Diagenetic History and Timing of Cu and Zn-Pb Sulfide Mineralization in the Permian Kupferschiefer System, Saale Subbasin, Eastern Germany

Stratigraphic redox boundaries play a key role in the formation of a number of sediment-hosted stratiform to stratabound massive sulfide deposits (Hitzman et al., 2010). This is primarily because base metal (Cu, Pb, and Zn) solubilities are enhanced under oxidizing conditions and greatly lowered under reducing conditions (Seward and Barnes, 1997). The sediment-hosted stratiform Cu (SSC) deposits at the southern margin of the Southern Permian basin in Germany and Poland represent some of the world’s most spectacular examples of sulfide mineralization associated with a stratigraphic redox boundary. This stratigraphic redox boundary is regionally extensive, extending from the United Kingdom to Poland, and records a transition from the terrestrially deposited Rotliegend sandstone (S1) to the organic-rich marine Kupferschiefer mudstone (T1). The T1 is overlain by the Zechstein limestone (Ca1). The highest Cu grades are in the T1, although sulfidemineralized rocks are also found in the S1 and hanging-wall units of the lowermost Ca1 and overlying anhydrite unit (A1) (Borg et al., 2012).

Base metal-mineralized rocks in the Kupferschiefer (T1) and adjacent stratigraphic units are generally believed to have formed when slightly acidic, oxidizing, metal-bearing fluids encountered the redox boundary between the S1 and T1 and mixed with slightly alkaline, saline, and reduced sulfurbearing pore fluids in the host rocks (e.g., Bechtel and Püttmann, 1991; Wedepohl and Rentzsch, 2006). One of the characteristic features associated with the high-grade ore is the “Rote Fäule,” a term given to epigenetic hematitic alteration (e.g., Piestrzyński et al., 2002; Pieczonka et al., 2008; Borg et al., 2012). The Rote Fäule was initially interpreted to be an oxidized shallow-water facies that was equivalent to the reduced Cu-bearing T1, with the contrasting redox potentials reflecting the different sedimentary and diagenetic conditions (e.g., Franz, 1965; Jung and Knitzschke, 1976). However, subsequent studies have described the Rote Fäule as crosscutting the stratigraphy, and the general consensus is that it has a postdepositional origin (Rydzewski, 1978; Oszczepalski, 1989).

Precise constraints on the timing of base metal sulfide formation in the Kupferschiefer district remain controversial. Several studies have suggested a syngenetic (e.g., Wedepohl, 1971) to early diagenetic (e.g., Wedepohl and Rentzsch, 2006) timing, i.e., base metal sulfides precipitated before major compaction and cementation. The preservation of negative δ³⁴S values (–40 to –25‰) in ore-stage sulfides has been used as evidence that reduced sulfur was generated via bacterial sulfate reduction (BSR) in an open system, during either syngenetic or early diagenetic mineralization (e.g., Sawlowicz, 1989). However, it has also been suggested that ore-stage sulfides could have replaced diagenetic pyrite and inherited the negative δ³⁴S values (Sun and Püttmann, 1997; Bechtel et al., 2001). Other studies have also proposed a late diagenetic timing of the mineralization, during tectonic hydrofracturing triggered by Triassic rifting (Jowett, 1986, 1987; Cathles et al., 1993). In these models, the reduced sulfur for the sulfide mineralization was derived from thermochemical sulfate reduction (TSR) (Heppenheimer et al., 1995; Sun and Püttmann, 2000). The preservation of higher δ³⁴S values (up to –8.4‰) in Cu sulfides from high-grade mineralized samples has also been linked with reduced sulfur derived via TSR at low temperatures (<130°C) during later diagenesis (Bechtel et al., 2001). These different models most likely reflect several stages of mineralization, from the diagenetic evolution of the Kupferschiefer in the late Permian to at least 100 m.y. after (Alderton et al., 2016).

The metal trap in sediment-hosted deposits may be influenced by a number of factors, including the depositional environment, diagenesis, and hydrothermal processes. In order to better understand how these different factors interact, it is necessary to generate different types of petrographic and geochemical data. For example, total organic carbon and bulk compositional data can be used to interpret the depositional redox conditions, which have a strong influence on organic matter preservation and sulfate reduction (Böning et al., 2012; Greenwood et al., 2013; Reinhard et al., 2014; Little et al., 2015; Magnall et al., 2018). Quantitative mineralogical data provide important constraints on the primary mineralogy of the host rocks, which is a first-order control on early (at nearsurface, lower temperatures <50°C) and burial (higher depths and temperatures >50°–300°C) diagenetic processes (Morad et al., 2000). In fine-grained rock types, it may be necessary to combine standard petrographic techniques (transmitted and reflected light) with high-resolution scanning electron microscopy (SEM) to establish the relative timing (paragenesis) of diagenetic and hydrothermal mineral assemblages.

This study investigates samples from the S1, T1, and Ca1 in three drill cores (Sangerhausen, Allstedt, and Wallendorf) in the Saale subbasin (Eastern Germany). Different petrographic (transmitted and reflected light, SEM), mineralogical (quantitative X-ray diffraction [QXRD]) and geochemical (X-ray fluorescence [XRF] and inductively coupled plasma-mass spectrometry [ICP-MS]) data sets have been generated to investigate which aspects of the host rock determined the style of mineralization in the different units and localities. This requires consideration of (1) the evolution of paleoenvironmental conditions during the late Permian marine transgression, (2) how this influenced the diagenetic evolution of the S1, T1, and overlying Ca1, and (3) the relative timing and style of sulfide mineralization with respect to different diagenetic processes.

The Southern Permian basin is part of the intracontinental Central European basin system, which formed in the late Carboniferous to early Permian, contemporaneous with the assembly of the Pangea supercontinent (Plein, 1990; Ziegler, 1990; Fig. 1A). The crust beneath the Southern Permian basin comprises a complex assemblage of orogenic terranes that were accreted to Baltica and the East European craton during the Caledonian and Variscan orogenies (Ziegler, 1990). The varying crustal rheologies had an important role for stress and strain localization during the post-Variscan evolution of the Southern Permian basin (Maystrenko et al., 2008). The development of intracontinental basins during the latest Carboniferous-earliest Permian followed the end of the Variscan orogeny, coinciding with wrench faulting, thermal thinning of the lithosphere, and widespread magmatism (Ziegler, 1990; Wilson et al., 2004; Breitkreuz et al., 2008; Pharaoh et al., 2010).

The early Permian Rotliegend Group overlies the crystalline basement and Carboniferous (Westphalian-Stephanian) formations in the Southern Permian basin. Deposition of the Lower Rotliegend and Upper Rotliegend I was associated with bimodal volcanism in transtensive pull-apart basins (Stollhofen et al., 2008). The mafic and felsic volcanic rocks in the Lower Rotliegend are found in most parts of the Southern Permian basin but are particularly voluminous in the eastern part of the North German basin and Polish basin (Fig. 1A). The Upper Rotliegend II, comprising thick continental siliciclastics (~2,500 m thick) and evaporites, was influenced by tectonics and climatic fluctuations and also short-term marine ingressions (Legler and Schneider, 2008). It is dominated by ephemeral fluvial (wadi), sabkha, lacustrine, and aeolian successions (Stollhofen et al., 2008). Weissliegend and Grauliegend are used to describe local color variations in the uppermost part of the Upper Rotliegend II, which are composed of white to gray aeolian, fluvial, and locally marine sandstones (Ehling et al., 2008). The color variations are the result of chemical reduction during diagenesis, meaning the Weissliegend and Grauliegend are not strictly chronostratigraphic units (Borg et al., 2012).

During the late Wuchiapingian (Lopingian, ~258 Ma; Brauns et al., 2003), the Permian intracontinental basins were flooded by a marine transgression from the Boreal Ocean (Smith, 1979; Ziegler, 1990; Legler and Schneider, 2008). Besides the local occurrence of the so-called “Zechstein conglomerate” as a reworked horizon and a basal carbonate layer, the marine transgression resulted in the deposition of the T1, Ca1, and Werra anhydrite (A1) of the Werra Formation (Fig. 1B, C). The T1 is composed of three subunits, where organic matter is highest in the lower T1 (T1-L) and carbonate increases upward into the middle and upper T1(T1-M and T1-U) (Grice et al., 1997). The T1 has been dated at 257.3 ± 1.6 Ma using the Re-Os system (Brauns et al., 2003), meaning deposition during the Wuchiapingian (Lopingian). This is supported by the depositional age (258 Ma) for the top of the underlying Elbe Group (Menning et al., 2006). This age is generally in agreement with a less precise Re-Os isochron age of 247 ± 20 Ma of Pašava et al. (2010) for unmineralized T1 samples from the Zdrada area in northern Poland. These ages also overlap with K-Ar ages of 277 to 249 Ma for authigenic illite from unmineralized T1 samples in northwest Poland (Bechtel et al., 1999).

The SW-NE–trending Saale subbasin is approximately 150 × 90 km and located to the southeast of the Harz mountains and northeast of the Thuringian subbasin (Fig. 1C). The crystalline basement is overlain by Visean to Westphalian sedimentary rocks (Schneider and Romer, 2010). The Stephanian Mansfeld Group is composed of (1) the basal fine-grained coal-bearing sedimentary rocks of the Grillenberg subformation and coarse-grained sedimentary rocks of the Gorenzen Formation, (2) unconformably overlying thick red beds of the Rothenburg Formation, and (3) braided river deposits of the Siebigerode Formation (Schneider et al., 1995; Schneider and Romer, 2010; Gebhardt and Hiete, 2013).

The overlying early Permian Rotliegend Group, cropping out in most marginal parts of the Saale subbasin, comprises chert-quartzite-conglomerates interbedded with lacustrine black shales of the Halle Formation with silica-rich volcanic rocks as sills and laccoliths (Fig. 1D). The Halle Formation is overlain by medium- to coarse-grained sandstones of the Hornburg Formation and braided river sandstones of the Eisleben Formation (Ehling et al., 2008). The T1 in the Saale subbasin is a thin (0.3–0.6 m), organic matter-rich mudstone unit and was historically mined for copper and other metals in the Sangerhausen and Mansfeld mining districts. The different districts in the Saale subbasin are separated by structural highs (e.g., Hornburger anticline) and major faults (e.g., Hornburg-Tiefen fault) (Fig. 1C). In addition, the Saale subbasin has also been affected by several northwest-southeast minor faults.

The Kupferschiefer system in Germany and Poland hosts the world’s second-largest SSC district, with combined resources of >60 Mt of contained copper (Borg et al., 2012). Copper and associated metals have historically been mined in different mining districts and are being actively mined in the Sieroszowice, Rudna, and Lubin districts in Poland (Oszczepalski et al., 2019).

The sulfide mineralization in the Kupferschiefer system is laterally extensive in the S1, T1, and Ca1 and is found locally in the Werra anhydrite A1. Different styles of sulfide mineralization have been described, including disseminated, vein-type, and detrital clast and fossil replacement sulfides (see summary by Borg et al., 2012). Metal zonation has been described as having a lateral component in the German Kupferschiefer districts and an oblique vertical component in the Polish Kupferschiefer districts (Borg et al., 2012). In either case, the zonation is characterized by (1) a Cu zone, adjacent to the Rote Fäule alteration, comprising chalcocite, bornite, chalcopyrite, and some other minor Cu sulfides, (2) an overlapping Zn-Pb zone, and (3) a noneconomic pyrite zone (Borg et al., 2012). The principal sources of copper in the Kupferschiefer system are unknown, but could be either the mafic volcanic rocks in the Rotliegend or the basement rocks of the Mid-European Crystalline High (Borg et al., 2012). Lead isotope data of galena ± sphalerite ± pyrite ores and vein galena from different parts of the Southern Permian basin suggest the basement rocks are a major source of Pb for the mineralization (Wedepohl et al., 1978).

In terms of geochronology, indirect paleomagnetic age data of hematite in the Polish basin suggest a Triassic age (250–220 Ma) of the Rote Fäule (Jowett et al., 1987). This paleomagnetic age is bracketed by the K-Ar ages of illites (~250 Ma) in mineralized samples close to the Rote Fäule (Bechtel et al., 1996) and by the estimated age of diagenetic illite (216–190 Ma) in the mineralized T1 samples in the Polish basin (Bechtel et al., 1999). A range of Re-Os ages have been derived for mineralized samples from the Polish basin, which correspond with different mineralization styles (Alderton et al., 2016). The Re-Os ages indicate that the fine-grained sandstone infill and massive ore that formed during the late Permian-Early Triassic and younger ages (Early to Middle Jurassic) correspond with sulfide veinlets. The paleomagnetic age data of hematite in the Rote Fäule in the Sangerhausen mining district (Eastern Germany) also suggest two distinct ages (Symons et al., 2011): 149 ± 3 Ma (Late Jurassic), which probably represent the disseminated, high-grade, Cu-Zn-Pb mineralization, and 53 ± 3 Ma (early Eocene) for the vein mineralization (Borg, 2017). The age discrepancy in the two basins could be related to either a local, younger postdiagenetic mineralizing event, such as structurally controlled (Rücken-type veins) mineralization associated with Carpathian-Alpine tectonism characterized by the presence of cobalt, nickel, and silver arsenide minerals (Wagner and Lorenz, 2002; Zientek et al., 2015) or a local fluid-flow event unrelated to the Cu-mineralizing system (e.g., Wilkinson et al., 2017).

Samples were selected from three different drill cores in the Saale subbasin (Eastern Germany; Figs. 1C, 2). Drilling took place in the 1980s for mineral exploration. The samples were selected from the Sangerhausen 102/80 (depth: 429.6–430.6 m), Allstedt 11/86 (depth: 945.4–946.4 m), and Wallendorf/Luppe 6/84 (depth: 359.1–360.1 m) drill holes. Each drill core contains the uppermost part of the Rotliegend sandstone (S1), Kupferschiefer (T1), and lowermost Zechstein limestone (Ca1) (Fig. 2A-C). These drill holes also intersect different sulfide zones: the Sangerhausen (SHN) drill hole intersects the Rote Fäule and Cu zone, the Allstedt (AST) drill hole intersects the Cu zone and overlapping Zn-Pb zone, and the Wallendorf (WDF) drill hole intersects the Zn-Pb zone (Fig. 2A-C). A total of 47 samples (SHN =17, AST = 17, and WDF = 13) were collected from the three drill cores, each every ~5-cm interval. The samples were split, with one-half prepared into polished thin and thick sections, and the other half powdered for mineralogical and geochemical analyses.

All analyses were carried out at the German Research Centre for Geosciences (GFZ-Potsdam). Hand sample photographs and binocular microscope images were taken to inspect variations in lithology, mineralogy, and organic matter. Standard polished thin sections (~25-mm thickness) and polished thick sections (~2-mm thickness) were prepared for petrography for all samples. Transmitted- and reflected-light petrography was performed on a BX51 polarizing microscope (Olympus Corporation). Organic petrography was performed on a DM4 P microscope (Leica) under an oil immersion objective lens using ultraviolet (UV) light and reflected white light illumination. Backscattered electron (BSE) and secondary electron (SE) imaging together with electron dispersive spectrometry (EDS) semiquantitative compositional analysis was performed using an Ultra Plus field emission scanning electron microscope (Carl Zeiss Microscopy GmbH). The EDS analysis was conducted using an accelerating voltage of 20 kV. False-color elemental composite maps were generated to determine the distribution of the different mineral phases.

Mineralogical and geochemical analyses (except total carbon and total organic carbon content) were performed at the Elements and Minerals of the Earth Laboratory (ElMiE-Lab) at GFZ-Potsdam. Forty-seven bulk-rock samples were crushed and powdered to a grain size of <62 µm. For quantitative mineral analysis, the powdered samples were further milled to a grain size of <10 µm using a micronizing mill (McCrone Group). Analyses were performed with an X-ray diffractometer (Malvern PANalytical) using a Bragg-Brentano geometry at 40 mA and 40 kV with CuKa radiation, and a PIXel3D detector at a step size of 0.013° 2θ from 4.6° to 85° 2θ and 60 s per step. The mineralogy was determined with the software EVA (version 11.0.0.3) by Bruker using an updated mineral library. Rietveld refinement for quantitative mineralogy was performed using the program BGMN (Bergmann et al., 1998) and the open-source software package Profex (version 4.3; Doebelin and Kleeberg, 2015), calibrated for the used diffractometer. The error is in the range of 3%.

Major and minor elements were determined on glass beads on <62-µm powdered sample split and analyzed by XRF using an Axios Advanced spectrometer (Malvern PANalytical). The quantification level was 0.02 wt % for major elements and 10 ppm for minor elements. Loss on ignition (LOI) was measured as weight loss during the ignition of the sample. Total S was measured by the Eltra elemental analyzer (Eltra CS 2000). For trace element analysis, samples were digested using the standard four-acid digestion method (concentrated AR/HF/HClO₄) using closed Savillex beakers on a hot plate at 180°C. After 24 h, the samples were checked for undissolved residue and evaporated to dryness. The digestion procedure was repeated until no residue was visible. Dissolved samples were diluted in 2 vol % HNO₃. All acids were prepared using reagents purified by subboiling distillation and diluted by bidistilled H₂O. Acid blanks of concentrated reagents are routinely tested. Trace elements were analyzed by high-resolution (HR-)ICP-MS using an Elements XR instrument (Thermo Fisher). Drift correction was performed by internal standard and addition of 1 µg/kg In. One procedural blank was included per 18 samples and was generally negligible compared to sample concentrations. Acid blanks are subtracted online. Measurements were repeated two to four times with repeatability generally within 5%.

The total carbon and total organic carbon contents were analyzed with an elemental analyzer (EA Isolink IRMS system, Thermo Fisher). For total carbon determination, about 5 mg of sample material was loaded into Sn capsules and burned in the elemental analyzer. Total organic carbon content was determined on in situ decalcified samples (3 mg). To remove the carbonate carbon, samples were loaded in Ag capsules and decalcified in two rounds by 3% HCl and 20% HCl, respectively, and heated for 3 h at 75°C. Both measurements were calibrated against urea and checked with an internal laboratory standard (TSK). The reproducibility for replicate analyses for total carbon and total organic carbon was better than 0.2 wt %.

The relative timing of the different mineral phases formed in each of the major units is presented in a paragenetic chart in Figure 3, and key paragenetic relationships are described below.

Rotliegend sandstones (S1): The analyzed samples comprise fine- to coarse-grained, poorly to moderately sorted, rounded to subrounded and randomly oriented detrital clasts composed of mostly submillimeter-size grains of quartz, feldspar, and lithic/rock fragments (Fig. 4A, B; App. Fig. A1A-D). Some larger (cm size) clasts have been preserved in the Sangerhausen drill core. Monocrystalline quartz clasts dominate over polycrystalline quartz (App. Fig. A1A). Plagioclase (albite) and K-feldspar mainly compose the detrital feldspar fraction. Rock fragments commonly have source-specific textures and compositions that can be used to identify them in thin sections (Ulmer-Scholle et al., 2015). Accordingly, the rock fragments include both felsic and mafic igneous, sedimentary, metasedimentary, and metamorphic rock fragments (App. Fig. A1A-D). The mafic rock fragments are mostly composed of quartz and albite where their groundmasses altered to, or were replaced by, early illite and chlorite (e.g., Fig. 4C, D).

Diagenetic assemblages in the sandstones include carbonate (calcite and minor dolomite) and sulfate (anhydrite and trace barite) cements, and clays (illite, chlorite, and kaolinite) (Figs. 4–6). The carbonate cement in the sandstones occurs as extensive, poikilotopic, intergranular calcite cement that forms overgrowths on minor rhombic dolomite (~5–10 μm) and also ingresses into intragranular pores (Fig. 4D-H). In samples from the Sangerhausen drill core, the calcite cement is zoned, and certain domains contain bladed hematite (Figs. 4F, H, 5B), producing the red color at hand-specimen scale (Fig. 2A). The hematite intergrowth mostly occurs in the outer parts of the intergranular calcite cement (Figs. 4F, H, 5B, C) but is also present as overgrowths on euhedral dolomite (Fig. 5C). The intergranular pore-filling calcite shows dissolution textures in all drill cores (Fig. 5B, C). These sandstones also contain poikilotopic anhydrite and trace barite cements (Figs. 4E, F; App. Fig. A2D). Anhydrite cement is mostly cogenetic with calcite cement and is more abundant in the Sangerhausen than in the Allstedt and Wallendorf drill core samples (Fig. 4E, F).

Mafic volcanic rock fragments have been cemented by calcite (± anhydrite), and their groundmasses are altered to or replaced by illite and chlorite (Figs. 4B-D; App. Fig. A2A, B). These altered clasts sometimes have serrated edges that allow inward growth of intergranular calcite cement (Fig. 4C). This interlayered illite with minor chlorite also occurs rarely in the intergranular pores (e.g., Fig. 5C). Rounded detrital albite clasts have been also replaced by calcite and minor anhydrite (± barite) cement intergrown with kaolinite (Fig. 5D, E). Kaolinite is only preserved in these assemblages in the intragranular pores (Fig. 5E, F). Detrital feldspar and quartz clasts have grain boundaries that are etched and corroded, and the released ions precipitated to intergranular late, meshwork illite (Fig. 4H; App. Fig. A2C), probably after the dissolution of intergranular calcite cement (Fig. 5B).

The sulfides in the sandstones include (1) accessory bornite and covellite in the Sangerhausen samples (Fig. 4F), (2) bornite and sphalerite in the Allstedt samples (Fig. 6A-F; App. Fig. A2F), and (3) galena and sphalerite with minor chalcopyrite and enargite in the Wallendorf drill core samples (Fig. 6G, H; App. Fig. A2G, H). These sulfides occur predominantly as a replacement of intergranular calcite cement, but also of detrital clasts (Fig. 6A-H; App. Fig. A2E-H). For example, the sulfides may have replaced clasts that have been first replaced by calcite (e.g., Fig. 6C). This is consistent with numerous examples of clasts that have been replaced by calcite (Figs. 4E, 5A, B). No pyrite was observed in the sandstones.

Kupferschiefer (T1): The T1 is composed of shale (lower T1 [T1-L]), carbonate-rich marlstone/mudstone (middle T1 [T1-M]), and mudstone (upper T1 [T1-U]) (Fig. 2A-C). The framework grains in these subunits are mainly clasts of quartz and albite. The matrix is composed of clay minerals, mainly illite and minor chlorite and kaolinite. The lower T1 contains alternating coarse and fine-grained laminations (Figs. 7A-C, 8A-H). The coarser-grained laminations are composed of detrital clasts of quartz (~50 μm) and albite (~30 μm), whereas the fine-grained laminations are dominated by organic matter-rich layers of illite ± chlorite ± kaolinite (Figs. 7A-C, 8A-H). The middle and upper T1 are composed of quartz and albite, cemented by calcite, dolomite, and minor anhydrite (Fig. 9A, B).

Organic petrographic and reflected UV and white (RW) light investigations of samples from the Sangerhausen and Wallendorf drill cores showed different macerals in the T1. These include vitrinite-like macerals, alginite, and inertinite (Fig. 7A, B). Vitrinite-like macerals are very thin (<10 μm) and conformable with clay laminations. They are abundant in the lower T1 in both drill cores (Fig. 7A, B). Framboidal pyrite (~5 μm) is mostly associated with these vitrinite-like macerals and is generally ubiquitous in clay-rich layers (Fig. 7A, B). Alginite with greenish-yellow fluorescent colors is partially degraded in the Sangerhausen drill core samples and better preserved in the Wallendorf drill core samples (Fig. 7C-E). Carbonates have a very high fluorescence and are mostly associated with sulfides both in the T1 and Ca1 (Fig. 7C-E).

Diagenetic assemblages in the T1 include carbonate (calcite and dolomite) and sulfate (anhydrite) cements and illite (Figs. 8A-H, 9A-E). In the lower T1, there is greater dolomite and less calcite cement in the coarser-grained laminations (Fig. 8A, B). In the middle and upper T1, fine-grained calcite (5–20 μm) forms a cement to euhedral to subrounded dolomite (~5 μm) and framework grains. The abundant calcite cement has occluded intergranular pores and is most abundant in the middle T1 (Fig. 9A-C). Trace anhydrite cement is also preserved as intergranular cement (Fig. 9A).

Carbonate cements and feldspars (albite and K-feldspar) have been affected by partial to complete dissolution in the T1, especially in the coarse-grained laminations in the lower T1 (Fig. 8C-F). In the middle and upper T1, calcite cement has been significantly affected by dissolution relative to albite and dolomite (Fig. 9A, B).

Framboidal pyrite is ubiquitous in the T1, especially in the fine-grained laminations in the lower T1 and foraminiferal tests at the boundary between the T1 and Ca1 (App. Fig. A3A-C). The ore-stage sulfides in the T1 have a style of replacement similar to that in the S1. In the lower T1, sulfides replace calcite and dolomite (e.g., bornite in the Sangerhausen, bornite and sphalerite in the Allstedt, and sphalerite and galena in the Wallendorf drill core samples) and, to a lesser degree, detrital clasts in coarse-grained laminations (Fig. 8C-F; App. Fig. A3E, F). In the middle T1 with high-grade Cu and Zn-Pb sulfide assemblages, the sulfides clearly replace the calcite cement (Fig. 9A-E). The post-ore-stage mineralization occurs parallel to bedding or as disseminated pyrite in T1 (e.g., App. Fig. A3D), but the abundance increases upward in each drill core. Furthermore, sphalerite and pyrite (not shown) occur in crosscutting calcite veins mostly in the Wallendorf drill core in the middle T1 (Fig. 9F). The sulfide assemblages in the middle T1 extend to the upper T1 though the type and abundance of sulfides may vary in different drill cores.

Zechstein limestone (Ca1): The Ca1 is largely composed of dolomite and a significant amount of quartz and phyllosilicates (illite and minor chlorite ± kaolinite) (Fig. 10A, B). Both euhedral and subhedral to anhedral (circular) dolomite (~5 μm) have been observed in the Ca1. The dominant intergranular cement in the Ca1 is highly fine grained calcite cement (<5 μm) with minor Fe oxide and anhydrite (Fig. 10A-C). The vuggy pores in calcite cemented dolomite are aligned in specific directions in hand-specimen (cm) scale and mostly filled with illite or sulfides (Figs. 7F, 10C, E). The sulfides in the Ca1, similar to the upper T1, include bornite, covellite, and post-ore-stage pyrite in the Sangerhausen drill core and sphalerite, galena, and post-ore-stage pyrite in the Allstedt and Wallendorf drill cores (e.g., Fig. 10D-F). The edges of bornite grains are altered to covellite in the Sangerhausen drill core samples (Fig. 10D).

Quantitative mineralogical data for the samples from the Sangerhausen, Allstedt, and Wallendorf drill cores are available in a supplementary data file (Mohammedyasin et al., 2022) and summarized in Figure 11A through C. Major mineral phases in the S1, T1, and Ca1 include quartz, plagioclase, K-feldspar, calcite, dolomite, phyllosilicates (muscovite, phlogopite, illite, and chlorite), and pyrite (Fig. 11A-C). Some of the minor to accessory minerals (<1%) confirmed in SEM analysis include biotite, anhydrite, barite, hematite, and Ti dioxide (rutile). The samples have been classified on the basis of quartz + feldspar, phyllosilicate, and carbonate content and subdivided according to lithological units (Fig. 12). The previous sedimentological classification of the T1 into three—T1-I (shale), T1-II (claystone), and T1-III (marlstone)—was generally based on assumptions that carbonate content increases upward. However, the carbonate in the T1 is an early diagenetic phase (i.e., is not primary) and is higher in the middle T1 than in the upper T1 in the drill cores. Hence, we reclassify the T1 mineralogically into three subunits: lower T1 (T1-L, shale), middle T1 (T1-M, mudstone to marlstone), and upper T1 (T1-U, mudstone) (Fig. 12). The QXRD data of the T1 samples from this study are in a good agreement with the T1 samples from the Mansfeld, Wettelrode, Polkowice, Rudna, and Lubin Kupferschiefer districts (Rahfeld et al., 2018; Fig. 12).

The quartz and feldspar content in the S1 samples is slightly lower in the Sangerhausen (mean = 57%) than in the Allstedt (73%) and Wallendorf (68%) drill cores (Fig. 11A-C), whereas the carbonate content in the Wallendorf samples (10%) is lower than in the Sangerhausen (21%) and Allstedt (18%) samples. The phyllosilicate content in the Allstedt samples (8%) is lower than in the Sangerhausen (21%) and Wallendorf (22%) samples. Minor bornite and accessory sphalerite were observed in the S1 samples (AST3 and WDF4) adjacent to the T1 from the Allstedt and Wallendorf drill cores. Furthermore, rare anhydrite occurs only in the Sangerhausen samples (max. 3%).

Samples from the lower T1 contain higher phyllosilicate (mean = 46%) and lower carbonate (12%) contents. The quartz + feldspar and phyllosilicate contents in the middle (T1-M, 31%, 31%) and upper (T1-U, 32%, 33%) T1, respectively, are generally proportional. However, the total carbonate content is slightly higher in the middle T1 (32%) than in the upper T1 (27%). There is a notable difference in carbonate mineralogy (calcite and dolomite) between drill cores. In samples from the Sangerhausen and Allstedt, calcite is dominant in the S1 and dolomite increases upward into the T1 (Fig. 11A, B; App. Fig. A4A, B). In contrast, in samples from the Wallendorf drill core, calcite is the dominant carbonate phase throughout the S1 and T1 (Fig. 11C; App. Fig. A4C). Sphalerite was quantified in the Allstedt and Wallendorf drill core samples, and bornite (AST6) and galena (WDF6 and WDF10) only in a few Allstedt and Wallendorf drill core samples. Pyrite is abundant in the lower and upper T1 in all drill cores (Fig. 11A-C; App. Fig. A4A-C). Samples from the Ca1 comprise high quartz, feldspar, and phyllosilicate contents (combined mean ~50%), which are proportional to carbonates.

The bulk-rock major and trace element composition of all samples from the Sangerhausen, Allstedt, and Wallendorf drill cores is available in a supplementary data file (Mohammedyasin et al., 2022).

Trends in major elements and base metal composition in the Rotliegend sandstone (S1): The SiO₂ concentration is slightly lower in the Sangerhausen (range = 50–65.5 wt %, mean = 58.9 wt %) compared to Allstedt (60–66.7 wt %, 64.2 wt %) and Wallendorf (59.2–67.8 wt %, 64.1 wt %) drill core samples. The MgO concentration is slightly higher in the Sangerhausen (1–3.5 wt %, 1.9 wt %) compared to the Allstedt (0.9– 1 wt %, 0.9 wt %) and Wallendorf (1.1–1.5 wt %, 1.3 wt %) drill core samples. The CaO concentration in the Sangerhausen (5.4–19 wt %, 12.2 wt %) and Allstedt (10.4–12.9 wt %, 11.8 wt %) drill core samples is nearly twice as high as the Wallendorf drill core samples (4.5–9.4 wt %, 6.5 wt %). The MnO concentration is similar in the Sangerhausen (0.2–0.7 wt %, 0.4 wt %) and Allstedt (0.3–0.4 wt %, 0.4 wt %) drill cores but is lower in the Wallendorf drill core samples (0.1–0.3 wt %, 0.2 wt %). The Fe₂O₃ concentration in the Sangerhausen (0.8–2.1 wt %, 1.4 wt %), Allstedt (1.2–1.6 wt %, 1.4 wt %), and Wallendorf drill core samples (1.6–2.2 wt %, 1.9 wt %) is comparable. The concentration of Al₂O₃ is slightly lower in the Sangerhausen and Wallendorf than in the Allstedt drill core samples (Fig. 13A, B). The minor oxides (TiO₂, mean ~0.4 wt %; P₂O₅, mean 0.1 wt %) have similar concentrations in all drill cores. The Al₂O₃ content is strongly correlated in all drill cores with TiO₂ or K₂O but not with Na₂O (Fig. 13A, C, D). However, the Al₂O₃ content is negatively correlated with CaO/TiO₂ (r²= –0.85; Fig. 13B). The concentration of base metals (Cu, Zn, and Pb) in the sandstone samples is generally very low, except for those samples adjacent to the T1 (Cu = 1% in SHN7, Cu = 0.9 wt % in AST3, and Cu = 0.7 wt % and Pb = 0.6 wt % in WDF4; Fig. 11A-C).

Trends in major elements and base metal composition in the Kupferschiefer (T1): The T1 samples preserve a very strong correlation between Al₂O₃ and TiO₂ (r² = 0.99; Fig. 13A). The lower T1 has a low CaO content in all drill cores. The highest CaO (19.1 wt % in WDF7) and MgO (7.5 wt % in SHN11) concentrations are in the middle T1 samples, but upper T1 samples from Allstedt (8 wt % in AST13) have similarly high values. All samples have a strong negative correlation between the CaO/TiO₂ ratio and Al₂O₃ (r² = –0.85; Fig. 13B).

Lower base metal concentrations are observed in the lowermost T1 samples in all drill cores, except samples with abundant layered sphalerite with pyrite in coarsergrained laminations such as in WDF5 (Zn = 3.1 wt %; Figs. 11, 13E). The entire T1 in the Sangerhausen drill core in the Cu zone has elevated Cu concentrations (max. 1.2 wt % in SHN11; Fig. 11A). The Allstedt samples are Cu rich in the lower T1 (max. 2.4 wt % in AST6), Zn rich in the middle T1 (max. 2.4 wt % in AST9), and Pb rich in the upper T1 (max. 0.7 wt % in AST13), reflecting the vertical sulfide zoning (Fig. 11B). The Wallendorf samples have higher Zn (max. 2.5 wt %) and Pb (max. 0.9 wt %) concentrations, both increasing toward the upper T1, in sample WDF10 (Fig. 11C). However, sample WDF5 in the lower T1 has exceptionally high Zn concentration (3.1 wt %) because of layered sphalerite (Fig. 11C).

The total sulfur, total organic carbon, and total inorganic carbon contents are presented in the supplementary data file (Mohammedyasin et al., 2022). The highest total sulfur and total inorganic carbon content, consistent with the abundant carbonate, is mostly observed in samples from the middle T1 in all drill cores (Fig. 13E, F). The total organic carbon content increases toward the lower T1 in all drill cores where fine-grained laminations are organic matter rich. The Allstedt drill core samples contain higher total organic carbon content (max. 20.3 wt % in AST5) compared to the Wallendorf (12.6 wt % in WDF6) and Sangerhausen (7.2 wt % in SHN9) drill core samples.

Trace element composition: The concentration of trace elements in the S1 is relatively lower and uniform compared to the T1 samples in all drill cores. Only U and Mo in samples from the Allstedt and Ga and Nb in samples from the Wallendorf have higher concentrations.

The concentration of trace elements varies in the different subunits of the T1. This has been further evaluated using the relative change in trace elements (TE_Δ) from the global median black shale (MBS) of all host rock types (Ketris and Yudovich, 2009) calculated by normalizing to Ti [TE_Δ = ((TE/Ti)_sample – (TE/Ti)_MBS)/(TE/Ti)_MBS) × 100]. The MBS is preferred here because other shale normalizations have very different trace element concentrations that may cause different paleoredox interpretations (Slack et al., 2021). Titanium is used here for normalizing because it has a nearly perfect correlation (r² = 0.99) with Al in the T1 samples, and Al was not reported in the MBS database of Ketris and Yudovich (2009).

Most of the trace elements (e.g., Mo, U, V, Ni, As, and Tl) have very large to extreme positive changes in the lower T1, except Mo and Tl in the Sangerhausen drill core (see Table 1). Cobalt and Cd, however, have very large to extreme positive changes in specific sulfide zones. For example, Co in the Cu zone from the Sangerhausen drill core and Cd in the Zn-Pb zone from the Allstedt and Wallendorf drill cores show extreme enrichments (Table 1; Fig. 14A-C). The highest TE_Δ values correspond to the highest base metal concentrations. These enrichments are found mostly in the middle T1, which corresponds with the high-grade sulfide mineralization. In contrast, the Cu zone shows depletion in Cd, i.e., the entire T1 in the Sangerhausen and lower T1 in the Allstedt drill core samples (Fig. 14A). Further investigation on the trace element composition of base metal sulfides is required to accurately constrain a genetic relationship associated with the hydrothermal input.

The relationship between several redox-sensitive trace elements (RSTEs; Mo, U, V, Cr, Ni, and As) and base metals has been tested (cf. Stüeken et al., 2020), and apart from Mo and Cu (Fig. 14D-G) there is no correlation. Rather, the concentration of several RSTEs covaries with total organic carbon content and increases toward the lower T1 (e.g., Fig. 15A-C). The lower T1 samples from the Sangerhausen drill core are depleted only in Mo, but not U (Fig. 15A). The enrichment factor (EF) of selected RSTEs (trace elements = U, Mo, Ni, V, Cr) was calculated relative to Post Archean Australian Shale (PAAS; Taylor and McLennan, 1985) on Al-normalized basis [TE_EF = (TE/Al)_sample/(TE/Al)_PAAS]. Enrichment factors of Mo, U, V, Cr, and Ni generally increase toward the lower T1 (Table 1; Fig. 15). However, Mo_EF and Ni_EF values are exceptionally low in the lower T1 in the Sangerhausen drill core samples (Table 1). The U_EF and Mo_EF values in the middle and upper T1 are similar in the Wallendorf drill core samples (Table 1). Most of the Ca1 samples, however, show Mo depletion relative to T1 samples (Fig. 15A-C).

The petrographic observations together with the mineralogical and geochemical trends have been used to unravel the relative timing and formation processes of diagenetic and hydrothermal mineral assemblages in the S1, T1, and Ca1 units (Fig. 3). These data sets will be combined to investigate the major diagenetic events and style of sulfide replacements in the main ore-bearing units from the Saale subbasin, Eastern Germany.

The primary composition of sandstone is controlled by the regional geologic setting, hydrology, and climate of the source area (Dickinson, 1970; Dickinson and Suczek, 1979; Dickinson et al., 1983). Similar to sandstones deposited in other semiarid to arid environments (Morad et al., 2010), the Rotliegend sandstones are mineralogically immature and characterized by high content of feldspar and rock fragments (Fig. 4A-D). The high proportion of silicate clasts means the Rotliegend sandstones were likely mechanically stable but chemically unstable (Worden and Morad, 2003), consistent with the lack of compacted grains (e.g., Fig. 4A-D).

Calcite (± dolomite) is the most abundant intergranular pore-filling carbonate cement between detrital clasts. The calcite cement likely formed shortly after deposition during the early stages of diagenesis, as it has occluded precompaction intergranular porosity (Worden et al., 2020). The random shape and orientation of clasts further indicates minimal compaction took place before calcite cementation (Fig. 4A-D). This is most clearly highlighted by the preservation of undeformed muscovite grains (e.g., Fig. 5B). The absence of clay rims means that porosity was occluded by the calcite cement before clasts could be altered by undersaturated pore fluids (e.g., Waldmann and Gaupp, 2016).

Following a marine transgression, early diagenetic carbonate cements can often develop in aeolian and fluvial sandstones (Ketzer et al., 2002, 2003b). The diffusion of Ca²⁺ and $HCO3−$ from overlying seawater is commonly enhanced by slow sedimentation rates associated with a sea-level highstand (e.g., Ketzer et al., 2002; El-Ghali et al., 2006, 2013). The patchy anhydrite cement that formed together with the calcite and barite cement may suggest the involvement of a partially evaporated seawater-derived fluid (Fig. 4E, F). Similarly, the strong enrichment of deuterium in n-alkanes (δD values up to –36‰) and isoprenoids (δD values up to –20‰) in T1 samples proximal to the Rote Fäule from the Spremberg area (Eastern Germany) has been used to show interaction between organic matter and evaporitic seawater at temperatures less than 110°C (Poetz et al., 2022). The evaporated seawater could derive from short-lived marine ingressions at the end of Rotliegend deposition (e.g., Stollhofen et al., 2008) or during the more widespread deposition of the overlying Zechstein salt (e.g., Sullivan et al., 1994).

There is a minor component of euhedral dolomite, which is commonly overgrown by calcite (e.g., Fig. 4E, F, H). The dolomite could have formed via a number of pathways—for example, mixing between meteoric, evaporative, and marine waters (Morad et al., 2000)—but it is also possible that it formed via Mg sources from the alteration of mafic rock fragments (e.g., Fig. 4C). The higher dolomite contents in the Sangerhausen and Allstedt samples relative to the Wallendorf samples may indicate differences in the Mg/Ca ratio of seawater between marginal marine and basinal settings, respectively. The precipitation of calcite rather than dolomite can be favored during significant seawater diffusion, which may be controlled by the primary porosity and permeability. The growth of trace hematite on euhedral dolomite shows that hematite formation postdates the euhedral dolomite (probably an initial phase; Fig. 5C). Therefore, the paragenetic relationship between dolomite and hematite suggests that the hematite did not form as the earliest phase i.e., as a detrital grain coating. The intergrowth of most of the hematite within later-formed zones of the calcite cement indicates that hematite mostly postdates the first extensive calcite cement but could overlap or predate the last phase of calcite cementation (Figs. 3, 4E-H, 5B, C). The origin of hematite in the Rote Fäule could be linked to ferrous Fe oxidation that generates acidity (H⁺) that would be buffered by the dissolution of carbonate in the host rocks. The ferrous Fe could be released during the dominant Cu sulfide mineralogy transitions from chalcopyrite to bornite and then to chalcocite (Zhang et al., 2021). Alternatively, the dissolution or breakdown of mafic rock fragments would have released Fe that could have been incorporated into hematite.

The other diagenetic feature in the Rotliegend unit is the alteration of feldspar grains. The alteration of detrital plagioclase follows cleavage planes and twinning surfaces (Fig. 4A), consistent with faster dissolution along weak surfaces due to excess surface energy (Morad et al., 1990). Detrital albite is also sometimes replaced by calcite with minor anhydrite and barite, mostly intergrown with kaolinite (Fig. 5E, F). These alteration assemblages may have originated from feldspar-undersaturated fluids, such as residual meteoric water that was trapped in pores before calcite cementation, or from organic acids that infiltrated from the overlying Kupferschiefer unit. However, several lines of evidence suggest that feldspar alteration mainly postdates carbonate cementation.

1. Dissolution of calcite cement and formation of a serrated dissolution contact with the altered volcanic rock fragment could be explained by local in situ acid generation either released from the overlying organic matter in the T1 or during clay precipitation (green arrow, Fig. 4B).

2. Precompaction dissolution of clasts would lead to greater susceptibility to deformation of clasts during burial, which is not observed, as clasts retain their original shape (Fig. 4A, B, D).

3. Clast dissolution and replacement was restricted to specific clasts under closed-system conditions, i.e., mafic volcanic rock fragments are more susceptible to dissolution and replacement (e.g., Fig. 4B, C).

The precipitation of illite can be explained by [a(K⁺)/a(H⁺)] ratio in the solution during fluid-rock interactions (Lanson et al., 2002). Textural evidence suggested that early illite and chlorite formed early in the diagenetic sequence as a result of alteration or replacement of mafic volcanic rock fragments and that these processes predate the extensive intergranular calcite cementation (Fig. 4B, C). The serrated edges of rock fragments support that the calcite cement postdates the early illite and chlorite formation (e.g., Fig. 4C). The late, intergranular pore-filling illite was formed after the dissolution of intergranular calcite and detrital feldspar clasts (e.g., K-feldspar) (Figs. 4E-H, 5B). Previous petrographic and modeling studies about the diagenesis of the Rotliegend sandstone also showed illite formation as a result of K-feldspar dissolution (Schöner and Gaupp, 2005; Waldmann et al., 2014; Waldmann and Gaupp, 2016). Similarly, the possible ions for late illite precipitation in the studied samples could have been derived from the partial dissolution of K-feldspar.

The deposition of the T1 represents a major change in the depositional paleoenvironment, which involved the flooding of the Permian intracontinental depression and the development of the Kupferschiefer Sea (Rentzsch, 1965; Legler et al., 2005; Fig. 1B). Importantly, the availability of reduced sulfur exerts a strong control on the behavior of a number of RSTEs, which can also be used as proxies for depositional paleoenvironments (Tribovillard et al., 2006; Algeo and Liu, 2020; Bennett and Canfield, 2020). The enrichment of some RSTEs (e.g., Mo, U, V, Ni, and Cr) in the organic matter-rich lower T1 (Fig. 15) is consistent with highly reducing conditions in the bottom waters or pore fluids during sediment deposition. The behavior of Mo is particularly useful; for example, Mo enrichment in modern anoxic systems has been used to distinguish noneuxinic (nonsulfidic, <25 ppm), seasonally/intermittently euxinic (25–100 ppm), and persistently euxinic (sulfidic, i.e., H₂S-bearing, >100 ppm) conditions, respectively (Scott and Lyons, 2012). The Mo concentration of samples from the T1 reflects persistently euxinic (lower T1, Mo > 100 ppm) to seasonally euxinic (middle T1 and upper T1, Mo = 25–100 ppm) bottom water or pore water conditions (Fig. 16).

Covariation between Mo and U is sensitive to both seawater Eh and water mass connectivity (Algeo and Tribovillard, 2009; He et al., 2021). Based on the U_EF and Mo_EF relationship (Fig. 16), the T1 was deposited in anoxic (nonsulfidic) to euxinic conditions, before a transition to suboxic and then to oxic conditions during the deposition of the Ca1. Furthermore, the T1 in the Saale subbasin preserves similar U_EF and Mo_EF systematics as samples from the adjacent Thuringian subbasin (located to the southwest of the Saale subbasin) and were interpreted to have been deposited under open marine conditions (Ruebsam et al., 2017; Fig. 16). The one exception is the group of samples from the Sangerhausen, which preserve relatively lower Mo values in the lower T1 and could be related to remobilization by the oxidizing fluid that produced the Rote Fäule alteration (Fig. 15A). Though the cause for Mo but not U removal in the lower T1 from Sangerhausen is not clear here, one possible explanation is that Mo and U have different redox potentials, so it may be that the fluids were not quite oxidizing enough to mobilize U. Alternatively, the other possible explanation could be that U was locally bounded in clay-organic nanocomposites (Kennedy et al., 2014; Löhr and Kennedy, 2014). It may also be related to biotic processes that are the prominent driver of U reduction in organic-matterrich sediments, and direct enzymatic U microbial reduction could increase the rate of U immobility (U⁶⁺ can be reduced to U⁴⁺ in such reactions) as evidenced in roll-front U deposits (e.g., Min et al., 2005).

Other RSTEs show relative enrichment in specific sulfide zones (e.g., Co in the Cu zone and Cd in the Zn-Pb zone; Fig. 14A-C). This probably shows either mobility of these elements by the mineralizing fluids or, more likely, the addition of these elements from the mineralizing fluids into the system. In contrast, the correlation between other RSTEs (e.g., Mo, U, V, Ni, and Cr) and total organic carbon (Fig. 15A-C) and their lack of covariation with base metals (Cu, Zn, and Pb) (Fig. 14D-G) indicate no significant mobility of these elements by the mineralizing fluids, and they can be used for paleoenvironmental interpretations (Stüeken et al., 2020).

In terms of diagenetic mineral phases, similar to the S1, textural evidence shows calcite and dolomite cements were the earliest phases in the T1. Early diagenetic carbonate cements can be formed under anoxic conditions in marine fine-grained siliciclastic sediments as a by-product of microbial sulfate reduction (Loyd and Smirnoff, 2022). However, understanding the source(s) of these fine-grained carbonate cements will require further investigations using in situ C and O isotope analyses (e.g., Denny et al., 2020; Cui et al., 2021).

Differentiating detrital and authigenic illite in the T1, especially in clay-rich fine-grained laminations, is very challenging. However, illite in carbonate-rich intervals was predominantly formed after the dissolution of carbonates (calcite and dolomite) and clasts (Fig. 9A, B). Similar to the Rotliegend sandstone, the dissolution of K-feldspar likely contributed ions to illite precipitation in the T1. Pore fluids may be another possible source of ions for illite precipitation; this cannot be ruled out from this study.

Our detailed petrographic observations showed both the sulfides and authigenic illite were formed after calcite dissolution, but the exact timing remains uncertain. Previous studies used illite crystallinity and K-Ar geochronology to discriminate between authigenic illite in mineralized and unmineralized zones. In general, illite in the mineralized zone in the Polish basin preserved lower crystallinity (Bechtel et al., 1999). The K-Ar age of the clay-size fraction (<2 μm) of the mineralized samples in this basin is ~250 Ma, whereas the age of diagenetic illite is estimated to range from 216 to 190 Ma (Bechtel et al., 1999). Furthermore, the δ¹⁸O and δD values of illite in the Rote Fäule and Cu zone were used to show the interaction between illite with the mineralizing fluids (Bechtel et al., 2000). For example, the shift in δ¹⁸O (decrease by ~4‰) and δD (increase by ~40‰) values of illite from the Rote Fäule to the Cu zone in the Polish basin and Richelsdorf ore district in Germany have been attributed to fluid-rock interaction with ascending, oxidizing, mineralizing fluids (Bechtel and Hoernes, 1993; Bechtel et al., 2000). These lines of evidence, together with our petrographic observation that sulfides and authigenic illite precipitated after calcite dissolution, suggest that some authigenic illite formation could at least overlap or postdate the mineralization.

The Ca1 shares key diagenetic features with the T1, including early diagenetic carbonate cementation, illite formation, and sulfide precipitation associated with carbonate dissolution (Fig. 10A, B). Similar to the T1, detrital quartz and feldspars are uncompacted and cemented by early diagenetic calcite cement. The calcite cement is restricted to intercrystalline pores without any precursor dolomite fabrics (e.g., Fu et al., 2008). The lack of textural evidence, meteoric diagenetic assemblages at this diagenetic stage (e.g., Al-Hashimi and Hemingway, 1973; Ronchi et al., 2004; Nader et al., 2008), and absence of local anhydrite unit at the Sangerhausen site as a Ca²⁺ fluid source (Schoenherr et al., 2018) imply dedolomitization is not the likely process for the formation of the vuggy pores in the Ca1. Dedolomitization is a slow process that can increase porosity (e.g., Lucia, 1995, 2004) or occlude porosity, depending on the presence (e.g., anhydrite dissolution) or absence of Ca²⁺-rich fluid sources (Escorcia et al., 2013; Schoenherr et al., 2018). In a hand-specimen (cm) scale, these vuggy pores are aligned in specific directions in the lowermost samples of the Ca1 and disappear upward. Therefore, the formation of these vuggy pores is related to selective dissolution of calcite cemented dolomite grains, probably by the mineralizing fluids (Figs. 7F, 10C).

The sulfides were mostly formed after the dissolution of highly calcite cemented areas of dolomite that have vuggy pores (Fig. 10E, F). Similar to S1 and T1, illite is formed after carbonate dissolution, with cations likely sourced from dissolution of K-feldspar (Fig. 10B).

The base metal sulfides in the three units include bornite in the Cu zone and sphalerite and galena in the overlapping Zn-Pb zone (Figs. 6A-H, 8C-H, 9C-E, 10C). The base metal data also reflect the vertical metal zonation in the studied drill cores (Fig. 11A-C). Zinc (sphalerite) is more proximal to the Cu zone than Pb (galena) in the Zn-Pb zone, and this could reflect the slight difference in their solubility during sulfide precipitation (Seward and Barnes, 1997; Huston et al., 2016). The distribution of certain trace elements also reflects the zonation of the major sulfide phases. For example, trace elements such as Bi, Ag, and Sb are primarily partitioned into galena and Cd into sphalerite (Cook et al., 2009; George et al., 2015). The enrichment of Cd in the Zn-Pb zone, therefore, may reflect its partitioning into sphalerite and/or galena (Table 1; Fig. 14A). Consistent with this interpretation, trace element analysis using laser ablation (LA)-ICP-MS on sulfides from the Polish Kupferschiefer showed higher Cd concentrations in sphalerite but not in bornite and chalcopyrite (Foltyn et al., 2022).

Different timings have been proposed for the high-grade Cu and Zn-Pb mineralization in the Kupferschiefer, from early diagenetic (e.g., Wedepohl and Rentzsch, 2006) to hydrothermal mineralization induced by tectonic fracturing (Jowett, 1986, 1987; Cathles et al., 1993). Most of the early diagenetic models are based on (1) the interpretation of early pyrite replacement and low δ³⁴S values of sulfides and (2) the laterally extensive stratiform style of the mineralization, especially in the fine-grained carbonaceous mudstones of the T1 (e.g., Wedepohl and Rentzsch, 2006). Pre-ore, framboidal pyrite in the T1 and Ca1 is preserved in organic matter- and clay-rich laminations and likely formed before carbonate cementation during early diagenesis or even in the water column. In contrast, the high-grade Cu and Zn-Pb sulfides are formed mostly as a replacement of early diagenetic intergranular calcite cement in S1 and T1 (Figs. 6D, 9B-E). The lack of grain collapse following calcite dissolution is consistent with synchronous calcite dissolution-sulfide precipitation. It may also suggest that the high-grade, disseminated, sulfide mineralization occurred at relatively shallow depths, i.e., grains were partially dissolved but lithostatic pressure was insufficient to cause the skeletal grains to collapse.

The other important feature is the replacement of detrital feldspar by sulfides, which is most readily observed in the S1 (Fig. 6A-C) and coarser-grained laminations in the lower T1 (Fig. 8C-H). For example, in the S1 from Allstedt, rounded detrital K-feldspar (and intragranular calcite) has been replaced by bornite (Fig. 6A-C). In the T1, petrographic comparison between unmineralized and mineralized domains shows abundant carbonate and subordinate feldspar replacement by sulfides, whereas quartz was less affected (Fig. 8A-D). The angular shape of sulfides (e.g., bornite) replacing detrital clasts in these coarse-grained laminations also provides indirect evidence for sulfide replacement of feldspar clasts (Fig. 8E).

Overall, our preferred model involves the influx of Cubearing hydrothermal fluids into the S1, T1, and Ca1 following early diagenesis, which resulted in the dissolution and replacement of calcite and feldspar by sulfides (Fig. 17). It is generally accepted that the Cu-bearing fluids that formed SSC deposits had a weakly acidic to neutral pH (Wedepohl and Rentzsch, 2006; Hitzman et al., 2010; Borg et al., 2012). As there is no evidence of widespread phyllosilicate alteration in the Rotliegend, therefore, the acidity that resulted in the dissolution of calcite and feldspar was likely to have been generated in situ during fluid-rock interaction and fluid mixing (e.g., Corbella et al., 2004; Zhang et al., 2021; Magnall et al., 2023).

The samples from the Saale subbasin contain fine-grained Cu and Pb-Zn sulfides that are hosted in the uppermost Rotliegend sandstone (S1), Kupferschiefer (T1), and overlying Zechstein limestone (Ca1). This stratigraphic interval records a major change in the depositional environment that followed the Zechstein marine transgression (e.g., Legler et al., 2005; Legler and Schneider, 2008). Importantly, the diagenetic evolution of units separated by major stratigraphic boundaries can have a major influence on reservoir heterogeneities (Morad et al., 2000, 2010, 2012; Al-Ramadan et al., 2012). For example, early diagenetic processes are primarily controlled by the mineralogy of detrital phases, pore fluid chemistry, organic matter content, and residence time of sediments under specific geochemical conditions (Ketzer et al., 2003a, b; Morad et al., 2000, 2010, 2012).

In the Saale subbasin, the immature primary sediment composition and transition toward seawater-dominated pore fluid chemistry across the S1-T1 boundary were key factors controlling the diagenetic evolution of the system. In particular, the distribution of early diagenetic carbonate, which was then subsequently replaced by sulfides, was a major factor controlling the influx of metal-bearing fluids through the S1 and T1 units. Reaction permeability developed when metal-bearing mineralizing fluids reacted with calcite in the host rock, resulting in fluid migration and sulfide precipitation. This model is similar to carbonate replacement models for clasticdominant (CD-type) Zn deposits, which are also hosted by carbonate-rich siliciclastic units (e.g., Magnall et al., 2021; Spinks et al., 2021). Future research may further elucidate the mineralizing fluid migration pathways in the host rocks by mapping the carbonate cement using techniques such as hyperspectral imaging and micro-X-ray fluorescence in cores, integrated with micro- to nanoscale imaging techniques (e.g., transmission electron microscopy) to better understand the interface reactions between sulfides and other mineral phases (e.g., Magnall et al., 2023).

Copper and Zn-Pb sulfide mineralized rocks in the Saale subbasin (Eastern Germany) formed primarily via the replacement of diagenetic carbonate cement in the uppermost Rotliegend (S1), Kupferschiefer (T1), and Zechstein limestone (Ca1). Early calcite cementation in the Rotliegend (S1) unit occluded high levels of precompaction primary porosity and was followed by diagenetic alteration of detrital feldspar clasts. In the overlying T1 unit, petrographic and mineralogical data show an increase in carbonate (calcite and dolomite) cement. Enrichment of RSTEs (e.g., Mo) and covariation with total organic matter in the lower T1 are consistent with the development of highly reducing depositional (euxinic) conditions. The styles of sulfide replacement in the S1 and T1 are similar, although the highest base metal grades (max. Cu = 2.4 wt %, Zn = 3.1 wt %, Pb = 1 wt %) are associated with carbonate in the middle T1. The most volumetrically important mineralization style involved the replacement of intergranular calcite cement in the S1 and in the middle and upper T1. A second style of sulfide mineralization involved the replacement of detrital feldspar clasts. In the Ca1 unit, sulfide mineralization has formed in vuggy pore spaces. Overall, the reduced nature of the T1 provided the necessary redox gradient for sulfide precipitation, and the dissolution of diagenetic carbonate was the main factor controlling reaction permeability and the lateral migration of hydrothermal fluids in the Saale subbasin.

Funding for this project was provided by the Helmholtz Recruitment Initiative grant to Sarah Gleeson. We acknowledge the Landesamt für Geologie und Bergwesen, Sachsen-Anhalt, for supplying core samples from their repository. We thank Uwe Dittmann for thin and thick section preparation, Hartmut Liep for sample powdering, Andrea Gottsche for assistance during sample preparation for XRD analysis, Ilona Schäpan for BSE imaging and EDS analysis, and Sylvia Pinkerneil for total organic carbon analysis, all at GFZ-Potsdam. We particularly acknowledge Simon Jones and Jonathan Cloutier for their thorough and thoughtful reviews of this paper.

Mohammed Seid Mohammedyasin is currently a Ph.D. student at GFZ Helmholtz Centre Potsdam and Free University Berlin, Germany. He completed his B.Sc. degree in applied geology at Arba Minch University (Ethiopia) and M.Sc. degrees in petroleum geoscience and mineral exploration geology at Norwegian University of Science and Technology and Addis Ababa University (Ethiopia). Mo’s Ph.D. project is investigating the diagenetic and hydrothermal processes involved in the formation of sediment-hosted stratiform Cu deposits in the Kupferschiefer district (Saale subbasin, Eastern Germany). Mo’s research interests focus on understanding fluid-rock interactions and fluid flow pathways in the formation of sediment-hosted mineral deposits.