Chemotrophy-based phosphatic microstromatolites from the Mississippian at Drewer, Rhenish Massif, Germany

Abstract The Drewer quarry located in the Rhenish Massif is a well-studied outcrop that comprises Upper Devonian (Famennian) to Lower Carboniferous (Viséan) strata. Within the Drewer deposits two black shale intervals have been described that are linked to two global oceanic anoxic events, the Hangenberg Event and the Lower Alum Shale Event. The black shales associated with the Middle Tournaisian Lower Alum Shale Event contain abundant phosphatic concretions, which were investigated using thin section petrography, powder X-ray diffraction, Fourier-transform infrared spectrometry and scanning electron microscopy. The concretions formed during several growth phases under anoxic and at least episodically sulphidic conditions within the sediment and served as a substrate for subsurface microbial mats that formed phosphatic microstromatolites. The microstromatolites occur either as partially branched columns of up to 600 µm in length attached to the phosphatic concretions or as smaller, bulbous aggregates surrounding the concretions. Element mapping identified the presence of pyrite and other metal sulphides within the phosphatic microstromatolites. The carbon and oxygen stable isotopic composition of phosphate-associated carbonate within the phosphatic microstromatolites suggests that the mat-forming microorganisms were probably anaerobic, chemotrophic microbial communities dwelling in the anoxic environment during the Lower Alum Shale Event. Such interpretation agrees with the deeper-water depositional setting of the Lower Alum Black Shale and its high content of organic matter, suggesting that chemotrophic microbial mats are potent agents of phosphogenesis in general, and of the formation of phosphatic stromatolites in particular.


Introduction
Black shales are organic-rich sediments that commonly formed coevally in space and time through Earth history and have been associated with perturbations in the carbon cycle, climate change and global extinction events (Schlanger & Jenkyns, 1976;Arthur & Sagemann, 1994;Sagemann et al. 2003;Jenkyns, 2010).They are characterized by rapid and efficient accumulation of organic matter in marine sediments, either by increased primary production or efficient preservation.Black shales commonly contain phosphorus-rich rocks referred to as phosphorites, which are defined by P 2 O 5 contents of 18 wt.%or higher (Föllmi, 1996).Phosphorites and black shales require similar formation conditions and are therefore commonly encountered together in the rock record.Phosphogenesis, which describes the precipitation of authigenic phosphorus minerals in marine sediments, occurs within a specific environmental spectrum where ocean circulation, sedimentation and the preservation of organic matter during early diagenesis allow for phosphorus to accumulate sufficiently in sedimentary pore waters (Glenn et al. 1994;Föllmi, 1996;Benitez-Nelson, 2000;März et al. 2008;Küster-Heins et al. 2010).In the modern ocean, such conditions exist in suboxic to anoxic marine sediments typified by a constant supply of organic matter, as observed in coastal upwelling zones, continental margin sediments or restricted marine basins (Filipelli & Delaney, 1996;Schenau et al. 2000;Filipelli, 2011;Lomnitz et al. 2016).Because phosphorus is highly mobile and is cycled efficiently during early diagenesis, persistent anoxic conditions are required to allow for phosphorus minerals to precipitate (Ruttenberg & Berner, 1993;Ingall & Jahnke, 1994;Föllmi, 1996;Kraal et al. 2012).Phosphogenesis typically requires multiple phosphorus sources including the anaerobic microbial degradation of organic matter, the reductive dissolution of iron oxides releasing adsorbed phosphate, as well as the degradation and dissolution of bone material and fish debris (Jensen et al. 1995;Schenau et al. 2000;Smith et al. 2015).Since phosphorus minerals such as carbonate fluorapatite are common in organic-rich black shales, these deposits are regarded as type sections for phosphogenesis that hold important clues on the driving processes enabling the formation of sedimentary phosphate minerals (Filippelli, 2011).
Phosphatic stromatolites have been found in Proterozoic and Phanerozoic carbonate, phosphorite and black shale lithologies (Krajewski et al. 2000;Caird et al. 2017;Sallstedt et al. 2018;Zoss et al. 2019).Stromatolites are lithified microbial build-ups that are the result of the interaction between various microbial metabolic processes and their sedimentary environment (Dupraz & Visscher, 2005;Allwood et al. 2007;Sallstedt et al. 2018).Stromatolites in the fossil record have commonly been interpreted as products of cyanobacteria due to their remarkable similarity to modern cyanobacterial mats (Stal, 2012).A cyanobacterial origin has also been suggested for phosphatic stromatolites, occurring in shallow marine settings due to the presence of preserved oxygen gas bubbles (Bosak et al. 2009;Sallstedt et al. 2018), laminated fabrics related to trapping and binding mechanisms, stable isotope analyses, mineralogy, as well as facies analyses of the host sediments (Rao et al. 2000(Rao et al. , 2002;;Lundberg & McFarlane, 2011;Drummond et al. 2015;Caird et al. 2017;Sallstedt et al. 2018).Modern examples of such cyanobacterial phosphatic stromatolites are scarce and have only recently been described from a low-phosphorus terrestrial environment (Büttner et al. 2021).In contrast to these photosynthesis-based cyanobacterial stromatolites, phosphogenesis can result from organic matter degradation by anaerobic, sulphate-reducing bacteria (Thamdrup & Canfield, 1996;Benitez-Nelson, 2000;Arning et al. 2009a;Berndmeyer et al. 2012).A further relationship between bacteria involved in the sulphur cycle and phosphogenesis has been suggested for various ancient and modern phosphorite and phosphorus-rich deposits (Williams & Reimers, 1983;Schulz & Schulz, 2005;Bailey et al. 2007Bailey et al. , 2013;;Arning et al. 2009a;Cosmidis et al. 2013;Zwicker et al. 2021), in particular with regard to the large, colourless chemotrophic sulphide-oxidizing bacteria.These bacteria take up and release phosphate into the pore waters during early diagenesis and are capable of storing polyphosphate within their cells (Schulz & Schulz, 2005;Sievert et al. 2007;Zopfi et al. 2008;Goldhammer et al. 2010).
Here we report chemotrophy-based phosphatic microstromatolites enclosed in black shales deposited after the Devonian-Carboniferous transition in the course of the globally coeval, transgressive Lower Alum Shale Event (Sobolev et al. 2000;Siegmund et al. 2002;Kaiser et al. 2011;Becker et al. 2016Becker et al. , 2021)).These black shales are rich in phosphorous concretions that host phosphatic microstromatolites of various size and morphology.Using a comprehensive approach combining petrography, mineralogy and isotope geochemistry, it is suggested that the stromatolite-forming microbial communities thrived in deepwater environments and were independent of sunlight but relied on chemotrophic metabolic pathways.

2.a. Tectonic setting
The Rhenohercynian Basin is part of the Avalonian Plate and was related to numerous active subduction zones associated to the closing Rheic Ocean (Oncken et al. 2000).The Rhenohercynian fold and thrust belt, which comprises the Rhenish Massif, are part of the Middle European Variscides that occur as well exposed and complete Middle Devonian to Lower Carboniferous successions (Oncken et al. 1999).The Rhenohercynian Zone has been characterized as an evolving rift system developing on Upper Devonian to Lower Carboniferous subsiding shelf sediments of the Old Red Continent (von Raumer et al. 2017).This zone, along with the Saxothuringian and the Moldanubian zones, divides the Central European Variscides from the northwest to the southeast (Kossmat, 1927;Brinkmann, 1948).The Rhenish Massif has been associated with the Avalonian Terrain that separated from Gondwana in the Early Ordovician (Oncken et al. 2000;Eckelmann et al. 2014).The resulting Rheic Ocean began to close in the Early Devonian up to the Carboniferous and involved the formation of several microplates separating the closing ocean from the Paleotethys and the Rhenohercynian Ocean by island arcs, which accreted to Avalonia in the Early Devonian (Oncken & Weber, 1995).The opening of the Rhenohercynian Ocean has been related to an active Laurussian continental margin that separated the Avalonian terranes (von Raumer & Stampfli, 2008;Zeh & Gerdes, 2010).These island arcs are documented by rocks of the Mid-German Rise, or 'Mid-German Crystalline Zone' (Altenberger et al. 1990;Dombrowski et al. 1995), which separates the Rhenohercynian from the Saxothuringian zones (Zeh & Gerdes, 2010).
The Rhenohercynian Basin developed as an elongated, narrow trough between the Mid-German Rise to the south and the London-Brabant High in the north as part of the Old Red Continent during the Middle Devonian (Königshof et al. 2016).The Upper Devonian sedimentary rocks of the Rhenohercynian Basin are part of the Hercynian Facies that comprise siltstones and red shales, reef carbonates and bioherms and drowned carbonate platforms (Becker et al. 2016).The Late Devonian reefs were subject to several global extinction events such as the Kellwasser crisis in the latest Frasnian and the Hangenberg Event at the Devonian-Carboniferous boundary (Becker et al. 2016).Deposits of the lowermost Carboniferous in the Rhenish Massif comprise cherts, organic-rich black shales and turbiditic limestones, which have been collectively associated with the 'Kulm-Facies'.In the Early Carboniferous, the Rhenohercynian Basin was successively closed and subducted under the Mid-German High, forming an active continental margin as evidenced by extensive local volcanism (Oncken & Weber, 1995;Siegmund et al. 2002).

2.b. Upper Devonian to Mississippian strata at Drewer
The Drewer quarry represents the northernmost locality of the Rhenish Massif cropping out within several anticlinal and synclinal structures (Becker et al. 2016), which are part of a larger thrust and fold belt with numerous synclinal and anticlinal systems from the Velbert and Remscheid-Altena anticlines to the west and the larger Warstein anticline to the south (Fig. 1).Drewer is located within the Belecke anticline, which formed an intrabasinal swell in a starved basin during the Early Carboniferous (Clausen et al. 1989;Korn, 2010).It represents a key lithostratigraphic and biostratigraphic locality for the Devonian-Carboniferous boundary displaying Milankovich cyclicity, several regressions and transgressions, as well as extinction events as recorded by the Hangenberg and Lower Alum Shale Events (Becker, 1992(Becker, , 1999;;Korn et al. 1994;Siegmund et al. 2002;Kumpan et al. 2015;Becker et al. 2016).These two black shale intervals have been correlated globally to coeval strata in Spain, France and Poland, as well as to Hangenberg Event strata from India (Ganai & Rashid, 2019), Vietnam (Komatsu et al. 2014), China (Zhang et al. 2019), Morocco (Kaiser et al. 2007), Italy (Spalletta et al. 2021) and the USA (Lu et al. 2019;Martinez et al. 2019;Barnes et al. 2020).Facies changes are recorded at Drewer starting with the latest Famennian nodular Wocklum Limestone and the calcareous, laminated Drewer Sandstone (Luppold et al. 1994), followed by the Hangenberg Black Shale and intercalated Hangenberg Sandstone (Clausen et al. 1989).The Devonian-Carboniferous boundary at Drewer is represented by the Stockum Limestone composed of alternating limestones and marls, followed by the Lower Tournaisian Hangenberg Limestone characterized as typical cephalopod limestone (Becker et al. 2016).A sharp transition to phosphorus-rich black shales associated with the global Lower Alum Shale Event (cf.Becker, 1992) represents a carbonate production and ecosystem collapse during a maximum flooding event (Korn, 2010;Becker et al. 2016).The abrupt transition from cephalopod-rich limestones to black shales marks the lower boundary of the German Kulm facies postdating the Hangenberg Event and the Devonian-Carboniferous boundary.
The uppermost strata at Drewer features grey, crinoidal limestones, which have been correlated to the Erdbach Limestone of the southern Rhenish Massif (Becker et al. 2016).

Material and methods
The northwestern face of the Drewer quarry was logged over an 8meter section, producing a detailed profile of the Lower Alum Black Shale facies.Thin sections of phosphorus-rich black shales were prepared.For bulk rock powder X-ray diffraction (XRD) analysis, samples were crushed to fine powder in an agate mortar.XRD analysis of carbonate-cemented background sediment was performed on a powder sample obtained with a handheld microdrill from a polished slab.XRD measurements were carried out at the Crystallography Department (University of Bremen), using a Philips X'Pert Pro MD X-ray diffractometer with a Cu-Kα tube (λ = 1.541; 45 kV, 40 mA).Scanning electron microscopy (SEM) was conducted on a conventional tungsten filament SEM (FEI Inspect S) and a field-emission-gun scanning electron microscope with integrated focused ion beam (FEI QuantaTM 3D FEG) and an energy-dispersive X-ray detection unit (EDAX Apollo XV) at the Institute for Geology of the University of Vienna.Data processing was conducted using the EDAX TEAM TM V3.1.1 software.The presence of pyrite was determined using a Cameca SX-100 electron microprobe at the Faculty of Geosciences, University of Bremen.Analytical conditions included an acceleration voltage of 20 kV, beam current of 10 nA and a defocussed beam of 1-2 μm diameter.Counting times were 20 seconds on peak and 10 seconds on background.For quantification natural minerals from the collections of the University of Vienna and the Smithsonian Institution were used (Jarosewich et al. 1980) and the built-in PAP matrix correction.
Fourier-transform infrared (FTIR) spectra were acquired from 370 to 4,000 cm −1 on a Bruker Tensor 27 FTIR spectrometer equipped with a glo(w)bar MIR light source, a KBr beam splitter and a DLaTGS detector at the Department of Mineralogy and Crystallography of the University of Vienna.A polished thin section was pressed on the 2 × 3 mm diamond window of a Harrick MVP 2 diamond attenuated total reflectance (ATR) accessory in such a way that either the apatite or the calcite components were predominantly probed.For comparison, spectra of blackboard chalk and a pure fluorapatite crystal from Durango, Mexico, were acquired.Sample and background spectra were averaged from 32 scans at 4 cm −1 spectral resolution.Background spectra were obtained from the empty ATR unit.Data handling was performed with OPUS 5.5 software (Bruker Optik GmbH, 2005).
Sample powders for carbon and oxygen stable isotope analysis of phosphate-associated carbonate (PAC) of carbonate fluorapatite were obtained from polished rock slabs using a handheld microdrill.A total of 14 samples were prepared; four from phosphatic microstromatolites, five from the phosphatic concretions and five samples from the host rock surrounding the concretions.Stable isotope analyses were conducted at IOW using a Thermo Scientific Gasbench II connected to a Thermo Finnigan MAT 253 gas mass spectrometer via a Thermo Scientific Conflo IV split interface following Böttcher et al. (2018).It was assumed that phosphoric acid reaction with apatite and calcite to release carbon dioxide is associated with the same kinetic 18 O/ 16 O fractionation effect (Kolodny & Kaplan, 1970;Loeffler et al. 2019).Isotope values given in '‰' are equivalent to 'mUr' (milli-Urey; Brand & Coplen, 2012).The carbon and oxygen isotope data are given with respect to the V-PDB standard with a reproducibility of better than ±0.10 mUr and ±0.15 mUr, respectively.Due to phase-specific sampling, the determined δ 13 C values from the phosphatic microstromatolites represent almost pure apatite-bound inorganic carbon.

4.a. Lithostratigraphy of the Devonian-Carboniferous boundary at Drewer
The Devonian-Carboniferous boundary section at Drewer exposes the Wocklum Limestone, Hangenberg Black Shale and Sandstone, Hangenberg Limestone, Lower Alum Black Shale, equivalents of the Erdbach Limestone and the siliceous Kulm siltstones (Fig. 2; cf.Korn et al. 1994;Becker et al. 2016).The Wocklum Limestone is grey to dark grey, nodular, partly sparitic and is well-bedded in the upper part with a sharp transition to the Hangenberg Black Shale, which is composed of darker grey to black, slaty to flaky bands of shale that are partly unconsolidated.The overlying Hangenberg Sandstone is fine grained and well sorted, with a gradational contact to the Hangenberg Limestone above.The transitional section of the lower part of the Hangenberg Limestone comprises fine-grained siltstone with several thin nodular limestone beds, whereas the main limestone is grey and massive.The shift to the Lower Alum Black Shale is marked by thin siltstone and limestone beds, which are 10-15 cm thick (Fig. 2).The Lower Alum Black Shale measures approximately 130 cm in thickness and is described in detail below.The grey to dark grey, massive, mostly sparitic Erdbach Limestone above the Lower Alum Black Shale is approximately 80 cm thick, followed by the black, very finegrained, siliceous Kulm siltstone.

4.b. Lithostratigraphy of the Lower Alum Black Shale interval at Drewer
The detailed profile of the Lower Alum Black Shale (Fig. 3) begins with the uppermost parts of the Hangenberg Limestone, a micritic, bioclastic wackestone (cf.Becker et al. 2016) that grades into weakly carbonaceous siltstones representing the last phase of deposition before the Lower Alum Shale Event.The Lower Alum Black Shale interval at Drewer has been divided into nine units, which are designated as units 5a to 5i (Fig. 3).Unit 5a comprises 40 cm of dark grey clay with pyrite-bearing intervals and 0.5-10 mm thick beds.Unit 5b includes 30 cm of a similar lithology as unit 5a with clay, as well as first occurrences of isolated, tabular phosphatic concretions up to 2 cm in width and 0.5 cm in thickness.Unit 5c is similar in appearance with more individual tabular concretions, which occur more frequently in unit 5d.In this unit, various concretion morphologies from tabular to platy to nodular are present (Fig. 3).Unit 5e shows wavy lamination within the finely bedded shales, as well as cracks and fissures filled with weathered material.Unit 5f exhibits a clear transition from unit 5e marked by numerous horizontal, tabular and partly overlapping phosphorite beds (Figs. 3,4a).This bed is characterized by its greyish colour and more abundant concretions (Fig. 3).Units 5g and 5h are replete in nodular to oval concretions measuring up to 10 cm in width that occur in close proximity to each other (Fig. 3).Unit 5i is the uppermost Lower Alum Black Shale unit at Drewer and features wavy lamination and tabular, elongated and occasionally irregular and fragmented concretions orientated with their long axes parallel to bedding (Figs. 3,4).Some smaller concretions are spherical.This unit is capped by a sharp transition to the pyrite-bearing, dark grey, Erdbach limestone (Fig. 3).

4.c. Petrography, mineralogy and stable isotope geochemistry
Lower Alum Black Shale unit 5h was chosen for sampling and detailed investigation due to its high content of phosphatic concretions (Fig. 5).Two microfacies were identified in these samples referred to as microfacies 1 and 2 from hereon (cf.Fig. 5a).Microfacies 1 is composed of silty shales and mudrocks featuring a pelitic matrix with poorly sorted components that comprises around 10% of the total volume.The components are mostly angular, detrital quartz grains with subordinate occurrences of muscovite and minor pyrite.X-ray diffraction patterns further revealed the presence of minor dolomite and ferroan dolomite.Microfacies 2 contains poorly sorted, angular quartz grains with subordinate dolomite and muscovite, as well as phosphatic minerals and large phosphatic concretions (Fig. 5).Microfacies 2 contains radiolarians and subordinate conodonts, and the portion of rock fragments and quartz components is estimated at Chemotrophy-based phosphatic microstromatolites 20-30 vol.%.The phosphatic concretions occur as spherical, oval and elongated forms that are between 1 and 4 cm in size.The phosphatic concretions can be distinguished in two different types.The first type of these concretions is oval to spherical and exhibits well-defined contours at their margins.These concretions contain a matrix of cryptocrystalline carbonate fluorapatite, as well as poorly preserved radiolarians that occupy an estimated 5% of the concretion volume.The second type of concretions grew around the first type of concretions exhibiting less well-defined contours at their margins (Fig. 5) and may reach up to 8 cm in diameter.The texture, fabric and mineral content of these concretions are similar to that of microfacies 2, but they contain more carbonate fluorapatite compared to quartz.
Phosphatic microstromatolites occur in two morphologies.The first type are aggregates of thin columns showing well-defined lamination (Fig. 6a, b), which measure up to 250 μm in diameter and reach 600 μm in length.Some columns are shorter and show a cauliflower-like morphology (Fig. 6b).Some of the microstromatolites of this type occur within the outer parts of phosphatic concretions (cf.Fig. 5d), where much of the space between individual columns is occupied by microcrystalline silica (Fig. 6a,  b).The second type of microstromatolite corresponds to smaller aggregates within the outer concretions (Fig. 6c, d).This type of microstromatolite does not exhibit any columnar or branching morphologies, yet it shows well-defined lamination within bulbous aggregates.Both types of microstromatolites contain unevenly distributed, authigenic pyrite (Figs. 6, 7).Most common are columnar and branching microstromatolites attached to the outer rim of the inner concretion (cf.Fig. 7a, c, d), with some concretions completely encased by microstromatolites (Fig. 7a-d).The individual stromatolite columns show a fine lamination (Fig. 7e, f).
Scanning electron microscopy confirms that the microstromatolites within the phosphatic concretions are composed predominantly of apatite revealing a light grey appearance distinct from purely siliciclastic components in backscatter imaging (Fig. 8).The alternating fine laminae within the microstromatolites contain different amounts of alumosilicates and microcrystalline silica.The columnar stromatolites attached to the inner concretion show varying degrees of siliciclastics (Fig. 8a, b) and engulf radiolarian tests that consist of silica (Figs.8c, 9c).Discernible laminae are composed mainly of alumosilicates, whereas the bulk of the stromatolites is composed of phosphorus minerals (Fig. 9b-d).Pyrite aggregates are prominent in backscatter mode by their strong white reflection (Fig. 8c, d).Electron microprobe analyses revealed that authigenic pyrite occurs within the secondgeneration concretions and within phosphatic microstromatolites (Fig. 8c, d), yet no pyrite was detected in the first-generation concretions.Pyrite within the second-generation concretions shows iron and sulphur contents between 40.3 and 47.1 wt% and 44.5 and 54.1 wt%, respectively (compared to 47.0 and 47.4 wt% iron and 53.8 and 54.3 wt% sulphur in the standard material), with average contents of 44.7 wt% iron and 51.4 wt% sulphur.Pyrite contains accessory elements including arsenic (averaging 0.19 wt%), cobalt (averaging 0.19 wt%), nickel (averaging 0.63 wt%), copper (averaging 0.12 wt%) and zinc (averaging 0.03 wt%).Pyrite analysed within the phosphatic microstromatolites shows average iron and sulphur contents of 44.9 and 52.8 wt%, respectively.This pyrite contains accessory elements including arsenic (averaging 0.08 wt%), cobalt (averaging 0.09 wt%), nickel (averaging 0.9 wt%), copper (averaging 0.07 wt%) and zinc (averaging 0.01 wt%).
FTIR ATR spectra of the mineral composition of the outer concretion, inner concretion and the phosphatic microstromatolites (cf.Fig. 5d) were compared to spectra of pure carbonate-free fluorapatite from Durango, Mexico (Becker et al. 2016), a carbonate fluorapatite from Limburg an der Lahn, Germany (RRUFF data base R050529; ca.20% of the phosphate sites is substituted by carbonate; Downs, 2006;Lafuente et al. 2015) and a calcite reference acquired from a piece of blackboard chalk (Fig. 10).The wavenumber region from 400 to 1,600 cm −1 was chosen because it displays all fundamental vibrations of the phosphate and carbonate anion groups (White, 1974;Böttcher et al. 1997;Penel et al. 1997).Regarding the exact peak positions, it must be emphasized that strong vibrational bands in ATR spectra are slightly red-shifted to lower wavenumbers due to the impact of the complex refractive index instead of pure absorption (Harrick, 1967).Whereas pure, carbonate-free fluorapatite shows only the overlapping stretching bands ν 1 and ν 3 (ca.1,100 to 930 cm −1 ) and bending modes ν 2 and ν 4 (ca.630 to 540, 470 cm −1 ) of the phosphate groups, with a flat baseline above 1,100 cm −1 , the outer and inner concretion and the microstromatolites, as well as the carbonate fluorapatite from the RRUFF database shows additional characteristic bands at approximately 1,425 and 1,450 cm −1 (double-headed arrows in Fig. 10); the latter bands have been assigned to the anti-symmetric stretching vibrations ν 3 of the carbonate group replacing phosphate (substitution type B) in the apatite structure (Hofmann, 1997).The spectra of these three Drewer apatites also reveal minor vibrational bands (ν 2, ν 3, ν 4 ) caused by the carbonate group, positioned at about 1,400, 860 and 710 cm −1 , respectively (cf.White, 1974;Böttcher et al. 1997).These results confirm the presence of carbonate groups in the structure of apatite from the Drewer concretions and microstromatolites.
The carbon stable isotope compositions of carbonate fluorapatite show some differences between the phosphatic microstromatolites, the phosphatic concretions and the host rock.δ 13 C values of phosphatic microstromatolites are lowest ranging between −3.3 and −2.2‰, whereas those of phosphatic concretions are less negative between −1.7 and −0.8‰ (Fig. 11).

5.a. Phosphogenesis and concretion growth
The Lower Alum Black Shale represents a period of reduced carbonate and clastic sedimentation, as well as high degrees of eutrophication and marine organic productivity leading to anoxic conditions in the sediments and possibly in the bottom waters (Siegmund et al. 2002;Rakocinski et al. 2023).These conditions are similar to present-day upwelling zones, where the lack of oxygen in sediments and bottom waters facilitates the accumulation of phosphorus released during organic matter remineralization (Lomnitz et al. 2016).The Lower Alum Shale Event has been interpreted as a global anoxic event leading to a severe regional extinction of many invertebrate groups (Becker, 1992;Walliser, 1996;Kaiser et al. 2011)  Phosphogenesis in black shales and other organic-rich deposits is commonly manifested in the form of laminated crusts, lithified hardgrounds and gravels, disseminated nodules and oval to spherical concretions (Glenn et al. 1994;Scasso & Castro, 1999;Arning et al. 2009a, Bradbury et al. 2015;Filek et al. 2021).Today, authigenic phosphate crusts, nodules and concretions precipitate in organic-rich sediments at shallow depth (Soudry, 2000;Arning et al. 2009a), deposited either below upwelling or oxygen minimum zones (Föllmi, 1996;Arning et al. 2008).The depletion of oxygen is a critical factor because it favours the preservation of organic matter, which is required as a phosphorus source during phosphogenesis (Filippelli, 2011).The Lower Alum Black Shale at Drewer exhibits two concretion types, which are interpreted as representing two different generations of growth (Fig. 5).The first generation is represented by spherical to oval phosphate aggregates, which are overgrown by a second generation that accreted a larger volume of phosphorite (Fig. 5c, d).The first generation of phosphatic concretions grew in organic-rich sediments corresponding to microfacies 1 and were possibly exposed by winnowing.After their exposure, the phosphatic concretions were apparently transported from their formation site into sediments corresponding to microfacies 2. The composition of first-generation concretions differs from microfacies 2, which also suggests that they did not grow in microfacies 2 sediments but were redeposited before the second generation of concretions formed.Different formation conditions for the two generations of concretions are further indicated by the absence of authigenic pyrite within the first-generation concretions, while pyrite is abundant within the second-generation concretions and in the microstromatolites.In contrast to the first-generation concretions, second-generation concretions incorporated more non-phosphate minerals including clay minerals during growth (Figs. 5,8,9).These non-phosphate minerals were most likely incorporated from microfacies 2, suggesting that the second-generation concretions grew in situ within this microfacies.In addition, the secondgeneration concretions exhibit several growth rims that suggest multiple phases of concretion growth within microfacies 2 sediments (Fig. 5b, d).The presence of pyrite dispersed within the second-generation concretion (Fig. 8d) suggests that conditions remained sulphidic within the sediment during the growth of the phosphatic concretions within the Lower Alum Black Shale at Drewer.

5.b. Phosphatic microstromatolites -Palaeoenvironment and palaeoecology
Three possible mechanisms for microbial phosphorus accumulation include the remineralization of organic matter by sulphatereducing bacteria, the release of internally stored (poly)phosphate by giant chemotrophic sulphide-oxidizing bacteria and the reductive dissolution of iron oxides that release adsorbed phosphorus (Froelich, 1988;Schulz & Schulz, 2005;Arning et al. 2009a;Berndmeyer et al. 2012).Microbial sulphate reduction is the quantitatively dominant anaerobic process in the remineralization of organic matter in modern continental margin sediments (Ferdelman et al. 1999;Brüchert et al. 2003).Sulphate reduction is considered as a main cause of liberating phosphorus from organic matter, enabling the formation of phosphorite crusts and nodules in modern upwelling sediments (Arning et al. 2009a(Arning et al. , 2009b)).The Lower Alum Black Shale would have provided much organic matter for sulphate-reducing bacteria to release dissolved phosphorus to sedimentary pore water.The distribution, abundance and morphology of authigenic pyrite are further indicators for sulphate reduction in ancient phosphate-rich deposits (Wilkin et al. 1996;Cosmidis et al. 2013;Rickard, 2021).Pyrite is abundant within the second-generation phosphatic concretions and within the microstromatolites (Figs.8a, c, d, 9f).The remineralization of organic matter and phosphorus release by sulphate reduction probably contributed substantially to phosphogenesis in the organic-rich sediments at Drewer, suggesting that phosphogenesis during early diagenesis was the primary process during growth of the second-generation concretions and microstromatolites.
Phosphorus-rich deposits in modern upwelling regions provide habitats for giant chemotrophic sulphide-oxidizing bacteria (Schulz & Schulz, 2005;Arning et al. 2008;Crosby & Bailey, 2012).The release of phosphate from internally stored (poly) phosphate compounds to pore waters has been demonstrated for

Chemotrophy-based phosphatic microstromatolites
Czaja et al. 2016); these chemotrophic bacteria are known to dwell in specific horizons at the redox boundary or in narrow zones at the sediment-water interface, contributing to the formation of phosphorite laminites and crusts on the seafloor (Jørgensen & Revsbach, 1983;Arning et al. 2008Arning et al. , 2009a)).Although evidence of the involvement of sulphide-oxidizing bacteria in phosphogenesis at Drewer is lacking, their possible contribution cannot be excluded.
An alternative microbial consortium that could have formed the phosphatic microstromatolites are microorganisms forming Frutexites, which are laminated, arborescent ferruginous or manganiferous microstromatolites that form in low-energy environments typified by low sedimentation rates and limited oxygen availability (Reitner et al. 1995;Woods & Baud, 2008;Lazar et al. 2013).Although it is not entirely clear which microbial consortia are involved in the formation of ancient Frutexites, iron enrichment appears to be a primary feature leading to the interpretation that iron-oxidizing bacteria in conjunction with either nitrate or sulphate reduction are key players in their development (Jakubowicz et al. 2014;Heim et al. 2017).Many Frutexites have been interpreted as stromatolites dwelling in deep-water or cryptic habitats that show no preferred phototactic growth direction (Böhm & Brachert, 1993;Gischler et al. 2021).Although the morphology and size of the Drewer microstromatolites are similar to reported Frutexites, the lack of iron oxides in the samples and the formation environment of the concretions argues against this interpretation (cf.Crosby et al. 2014).If bottom and pore waters were indeed anoxic to sulphidic at Drewer as suggested by trace element analyses (Siegmund et al. 2002;Becker et al. 2016) and by the presence of pyrite, an accumulative mechanism by the formation and subsequent dissolution of iron oxides would be difficult to achieve and attain, although this may depend on the intensity and expansion of reducing bottom water conditions (Dellwig et al. 2010).

5.c. Stable isotopes and a scenario of microstromatolite and concretion formation
The δ 13 C PAC values of the phosphatic microstromatolites between −3.5 and −2‰ (Fig. 11) probably do not record early-to mid-   Tournaisian seawater dissolved inorganic carbon (DIC).Carbon isotope stratigraphy from the Devonian-Carboniferous boundary suggests high rates of organic carbon burial and drawdown of atmospheric carbon dioxide as evidenced by positive δ 13 C excursions across this interval from localities in Europe, China and the USA (Buggisch & Joachimski, 2006;Myrow et al. 2011;Kumpan et al. 2014;Kaiser et al. 2015;Qie et al. 2015).These excursions show a variety of δ 13 C peaks between þ2 and þ6‰ and are difficult to correlate globally as they differ between various regions and basins, which makes a reliable reconstruction of the carbon isotope composition of global ocean DIC difficult (Kaiser et al. 2015;Paschall et al. 2019;Barnes et al. 2020).A lack of positive δ 13 C excursions during the Middle and Late Devonian of the Drewer section was noted by Buggisch & Joachimski (2006), who attributed this lack to a stratigraphic gap where no carbonates were deposited.Kumpan et al. (2015) investigated the δ 13 C record at Drewer, yet concluded that diagenetic alteration had compromised any reliable δ 13 C record of seawater.However, a positive δ 13 C excursion was documented in the Hasselbachgraben section near Oberrödinghausen, which is in the vicinity of the Drewer locality (Kaiser et al. 2006).These δ 13 C excursions were assigned to the Hangenberg Shale Event and interpreted as a sea-level highstand at the peak of the Hangenberg crisis (Kaiser et al. 2006).Although chemostratigraphic data directly related to the Lower Alum Shale Event are sparse (e.g.Saltzman et al. 2004), Buggisch et al. (2008) reported a positive carbon isotope excursion during this event from coeval sections across Europe, reporting δ 13 C carbonate values as high as þ5‰.
The δ 13 C composition of seawater DIC during the Late Devonian Hangenberg Shale Event until and after the Lower Alum Shale Event was strongly influenced by positive δ 13 C excursions reflecting perturbations in the global carbon cycle.Moreover, carbonate δ 13 C values from the latest Devonian through to the mid-Tournaisian never dropped below 0‰, which is distinct from δ 13 C values determined for the phosphatic stromatolites and concretions at Drewer (cf.Fig. 11).The δ 13 C values of the Drewer phosphatic microstromatolites are more negative compared to the δ 13 C value of Early Carboniferous seawater DIC (Saltzman et al. 2004;Buggisch et al. 2008).Moreover, the δ 13 C values of phosphatic microstromatolites are also more negative than δ 13 C values of the phosphatic concretions and the host material of microfacies 2 (Fig. 11).Inorganic carbon incorporated by phosphate minerals during phosphogenesis was likely a mixture of seawater DIC and DIC that was formed during the anaerobic microbial remineralization of organic matter in the sediment.
The lower δ 13 C values of the Drewer microstromatolites compared to phosphatic concretions and host rock suggest that the metabolism of the mat-forming microorganisms may have been responsible for local fractionation of stable carbon isotopes within the microbial mats that formed the microstromatolites.All δ 13 C values are substantially lower than what would be expected by fractionation during Calvin cycle carbon fixation mediated by oxygenic phototrophic microbial communities (O 'Leary, 1988;Laws et al. 1995).Although a covariance of δ 13 C and δ 18 O values of microstromatolites (r 2 = 0.94) and the phosphatic concretions (r 2 = 0.97) may argue for some degree of diagenetic overprint with meteoric waters during burial diagenesis (cf.Brand & Veizer, 1981;Banner & Hanson, 1990;Heydari et al. 2001;Tong et al. 2016), the δ 13 C values of the microstromatolites are still more negative than those determined for the phosphorous concretions and the host rock, suggesting that the signal of microbial fractionation has been preserved to some degree.Moreover, environmental perturbations related to climatic warming may cause shifts of δ 13 C and δ 18 O to more negative values as observed for the Paleocene -Eocene Thermal Maximum and several oceanic anoxic events (Zachos et al. 2006;Ullmann et al. 2014).Therefore, the observed covariance between δ 13 C and δ 18 O values of the phosphatic concretions and microstromatolites may additionally be impacted by climatic warming, elevated organic matter burial and increasing seawater temperatures from the late Famennian to the early Tournaisian (cf.Kaiser et al. 2008).Based on the temperature dependence of oxygen isotope fractionation between carbonate minerals, carbonate-bearing apatite and water (O'Neil et al. 1969;Loeffler et al. 2019) and neglecting site-specific isotope fractionation (Zheng, 2016;Aufort et al. 2017), the observed maximum isotope variation (Fig. 11) would indicate a temperature rise by up to 8 and 11°C.This seems to be unrealistically high, therefore, at least parts of the oxygen isotope shift may be caused by a lowering in the oxygen isotope composition and therefore salinity of the pore water.Taken together, it is likely that the δ 13 C values of the phosphatic microstromatolites do not reflect ambient seawater composition, suggesting that the incorporated carbon within the carbonate fluorapatite reflects microbial fractionation related to a non-phototrophic metabolism.
The phosphatic concretions formed in anoxic sediments with the oxic-anoxic interface likely present in the water column in a regime of low water energy and low sedimentation rates.Some first-generation concretions are completely surrounded by the phosphatic microstromatolites (Figs. 5c,d,6a,c,d), with the concretions as a substrate for the growth of stromatolite-forming microbial mats.The growth direction of the microbial mats was highly variable, showing no unidirectional upward growth direction as would be expected for phototactic growth as observed for photosynthesis-based cyanobacterial stromatolites (Allwood et al. 2006(Allwood et al. , 2007)).Interestingly, chemotrophic microbial mats growing around carbonate concretions have also been observed at methane seeps in the Black Sea (Reitner et al. 2005).There, the mat-forming methanotrophic archaea and sulphate-reducing bacteria form thin layers around the concretions, showing a growth mode similar to the Drewer microstromatolites.It cannot be ruled out that some of the microstromatolites formed during reworking of the concretions at the sediment water interface, yet it is unlikely that the delicate branching stromatolites (Fig. 7) would have been preserved during physical reworking, sedimentation and burial of the concretions.The envisaged scenario of microstromatolite formation on phosphatic concretions is depicted schematically in Fig. 12.The first-generation concretions formed within deep-water anoxic, organic carbon-rich sediments during the global sea-level highstand of the Lower Alum Shale Event.The high content of organic matter and anoxic conditions allowed for the accumulation of phosphorus and the formation of phosphatic concretions (Fig. 12a).These first-generation concretions were subsequently moved from their original locus of formation, either by winnowing, bottom currents, or mass wasting and transported into a host sediment consisting of microfacies 2 sediments (Figs. 5,12b).The phosphatic concretions were then colonized within the soft and porous sediment by microbial mats, probably including sulphate-reducing bacteria.These bacterial mats used the firstgeneration concretions as templates for biofilm attachment, growing on all sides of the concretions by displacive growth, pushing the soft surrounding sediments outwards, away from the concretion (Fig. 12c).The stromatolite-forming microbial mats showed no preferred growth direction that would suggest phototactic growth, which is in line with the interpretation that Chemotrophy-based phosphatic microstromatolites the phosphatic concretions and the microstromatolites formed within the sediments.The metabolic activity of sulphatereducing bacteria gradually shifted the environmental conditions in the sediments from anoxic to sulphidic during a second phase of phosphatic concretion growth (Fig. 12d), as suggested by the occurrence of pyrite within the secondgeneration concretions (Fig. 8a, d).

Conclusions
The Lower Alum Shale Event is archived within black shales deposited at the base of the Carboniferous, cropping out in the Rhenish Massif at Drewer.These black shales are rich in phosphate minerals, specifically apatite, and contain abundant phosphatic concretions that grew in two generations under anoxic to sulphidic conditions within the sediments during a sea-level highstand.The phosphatic concretions also served as a substrate for phosphatic microstromatolites, which are present in various morphologies including branched and cauliflower-shaped types attached to the first-generation concretions, as well as individual aggregates present within the second-generation concretions.Carbon stable isotope analyses of PAC derived from the phosphatic microstromatolites reveal low δ 13 C values compared to dissolved inorganic carbon of Early Carboniferous seawater, possibly pointing to a chemotrophic metabolism of the matforming microorganisms.Together with the inferred environmental conditions during black shale deposition and the formation of phosphatic concretions, these results suggest that the stromatolite-forming microbial mats thrived in an aphotic environment.Stromatolites through Earth's history have mostly been interpreted as being the remnants of photosynthetic cyanobacteria, and therefore seen as indicators for shallow water environments.This study, however, shows that microbial communities forming stromatolites can also inhabit deep-water settings, and that the mat-forming microorganisms do not necessarily depend on photosynthesis as primary metabolic pathway.The reconstruction of environmental conditions and microbial metabolisms during stromatolite growth should therefore always regard chemotrophy as a feasible alternative to the common interpretation of stromatolites formed by cyanobacteria in shallow water settings within the photic zone.

Figure 1 .
Figure 1.(Colour online) Geologic overview of the Rhenish Massif east of the Rhine river with a focus on Upper Devonian to Lower Carboniferous strata and the location of the Drewer quarry (see red box); after Königshof et al. (2016).

Figure 2 .
Figure 2. (Colour online) Outcrop photograph and sedimentary log of the investigated outcrop in the Drewer quarry; person for scale.

Figure 3 .
Figure 3. (Colour online) Detailed lithostratigraphic log and an outcrop photograph showing the corresponding beds of the Lower Alum Black Shale at Drewer; hammer for scale.

Figure 4 .
Figure 4. (Colour online) Outcrop photographs of the Lower Alum Black Shale at Drewer.(a) Elongated, flat phosphatic concretions within the transition of beds 5f to 5g.(b) Bed 5h with larger, oval to spherical phosphatic concretions overlain by bed 5i and the Erdbach limestone; folding rule for scale.(c) Detail corresponding to the left red rectangle in (b) showing numerous oval phosphatic concretions within bed 5h; coin for scale.(d) Detail corresponding to the right red rectangle in (b) with spherical phosphatic concretions within laminated black shales from bed 5h; coin for scale.

Figure 5 .
Figure 5. (Colour online) Thin section scans and photomicrographs of phosphatic concretions from bed 5g of the Lower Alum Black Shale; mf = microfacies, 1 = first-generation concretion, 2 = second-generation concretion.(a) Thin section scan showing two microfacies of background sediment with numerous smaller spherical, oval and elongated concretions and a large concretion floating within microfacies 2. (b) Close-up view of the larger phosphorous concretion in (a) with the first-and second-generation concretions, whereby the latter completely surrounds the former.White arrows show growth rims in the second-generation concretion.(c) Thin section scan showing microfacies 1 and 2 with numerous smaller concretions and a large concretion floating within microfacies 2. (d) Detailed view of the large concretion in (c) showing the second-generation concretion not completely surrounding the first-generation concretion.Arrows denoting a vague laminated growth pattern in the outer concretion, and the red rectangle highlighting a large aggregate of columnar branching phosphatic microstromatolites within the second-generation concretion.

Figure 6 .
Figure 6.(Colour online) Photomicrographs of phosphatic microstromatolites; ps = phosphatic microstromatolite, ms = microcrystalline silica, cf = carbonate fluorapatite, py = pyrite.(a, b) Large aggregates of columnar branching microstromatolites floating within the second concretion with abundant microcrystalline silica between individual columns.(b) An aggregate of more bulbous, cauliflower-shaped microstromatolites with microcrystalline silica and pyrite (opaque minerals); arrow denoting a radiolarian test.(c) Small aggregates of microstromatolites floating within the phosphatic second-generation concretion, arrow denoting a smaller aggregate.(d) Close-up view of a microstromatolite aggregate exhibiting small-scale darker and lighter laminae; arrows denoting authigenic pyrite within the phosphatic microstromatolite.

Figure 7 .
Figure 7. (Colour online) Photomicrographs of columnar branched phosphatic microstromatolites attached to the first-generation concretion; 1 = firstgeneration concretion, 2 = second-generation concretion, cv = carbonate vein, cm = clay minerals, qtz = quartz grains.(a) Columnar microstromatolites attached to the first-generation concretion.(b) Columnar, branched microstromatolites attached to the firstgeneration concretion, arrow denoting pyrite aggregates.(c, d) Columnar microstromatolites among clay minerals from a carbonate vein on the surface of the first-generation concretion.(e, f) Close-up view of columnar microstromatolites showing fine alterations of darker and lighter laminae among dispersed clay minerals and pyrite.

Figure 8 .
Figure 8. (Colour online) Scanning electron microscopy images of phosphatic microstromatolites; 1 = first-generation concretion, 2 = second-generation concretion, ps = phosphatic microstromatolite, ms = microcrystalline silica.(a) Bulbous microstromatolite projecting from the first-generation concretion outwards and showing thin dark laminae of mostly alumosilicates, as well as finely dispersed pyrite (white reflecting minerals).(b) Columnar microstromatolite attached to the first-generation concretion with more intense, frequent interlayering of alumosilicates.(c) Bulbous to cauliflower-shaped microstromatolite with little lamination; red rectangles denoting authigenic pyrite, arrows denoting radiolarian tests.(d) Apparently free-floating microstromatolite aggregate within the second-generation concretion among abundant microcrystalline silica; red rectangles highlighting larger pyrite aggregates within the microstromatolite and the second-generation concretion.

Figure 10 .
Figure 10.Fourier-transform infrared (FTIR) attenuated total reflectance (ATR) spectra.Spectra of (1) firstgeneration concretion, (2) second-generation concretion, and (3) phosphatic microstromatolite, as well as blackboard chalk, carbonate-bearing fluorapatite (RRUFF database; Downs, 2006; Lafuente et al. 2015), carbonate-free fluorapatite from Durango, Mexico (Becker et al. 2016); broken lines and shaded areas indicate the positions and areas of certain vibrations of the anion groups.The two double arrows at ca 1,425 and 1,450 cm -1 indicate the anti-symmetric stretching ν 3 modes of B-type carbonate groups in apatite.Spectra have been vertically normalized and offset for better visibility.See text for details.

Figure 11 .
Figure 11.(Colour online) Cross plot showing carbon and oxygen stable isotope compositions of phosphate-associated carbonate (PAC) of the phosphatic microstromatolites, phosphatic concretions and the phosphatic microfacies 2 of the host rock.

Figure 12 .
Figure 12. (Colour online) Schematic cartoon illustrating the formation of phosphatic concretions and phosphatic microstromatolites at Drewer, SWI = sediment-water interface.(a) Autochthonous formation of first-generation concretions in anoxic, organic-rich sediments.(b) Transport and redeposition of first-generation concretions.(c) Growth of phosphatic microstromatolites on first-generation concretions within sulphidic sediments.(d) Formation of second-generation concretions around first-generation concretions during sulphidic conditions in the sediment.
(Kaiser et al. 2011;Becker et al. 2021)ring the Mississippian(Kaiser et al. 2011;Becker et al. 2021).The Lower Alum Shale Event is also represented by black shales and cherts in the Holy Cross Mountains in Poland and from the Lydiennes Formation in the (Siegmund et al. 2002it suggests that the prolonged sea-level highstand with eutrophication in the photic zone (cf.Siegmund et al. 2002)favoured phosphogenesis at Drewer.The scarcity or the lack of bioturbation in the black shales at Drewer(Siegmund et al. 2002; own observation) agrees with episodically anoxic bottom water conditions.