoisite component of epidote-group minerals. Sn is more enriched in Fe-rich epidote. (D) Analyses of amphibole with Sn per formula unit plotted against the Mg number. Analyses with intermediate Mg number but high Sn are amphiboles of the substage IIA.

The oldest garnets of stage II (Grt IIa) cover a compositional range of And_2–39Grs_55–92 with F concentrations up to 0.51 apfu. Sn is still low (0.02 apfu), but consistently above the LOD of the EPMA. The concentration of Fe increases in garnet IIb (And_29–99Grs_0–65), accompanied by low F and high Sn concentrations (up to 0.11 apfu; Fig. 8A). Garnet IIc has a wide range of compositions between And₁₃Grs₇₇ and And₉₀Grs₅, with highly variable F concentrations between LOD and 0.55 apfu, and Sn concentrations ranging between 0.01 and 0.12 apfu (Fig. 8A). The final garnet generation of stage II (Grt IId) has again a large range of compositions (And_36–100Grs_0–55), with low F contents but (on average) highest Sn concentrations (up to 0.22 apfu; Fig. 8A), which corresponds to ~ 7 wt% of Sn. In all these metasomatic garnet generations, F concentrations increase with increasing grossular component, while Sn concentrations increase with increasing andradite component.

Among the trace elements, REE provide the most remarkable results. Different garnet generations have rather distinct REE compositions – and these differ strikingly between the different paragenetic associations. REE patterns of different garnet generations from Geyer SW are shown in Fig. 9A. The corresponding LA-ICP-MS data is collated in ESM 5. Grt Ia of stage I exhibits a distinctive left-sloping REE distribution, which is characterized by depletion of light REEs (La–Nd; LREE), relative to elevated middle (Sm–Ho; MREE) and heavy REEs (Er–Lu; HREE), and an insignificant to negative Eu anomaly (δEu = 0.2–1.8). In contrast, the REE patterns of Grt Ib are relatively flat, with a slight deficit in LREE, a more variable distribution of HREEs and small negative Eu anomalies (δEu = 0.4–1.3).

Fig. 9

(A) REE distribution normalized to C1 chondrite (McDonough and Sun 1989) in garnet from Geyer SW. See text for more details. (B) Trace element contents of W and In in µg/g and HFSE content represented by Ti and Zr of different garnet generations. Note the differences in scale

Grt IIa, the first garnet generation of stage II, exhibits a LREE-depleted pattern broadly similar to Grt Ia but it is less depleted in LREEs, a minor decrease in HREEs and moderate negative Eu anomaly (δEu = 0.2–2.1). More andradite-rich Grt IIb displays a relatively flat REE pattern. It features low La and HREE, but high Sm and Nd concentrations, and a strong positive Eu anomaly (δEu = 1.3–29.7). The REE_N pattern of Grt IIc shows a significant decrease in La and an increase in Ce–Gd, along with no or an only slight negative Eu anomaly (δEu = 0.3–1.75). HREE concentrations are variable, but generally lower than LREEs, except for La. Grt IId, the final garnet generation, has the highest La concentrations of all garnets. LREEs are high, but decrease rapidly towards HREEs, except for a strongly positive Eu anomaly (δEu = 0.9–25.4).

High field strength elements (HFSE), such as Ti, Zr (Fig. 9B), Y, Nb, and Ta, are typically higher in garnets with a higher grossular component (e.g., Grt Ia, Grt IIa, Grt IIc). Sn, W and In, on the other hand (Fig. 9B), Fig. 10, 11, 12are highest in garnets with an elevated andradite component. The content of Li is generally low in stage I garnet but increase in Grt IIc and IId of substage IIA. Within the garnet generations, a trend from low W concentrations with high HREEs to high W with high LREEs can be observed (Fig. 13A). This coincides with deviations towards higher ΣREEs against Y (Fig. 13B), while the Y/Ho ratio stays rather constant (Fig. 13C).

Pyroxene Group

Clinopyroxene of stage I (Cpx I) is variable in composition (Di_21–88Hd_11–73Jhn_1–6) with a distinct positive correlation between the hedenbergite and johannsenite components. Stage II clinopyroxene (Cpx II) has a similarly variable composition of Di₁₄–₉₉Hd_1–79Jhn_0–7. The most diopside-rich (i.e. iron-poor) clinopyroxene of stage II are typically present in magnetite-bearing skarns. The major element compositions of Cpx I and Cpx II thus occupy a similar broad range and can only be subdivided petrographically. Trace element contents and “typical” clinopyroxene minor elements Na (up to 200 µg/g), Li (up to 40 µg/g) or Ti (up to 480 µg/g) are low and do not differ significantly between the two skarn stages. The only exception is Zn, which shows elevated concentrations in both stage I (30–2150 µg/g) and II (50–1600 µg/g). In contrast, Cpx II intergrown with magnetite, yields significantly lower Zn concentrations with a range between 20 µg/g and 85 µg/g. These clinopyroxenes are also the only ones to show somewhat elevated concentrations of REEs with a significant negative Eu anomaly, while all the other pyroxenes have remarkably low REE concentrations.

Vesuvianite (Ca₁₉(Al, Mg, Fe)₁₃(SiO₄)₁₀(Si₂O₇)₄(OH, F,O)₁₀).

Vesuvianite of stage I (Ves I) displays the highest concentrations of Mg (4.00 apfu) and Ti (1.14 apfu; 3 wt%), while Sn is generally below the detection limit of the EPMA. The REE pattern of vesuvianite I is marked by depletion of HREEs with enriched LREEs and without a significant Eu anomaly. Concentrations of HFSE, B and Sb are high, other elements like In and W are present in low concentrations. Stage II vesuvianite (Ves II) contains somewhat higher Fe and Mn contents and Sn concentrations of up to 0.21 apfu, corresponding to 1 wt% Sn. The REE concentrations in vesuvianite define a consistent chondrite-normalized fractionation trend with depleted LREE and a positive Eu anomaly. HFSE, Y, Ho and B contents in Ves II are lower compared to Ves I, whereas Li, In, and W contents are higher.

Titanite - Malayaite solid solution series Ca(Ti, Sn)(SiO₄)O.

Titanite with Sn concentrations below the detection limit (Ttn I) of the electron microprobe is found exclusively in association with stage I garnet (Fig. 8B). This titanite has a composition of Ttn₁₀₀Mly₀ to Ttn₉₄Mly₀ where Ti is substituted by Al. The resulting imbalance in charge is likely accommodated by the following coupled substitution: (Al, Fe)³⁺ + (OH, F)⁻ ⟺ Ti⁴⁺ + O²⁻ (Ribbe 1982). The REE patterns of Ttn I are generally flat, with a minor negative Eu anomaly (δEu = 0.4–0.9). and exhibit high ΣREE concentrations ranging from 51 to 1168 µg/g (Fig. 10A). Concentrations of elements such as W (1–28 µg/g), In (below LOD–26 µg/g), and Sn (17–324 µg/g) are relatively low, while Y (43–1026 µg/g) and Zr (5–1114 µg/g) concentrations are rather high.

In stage II, increasing amounts of Sn are incorporated into the crystal lattice of the solid solution series titanite – malayaite - until the malayaite component becomes predominant (Ttn IIa). A distinct miscibility gap is observed within the compositional range between Ttn₆₃Mly₃₇ and Ttn₄₂Mly₅₈, corresponding to 0.55 to 0.35 apfu Ti in malayaite and vice versa (Fig. 8B). This has also been described in previous studies (Aleksandrov and Trenova 2007). With a higher malayaite component, REE patterns remain flat, but exhibit more pronounced HREE depletion, a stronger negative Eu anomaly, and an overall decrease in ΣREE (Fig. 10A). Higher Sn concentrations in the solid solution correlate with increased contents of In (0.4–1751 µg/g) and W (2–760 µg/g). In contrast, higher titanite contents are associated with elevated concentrations of Y and Zr, along with higher ΣHREE.

Malayaite that contains only an insignificant titanite component (ranging from Ttn₂Mly₉₈ to Ttn₀Mly₁₀₀) is primarily found in assemblages relating to late stage II (Ttn IIb). Ttn IIb is mainly enclosed by garnet IIc. No coupled substitution involving Al, Fe, or other elements is observed. In this group, In (667–2589 µg/g) and W (0.5–140 µg/g) exhibit the highest concentrations in members of the titanite-malayaite solid solution series, while Zr (0.8–221 µg/g) and Y (0.5–58 µg/g) are comparatively low in concentration. Malayaite has the lowest ΣREE concentrations (up to 103 µg/g) and displays a REE pattern characterized by depleted HREE, a clear negative Eu anomaly (δEu = 0.04–0.72), and in some cases, a clear positive anomaly of Gd and Tb (Fig. 10B), often referred to as the third tetrad anomaly (Masuda et al. 1987).

Epidote Group

Epidotes in stage I (Ep I) occur together with calcite and quartz, preferentially at the transition between skarn and altered mica schist. Compositionally, Ep I occupy a wide range between Czo₈₄Ep₁₄Pmt₂ and Czo₁₆Ep₈₃Pmt₂ with generally low Sn concentrations (Fig. 8C). Because of the crystal size and the tight intergrowth with garnet, only a few LA-ICP-MS measurements could be conducted on Ep I. The REE_N pattern of Ep I is relatively flat with an insignificant positive Eu anomaly (δEu = 1.9–2.3) and a gradual decrease of the HREEs. Most trace element abundances in Ep I are also low (see ESM 2 for details).

The bulk of the epidote at Geyer SW is associated with substage IIB (Ep II). The compositional range is similar to Ep I, spanning a range between Czo₅₉Ep₃₉Pmt₂ and Czo₀Ep₉₄Pmt₆, but with Sn concentrations (Fig. 8C) reaching up to 0.27 apfu (equivalent to 7.8 wt% Sn). Concentrations of Ti or REEs, on the other hand, are only minor. The REE pattern of Ep II is strongly depleted in HREEs and shows a distinct positive Eu anomaly (δEu = 0.9–18.6). High In concentrations are associated with elevated Sn contents, while other trace elements exhibit insignificant concentrations (see ESM 2 for more details).

Amphibole Supergroup

Amphiboles were only recognized in stage II, and they belong to the Ca-amphibole subgroup (Hawthorne et al. 2012; Locock 2014). Amphiboles are scant in mineral association IIA (Amp IIa); they occur together with vesuvianite and clinopyroxene and are euhedral, with well-developed chemical zonation. Most commonly, they belong to the pargasite-hastingsite series, with variable amounts of K, Mg, and Fe²⁺, as well as intermediate to high concentrations of Sn (0.05–0.14 apfu; Fig. 8D). Amphibole IIa also has the highest concentrations of F (0.46–0.64 apfu) among the analyzed amphiboles. The REE_N patterns of Amp IIa (Fig. 10B) are characterized by low La and relatively high Ce–Nd concentrations, and a decrease in HREEs with a discernable negative Eu anomaly (δEu = 0.55–0.97).

Fig. 10

REE distribution normalized to C1 chondrite (McDonough and Sun 1989) of (A) titanite-malayaite_ss and (B) amphibole

Most amphibole occurs in mineral association IIB (Amp IIb and Amp IIc). These amphiboles are of heterogeneous composition, but mainly belong to the root groups of tremolite, actinolite, hornblende, pargasite, and hastingsite. Amphibole IIb occurs in veins or as replacement product of clinopyroxene and garnet; it is commonly enriched in ΣFe and can contain up to 0.21 apfu Sn (2.7 wt%). Tremolite-rich Amp IIb is restricted to clinopyroxene skarn with a substantial proportion of diopside (Fig. 8D) and contains only low concentrations of Sn. The REE signature of Amp IIb is characterized by depletion of La, high Ce–Nd concentrations, and only a minor positive Eu anomaly (δEu = 0.79–2.37). Further, Amp IIb is marked by an enrichment of the heaviest REEs whilst maintaining a consistent Y/Ho ratio. Amp IIc, in contrast, is intergrown with or replaced by base metal sulfides. It belongs to the (ferro)-actinolite group and commonly contains only low Sn concentrations (Fig. 8D). Amp IIc is, however, enriched in HREEs, depleted in LREEs, and may display a negative Eu anomaly (δEu = 0,36–1,05). The Y/Ho ratio in Amp IIc deviates strongly from that of the other amphibole generations – which is due to lower Ho concentrations (Fig. 13F). Details about the endmember calculations, major and trace element data for amphiboles can be found in ESM 1, 2 and 5, respectively.

Chlorite Group

Minerals of the chlorite group are an important constituent of the late part of the substage IIB assemblage; they belong to the chamosite (Fe-rich) and clinochlore (Mg-rich) solid solution series, with only a minor presence of the pennantite (Mn) component (ESM 1, 2). Where chlorite replaces clinopyroxene, amphibole or epidote, its composition largely depends on the precursor mineral. In cases where chlorite coexists with newly grown cassiterite as infill of veinlets, it tends to be Fe-rich (chamosite-rich). In all cases, Sn contents in chlorite were found to be below the detection limit of both EPMA and LA-ICP-MS. Other minor and trace elements (including the REEs) also showed low to very low concentrations (ESM 5).

Cassiterite

Well-developed euhedral cassiterite is only common in veinlets of substage IIB, where it commonly displays optical and chemical zonation. Cassiterite occurring in aureoles around these veinlets is of smaller, anhedral appearance and typically shows less pronounced zoning. Fe is the most common substituent in cassiterite with up to 0.12 apfu (ESM 1, 2). Elements such as Al and Si are present in smaller quantities (0.02 and 0.07 apfu, respectively). Despite the variable textural context in which cassiterite is found, its trace element composition is remarkably uniform. Compared to whole rock data of the Geyer granite (Förster et al. 2021), cassiterite is depleted in Li, Rb, Y, Zr, Pb, enriched to a certain degree in V, Co, Ni, Zn, Nb, Mo, and notably enriched in Sb, W, and In (Fig. 6B). Indium may be enriched in cassiterite by up to four orders of magnitude (note that In in Fig. 6B is divided by 1000). Furthermore, the normalized REE patterns are strikingly consistent, all marked by low ΣREE (6.6 to 57.7 µg/g), low LREE and HREE concentrations, an insignificant Eu anomaly, but a significant third tetrad anomaly (Fig. 6C).

Garnet and clinopyroxene from the crown mine skarn

An introduction of the regional and local geology is summarized in the ESM 1. The mineral chemistry of clinopyroxene and garnet from the Crown Mine skarns are presented below; the complete dataset of all important skarn phases can be found in the ESM 3.

Garnet occurs in two different generations (stage I and II). The first garnet stage is accompanied by titanite, chlorite or magnetite. Garnet stage I shows complex extinction features, but only a very weak chemical zonation and has a composition of And_15–45Grs_51–85 with Sn invariably below the detection limit of the EPMA. Associated with the first garnet generation is clinopyroxene, which commonly shows a bright core and a greenish rim. This is further reflected in the chemical composition, where the core is more diopside-rich compared to the hedenbergite-rich and johannsenite-bearing rim. The composition of clinopyroxene ranges from Hd₂₃Di₇₄Jhn_1,6 to Hd₆₃Di₃₃Jhn₄. Sn concentrations are always below the detection limit of the EPMA.

The second garnet generation (stage II) occurs in monomineralic, vein-like structures cross-cutting older garnet generations and other skarn minerals. Garnet II is mainly associated with stanniferous titanite, malayaite, and axinite; clinopyroxene is distinctly lacking. The composition of garnet II ranges from And_0–60Grs_39–73 with, in parts, a pronounced spessartine and almandine component and elevated Sn concentrations up to 0.03 apfu, corresponding to 1.58 wt%.

Garnet and clinopyroxene from the hämmerlein skarn

Information about the local geological setting of the Hämmerlein skarn is summarized in the ESM 1. Mineral chemical analyses were performed by EPMA on major skarn-forming minerals of the Hämmerlein skarn (details results are provided in ESM 4). Here, only data of garnet and clinopyroxene are presented, as these are the main constitutes of the skarns. A subdivision between different generations of these skarn-forming minerals was not possible because analyses were carried out on grain mounts rather than polished thin sections. However, a general petrographic description of calc-silicates from this skarn can be found in Kern et al. (2019), Korges et al. (2018), Lefebvre et al. (2019) and Schuppan and Hiller (2012).

Garnet from the Hämmerlein skarn has a composition between And₁₄Grs₈₆ to And₉₅Grs₅ with only minor spessartine and pyrope component. The Sn concentration ranges from below the detection limit of the EPMA up to 0.04 apfu, corresponding to 2.25wt.% Sn. Although the Sn content increases with increasing andradite component, andradite-rich garnets with very low tin concentrations are also present. Clinopyroxene, commonly associated with garnet, has a composition between almost endmember diopside (Hd_0.25Di_99.75Jhn₀) and Hd₃₄Di₆₂Jhn₄. In none of the analyses did Sn contents exceed the detection limit of the EPMA.

Discussion

We will start the discussion by a general overview of mineral data from tin skarns by comparing the detailed and petrographically well-based major element composition of garnet and clinopyroxene from Geyer SW to the data from the Hämmerlein and Waschleithe skarns (Erzgebirge, Germany), the Crown Mine (Cornwall, England) and further Sn skarns (worldwide) compiled by Meinert (1992). We will then present a mineral system model of tin skarns based on which, as a last step, our major, minor and trace elements will be interpreted within the context of the multi-stage origin of the skarn mineral system at Geyer SW. A particular emphasis will be placed on the abundance and distribution of Sn in the skarn-forming minerals at Geyer SW.

Classification and comparison of skarns in the Erzgebirge and worldwide

Skarn stage I

Early stage skarnoid skarn alteration is documented for most of the skarn deposits in the Erzgebirge (Burisch 2019; Meyer 2024). Furthermore, they are also often reported from skarn systems elsewhere (Cornwall; Alderton and Jackson (1978), Burisch et al. (2023), Canada; Layne and Spooner (1991), China; Huang et al. (2022). Commonly, such early skarn stages are barren or only host significant mineralization if overprinted by subsequent skarn stages, as they are mainly related to the contact metamorphic overprint, which pre-dates the main magmatic-hydrothermal event. For this reason, we exclude the mineral data of skarn stage I from the following comparison.

Skarn stage II

In the following, the second skarn stage at Geyer, which is genetically related to tin mineralization, is compared to other Variscan tin skarns. The data comprise major and minor element concentrations of garnet and clinopyroxene of the Hämmerlein skarn, Erzgebirge (Lefebvre et al. 2019; this study), the Waschleithe skarn, Erzgebirge (Reinhardt et al. 2021), and the Crown Mine skarn, Cornwall (this study).

Meinert (1992) suggested that major element compositional trends of both garnets and clinopyroxenes can be linked to metal endowment in skarn systems (see Fig. 11A). Differences between the various skarn types, however, are actually only recorded in the variable proportions of spessartine + almandine (in garnet), and johannsenite (in clinopyroxene) endmembers. The data presented here for Variscan Sn skarn deposits can be evaluated in the context of Meinert (1992).

Fig. 11

Ternary plots of garnet (left) and clinopyroxene (right) composition from individual deposits of this study (Geyer SW (blue and green), Hämmerlein (orange) and Crown mine (brown)) compared with literature data from Hämmerlein skarn, Erzgebirge (yellow; Lefebvre et al. 2019) and the Waschleithe skarn, Erzgebirge (purple; Reinhardt et al. 2021). Grey areas are garnet and clinopyroxene compositions observed in association with skarn deposits with distinct metal tenor (Meinert 1992)

As shown in Fig. 11, each of the Variscan Sn skarn deposits shows two different groups of garnet compositions: one very (Sps + Alm)-poor spanning most of the range between Grs and And, and another group with significantly higher (Sps + Alm) content. In the case of Hämmerlein and Crown mine, the (Sps + Alm) enrichment closely follows the trend shown by Meinert (1992), while in Geyer SW and Waschleithe (a distal Zn-skarn in the Schwarzenberg District, Reinhardt et al. 2021), this enrichment is indistinct.

Accordingly, the major element composition of garnet and clinopyroxene from Variscan Sn skarns does not correspond well to the trends proposed by Meinert (1992). However, high Sn concentrations in most skarn-forming minerals in combination with an abundance of sphalerite suggests that Geyer SW may be classified as an intermediate Sn-Zn skarn with Sn dominance (note, that Sn concentrations in garnet from Geyer SW are higher than elsewhere in the Erzgebirge and in Cornwall; Lefebvre et al. 2019; Reinhardt et al. 2021; this study). Similarly, the Waschleithe skarn with its much higher amounts of sphalerite and only negligible known Sn mineralization can be regarded as an intermediate Sn-Zn skarn with Zn dominance (as also suggested by Reinhardt et al. 2021), while the Crown Mine skarn is a more or less “pure” endmember Sn skarn. The Hämmerlein skarn deposit, however, does not really fit into this classification scheme. There, the skarn shows extensive sphalerite mineralization together with abundant cassiterite and other Sn-bearing minerals (Korges et al. 2020). Despite its enrichment in Zn and Sn, the Hämmerlein skarn shows similar patterns of garnet and clinopyroxene compositions as the Crown Mine Sn-skarn (Fig. 11B). This leads to the conclusion that – in contrast to Meinert (1992) – garnet and clinopyroxene major element compositions do not allow to unequivocally constrain the metal endowment of skarns in the Variscan orogen. Obviously, a multitude of complex variables control skarn mineral composition and a direct inference of mineralization type based on mineral chemistry alone is often not possible. These variables include the composition of the precursor rocks, formation temperature, and fluid- or rock-buffering of the system. As these factors can vary significantly, not just within one skarn category, but even within one single deposit, the resulting compositional trends are very wide and any overlap with the fields defined by Meinert (1992) is thus of limited significance, especially if only small datasets are considered.

A concise mineral system model

Detailed petrographically work, microthermometric data, precise in-situ geochronology (Meyer et al. 2024), together with major, minor and trace element mineral chemistry of important skarn forming minerals allow the establishment of a coherent mineral system model for the Geyer SW skarn system which is depicted in Fig. 12.

Fig. 12

Composite sketch showing the main characteristics of calc-silicate formation in the different steps of skarn formation with typical elements, REE patterns and textures

Fig. 13

(A) W (µg/g) against the ratio of La (µg/g) vs. Lu (µg/g) of garnet showing the development from rock-buffered (stage I and substage IIA) to fluid-buffered (substage IIB) signatures. (B) Y (µg/g) vs. total REEs (µg/g). Lower Y values with relatively constant total REE suggests an open system while constant ratios are evident for a closed system. (C) A constant ratio of Ho (µg/g) vs. Y (µg/g) in garnet over a wide range of concentrations indicates that the influx of meteoric fluid in the prograde skarn stages I and II is very limited. (D) Similarly, the Y vs. Ho ratio of epidote remains constant, suggesting an absence of fractionation which would indicate an influx of meteoric water. Note also that decreasing Y and Ho concentrations in epidote are accompanied by increasing Sn contents, (E). Influx of meteoric water is evident in Y/Ho ratios of cassiterite by the variations of Y vs. Ho. (F) The deviation of the Y to Ho ratio of amphibole intergrown with sulfides indicates an enhanced influx of meteoric water while stanniferous amphibole grows under hydrothermal conditions with limited meteoric water influx

The first stage of skarn formation at Geyer is characterized by the development of fine-grained calc-silicate assemblages dominated by garnet, clinopyroxene and vesuvianite with a skarnoid texture (Meinert et al. 1992; Meyer et al. 2024), i.e., the skarn preserves the metamorphic fabric of the protolith, and calc-silicate minerals are not chemically zoned. U–Pb ages for garnet of stage I coincide in time with the emplacement of the Ehrenfriedersdorf batholith, which is known to immediately underlie the skarn units (Bolduan 1963; Hösel et al. 1996). Fluid inclusions in pyroxene constrain the formation temperature of skarn I to about 450–470 °C under lithostatic conditions (Meyer et al. 2024), which is consistent with p-T-X(CO₂) conditions predicted by component reaction calculation in Einaudi et al. (1981) and Hochella et al. (1982). Such skarnoids are widely believed to form under rock-buffered conditions (Gaspar et al. 2008; Jamtveit et al. 1993; Meinert 1992), which is proved by the composition (including REE pattern) of stage I skarn minerals, which provide evidence for a mainly rock-buffered system without tin mineralization. Only in the latest stage, a minor influence of a magmatic-hydrothermal fluid is discernible.

U–Pb ages of garnet and cassiterite unequivocally show a temporal gap of more than 10 million years between stage I (~ 320 Ma) and stage II (~ 305 Ma) of skarn formation at Geyer SW, indicating that they are temporally and genetically clearly distinct. Stage II skarn also shows characteristics indicative of formation at a relatively shallow depth (1.5–2 km) under hydrostatic conditions (Meyer et al. 2024), probably related to a small, highly evolved following intrusion at a shallower level. Mineral reactions (following Einaudi et al. 1981; Hochella et al. 1982) and microthermometric data for clinopyroxene (Meyer et al. 2024) additionally confirm a lower temperature of skarn formation for stage II (320–335 °C). Textures along with mineral compositions and REE signatures record an overall trend from a rock-buffered, nearly closed system towards a fluid-buffered, open skarn environment corresponding to increasing fluid/rock ratios in stage II at Geyer SW (Fig. 12). However, oscillations in the fluid/rock ratio are indicated by garnet IIc which records lower fluid/rock ratios than garnet IIb. Hence, we propose that at least two distinct fluid pulses occurred during the prograde stage II skarn alteration. High Sn concentrations in minerals are invariably related to magmatic-hydrothermal fluids in a fluid- buffered environment. Here, garnet is host to the bulk of Sn in the system.

The late stage II is characterized as the retrograde phase during skarn alteration in a purely fluid-buffered environment under hydrostatic conditions and brittle deformation (veining). The onset of this stage is marked by a change of silicate assemblages from tin-bearing garnet (Grt IId) to tin-bearing epidote (Ep II) and tin-bearing amphibole (Amp IIb). Amphibole (Amp IIb and IIc) and chlorite replace existing Fe-Mg-silicates; their formation suggests lower temperatures and the presence of an aqueous fluid as compared to the prograde stage II (substage IIA). Formation temperatures calculated from the mineral chemistry of chlorite after Cathelineau (1988) (210–350 °C; ESM 2) are in good agreement with fluid inclusion data associated with this stage (255–340 °C; Meyer et al. 2024) that also show a prolonged dilution of a more saline magmatic fluid with a low-salinity meteoric fluid. During retrograde alteration, the majority of cassiterite precipitates, leading to the formation of the important and extractable Sn mineralization.

The following chapter will in detail discuss the variable fluid/rock ratios, the impact of magmatic-hydrothermal and meteoric fluids and the (re-)distribution of Sn in the framework of the model presented above. Some of the trace element characteristics, especially during stage I skarn alteration, are inherited from the replaced host rock. Nevertheless, systematic changes can be observed during the development of the skarn system, as elaborated below.

Rock vs. fluid buffered: the evolution of the skarn system reflected by mineral chemistry

Stage I: A rock-buffered start

Garnet Ia has a grossular-rich composition with elevated concentrations of HFSE. Similar is true for clinopyroxene and vesuvianite associated with Garnet Ia – both have a distinct Mg-component. The enrichment in Al, Mg and HFSE in early-stage calc-silicates is indicative for a rock-buffered environment as those elements are abundant in the metamorphic host rocks but are typically depleted in most hydrothermal fluids (Corfu and Davis 1991; Jamtveit and Hervig 1994). Under such rock-buffered conditions, pore fluid may enhance local element mobility, promoting diffusion of e.g., Al, Mg and HFSE like Zr or Ti during skarn formation (Burt 1981; Capitani and Mellini 2000; Chu et al. 2023; Jamtveit and Hervig 1994). The absence of a significant hydrothermal input from a fertile (e.g. Sn, W, Li or In-rich) magmatic source is furthermore supported by invariably low concentrations of such metals in the calc-silicates of stage I – in combination with the absence of cassiterite, malayaite or other distinct ore minerals.

The REE pattern of Garnet Ia is strongly depleted in LREEs and enriched in HREEs, which is typical for slow growth rates that promote trace element incorporation controlled by the crystal structure (Røhr et al. 2007; Yang et al. 2013). Divalent Eu is considered to be the stable and predominant species at temperatures higher than 250 °C (Bau 1996; Sverjensky 1984), which is the case in stage I at Geyer. The negative Eu values may also be inherited from the host rock. Both might explain the deviation of Eu from the normal distribution pattern of REEs in minerals. Major and trace element systematics of garnet Ib mark a slight but distinct change towards less depleted LREE and less enriched HREE compositions as well as a less pronounced negative Eu anomaly. This is tentatively interpreted to reflect the influence of at least some magmatic-hydrothermal fluid and thus a higher fluid/rock ratio during the waning of stage I.

Substage IIA: oscillations between rock and fluid buffering

Skarn stage II comprises four garnet generations which range in textures from paragenetically early fine-grained poikilitic (Grt IIa and c) garnet to late coarse-grained euhedral garnet with prominent oscillatory zoning (Grt IIb and d). Consistent with its poikilitic texture, the first generation of garnet (Grt IIa) is grossular-rich, high in HFSE, depleted in LREE, and enriched in HREE suggesting growth under low fluid/rock ratios analogous to garnet related to stage I. However, Grt IIa is marked by elevated concentrations of Sn, Li, and W, which may indicate some involvement of a magmatic-hydrothermal fluid endowed in these elements.

The textures of Grt IIb and associated vesuvianite (Ves II) and amphibole (Amp IIa) mark a distinct change in mineral and skarn textures as they form coarser grained euhedral crystals with well-developed oscillatory zoning. Such textures generally indicate a progressive increase in fluid flow and permeability from fluid migration along interconnected pore space to fractures and veins (Gaspar et al. 2008; Huang et al. 2022; Jamtveit et al. 1993; Jamtveit and Hervig 1994; Meinert 1992; Park et al. 2017a). Concomitant to the textural change, Grt IIb is andradite-rich, enriched in LREE relative to HREEs and has elevated concentrations of Li, W, and In. All of this suggests fluid-buffered conditions with the involvement of a magmatic-hydrothermal fluid (Ayers and Eggler 1995; Bai and Koster van Groos 1999; Douville et al. 1999; Jamtveit et al. 1993; Wood 1990). This notion is further supported by the first occurrence of magnetite in the skarn assemblage, typically requiring externally derived Fe from a magmatic-hydrothermal fluid source (Meinert 1992). Further evidence for an influx of a magmatic-hydrothermal fluid is the high F concentrations of amphibole IIa (intergrown with Grt IIb), as fertile Sn granites in the Erzgebirge have invariably high concentrations of F (Förster et al. 1999; Lehmann 2021). High F activities may also promote precipitation of amphibole IIa early in the prograde skarn stage, whereas it is usually related to the retrograde alteration stage (Dobson 1982; Meinert et al. 2005).

The positive Eu-anomaly in Grt IIb marks a sharp contrast to previous garnet generations that are all marked by negative Eu-anomalies. This anomaly is most likely related to a new batch of magmatic-hydrothermal fluid without buffering by the host rock (Smith et al. 2004). The observation that prograde garnet formed in a fluid-buffered environment shows a positive Eu-anomaly, is widespread in literature (e.g.Chen et al. 2022; Gaspar et al. 2008; Jamtveit and Hervig 1994; Park et al. 2017a, b; Smith et al. 2004) but has not been explained so far. As crystal chemical effects can be ruled out (else, Grt Ib should have a positive Eu anomaly), we suggest that the magmatic-hydrothermal fluid probably itself had a positive Eu anomaly at this stage. This could be due to its residual character after extensive crystallization of minerals with negative Eu anomalies (Breiter et al. 1999; Terekhov and Shcherbakova 2006) or the positive Eu-signature was obtained by dissolution of plagioclase in the rock units it percolated through (Möller et al. 2021). However, positive Eu anomalies were identified in whole-rock data of apical parts of the Ehrenfriedersdorf deposit adjacent to Geyer SW (Lehmann and Seltmann 1995) and in highly evolved, peraluminous granites in France (Raimbault and Azencott 1987). This supports the interpretation that the origin of these garnets displaying a distinct positive Eu anomaly is genetically associated with a magmatic fluid from a highly evolved source.

Following the fluid-dominated formation of garnet IIb, the system evolved towards an intermediate stage, which shows ingress of magmatic fluid but also rock buffering. Under these conditions, Grt IIc formed again with poikilitic texture. Grt IIc is intergrown with fluorite, malayaite and remnant clinopyroxene. The following mineral chemical arguments support an increased importance of rock-buffering during the formation of Grt IIc: higher contents of fluid-immobile elements (Al, Ti and Zr) in garnet, and a REE pattern that is less depleted in HREE. The REE pattern is governed by garnet crystal chemistry (Marks et al. 2008), which indicates the amount of fluid was low enough to allow rock buffering, similar to the observations of Jamtveit and Hervig (1994) and similar to the conditions of Grt IIa growth, which is exemplified by the similar, slightly negative Eu anomaly (Fig. 9A). The fact that the garnets have elevated contents of W and In indicates, however, shows that a magmatic fluid was still present at this stage.

This intermediate stage was followed by the most fluid-dominated stage, in which Grt IId grew with an andradite-rich composition, the highest concentrations of Sn, Li, W and In, high LREE and low HREE. This type of REE pattern and the again strongly positive Eu anomaly (even stronger that Grt IIb; see Fig. 9A) support the fluid-dominated nature of the system at this stage (Gaspar et al. 2008; Jamtveit and Hervig 1994).

Lanthanum concentrations in Grt IId are ten time higher than in Grt IIc and five times higher than in Grt IIb, with overall similar REE patterns in all three garnet generations. This increase in La may be related to higher fluid salinities, as high chlorinity increases La solubility (Allen and Seyfried 2005). This is supported by fluid inclusion observations (Meyer et al. 2024). Thus, we propose that the high La concentrations and the tetrad effect visible in measurements of garnet and malayaite (Fig. 10A) are related to phase separation, salinity increase and REE fractionation of the magmatic-hydrothermal fluid (Flynn and Burnham 1978; Bai and Koster van Groos 1999; Gaspar et al. 2008; Lehmann 2021; Monecke et al. 2011).

Substage IIB: a fluid-buffered finish

Both epidote II and amphibole IIb of substage IIB precipitated from a cooling magmatic-hydrothermal fluid, indicated by the still strong magmatic-hydrothermal REE signature (slightly positive Eu anomaly and a small tetrad effect; see Fig. 10B). Similar epidote textures have been described in the skarn system of in Petrovitsa, Bulgaria, by Hantsche et al. (2021).

Chlorite commonly replaces both clinopyroxene and amphibole during the waning phase of substage IIB, suggesting further cooling of the system (Meinert 1992). Chlorite shows partly elevated concentrations of Li, B and Zn, indicating that it also likely precipitated in the presence of a magmatic-hydrothermal fluid, albeit parts of these traces may originate from the precursor minerals. The most common mineral accompanying chlorite is cassiterite, both in veins and alteration halos. The concentration of trace elements in cassiterite is strongly dependent on crystallographic properties (Schneider et al. 1978). The strong enrichment of In and W suggests a strong influence of a magmatic-hydrothermal fluid. This is supported by the obvious third tetrad anomaly, which is considered to be a primary feature of a magmatic-hydrothermal fluid from a highly evolved granite (Bau 1996; Irber 1999; Förster and Tischendorf 1994; Gemmrich et al. 2021; Monecke et al. 2011).

Magmatic vs. meteoric fluids

Apart from isotopic signatures (which are beyond the scope of this study), the Y/Ho ratio is arguably the best parameter to constrain the source of a hydrothermal fluid from whole rock compositions (Bau and Dulski 1995; Bau 1996) or mineral chemistry (Chu et al. 2023) in a fluid-dominated hydrothermal system. Neither fractional crystallization of a granitic system nor modifications of the fluid like boiling or cooling lead to fractionation between Y and Ho (Bau and Dulski 1995). Variations of the Y/Ho ratio have thus been attributed to the dilution of a magmatic-hydrothermal by a meteoric fluid (Bau 1996).

Figure 12 shows that all garnet generations, both epidote generations and the first two amphibole generations (Amp IIa and IIb) show constant Y/Ho ratios over a wide range of concentrations. This suggests that all these minerals formed from a magmatic fluid of constant Y/Ho composition (Ayers and Eggler 1995; Bai and Koster van Groos 1999; Bau 1996; Park et al. 2017b). However, amphiboles that coexist with sulfides (Amp IIc) show a deviation in the Y/Ho ratio (Fig. 13F) and a different REE distribution (Fig. 10B), suggesting the influx of meteoric water (Bau and Dulski 1995). This is consistent with the abundance of sulfides in this assemblage, which is a typical response to enhanced influx of meteoric water into a retrograde skarn system (Gao et al. 2020; Meinert et al. 2003; Öztürk et al. 2008). Later phases (chlorite and cassiterite) probably formed from a mixture of magmatic and meteoric fluids as prominently shown in the Y/Ho systematics of cassiterite (Fig. 11E).

The fate of tin

Skarn stage I at Geyer SW lacks discrete tin minerals and maximum concentrations of Sn in calc-silicates do not exceed 900 µg/g (Table 1). This leads to two important conjectures, namely (a) that the precursor rock of the skarnoid did not contain significant concentrations of Sn and (b) that the skarnoid formed under rock-buffered conditions as stipulated above.

Conversely, the major skarn minerals malayaite-titanite, garnet, vesuvianite, and amphibole related to stage II contain considerable amounts of Sn (Table 1), of which garnet can be considered as the main host of the bulk Sn due to its high abundance and high Sn contents (Table 1).

Mineral chemistry data (Table 1) document clearly that during stage II, the influx of Sn-bearing magmatic-hydrothermal fluids progressively increased, which resulted in an overall trend of increasing Sn concentrations in skarn minerals and also the occurrence of malayaite. The observed increase in Sn concentration in calc-silicates obviously coincides with high fluid/rock ratios and a fluid-dominated system of skarn mineral formation (Table 1).

Table 1 Average and maximum concentrations of Sn and F and the ratio between the LREEs (La–Nd) and HREEs (Er–Lu)

During the waning of stage II (substage IIB), cassiterite, epidote and amphibole are the main hosts of Sn (Table 1). Cooling of the system displayed by the stabilization of amphibole and chlorite (Meinert 1992), and the dilution of the magmatic fluid with meteoric water, indicated by the diverging Y/Ho ratio in cassiterite, are most likely reason for cassiterite and sulfide precipitation.

The mineral chemical data of all analyzed silicates in Geyer suggests an incorporation of Sn⁴⁺ via isovalent substitution (titanite-malayaite; Deer et al. 2013) or coupled substitution with Fe²⁺ Mg, and Mn²⁺, as it was shown for garnet (Amthauer et al. 1976; Chen et al. 1992; Plimer 1984; Yu et al. 2020) and epidote (Armbruster et al. 2006; Van Marcke de Lummen 1986). This indicates, in contrast to common belief and in accordance with recent experiments (Chou et al. 2021; Schmidt 2018; Wang et al. 2021) as well as with thermodynamic modelling (Liu et al. 2023), that a considerable amount of tin can be transported as Sn⁴⁺-complexes. An oxidized environment during the formation of stanniferous calc-silicates is also supported by the high concentration of Fe³⁺ compared to reduced Fe²⁺ in minerals such as andradite and epidote of stage II. Following this argumentation, an external oxidizing agent is not necessary to either incorporate tin into silicates or to precipitate cassiterite – oxidizing of the system by temperature decrease and fluid mixing are sufficient (see also Schmidt 2018).

During the evolution of the skarn system, two different stages of tin remobilization can be distinguished. The first one is associated with substage IIA, in which Al-rich garnet (garnet IIc), replaces stanniferous calc-silicates. Due to the lower ability of grossular-rich garnet to incorporate Sn (Sonnet and Verkaeren 1989; Park et al. 2017b; Plimer 1984), malayaite forms as a Sn-rich mineral. Relatively high temperatures and a high silica activity, which is assumed at this stage (Meyer et al. 2024), stabilized malayaite rather than cassiterite (Aleksandrov and Troneva 2007). The second remobilization of Sn occurred during the waning of substage IIB, when most skarn minerals were replaced by hydrothermal calcite, quartz and prehnite, implying a further cooling of the system. During this process, cassiterite is replaced by stokesite. A similar late-stage occurrence of stokesite has been described by Kern et al. (2019) for the Hämmerlein skarn.

Conclusions

Detailed petrography, microthermometry, and precise geochronology are sufficient to characterize the general structure of a Sn skarn system. However, the integration of major, minor and trace elements (especially REEs) of relevant calc-silicate minerals is necessary to understand the details of the physicochemical evolution of skarn systems. Only a combination of these observations allows the establishment of a concise mineral system model.

The present contribution has combined careful petrography with major, minor and trace element compositions of important rock-forming and ore minerals of the Geyer SW skarn to show the importance of such studies in a textbook example. Importantly, the oscillation between fluid-buffered and rock-buffered conditions during skarn mineralization and alteration governs the texture of minerals and determines their chemical composition. Changes in the fluid/rock ratio directly respond to discontinuous fluid pulses exiting the intrusion at depth. The invariably high concentrations of Sn, In, and W in minerals which formed under high fluid-rock ratios strongly supports that these elements were introduced by the magmatic-hydrothermal fluid rather than being remobilized from the surrounding metasediments.

This investigation and comparison with other Variscan tin skarn data showed that the metal classification of Variscan Sn-Zn skarns solely based on the major mineral chemistry of garnet and clinopyroxene is inadequate and does not help to constrain the dominant metal tenor of a skarn system; therefore, it is of limited use as an exploration vector.