Introduction

The Rhenish Massif in western Germany with its complex Paleozoic evolution represents a piece of puzzle in the understanding of the processes that led to the polyphase collision of Gondwana and Laurussia terranes and the Devonian-Carboniferous Variscan orogeny (see review by Franke et al. 2017). During the Upper Devonian, the area of the recent Rhenish Massif represented a part of the Rhenohercynian Basin, where sedimentation was controlled by sea level fluctuations, synsedimentary tectonic activity, and a changing supply of siliciclastic material (e.g., Paproth 1986; Königshof et al. 2016). Sedimentological, paleogeographic, and petrological analyses of Upper Devonian sedimentary rocks of the northern Rhenish Massif (e.g., Paproth 1986; Bless et al. 1993) indicated their derivation from southern sources (adjoining highs/shoals). However, Schulz-Dobrick and Wedepohl (1983), Press (1986), and Haverkamp (1991) suggested that the main detritus was shed from the north (“Old Red Continent”) into the Rhenohercynian Basin during the Lower and Middle Devonian. Later studies indicated that the Silurian and Devonian siliciclastic material of the neighbouring Mid-German Crystalline Rise and Harz Mountains was supplied from northern sources, e.g., Baltica as a part of the Laurussia paleocontinent (e.g., Geisler et al. 2005; Zeh and Gerdes 2010). However, in a recent paper, Dörr et al. (2017) excluded a Baltic source for detrital zircons from Upper Devonian clastic (meta)sediments of the Böllstein Odenwald (Mid-German-Crystalline-Zone). A potential source area for Upper Devonian detritus in the Rhenohercynian Basin could also have been shoals known from the Rügen area (northestern Germany; e.g., McCann 1999). The provenance of Upper Devonian sedimentary rocks of the northern part of the Rhenish Massif has not been studied in detail before. Detrital zircon provenance data exist for Devonian and Carboniferous sedimentary rocks of the southeastern and eastern part of the Rhenish massif though (e.g., Haverkamp et al. 1992; Eckelmann et al. 2014; Franke and Dulce 2017) and can be used for detailed correlations.

Provenance analysis based on U-Pb dating of detrital zircon is nowadays a routine method for resolving paleogeographic and plate tectonic settings. Supplemented with mineralogical and geochemical data, it may provide insights into the geodynamic processes at different time scales. The main aim of this paper is to define the source areas of uppermost Fammenian siliciclastic rocks (sandstone, siltstone and mudstone; see below for explanation) from the northern Rhenish Massif based on U-Pb dating of detrital zircon and mineralogical and geochemical analysis of the sedimentary rocks. These data also allow to test the presence of the German-Polish (Rügen-Pomeranian) Caledonides as an uplifted land in the Rügen area during the Upper Devonian.

Geological setting

The Paleozoic Variscan orogen is one of the most significant examples of large-scale collisional tectonics with complicated interactions between different terranes or microplates and several narrow oceans (e.g., Franke and Oncken 1990, 1995; Franke 2000; Franke et al. 2017). The classical sub-division of the Variscan mountain belt in central Europe follows the early work of Kossmat (1927). From northwest to southeast, three successive zones with different metamorphic history were defined: the Rhenohercynian, the Saxothuringian, and the Moldanubian units (Fig. 1a). An additional zone, the strongly metamorphosed Mid-German Crystalline High, situated between the Rhenohercynian and Saxothuringian zones, was defined by Brinkmann (1948).

Fig. 1
figure 1

a Geological overview of the European Variscides (after Grad et al. 2002; Linnemann et al. 2012; Franke 2014) with major tectonic structures. A black rectangle marks the study area shown in c. NPZ northern phyllite zone, MGCH mid-German crystalline high, Mod. Nappes moldanubian nappes, S-Port. south Portugal, Rheno-Herc. rheno-hercynian, Sax. saxo-thuringian, M. massif. b Sketch map of the Rhenish Massif showing the main regions. c Geological map of the northern Rhenish Massif modified from Korn (2004) with marked positions (red points) of the studied sections

Full size image

The Rhenohercynian unit is a typical example of a fold-and-thrust belt and is best exposed in the Rhenish Massif (e.g., Oncken et al. 1999; Franke 2000), with continuations westwards into the Ardennes, to southwest England, and to the Harz Mountains in the east. It is generally accepted that the basement of Rhenish Massif belongs to the Avalonia Terrane (e.g., Franke 2000). Separation of Avalonia from Gondwana started with a Late Cambrian rifting episode, followed by Lower Ordovician drift northwards that led to the opening of the Rheic Ocean in the south (e.g., Oncken et al. 2000; Linnemann et al. 2008, 2010; Eckelmann et al. 2014). The docking of Avalonia to Baltica during the Upper Ordovician to middle Silurian (~ 450–430 Ma) resulted in the progressive closure of the oceanic Tornquist Sea. During the upper Silurian (~ 420 Ma), the Iapetus Ocean finally closed due to the collision of Baltica and Avalonia with Laurentia. As a result, the Laurussia paleocontinent, i.e., Old Red Continent (Linnemann et al. 2008; Nance et al. 2010), was formed. Incipient stages of the Variscan orogeny were connected with the closure of the Rheic Ocean by the northward migration of the Armorican Terrane Assemblage (Franke 1995) during the upper Silurian-Lower Devonian (Sánchez Martínez et al. 2007; Nance et al. 2010). This was followed by the opening of the narrow Rhenohercynian Ocean southeast of the Rheic Ocean (Franke et al. 2017). Subduction-related volcanism lasted into the Lower Carboniferous (e.g., Franke et al. 2017).

The Rhenish Massif

The Rhenish Massif is composed of an autochthon onto which several allochthonous tectonic units are superimposed (e.g., Franke 1995; Eckelmann et al. 2014; Franke et al. 2017). The oldest deposits known from the northern autochthonous part of the massif are represented by Ordovician and uppermost Silurian sedimentary rocks exposed in the cores of the Ebbe and Remscheid-Altena anticlines (Fig. 1b; Langenstrassen 1983). Infrequent outcrops of pre-Devonian rocks also occur in the southern Rhenish Massif including the Giessen, Marburg and Kellerwald areas (Fig. 1b, e.g., Langenstrassen 1983). The autochthonous Devonian sedimentary sequence is composed of shallow shelf deposits (“Rhenish Magnafacies” and its subfacies, Jansen 2016) and deeper-water, pelagic sedimentary rocks of more distal shelf basins, seamounts, and synsedimentary structural highs (“Hercynian Magnafacies”; e.g., Krebs 1979; Franke 1995; Stets and Schäfer 2002; Becker 2008). The allochthonous units, represented by ocean floor and active margin deposits, are exposed in the southern Rhenish Massif and can be divided into belts known as the Lohra, Steinhorn, Hörre, and Giessen nappes (Franke 1995). A Gondwanan source of detritus for Devonian and Lower Carboniferous allochthonous sedimentary rocks was determined by Huckriede at al. (2004) and Eckelmann et al. (2014) based on muscovite and detrital zircon ages. This suggests a significant late Variscan northwards transport from an original position south of the Rheic Suture and north of the preserved part of the Mid-German Crystalline High (Franke et al. 2017).

During the Early Devonian, the Rhenohercynian Basin was located at the southern margin of Laurussia (e.g., Franke 2000). The southern boundary of the Rhenohercynian Basin was formed by the source regions of the allochthonous units and by the Armorican Terrane Assemblage of peri-Gonwanian origin that later formed the Mid-German Crystalline High. The siliciclastic material filling of the Rhenohercynian Basin derived mostly from northern sources, i.e., Laurussia (e.g., Langenstrassen 1983; Bełka et al. 2010), but also from southern sources, i.e., from the exposed Silurian magmatic arc related to the Rheic Ocean, from the northern Armorica terranes (“Franconia”) that are only partially preserved in the Taunus (the southern Rhenish Massif; Fig. 1b), and also from the Mid-German Crystalline High (Hahn 1990; Hahn and Zankl 1991; Franke et al. 2017). The opening of the narrow Rhenohercynian Basin generated a barrier in the form of a trap for northward transported detritus. The Lower Devonian sedimentary rocks of the Rhenish Massif are mostly sandstone and siltstone (review in Jansen 2016), whereas Middle Devonian rocks are represented by silty sandstone, limestone, and shale (e.g., Langenstrassen 1983; Königshof et al. 2016). The formation of partly large Givetian reef carbonates reflects directly the decrease of detritus influx into the basin (e.g., Franke 2000). This may have been controlled by the change of climate and weathering regime on the Old Red Continent (Becker et al. 2016). Magmatic activity is represented by upper Emsian felsic volcanism in the southern Rhenish Massif (Kirnbauer 1991; Franke 2000), intensive Middle/Upper Devonian (middle Givetian to middle Frasnian) basaltic volcanism recorded in the Lahn-Dill Syncline and Sauerland (Fig. 1b, e.g., Sunkel 1990), and Lower/Middle Devonian (Emsian–Eifelian) keratophyre volcanism recorded in the Siegen and Ebbe anticlines in the central Rhenish Massif (Fig. 1b; Loske and Miller 1987; Winter 2006).

Thin but wide-spread layers of subaerial, explosive volcanic deposits of rhyolitic to trachyandesitic composition characterize the Upper Devonian (Frasnian to lower Famennian) and provide evidence for ongoing subduction of oceanic crust below the Laurussia paleocontinent (Winter 2015). The Upper Devonian sedimentary stack consists of shale, limestone (drowning reef complexed and their talus, followed by condensed, pelagic seamount deposits), and turbiditic sandstone (e.g., Königshof et al. 2016). During Upper Devonian time, eustatic regressions led to episodic sand influxes into the basin (Nehden, Hemberg, Dasberg, Drewer and Hangenberg Sandstones; Krebs 1979; Königshof et al. 2016). It is believed that the Remscheid-Altena shoal that subsequently formed the Variscian Remscheid-Altena Anticline (Fig. 1b) was a sediment source of terrigenous detritus during the Upper Devonian, with fluvial shedding to the north (Keupp and Kompa 1984; Paproth 1986, 1991; Paproth et al. 1986). From the Upper Devonian onward, the Mid-German Crystalline High, which was then the active southern margin of the Rhenohercynian Basin, became a source area for the detrital material in the south (e.g., Franke and Engel 1986; Franke and Dulce 2017). The Lower Carboniferous sedimentary rocks of the northern Rhenish Massif represent passive margin deposits (chert, shale, neritic, pelagic and turbiditic limestone of the “Kohlenkalk” and Culm facies) whilst synorogenic deposition of immature sand of polymictic greywacke composition prograded from the south to north (e.g., Königshof et al. 2016).

Samples

The Famennian sequence of the northern Rhenish Massif contains five discrete sandstone horizons: the Nehden (= “Plattensandstein”), Hemberg, Dasberg, Drewer, and Hangenberg Sandstone. In this study, we focus on the two youngest units, i.e., the Drewer and Hangenberg Sandstones (Figs. 2, 3).

Fig. 2
figure 2

Generalized stratigraphy of the Devonian/Carboniferous boundary beds in the northern part of the Rhenish Massif with sample positions (Kaiser et al. 2015; Becker et al. 2016a, b, c, modified). WA and WB—sections in Drewer described by Korn (1991). (1) Limestone, (2) sandstone and siltstone (3) black mudstone, (4) greenish-grey mudstone. C carboniferous, T tournaisian, L lower, M middle, U upper

Full size image
Fig. 3
figure 3

Generalized lithostratigraphy of the Rhenish Massif to Saxothuringian Saalfeld area (Southeast German) with stratigraphic position of the Famennian–Tournaisian sandstone units after (1) Bełka et al. (2010), (2) Crônier (2007) and (3) Pfeiffer (1954) and https://www.thueringen.de/imperia/md/content/tlug/geologie/geotope/11_5334_anlage.pdf

Full size image

Sand and silt fractions are crucial as potential source of accessory, detrital minerals grains, opposite to finer fractions, which are very poor in accessory minerals. Sand- and siltstones were identified by both macro- and microscopic observations. The Drewer and Hangenberg Sandstones, Hangenberg Black Shale (HBS) and Hangenberg Shale are established lithostratigraphic units with almost isochronous boundaries (Becker et al. 2016a, b, c), controlled by glacio-eustatic changes of sea-level. The lithological inventory of the Drewer and Hangenberg Sandstones in the studied sections covers siliciclastic rocks with different fractions from clay up to gravel (according to the classification scale of Wentworth 1922). In particular sections, silt is the coarsest fraction found (Drewer and Effenberg), while in others, the Hangenberg Sandstone includes true sandstones (e.g., Oese). The HBS and Hangenberg Shale consist mainly of mud and clay fractions. The exception is Effenberg section, where the HBS is represented by mudstone intercalated with thin layers of siltstone up to very fine sandstone. To emphasize the dominant fraction of the studied sedimentary rocks, three simplified types were distinguished: sandstone, siltstone, and mudstone. To avoid stratigraphic confusion, the formal lithological names of the studied units (Drewer and Hangenberg Sandstones, Hangenberg Black Shale and Hangenberg Shale) are used even if the rocks are locally variably siltstone or mudstone.

The Drewer Sandstone is restricted to its type locality of the Belecke Anticline (Figs. 1b, 3) and is a laminated, mica-rich and pyrite-rich calcareous siltstone with a low-diversity pelagic fauna that pinches out laterally (Schmidt 1924; Korn 1984; Korn et al. 1994; Kumpan et al. 2015; Becker et al. 2016). It accumulated during the topmost Famennian (Upper Devonian VI-D) and marks the onset or regressive prelude of the global Hangenberg Crisis (Kaiser et al. 2015), slightly below the famous Hangenberg Black Shale, which falls in Upper Devonian VI-E.

The Hangenberg Sandstone is composed of sandstone and siltstone and represents a basin-floor fan (Kaiser et al. 2015). It is rich in mica, with infrequent quartzite and chert pebbles and plant debris (Keupp and Kompa 1984). It represents the regressive peak in the middle Hangenberg Crisis Interval (Kaiser et al. 2015, lower part of Upper Devonian VI-F).

Sandstone and siltstone samples were collected at Drewer, Effenberg, and Oese located in the northern part of the Rhenish Massif (Figs. 1c, 2). The Oese section lies on the northern flank of the Remscheid-Altena Anticline (e.g., Paproth and Streel 1982; Becker et al. 1993), Effenberg on its anticlinal crest (e.g., Luppold et al. 1994; Hartenfels 2011; Hartenfels and Becker 2016), and Drewer represents the Belecke Anticline (Klein 2016; Becker et al. 2016). In the abandoned quarry at Drewer (N51°29.616′; E8°21.360′) one siltstone sample of the Drewer Sandstone (DR_DS; Section WB in Korn 1991) and four siltstone samples of Hangenberg Sandstone (samples DR_HS1, DR_HS2 and DR_HS3 from Section WA and sample DR_HS from Section WB in Korn 1991) were collected from its western wall. Sandstone and siltstone samples of the Hangenberg Sandstone were also obtained at Oese (N51°24.090′; E7°47.262′) from the small abandoned quarry along the road B7 between Menden and Hemer (sample OE_HS) and from the large, active Effenberg Quarry (N51°23.400′; E7°57.583′; sample EFF HS).

For comparison, mudstone representing the Hangenberg Black Shale (HBS) and the overlying greenish-grey Hangenberg Shale were collected at Drewer, Oese, Effenberg, Oberrödinghausen (N51°23.750′; E7°50.383′), and Kattensipen (N51°27.950′; E8°23.616′) sections (Fig. 1c).

Methods

LA-ICP-MS U-Pb zircon dating

93 to 116 zircon grains with an average grain size of 40 to 120 µm from four samples from the Drewer and Hangenberg Sandstones were dated by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). The samples were crushed in a metal mortar and sieved for the fraction < 315 µm. Magnetic separation (Frantz isodynamic separator) and heavy liquid separation (methylene iodide of 3.32 g/cm3 density) were used to obtain the zircon concentrate. Zircon grains were hand-picked under a binocular microscope, mounted in epoxy resin, and polished to about half-thickness. Prior to U-Pb analysis, the zircon mounts were imaged with cathodoluminescence (CL) on a scanning electron microscope at the Institute of Petrology and Structural Geology, Charles University in Prague. A Thermo Scientific Element 2 sector field ICP-MS coupled to a 193 nm ArF excimer laser (Teledyne Cetac Analyte Excite laser) at the Institute of Geology of the Czech Academy of Sciences, Prague, Czech Republic was used to measure the Pb/U and Pb isotopic ratios. The laser was fired at a repetition rate of 5 Hz and fluence of 4.51 J/cm2, with a 20 µm spot size. The He carrier gas was flushed through the two-volume ablation cell at a flow rate of 0.7 l/min and mixed with 0.66 l/min Ar and 0.004 l/min N prior to introduction into the ICP. The in-house glass signal homogenizer (design of Tunheng and Hirata 2004) was used for mixing all the gases and aerosol resulting in a smooth, spike-free signal. The signal was tuned for maximum sensitivity of Pb and U, a Th/U ratio close to unity and low oxide level, commonly below 0.2%. Typical acquisitions consisted of 15 s measurement of blank followed by measurements of U, Th and Pb signals from the ablated zircon for another 35 s. The total of 420 massscans was acquired in the time resolved—peak jumping—pulse counting/analog mode with 1 point measured per peak for masses 204Pb + Hg, 206Pb, 207Pb, 208Pb, 232Th, 235U and 238U. Due to a non-linear transition between the counting and analog acquisition modes of the ICP instrument, the raw data were pre-processed using a dedicated Excel macro. As a result, the intensities of 238U were left unchanged if measured in a counting mode and recalculated from 235U intensities if the 238U was acquired in analog mode. Data reduction was then carried out offline using the Iolite data reduction package version 3.4 with VizualAge utility (Petrus and Kamber 2012). Full details of the data reduction methodology can be found in Paton et al. (2010). The data reduction included the correction for gas blank, laser-induced elemental fractionation of Pb and U, and instrument mass bias. For the data presented here, blank intensities and instrumental bias were interpolated using an automatic spline function while down-hole inter-element fractionation was corrected using an exponential function. No common Pb correction was applied to the data due to the high Hg contamination of the commercially available He carrier gas, which precludes accurate correction of the interfering 204Hg on the very small signal of 204Pb (common lead). Residual elemental fractionation and instrumental mass bias were corrected by normalization to the Plešovice natural zircon reference material (Sláma et al. 2008). Zircon reference material GJ-1 {no. 63} (Jackson et al. 2004) and 91500 (Wiedenbeck et al. 1995) were periodically analyzed during the measurements for quality control. The values obtained from analyses performed over many days (discordant GJ-1: mean concordia age of 600 ± 3 Ma (2σ), mean 207Pb/206Pb age of 608 ± 2 Ma (2σ); near-concordant 91500: mean concordia age of 1065 ± 5 Ma (2σ)) correspond perfectly and are less than 1% accurate within the published reference values (GJ-1: 206Pb/238U age of 600.5 ± 0.4 Ma, Schaltegger et al. 2015 and 207Pb/206Pb age of 608.53 ± 0.4 Ma; Jackson et al. 2004, respectively; 91500: 207Pb/206Pb age of 1065.4 ± 0.3 Ma; Wiedenbeck et al. 1995). The zircon U-Pb ages are presented as probability density plots generated with the ISOPLOT program v. 3.70 (Ludwig 2008). Only analyses showing discordance less than 10% were taken into account. Of the ages determined, 87 of 100 (DR_DS), 94 of 115 (EFF) and 102 of 109 (OE) ages are concordant (exception: DR_HS with 64 of 93 concordant ages). For the data interpretation, the 207Pb/206Pb age was used for zircon older than 1 Ga, whereas the 206Pb/238U age was used for zircon younger than 1 Ga. A summary of the isotope ratio measurements and dating results is presented in Supplement 1. Age errors are quoted at 2 sigma level.

Bulk chemical analysis

The studied samples were powdered in an agate mill for bulk chemical analysis. The major and trace elements concentrations’ measurements were performed on an ICP-MS at Bureau Veritas Mineral Laboratories, Vancouver, Canada. The accuracy and reproducibility of the results were monitored by the repeated analyses of international standard reference materials along the analytical session. Precision of the results was better than ± 0.05% (mostly ± 0.01%) for the major elements and generally better than ± 1 ppm for the trace elements.

The chemical index of alteration (CIA) was calculated using molecular proportions as given in the equation: CIA = [Al2O3/(Al2O3 + CaO* + Na2O + K2O)] × 100 (Nesbitt and Young 1982), where CaO* is the amount of CaO incorporated in the silicate fraction of the rock. The CaO* content was calculated based on XRD analyses: carbonate CaO in ankerite and calcite was subtracted from total CaO. If result of this equation was lower than zero, then the CaO* content was set to zero. The mineralogical index of alteration was calculated as: MIA = [quartz/(quartz + K-feldspar + plagioclase)] × 100 (Nesbitt et al. 1996, 1997), using proportions of mineral weight percentages derived from XRD analyses.

Bulk rock mineral XRD quantification

The sample preparation procedure for mineral quantification involved gentle crushing of sample aliquots in a mortar to pass through a 0.4 mm sieve. An ZnO internal standard was added in the amount of 10%, followed by McCrone milling with methanol (Środoń et al. 2001).

X-ray diffraction (XRD) analysis was performed using a side-loading procedure to maximize random arrangement of mineral grains. Diffractograms were recorded on a Thermo ARL X’Tra system, CuKα radiation, equipped with a Peltier-cooled solid-state detector. The quantitative mineralogy was derived from a whole pattern fitting of experimental diffractograms, with a collection of registered pure standards using an in-house software (Q-Min). During the matching procedure, some parts of diffraction pattern are given higher weighting factors. These are typically ranges, for which stable reflections of internal standard and clay minerals are located.

Results

Detrital zircon morphology

Zircon grains have variable shapes and degrees of roundness, from euhedral or slightly abraded to completely rounded grains. There is no visible correlation between the crystallization age of zircon grains and the degree of roundness (Supplement 2). The typical grain size is 40 to 120 µm, with a majority below 80 µm and with most grains characterized by oscillatory zoning (Supplement 1), which is a diagnostic feature of zircon of magmatic origin (Corfu et al. 2003). Some zircon grains reveal a faint or homogenous zonation that may suggest metamorphic growth or modification (Corfu et al. 2003).

Detrital zircon U-Pb ages

The detrital zircon age distributions of all studied Hangenberg and Drewer Sandstone samples are very uniform. The dominant age populations are ca. 2.0 Ga to 900 Ma and 500–400 Ma (Fig. 4). DR_DS also includes a subordinate group at 530–510 Ma. A further difference is the presence of an older zircon population of ca. 2.9–2.6 Ga in sample DR_DS (Fig. 4). The Th/U ratios in the dated zircon grains are mostly higher than 0.1 (Supplement 2).

Fig. 4
figure 4

a Probability density plots of detrital zircon U-Pb ages from four samples of top Famennian sandstone. b The youngest ages of the studied samples < 800 Ma. Neoprot. neoproterozoic, Mesoprot. mesoproterozoic, Paleoprot. paleoproterozoic

Full size image

Mineralogy and petrography

The samples from Drewer (DR_HS1, DR_HS2, DR_DS) and Effenberg (EFF_HS) are coarse siltstone with grain sizes of 30–50 µm (Supplement 3). The sample from Oese (OE_HS) represents fine-grained sandstone with grain size 50–300 µm (dominated by ~ 100 µm). All the siltstone and sandstone rocks are fossil free and generally texturally immature. Framework-forming quartz and feldspar grains are mostly sub-angular to occasionally sub-rounded. Weathering-resistant accessory minerals, especially zircon, have a wider spectrum of roundness, from euhedral to well rounded. The sorting is generally moderate but is better in siltstone.

Mineralogically, quartz and muscovite are the dominant compounds in the sandstone and siltstone samples with 33–50% and 15–43%, respectively. Among the heavy minerals, anatase and/or rutile were identified in all samples. The content of zircon is below sensitivity of the XRD analysis of ca. 0.5%. Other components are variably represented (Table 1). The siltstone samples DR_DS, DR_HS1 and DR_HS2 from Drewer contain 2–5% sodium plagioclase and < 1% potassium feldspar. Ankerite is the dominant carbonate. Pyrite is present in all studied samples but dominantly in the siltstone of the Drewer Sandstone. All the samples contain a small percentage of 1M illite of 1–6% but it is the only clay mineral in DR_DS and DR_HS2. DR_HS1 also contains 2% layered illite–smectite with an illitic component above ca. 80%. The siltstone samples DR_DS and DR_HS from Drewer are quite inhomogenous, with laminae defined by the accumulation of muscovite flakes. The pyrite is present as framboidal crystals.

Table 1 Mineralogical composition of the studied samples based on XRD analysis
Full size table

Differently to Drewer, the siltstone and sandstone from Effenberg (EFF_HS) and Oese (OE_HS) contain 7% and 9% potassium feldspar, respectively (Table 1). Calcite is the most common carbonate mineral with 11% and 5%, respectively. The abundance of ankerite is less than 2%. In addition to illite, the content of illite–smectite is 3 and 4%, respectively. Furthermore, EFF_HS contains 3% kaolinite and OF_HS 3% chlorite. The Effenberg sample (EFF_HS) is heterolithic with an alternation of thicker siltstone and very thin mudstone layers. Chromian spinel was determined by SEM analysis of the heavy-mineral spectrum of the sandstone from the Hangenberg Sandstone (OE_HS; Fig. 5).

Fig. 5
figure 5

Scanning electron microscope images showing morphology of chromian spinel from the sample OE_HS

Full size image

A similar general composition with dominance of quartz and muscovite and, locality-based modal composition for feldspar, carbonate and clay minerals is observed from the Hangenberg Black Shale and Hangenberg Shale from all studied localities (Table 1).

Chemical composition

The analyzed siltstone and sandstone display compositions with SiO2 contents of 56–75 wt.% (average 62 ± 4%), Al2O3 ranges from 10 to 13 wt.% (average 12 ± 1%), Na2O from 0.4 to 1.6 wt.% (average 0.7 ± 0.3%), K2O from 2.9 to 3.7 wt.% (average 3.2 ± 0.2%) and Fe2O3 contents of 2.4 to 5.5 wt.% (average 4.2 ± 0.8%; Table 2; Fig. 6). Both the Hangenberg Black Shale and Hangenberg Shale are characterized by narrow compositional changes for SiO2, Al2O3, K2O and TiO2, which have concentrations close to those of post-Archean Australian Shale (PAAS; Condie 1993; McLennan 2001; Table 2; Fig. 6). CaO shows highly variable concentrations ranging from 0.5 to 22.2 wt.% (average 3.8 ± 2.1% for the HBS and 7.6 ± 5.2% for the Hangenberg Shale, respectively). Fe2O3 is also variable, with concentrations ranging from 2.4 to 7.2 wt.% (average 4.7 ± 1.0%). Na2O is strongly depleted relative to the upper continental crust (UCC; McLennan 2001) and PAAS (average 0.4 ± 0.1%; Fig. 6).

Table 2 Major, trace element and ratios data for studied samples
Full size table
Fig. 6
figure 6

Major, trace and REE element compositions of a sandstone and siltstone of the Drewer and Hangenberg Sandstones and b mudstone of the Hangenberg Black Shale and Hangenberg Shale normalized to average upper continental crust (UCC; McLennan 2001). The post-Archean Australian Shales (PAAS; Condie 1993; McLennan 2001) composition is shown for comparison

Full size image

Most trace element concentrations (Cr, Ni, Pb, Sr, Zn, Zr) of the studied topmost Famennian rocks show higher values than those of the UCC and PAAS (Table 2; Fig. 6). Moreover, the Drewer samples are enriched in Ba. The concentrations of Ba, Cr, Cu, Ni, Pb, V, and Zn show a poor correlation with Al (R ≤ 0.5). In the studied samples, only Ba and V contents correlate with TOC (= total organic carbon) concentration (R ≥ 0.8), while Cd and Zn show positive correlation with Fe2O3 (R = 0.7). A good correlation occurs between Cr and TiO2, Cr and V (R = 0.6–0.7), while there is a poor correlation between Cr and Al (R = 0.4). Rare earth element (REE) values and chondrite-normalized LaN/YbN ratios, with average 8.3 ± 0.6 for the samples from the Drewer and Hangenberg Sandstones and 8.7 ± 0.9 for mudstone of the HBS and Hangenberg Shale, are similar to those of the UCC and PAAS (Table 2; Fig. 6). The values of the Eu anomaly (Eu/Eu*) are 0.53–0.76 (average 0.61 ± 0.05; Table 2) and they are slightly lower than the reference values of the UCC and PAAS. The provenance indicative element ratios Th/Sc, Th/Co, La/Sc and Cr/Th show higher values than those for the UCC for siltstone and sandstone, whereas mudstone ratios also lie above those for the PAAS (except Th/Sc; Table 2; Fig. 7). The values of Zr/Sc and Hf/Sc are 19–35 and 0.5–0.9, respectively, for sandstone and siltstone (Table 2), and 6–25 and 0.2–0.7, respectively, for mudstone.

Fig. 7
figure 7

Variations of Th/Sc, La/Sc, Th/Co, Th/Cr and Cr/Th ratios in the siliciclastic sedimentary rocks of the Drewer and Hangenberg Sandstones, Hangenberg Black Shale and Hangenberg Shale of the northern Rhenish Massif. The upper continental crust (UCC; McLennan 2001) and the post-Archean Australian Shales (PAAS; Condie 1993; McLennan 2001) composition is shown for comparison

Full size image

Discussion

Paleo-weathering, sorting and recycling

The comparison of the mineralogical data from the Hangenberg Sandstone samples with data from Lower and Upper Devonian sediments from the RWTH-1 drill hole (northern Eifel region near Aachen, northwestern Rhenish Massif; Sindern et al. 2008) shows higher abundances of ankerite, illite and detrital muscovite and less plagioclase, K-feldspar and chlorite in the uppermost Famennian samples from the northern Rhenish Massif. The sandstone and siltstone of the Hangenberg Sandstone have a more similar mineralogical composition to the Upper Carboniferous siliciclastic rocks of the northwestern Rhenish Massif than to Lower and Upper Devonian sedimentary rocks (Table 1 in Sindern et al. 2008). The SiO2, Al2O3, K2O and Fe2O3 contents are similar to values of Upper Devonian siliciclastic rocks in the northwestern Rhenish Massif (Sindern et al. 2008) but different from the major element composition of the Middle-Upper Devonian Giessen greywacke (Floyd et al. 1991) of the southeastern Rhenish Massif.

The topmost Famennian sedimentary rocks of the northern Rhenish Massif have been affected by a moderate degree of chemical alteration (a proxy of the intensity of weathering), as indicated by the chemical index of alteration values of 61–71 (sandstone and siltstone) and 61–74 (mudstone), respectively (Fig. 8; Nesbitt and Young 1982). A higher degree of alteration is given by the mineralogical index of alteration with MIAsandstone and siltstone = 71–94 and MIAmudstone = 75–96. This is mainly due to very low feldspar contents (Fig. 8; except sample OE_HS). The CIA and MIA values in studied Drewer and Hangenberg Sandstones, Hangenberg Black Shale and Hangenberg Shale samples are different from the values of older Famennian siltstone and sandstone (considered as an equivalent of the Condroz Sandstone; Fig. 3) and the upper Famennian Shale of the northwestern Rhenish Massif close to Aachen (CIA = 43–74, average 62; Sindern et al. 2008). The CIA and MIA values of the studied sedimentary rocks suggest intense weathering in the source area under steady-state conditions during which the material removal rate matches the production of mineralogically uniform weathering products (Nesbitt et al. 1997). Higher plagioclase and K-feldspar contents (Plag + K-fsp) of the sandstone and siltstone of the Hangenberg Sandstone in the Oese (20.7%) and Effenberg (11.2%) sections indicate the delivery of less weathered detritus in comparison to the highly matured mineralogical composition of the siltstone of the Drewer and Hangenberg Sandstones at Drewer (1.9 and 5.0%, respectively; Table 1). Generally, the topmost Famennian sedimentary rocks of the northern Rhenish Massif show a higher degree of chemical and mineralogical alterations than the older Famennian terrigenous sedimentary rocks of the northwestern Rhenohercynian Basin (Sindern et al. 2008).

Fig. 8
figure 8

Mineralogical and chemical composition of samples of the Drewer and Hangenberg Sandstones, Hangenberg Black Shale and Hangenberg Shale plotted in quartz (Q), plagioclase (Plag), potassium feldspar (K-fsp) diagram and in Al2O3−CaO* + Na2O−K2O compositional space (Nesbitt and Young 1984). Note that the lower part of the diagrams with Al2O3 and Q < 40 is not shown. MIA is mineralogical index of alteration (MIA = [quartz/(quartz + K-feldspar + plagioclase)] × 100) and CIA (Nesbitt and Young 1982; Nesbitt et al. 1997) is chemical index of alteration (CIA = [Al2O3/(Al2O3 + CaO* + Na2O + K2O)] × 100). Middle to Upper Devonian Gramscatho and Giessen Greywacke data after Floyd et al. (1991); Famennian sandstone and mudstone data from the RWTH-1 well (northwestern Rhenohercynian close to Aachen; Sindern et al. 2008) are shown for comparison

Full size image

Similar Cr, Cu, and Zr concentrations of the studied sedimentary rocks were reported from Upper Devonian siltstone, sandstone and shale of the Rhenohercynian Basin (Schulz-Dobrick and Wedepohl 1983; Sindern et al. 2008). The Giessen greywacke shows similar Ba, Hf, Ni, Zn and Zr concentrations (Floyd et al. 1991) as the Hangenberg Sandstone. The contents of Cd and Ba in some Famennian siltstone and mudstone from the Rhenish Massif are the highest among Upper Devonian siliciclastic rocks from the Rhenohercynian Basin (Schulz-Dobrick and Wedepohl 1983; Floyd et al. 1991).

Ba, Cr, Cu, Ni, Pb, V and Zn are essential micro-nutrients for phytoplankton and thus intimately related to the growth of oceanic phytoplankton. Organic matter-rich mudstone, such as the Hangenberg Black Shale, deposited in anoxic and sulfidic conditions is typically enriched in trace elements associated with organic carbon burial fluxes (e.g., Cd, Cu) and/or with sulphides (e.g., Ni, Zn; Tribovillard et al. 2006). Studies of concentrations of Zn in iron formations (Robbins et al. 2013) revealed that elevated Zn values may reflect a local Zn enrichment in near-vent hydrothermal exhalative sediments. Similarly, high Cu, Fe and Ba concentrations indicate a contribution from hydrothermal sources (e.g., Robbins et al. 2013; von Raumer et al. 2017). Therefore, although the enrichment of Ba, Zn, Cu, Ni, Cd, and TOC reflects the primary productivity in the Hangenberg Black Shale (Kaiser et al. 2015), it cannot be excluded that elevated concentrations were induced by hydrothermal processes connected with terminal Devonian synsedimentary volcanism (e.g., Iberian Pyrite Belt; Oliveira et al. 2005) or later, during postdepositional Variscan deformation, as in other parts of the northwestern Rhenohercynian Basin (Sindern et al. 2008).

The relatively high contents of Cr and Ni are typical for Lower and Middle Devonian sedimentary rocks of the Rhenish Massif that are indicative of basic and ultrabasic rocks and ophiolite in the source regions of these sedimentary rocks (Schulz-Dobrick and Wedepohl 1983; Press 1986). The elevated Cr concentrations of up to 130 ppm for the siltstone and 147 ppm for the mudstone are in accordance with the Hangenberg siliciclastic rocks containing detrital chromian spinel, as observed is OE_HS, as well as in Lower Devonian sedimentary rocks of the northern Eifel Mountains (Haverkamp 1991; Fig. 1b) and Middle Devonian siltstone and sandstone of the Rhenish Massif (Press 1986). Furthermore, Middle Devonian sandstone of the Rügen Depression in northeastern Germany has elevated Cr values (McCann 1999), likely genetically linked with chromian spinel grains. A probable primary crystalline source of the detrital chromian spinel is pre-Givetian Alpine-type ophiolite (Press 1986; Sindern et al. 2008) of the Middle-Upper Ordovician Rügen greywacke (Giese et al. 1994). Hence, Cr spinel originally most probably derived from uplifted Caledonian orogen parts and from the exposed parts of Baltica in the north and northeast from where it was transported into Rhenohercynian Basin during the Lower-Middle Devonian (Schulz-Dobrick and Wedepohl 1983; Press 1986). The source terrain could be located in the Arkona High (Rügen area; Northeastern Germany) inside the Caledonide deformation zone, where Mesozoic strata directly cover Lower Paleozoic rocks of the Caledonian structural stage (without Upper Paleozoic sedimentary units; Katzung et al. 1993). The presence of Cr spinel in the Hangenberg Sandstone suggests that sediment of the northern part of the Rhenohercynian Basin still derived from a northern source area during the terminal Devonian and they could have been recycled from Ordovician Rügen greywacke. Recycling from the greywacke is supported by the rounded zircon grains in the studied samples that indicate a long-range detritus transport paths or/and input of detritus from recycled sources. This is further supported by the high Zr/Sc and Hf/Sc values.

Detrital zircon provenance

The oscillatory zoning and high Th/U in the detrital zircon indicate that originally most of them were sourced from magmatic rocks. The uniform LaN/YbN ratios suggest that the sediment had a common source or several ones with mixing being efficient during transport and deposition. The provenance proxies suggest that the detritus originated mainly from continental, felsic material (Fig. 7; compare Augustsson and Bahlburg 2008). More specifically, the geochemical characteristics point towards sources with a granitic to granodioritic composition (Fig. 9). On the other hand, the presence of chromian spinel indicates also mafic to ultramafic rocks in the primary (first order) crystalline source.

Fig. 9
figure 9

La-Th-Sc plot (Bhatia and Crook 1986) for Famennian silt-, sand- and mudstone from the northern Rhenish Massif. The areas of typical rock types after Perri et al. (2012). The Gramscatho and Giessen Greywacke data are after Floyd et al. (1991). The upper continental crust after McLennan (2001), the post-Archean Australian Shales after Condie (1993) and McLennan (2001)

Full size image

In accordance with the geochemical composition, the uniform zircon age distributions of the studied rocks also indicate a similar felsic source for all samples. The small differences in zircon age distributions may be attributed to grain size sorting and break-up of zircon during sediment erosion, transport and re-sedimentation (Sláma and Kosler 2012). Our new zircon ages indicate that the top-Famennian detrital material was probably not derived from southern sources, as proposed in the past (Paproth 1986). There are several lines of evidence indicating that the transport of detritus to the Rhenish Massif was rather from the northern sources. First, the dominance of Mesoproterozoic and Paleoproterozoic zircon is typical for Baltica-derived grains. The range of ages corresponds to those known from magmatic rocks of Sveconorwegian and Svecofennian ancestry and of the Transscandinavian Igneous Belt, occupying large extents of Norway, Sweden, Poland, Denmark, and northern Germany (e.g., Söderlund et al. 1999; Andersen et al. 2002; Bingen et al. 2003).

Secondly, the other prominent detrital zircon group of Cambrian-Silurian age (500–400 Ma) matches transport from the Caledonian orogen of Scandinavia (e.g., Gee and Sturt 1985; Rey et al. 1997). The geographically closer German-Polish (Rügen-Pomeranian) Caledonides are considered as minor source area only due to limited availability. Devonian siliciclastic rocks from Pomerania in northwestern Poland contain detrital zircon with Caledonian crystallization ages and detrital muscovite with mostly Caledonian cooling ages (Paszkowski, unpublished data). This indicates that the German-Polish Caledonides were not fully submerged during the Middle Devonian as postulated e.g., by Franke (1989), but that small shoals/highs composed of Ordovician sequences and Caledonian magmatic rocks were subjected to erosion. Such structural highs have been confirmed by geophysical studies in the Rügen area with the spinel-rich greywacke (e.g., Piske et al. 1994; Hoffmann et al. 1998). This is in accordance with data from the Devonian of Pomerania (northwestern Poland), which is known to have been sourced from German-Polish Caledonides and where Caledonian zircon grains represent a smaller proportion of the detrital zircon population than in our samples (Paszkowski, unpublished data). Additionally, Caledonian ages are also known from the Silurian magmatic arc in the Mid-German Crystalline High and the Northern Phyllite Zone in the south (Fig. 1a; Franke 2000; Zeh and Will 2010; Franke and Dulce 2017). However, the dominance in our samples of Mesoproterozoic and Paleoproterozoic zircon ages that is typical for Baltica indicates that also the Caledonian zircons were most probably derived from northern sources. Besides, the Mid-German Crystalline High and the Northern Phyllite Zone contain a restricted population of Caledonian ages not older than ~ 430 Ma, at odds with the wide range of Caledonian ages of ca 400–500 Ma found in the studied sedimentary rocks. Derivation from southern sources also can be ruled out due to the lack of Neoproterozoic and Cambrian (700–500 Ma) zircon. The 700–500 Ma component is characteristic of Devonian and Lower Carboniferous sedimentary rocks from the allochthonous units in the southern Rhenish Massif that were fed from the northern margin of the Saxothuringian Zone bearing Gondwana zircon signatures (Eckelmann et al. 2014). Potential other southern clastic sources are a system of shoals that developed during regressions in the central to northern Rhenish Massif area (e.g., cores of the future Ebbe, Remscheid-Altena and Siegen anticlines; Paproth 1986), from which recycled detrital material with Baltica signatures could have supplied Famennian siliciclastic rocks. However, the scarcity in the studied sedimentary rocks of ca. 500 Ma ages typical for the Lower Devonian from the Ebbe Anticline (Wagener 1989) provides further evidence against a southern transport. Also, zircon of Lower/Middle Devonian (Emsian and Eifelian) age, which is present in the Siegen and Ebbe anticlines and related to the wide-spread keratophyre volcanism (Loske and Miller 1987; Winter 2006), is missing in the studied samples. Besides, the narrow, synsedimentary active Rhenohercynian Basin formed during the Upper Devonian a trap for detritus from southern areas.

For the studied siliciclastic rocks, a derivation of detritus from the Avalonian basement of the Rhenish Massif can also be ruled out. This is supported by the lack of Neoproterozoic age, which are typical for Avalonian sources (Linnemann et al. 2012) in the Drewer and Hangenberg Sandstones. Small Middle/Upper Devonian exposures of Caledonian basement occurred also in the present-day subsurface northwestern Rhenish Massif, from where detritus including phyllite and quartzite was shed into Givetian neritic to reefal and Frasnian/lower Famennian post-reefal successions (e.g., Ribbert 1982; Neumann-Mahlkau and Ribbert 1998; Ahrendt et al. 2001). It is currently unknown, whether these very local deposits share the Avalonian signatures.

Our interpretation is in line with that of Zeh and Gerdes (2010), who concluded a northern Baltica source of detritus for the Rhenohercynian Basin, from the upper Silurian to the Upper Devonian/Lower Carboniferous. Detrital zircon populations with Baltican affinities are known from Lower Devonian to Upper Carboniferous sandstone in other parts of the Rhenish Massif (Haverkamp et al. 1992; Huckriede et al. 2004; Eckelmann et al. 2014). Such similarities in detrital zircon age distribution suggest that the transport of detritus from the north was not a local event but rather a regional phenomenon that lasted through the Devonian/Carboniferous.

Conclusions

Hydrothermal processes connected with terminal Devonian synsedimentary volcanism or post-depositional Variscan deformation are the likely reason for Ba, Cu, Pb, and Zn enrichments in topmost Devonian sedimentary rocks of the northwestern Rhenish Massif as well as in some other sedimentary sequences in the Rhenish Massif.

The chemical and mineralogical composition of the studied siltstone and sandstone indicate an advanced stage of sediment maturity, with detritus derived predominantly from felsic sources. The detrital chromian spinel in the Hangenberg Sandstone indicates mafic to ultramafic rocks in the source area or, more probably, the recycling of uplifted Caledonian orogen parts, and/or were derived from the exposed parts of Baltica in the north and northeast.

Based on detrital zircon spectra, the studied top-Famennian siliciclastic sedimentary rocks from the northern Rhenish Massif have Baltic provenance signature with a majority of Paleo- and Mesoproterozoic (~ 2000–1000 Ma) and lower Paleozoic ages (~ 500–400 Ma) of Scandinavian Caledonides affinity, and the Rügen-Pomeranian Caledonides (Arkona High) as a minor contributor to the detritus budget. In conclusion, our provenance study enabled to settle previous speculations regarding the source areas for Upper Devonian rock of the northern Rhenish Massif.