Metamorphic Evolution of Chloritoid-Bearing Micaschist from the Variscan Elstergebirge: Evidences for Stacking of High-Pressure Rocks in the Saxothuringian Zone of Central Europe

Significant progress was made in the last decade to better quantify the pressure (P)-temperature (T)-time (t) evolution of medium- to high-grade metapelites especially when they have experienced polymetamorphic events (e.g., Li and Massonne, 2018; Massonne et al., 2018a; Likhanova et al., 2015; Massonne, 2014; Gaidies et al., 2006). These rocks frequently contain garnet, potassic white mica, and accessory monazite. Mainly on the basis of the former rock-forming minerals, which are usually chemically zoned, a P-T path can be constructed using isochemical phase diagrams, called pseudosections (e.g., Massonne, 2013; Zuluaga et al., 2005; Johnson and Brown, 2004). Monazite, approximately (La, Ce)PO₄, introduces relatively large quantities of Th and U (Parrish, 1990), so that much effort has been devoted over the last decades to develop its dating methods, including thermal ionization mass spectrometry, ion microprobe analysis, and electron microprobe (EMP) analysis (Harrison et al., 2002; Cocherie and Albarède, 2001). In particular, the EMP dating was improved to obtain fairly accurate ages for small grains or chemical domains of monazite (e.g., Waizenhöfer and Massonne, 2017; Schulz and Schüssler, 2013; Kelly et al., 2006; Braun et al., 1998). Furthermore, monazite in metamorphic rocks has also been shown to sensitively and robustly record prograde metamorphic processes and reactions (e.g., Kohn and Malloy, 2004; Pyle and Spear, 2003; Foster et al., 2002). In particular, the coexistence of garnet and monazite, both incorporating significant quantities of Y, allows petrogenetic interpretations upon which chronological relations are based on (e.g., Massonne, 2016a; Tomkins and Pattison, 2007; Yang and Pattison, 2006; Catlos et al., 2002; Pyle et al., 2001).

We have studied monazite-bearing metamorphic rocks from the Elstergebirge, an area of the Saxothuringian zone (SZ) of the Variscan Orogen (Fig. 1), for which no P-T path has been derived so far. However, similar micaschists in adjacent areas were investigated for their metamorphic evolution (southwesternmost Erzgebirge crystalline complex: Faryad and Kachlík, 2013; northeastern Fichtelgebirge crystalline complex: Rahimi and Massonne, 2018). In both cases, these rocks experienced high-pressure (P≥10 kbar) conditions, but showed a difference in peak pressure of 5–6 kbar.

Figure 1.  (a) Schematic map of the Variscan zones in Europe (modified after Keppie et al., 2010; Murphy et al., 2009; Franke, 1989). (b) Geological features of the northwestern Bohemian Massif, based on Willner et al. (2000) and Faryad (2011) including the location of the studied area (red rectangle): 1. medium- to high-grade units including rocks with high-pressure (HP)-high-temperature (HT) metamorphism around 340 Ma and mineral cooling ages > 320 Ma; 2. units with low-pressure (LP)-HT metamorphic imprint; 3. medium- to high-grade units including rocks with HP-HT metamorphism around 400 Ma and mineral cooling ages > 370 Ma; 4. mainly low-grade units (tectonically transposed) of the Saxothuringian and the Sudetes including low-temperature (LT)-HP rocks; 5. unmetamorphosed to very-low grade Neoproterozoic to Lower Carboniferous sediments (supracrustal units); 6. medium-grade units of the Sudetes and Moravo-Silesian (undifferentiated). (c) Simplified geological map of the Fichtelgebirge-Elstergebirge (based on Crender, 1902; Bernstein et al., 1973) and approximate sample location. Fr. Frankenberg; MGCH. Mid German Crystalline High; MDZ. Moldanubian zone; MS. Moravo-Silesian; Wd. Wildenfelds; WS. West Sudetes.

The micaschists from the Elstergebirge locally contain chloritoid. This mineral in HP micaschists can provide information on P-T conditions in addition to potassic white mica and garnet as was demonstrated, for instance, by Stöckhert et al. (1997), Rötzler et al. (1998), and Negulescu et al.(2018, 2009). Thus, in order to better understand the evolution of the medium-grade HP unit(s) of the SZ of the Bohemian Massif and its/their geodynamic position in the complex Variscan Orogen, a P-T-t path for chloritoid-bearing micaschist of the Elstergebirge was derived.

The northeastern part of the Variscan Orogen is exposed in the Bohemian Massif which is a collage of several basement complexes differing in age and metamorphic evolution (Fig. 1a). Traditionally, the Bohemian Massif and portions of the Variscan Orogen exposed to its west are subdivided into SZ and Moldanubian zone (e.g., Grandmontagne et al., 1994; Zulauf, 1993). These terms are still used in recent literature on the Bohemian Massif (e.g., Faryad, 2011; Hajná et al., 2011; Schulmann et al., 2009), which additionally consists of the Moravo-Silesian and Lugian-Sudetian zones in the east and northeast, respectively. The Moldanubian zone is bounded to the north by the SZ, which is one of the peri-Gondwanan fragments (microplates) of the lithosphere that became incorporated into the Central European part of the Variscan orogenic belt (see, e.g., Linnemann et al., 2000). The autochthonous units of the SZ, including nonmetamorphic nappes (Franke, 1995), represent Cadomian basement composed of Neoproterozoic to Early Cambrian arc-related volcano-sedimentary low-grade rock complexes and plutonic bodies, transgressed by Cambro–Ordovician overstep sequences with passive margin signatures. The eastern margin of the Moldanubian zone was thrust over the Moravo-Silesian zone, which probably represents the margin of Baltica, during the Lower Paleozoic (Schulmann et al., 2005).

The SZ of the Bohemian Massif, including the Sudetian zone, contains blueschist-facies and high-grade rocks in several distinct lithotectonic units (Faryad, 2011). As a major rock complex of the SZ of the Bohemian Massif, the Erzgebirge crystalline complex (Fig. 1b) is composed of several lithotectonic units characterized by different metamorphic conditions. From the bottom to the top, these units are the Cadomian basement, the gneiss-eclogite unit, the micaschist-eclogite unit, the (garnet)-phyllite unit and the uppermost allochthon represented by several klippes (Wildenfels, Frankenberg; Fig. 1b).

The Cadomian basement consists of medium-pressure orthogneisses and migmatites with no evidence of HP metamorphism (Willner et al., 2000). The gneiss-eclogite unit is composed of HP gneisses and migmatites (Willner et al., 1997) with lenses of various types of HP to ultrahigh-pressure rocks such as coesite-bearing eclogite (Massonne, 2001), garnet peridotite (Massonne and Neuser, 2005; Schmädicke and Evans, 1997), and diamondiferous quartzofeldspathic rock (Massonne, 2003). The micaschist-eclogite unit is composed of chloritoid- and garnet- bearing micaschists with intercalations of quartzites, marbles, and metabasites. Peak pressures close to ultrahigh-pressure conditions (2.6 GPa at 650–700 ℃) were obtained for eclogites (Massonne and Kopp, 2005; Klápová et al., 1998), siliceous marble (Gross et al., 2008), and eclogitic micaschist (Konopásek, 2001). The surrounding metasediments gave substantially lower pressures of 1.2–1.3 GPa (Rötzler et al., 1998). The (garnet)- phyllite unit is represented by phyllites and schists with intercalations of quartzite and metabasite. Blue amphibole has been reported from some metabasites in the basal part of this unit (Rötzler et al., 1999; Holub and Souček, 1994).

Our study area is located in the Elstergebirge (Halštrovské Hory in Czech), a hilly area in the border region of the northwestern Czech Republic and southwestern Saxony. Phyllites, interpreted to have Cambro–Ordovician protolith ages (e.g., Bernstein et al., 1973), and micaschists are the dominant rocks of the Elstergebirge (Fig. 1c; Stumm, 2002). These rocks show similarities to those of the micaschist-eclogite and (garnet)-phyllite units of the Erzgebirge crystalline complex, being adjacent to the northeast, although occurrences of eclogite were never reported from the Elstergebirge. In addition, rocks of the Arzberg series in the Fichtelgebirge crystalline complex, situated in southwest of the Elstergebirge, also show such similarities. The Arzberg series is mainly composed of metasediments (phyllite, quartzite, graphite schist, marble) with some minor layers of amphibolite and greenschist. These metasediments underwent a low-pressure greenschist- and amphibolite-facies metamorphism, characterized by the assemblage quartz-muscovite-chlorite- biotite-garnet-albite, during the Variscan orogeny (Mielke et al., 1979).

The Erzgebirge and Fichtelgebirge crystalline complexes host several post-tectonic granitic plutons, which yielded ages either around 320 Ma or between 280 and 300 Ma (Siebel et al., 2010; Carl and Wendt, 1993; Richter and Stettner, 1979). Based on the composition of magmatic muscovite, the intrusion level of these granites was in the depth range of 15–20 km (around 4.3 kbar: Massonne, 1984). Such a pluton (Fichtelgebirge granite, Fig. 1c) bounds the Elstergebirge to the south. Between this pluton and the micaschists and phyllites in the north, para- and orthogneisses were mapped. These gneisses can contain andalusite (Stumm, 2002). Locally, metapelitic layers with aegirine occur in these gneisses (Freyer and Tröger, 1965).

Various dating methods (Sm-Nd isochrons, Ar-Ar in phengite and amphibole, U-Pb in zircon and monazite) were used to determine the age of metamorphism in the SZ of the Bohemian Massif. U-Pb dating on zircon in medium- to high-grade metamorphic rocks yielded ages of c. 340 Ma (Romer and Rötzler, 2001; Kröner and Willner, 1998; von Quadt and Gebauer, 1998) or somewhat younger (Liati and Gebauer, 2009; Massonne et al., 2007). Ar-Ar dating on phengite of two eclogites from the Erzgebirge crystalline complex revealed ages of 348±2 and 355±2 Ma (Schmädicke et al., 1995). The same dating method applied to white mica and hornblende from 68 micaschists and gneisses of the EC yielded two age clusters at 340±2 and 329.7±1.5 Ma (Werner and Lippolt, 2000) confirming metamorphism in the Early Carboniferous. On the other hand, rocks from the Fichtelgebirge crystalline complex underwent low-pressure regional metamorphism in the time interval 330–320 Ma (Okrusch et al., 1990). Kreuzer et al. (1989) reported K-Ar ages of 316±3 Ma obtained on muscovite in gneiss and schist, which were sampled ~2.5 km northeast of the town of Selb (Fig. 1c). In addition, these authors applied K-Ar dating of amphibole from schistose amphibolites of the Fichtelgebirge crystalline complex and revealed ages of 332 and 299 Ma.

Dating of zircon and monazite also yielded information on protoliths of medium- to high-grade metamorphic rocks from the SZ of the Bohemian Massif. Tichomirowa (2003) and Tichomirowa et al. (2012) determined age clusters at ca. 575, 540–530 and 500–470 Ma on zircon in ortho- and paragneisses from the Erzgebirge crystalline complex indicating Cadomian and Cambro–Ordovician magmatic events at the margin of Gondwana or peri-Gondwanan terranes (see Linnemann et al., 2010, 2000). Teufel (1988) suggested Ordovician or younger protolith ages based on dating of monazite of paragneiss from the southern Fichtelgebirge crystalline complex and adjacent Moldanubian units. Waizenhöfer and Massonne (2017) (mylonitic migmatite from the Münchberg metamorphic complex) and Rahimi and Massonne (2018) (micaschist from the northern Fichtelgebirge crystalline complex) found age clusters in the time interval between 575 and 455 Ma and 472 and 405 Ma, respectively, by dating monazite with the EMP and related them to the provenance area of detrital monazite.

Major and trace element concentrations in the bulk rock were analyzed using a rock slab, which was ground to rock powder. A glass disk was prepared by fusing this powder with Spectromelt® (ratio 1 : 6). This disk was analyzed with a PHILIPS PW2400 X-ray fluorescence (XRF) spectrometer with a wavelength-dispersive (WD) system considering certified geostandards. The contents of H₂O and CO₂ in the rock powder were determined with a LECO RC-412 C-O-H multiphase determinator after drying the powder at 110 ℃ for several hours.

The chemical compositions of minerals in polished thin sections were analyzed with a CAMECA SX100 EMP with five WD spectrometers applying the PaP correction procedure provided by CAMECA. For silicates and oxides, the concentrations of F, Na, Mg, Al, Si, K, Ca, Ti, Cr, Mn, Fe, and Ba were determined using counting times of 20 s at the peak and on the background, except for F (30 s). Synthetic and natural minerals, glasses (e.g., Ba glass for the BaLα₁ radiation), and pure oxides were used as standards. The applied acceleration voltage and electric current were 15 kV and 30 nA, respectively, for analyzing garnet. For other minerals, an electric current of 10 nA was used. The beam diameter was usually 3–5 μm. The errors of this analytical method were reported by Massonne (2012) for a setting with 15 nA.

Full analyses of monazite with the EMP included the elements Si, P, S, Ca, Y, La, Ce, Pr, Nd, Sm, Eu, Gd, Pb, Th, and U. The procedure for a c. 17 minutes lasting analysis was described by Massonne et al. (2012); in the present case, the usually used beam current of 180 nA was reduced to 150 nA for the analysis of small monazite inclusions in garnet. The applied analytical conditions had been tested against various Paleozoic monazite (Waizenhöfer and Massonne, 2017) independently dated by other geochronological methods (see Langone et al., 2011; Massonne et al., 2007). For the calculation of ages and their errors, the program MINCALC-V5 (Bernhardt, 2007) was used. The age data were further processed with the Isoplot/Ex program by Ludwig (1999).

Structural formulae of all minerals were calculated with the CalcMin program (Brandelik, 2009). According to the five WD systems of the EMP up to five element concentration maps for specific elements (garnet: Ca, Mn, Fe, Mg; monazite: Y, Th, Ca, Ce, Nd; mica: Ba, Na, Mg, Fe, Ti) were simultaneously prepared by step-wise movement of the thin section under the electron beam of the EMP and subsequent computer aided evaluation (Bernhardt et al., 1995). Counting times per step were 80 ms. An electric current of 40, 20, and 150 nA was applied for areas with garnet, potassic white mica and monazite, respectively. Back-scattered electron (BSE) images were taken with the EMP to document specific textural features.

In the area between the towns of Asch (Aš in Czech) and Bad Brambach and the village of Neuberg (Podhradí in Czech) several micaschists were sampled (for sample locations and characteristics see Table 1). From these rocks we selected micaschist sample 14AS6 because it contains garnet and chloritoid. In addition, this sample is relatively fresh with well defined mineral textures (Fig. 2a). Sample 14AS6 was taken from a 5–6 m high cliff in the forest c. 1 km southeast of the village of Raun (Fig. 1c). Macroscopically, the rock is light grey-green in colour with mm-sized dark-green chloritoid and red-brownish garnet porphyroblasts and a pervasive, crenulated foliation which is mainly defined by phyllosilicates.

Figure 2.  (a) View of the outcrop of the chloritoid-garnet bearing micaschist between Bad Brambach and Bad Elster. (b)–(e) Photomicrographs showing microstructural features of the chloritoid-garnet-bearing micaschist under plane-polarized light or (not perfectly) crossed polarizers: (b) thick phyllosilicate layer containing potassic white mica (Ms), biotite (Bt), chlorite (Ch), and accessory rutile (Rt); in the adjacent quartz (Qz) layers some chloritoid (Ctd) occurs; (c) garnet porphyroblast (Gt1) in a quartz-rich layer showing spaced crenulated foliation defined by phyllosilicates; (d) garnet porphyroblast (Gt2) showing skeletal structure in a quartz-rich matrix; (e) a quartz-rich layer with chloritoid and a phyllosilicate-rich layer with oriented potassic white mica and chlorite, which also forms a fan perpendicular to the main foliation. (f)–(h) Back-scattered electron images showing: (f) a garnet porphyroblast (Gt2) with skeletal rim and quartz, chlorite, and potassic white mica in strain shadows; rutile, monazite (Mz), and zircon (Zr) occur as inclusion in Gt2; (g) subhedral rutile grain, partially replaced by ilmenite (Im), in phyllosilicates; (h) similar rutile grain surrounded by phyllosilicates. Apt. Apatite.

Quartz (43 %), potassic white mica (30%–32%), biotite (5%), chlorite (10%–11%) and chloritoid (5%–6%) form the rock matrix of sample 14AS6 (Table 1). In addition, millimetre-sized porphyroblasts of garnet (3%) and accessory phases (1%–2%) of ilmenite, rutile, zircon, apatite, and monazite are present in the sample. Layers with thickness between 0.5 and 5 mm, consisting mainly of quartz (mean grain size 0.2 mm) slightly elongated along the main foliation, alternate with phyllosilicate layers (Figs. 2b–2e) composed of oriented white mica (mean grain size 2 mm×0.3 mm) and chlorite (mean grain size 1 mm×0.1 mm).

Two groups of garnet can be distinguished according to grain size and textural setting. Rare euhedral garnet with intermediate size (mean 1 mm, Gt1) occurs in phyllosilicate layers (Fig. 2c), whereas large and subhedral garnet (3.5 mm, Gt2) appears in quartz domains and is elongated with the longer axis being at a low angle with the main foliation of the rock (Fig. 2d). The rim of Gt2 frequently shows a foam microstructure with quartz (Fig. 2f), i.e., the garnet has crystallized along the intergranular boundaries of a quartz aggregate as the size (0.1–0.7 mm) and fabric of the quartz grains are similar to those in the matrix (see also Hawkins et al., 2007; Stöckhert et al., 1997). According to Engvik et al. (2000) and Hoschek (2001), we use the term skeletal for garnet in this intergrowth texture.

Inclusions in garnet (mostly Gt2) comprise single grains of zircon (up to 20 μm large), monazite, apatite, plagioclase, opaque phases, quartz, and rutile frequently surrounded or almost replaced by ilmenite (Figs. 2d, 2f). Garnet (mainly Gt1) is marginally replaced and locally entirely wrapped by potassic white mica, chlorite, and quartz. Chlorite and potassic white mica also fill garnet fractures.

Both Gt1 and Gt2 are rich in almandine component (X_Fe2+=Fe²⁺/(Fe²⁺+Mn+Mg+Ca), > 0.74) and show the same prograde concentric zonation from core to rim (Fig. 3). The garnet core is characterized by pyrope, grossular (+andradite), and spessartine components of X_Mg=0.035–0.04, X_Ca=0.09, and X_Mn=0.11, respectively. Towards the rim, grossular (+andradite) and spessartine components considerably decrease to X_Ca=0.02 and X_Mn=0.05, whereas pyrope contents increase to X_Mg=0.065 (Fig. 3g, Table 2).

Figure 3.  X-ray concentration maps for Mg (a), (b), Ca (c), (d), and Mn (e), (f) in Gt1 (a), (c), (e) and Gt2 (b), (d), (f). Cold to warm colours indicate increasing element concentrations. (g) EMP analyses of garnet in terms of molar fractions of pyrope (X_Mg) and spessartine (X_Mn) versus that of grossular (+andradite) as X_Ca with trends from core to rim.

White mica shows a compositional variability for Si from 3.00–3.31 per formula unit (pfu, Figs. 4b–4d) with the highest values observed in inner parts of large grains; the rims are significantly poorer in Mg than the cores (Fig. 4a). Thus, there is a clear trend, which developed during metamorphism, from the Si- richest (Si ~3.31 pfu) to the Si-poorest (Si ~3.00 pfu) potassic white mica. According to spot analyses (Figs. 4b–4d), the Mg and Fe²⁺ poorer domains are richer in Na and Al and poorer in Si with values of Mg/(Mg+Fe)=Mg^# and X_Na=Na/(Na+K) varying between 0.25 and 0.60 and 0.05 and 0.10, respectively. The Fe contents first decrease as do the Mg contents, according to the (inverse) Tschermakʼs substitution, but at low Si contents there is a clear increase of the Fe contents leading to the lowest Mg^# values. The Ti and Ba contents (0.01–0.02 pfu and around 0.005 pfu, respectively) remain fairly constant (Fig. 4, Table 3).

Figure 4.  X-ray concentration maps for Mg of (a) white mica and (e) chloritoid. Black to red colours indicate increasing element concentrations. Compositional diagrams for potassic white mica in sample 14AS6 refer to (b) Mg and Na vs. Si per formula unit (pfu), (c) Fe_total vs. Si pfu; (d) Al_total vs. Si pfu.

Analyses of pale to dark green chlorite in the matrix yielded Si=2.52–2.70 pfu, Al^VI=1.38–1.49 pfu, and Mg^# values around 0.32 (Table 3). Locally, chlorite forms fan-like aggregates perpendicular to the main foliation indicating that this mineral also grew post-kinematically at a late metamorphic stage (Figs. 2b, 2e).

Chloritoid exhibits a poor cleavage and a greenish gray to blue pleochroism, which is locally elongated along the main foliation, and up to 1 mm in length (Fig. 2b). The composition of chloritoid is only slightly variable with Mg^# values between 0.08 and 0.10 and MnO contents between 0.20 wt.% and 0.40 wt.% (Table 3, Fig. 4e). Contents of Ti are below the detection limit. Although no clear zonation of chloritoid was noted from elemental mapping, it seems to be that the Mg concentrations decrease from core to rim (Fig. 4e).

Biotite is reddish-brown and forms up to 1 mm-sized grains in the matrix. It also occurs enclosed in potassic white mica. The contents of TiO₂ in biotite are around 1.7 wt.%. The Mg^# values are also fairly constant (0.46–0.48, Table 3). Rare plagioclase with X_An=Ca/(Ca+Na+K) of 0.01 and grain size between 0.1 and 0.4 mm occurs along boundaries of white mica flakes. Rutile grains, which usually occur in the matrix, are subhedral with lengths up to 0.5–1 mm; those enclosed in garnet are smaller in size (up to 0.2–0.3 mm) (Fig. 2f). Sub-anhedral ilmenite occurs mostly in phyllosilicates and rarely included in garnet ranging in size of 0.1–0.3 mm (Fig. 2g–2h). Compositionally, ilmenite is characterized by an average of 4.4 wt.% MnO (Table 3).

P-T pseudosections and their contouring with modal and compositional isopleths were computed with PERPLE_X (see Connolly, 2005; update in August 2011 downloaded from the internet site http://www.perplex.ethz.ch/) and the internally consistent thermodynamic data set (Holland and Powell, 1998; and updates) for minerals and H₂O (model CORK: Holland and Powell, 1991) for the P-T range of 2-22 kbar and 350-650 ℃ (Fig. 5a). The following solid-solution models (see file solution_model taken from the PERPLE_X software and stored in Massonne et al., 2018b) were used: GlTrTsPg for amphibole (White et al., 2003), TiBio(HP) for biotite (see below), Carp(M) for carpholite (Massonne, 2010), Chl(HP) for chlorite (Holland et al., 1998), Ctd(HP) for chloritoid (see White et al., 2000), feldspar for plagioclase and K-feldspar (Fuhrman and Lindsley, 1988), Gt(HP) for garnet (Holland and Powell, 1998), Opx(HP) for orthopyroxene (Powell and Holland, 1999), Omph(HP) for clinopyroxene (Holland and Powell, 1996), Mica(M), restricting the contents of the muscovite and margarite components to 30 mol%, for paragonite (Massonne, 2010), Pheng(HP) for potassic white mica (see Powell and Holland, 1999, with a maximum paragonite content of 50 mol%), St(HP) for staurolite (details in http://www.perplex.ethz.ch/perplex_solution_model_glossary.html), and Stlp(M) for stilpnomelane (Massonne and Toulkeridis, 2010). For cordierite, ilmenite, and talc, the ideal solid-solution models hCrd, IlGkPy, and T, respectively, were used with thermodynamic data for corresponding end-members given by Holland and Powell (1998). Quartz, rutile, titanite, Al-silicates, zoisite, and lawsonite were considered as pure phases. The amphibole components cummingtonite and grunerite and the feldspar components (ab, mic) that are alternative to those in the model "feldspar" were excluded. In addition, the mica components tip, tbi, and ann1 were neglected owing to their limited reliability (Massonne et al., 2018b). Thus, the used solid-solution model of biotite is actually that applied in former versions of PERPLE_X (Powell and Holland, 1999: Bio(HP)).

Figure 5.  P-T pseudosections calculated for (a) the slightly modified composition of metapelite 14AS6 (Table 4, bulk1); (b) an effective bulk-rock composition (Table 4, bulk2) with the computer software package PERPLE_X (see text). The grey tones of P-T fields are related to the variance (the darker the higher) of the corresponding mineral (+H₂O) assemblage; very small P-T fields are not labeled. The pseudosection in (b) is contoured by isopleths of X_Mg, X_Mn, and X_Ca in garnet and Si content (pfu) in K-white mica. The P-T field with red boundaries in the pseudosection of (a) refers to the mineral assemblage of the early metamorphic stage. And. Andalusite; Bt. biotite; Ch. chlorite; Crd. cordierite; Crp. carpholite; Ctd. chloritoid; Gt. garnet; Im. ilmenite; Kf. K-feldspar; Ky. kyanite; Lw. lawsonite; Ma. margarite; Ms. potassic white mica; Om. omphacite; Pa. paragonite; Pl. plagioclase; Qz. quartz; Rt. rutile; Sil. sillimanite; St. staurolite; Stl. stilpnomelane; Zo. zoisite.

The bulk-rock composition (Table 4, bulk1) was slightly modified for the PERPLE_X calculations since they were undertaken in the system Na₂O-K₂O-CaO-FeO-MnO-MgO-Al₂O₃- SiO₂-TiO₂-H₂O: (1) The H₂O content was set to 5 wt.% to permit the formation of a free hydrous fluid phase already at relatively low temperatures to simulate the release of water during prograde metamorphism. (2) All iron was considered to be divalent (O₂ content=0) because (i) magnetite is absent, (ii) the amount of ferric iron in minerals is very low, and (iii) rutile+ilmenite (±pyrite) indicate low oxidation conditions (Diener and Powell, 2010; Groppo et al., 2010). (3) For apatite-bearing rocks, it is suggested (e.g., Massonne, 2012; Massonne and Toulkeridis, 2012) to reduce the CaO content according to the analyzed P₂O₅ content in the bulk rock. However, it was decided here to subtract an equivalent of CaO related to 30% of the P₂O₅ content since monazite is present and apatite is rare in the rock. Because of this arbitrary decision and the low CaO content in sample 14AS6 (Table 4), the application of the X_Ca isopleths for garnet, especially for low X_Ca values, can be flawed (cf., Massonne et al., 2018b).

As the equilibration volume is the volume of rock at a set of P-T conditions in chemical equilibrium (Spear, 1995), early formed minerals, such as the core of garnet shielded by its own overgrowth domain, must be removed from the bulk rock for PERPLE_X calculations of later metamorphic stages. Such effective bulk rock compositions are difficult to assess as it is not always clear which the minerals or mineral domains have to be subtracted from the bulk rock. Commonly, garnet domains are stepwise removed to account for that (see Konrad-Schmolke et al., 2008; Evans, 2004). Occasionally, potassic white mica is considered as well (e.g., Waizenhöfer and Massonne, 2017). However, the removal of inner garnet domains from the bulk rock (or a previously defined effective bulk-rock composition) can have only a minor effect on the P-T position of isopleths for the chemical parameters of garnet and potassic white mica especially for the here relevant metapelitic compositions (see Massonne et al., 2018a, b; Rahimi and Massonne, 2018; Massonne, 2014). Nevertheless, we calculated a P-T pseudosection for one effective bulk-rock composition for the P-T range 2–9 kbar and 530–620 ℃ (Fig. 5b), following the procedure used by Marmo et al. (2002) and the garnet rim. According to the X-ray elemental mapping (see Figs. 3a–3f), we defined the garnet portion (excluded garnet core+inner mantle volume amounts to 1.5%) to be included in the effective bulk-rock composition and considered the average chemical compositions of the garnet domains, based on the EMP analyses, and the higher density of garnet with respect to the main silicates in sample 14AS6. The results of this procedure after addition of H₂O and normalization to 100% are given in Table 4 (bulk2).

The P-T pseudosections (Figs. 5a–5b) are dominated by quadri- and tri-variant fields with pentavariant fields mostly confined to HP-high-temperature conditions, and some minor divariant fields. Only quartz and H₂O are present over the entire P-T range. Potassic white mica appears in the nearly entire investigated P-T range whereas a lower pressure limit of 11.2 kbar at temperatures below 620 ℃ exists for paragonite (Fig. 6d). The Si content in potassic white mica strongly depends on pressure: at 400 ℃, which is only 3.03 pfu at 2 kbar, but rises to 3.43 pfu at 22 kbar (Fig. 6b).

Figure 6.  Contouring of the P-T pseudosection of Fig. 5a was undertaken by isopleths for (a) X_Ca (blue), X_Mg (red), and X_Mn (green) in garnet (Gt); (b) the Si content (pfu) in potassic white mica (Ms) (blue) and the modal content of garnet; (c) X_Mg in chloritoid (Ctd) and the modal content of biotite (Bt) (red); (d) X_Mg in chlorite (Ch). The orange line in (d) shows the limit of the paragonite (Pa) field. Mineral boundary lines were taken from Fig. 5.

Garnet appears at 370 and 460 ℃ at ~20 and 10 kbar, respectively, in our P-T pseudosections. With rising pressure and temperature, increasing amounts of garnet occur in various mineral assemblages. Maximum calculated garnet contents are about 12 vol% reached at 650 ℃ and ~14 kbar (Fig. 6b). X_Mg in garnet increases with rising temperature and, only at low pressures, increasing pressure (Fig. 6a). The here relevant X_Mg (0.03–0.07) occurs in a narrow temperature range around 540 ℃ at pressures above 4 kbar. X_Ca in garnet decreases with rising temperature, but also with increasing pressure (Fig. 6a). For instance, at elevated pressures (> 6 kbar) there is a clear drop in X_Ca at here relevant temperatures from X_Ca=0.11 to 0.03 within about 50 ℃.

Multivariant fields at HP-LT conditions are characterized by the presence of Na-rich clinopyroxene and carpholite at temperatures below 520 ℃ and pressures higher than 12 kbar. Chlorite occurs at temperatures and pressures below 613 ℃ (9.3 kbar) and ~17.5 kbar (500 ℃), respectively, whereas assemblages with staurolite appear at temperatures above 537 ℃ and pressures below 14.5 kbar. The maximum Mg^# in chlorite is 0.32 (see Fig. 6d). Plagioclase is present at pressures below ~6.4 kbar (650 ℃, 2.2 kbar at 420 ℃). At lower P-T conditions (P ~2 kbar and T ~415 ℃), anorthite (X_An ~0.97) is the stable plagioclase. Kyanite appears at temperatures above 600 ℃ and at pressures above 10 kbar. Rutile occurs in a small P-T field close to 350 ℃ and 2 kbar, but mainly at high-pressures above 10 kbar. Chloritoid is limited to temperatures between 350 and 610 ℃ in the pressure range 2–22 kbar (Fig. 6c). The highest Mg content of chloritoid (Mg^#=0.16) is reached at this temperature limit (Fig. 6c). Biotite appears at T > 510 ℃ and P < 9 kbar. Contents of biotite above 10 vol% occur only at pressures below 5.6 kbar and temperatures above 515 ℃ (Fig. 6c).

The metamorphic assemblage of the studied micaschist (garnet+chloritoid+chlorite+white mica+quartz) is compatible with multivariant P-T fields appearing in Fig. 5a between 470 and 590 ℃ and 6.2 and 17.8 kbar. These fields are located at temperatures below those with staurolite and kyanite.

Isopleths of molar fractions of garnet components, Mg^# of chlorite and chloritoid, Si in phengite, and modal amounts of garnet and biotite were used to estimate P-T conditions of various metamorphic stages. These stages are defined by a specific garnet domain each and related minerals and their compositional domains, which were selected based on mineral inclusions, contact relations and/or compatibility of observed with calculated assemblages at estimated P-T conditions. Error ellipses of the P-T determinations (Fig. 8) were estimated on the basis of analytical errors of mineral compositions and the sensitivity of relevant isopleths to changes in pressure and/or temperature. In addition, small angles between intersecting isopleths, for example, result in ellipses strongly elongated along the bisecting line (see also Li and Massonne, 2018).

Figure 8.  BSE images (a)–(f) showing monazite (Mz) in different textural positions: (a), (b) Mz 6, Mz 6a, and Mz 29 occurring in a domain rich in potassic white mica (Ms); (c) Mz 13 surrounded by chlorite (Ch); (d) Mz 3 and Mz 4 in a quartz-rich matrix; (e) Mz 5 in contact with chloritoid (Ctd) in a quartz-rich matrix; (f) Mz 18 enclosed in Gt2. (g), (h) X-ray concentration maps for Th and Y in monazite grains included in white mica and chlorite. Cold to warm colours indicate increasing element concentrations. The determined Th-Pb ages are reported with 2σ errors. Abbreviations as in Fig. 5.

The isopleths for the inner-most core of garnet (X_Ca=0.095, X_Mg=0.037, X_Mn=0.11) and that for high Si contents in early potassic white mica (Si=3.30 pfu: average of 12 analyses) intersect at 16 kbar and 510 ℃ (Fig. 7). At this HP stage, the calculated mineral assemblage is paragonite (2 vol%), chlorite (1.5 vol%, Mg^#=0.32), rutile, phengite, garnet (~1 vol%), chloritoid (10.5 vol%), and quartz (Fig. 5a). A later metamorphic stage is defined by intersections of isopleths for the garnet outer-core (X_Ca=0.08, X_Mg=0.045, X_Mn=0.08) with that of potassic white mica showing a lower Si content than the mica cores (Si=3.17 pfu, average of 7 analyses) at 10–11 kbar and 535–540 ℃ (Fig. 7). We calculated garnet (1.5 vol%), chlorite (4.5 vol%), chloritoid (7 vol%), paragonite (0.5 vol%) coexisting with potassic white mica, quartz, ilmenite, and rutile for this stage. Further P-T data of the metamorphic track were reconstructed, for instance, by intersecting isopleths for the garnet inner mantle (X_Ca=0.07, X_Mg=0.05, X_Mn=0.07) and a Si-poor (3.11 pfu) potassic white mica at 8 kbar and 550 ℃ (Fig. 7). Along the path from 11 to 8 kbar, the calculated modal content of garnet had increased to almost 2 vol%. The calculated mineral assemblage at 8 kbar and 550 ℃ contains chlorite (~5.1 vol%, Mg^#=0.34) and still some chloritoid (5.5 vol%, Mg^#=0.14) and paragonite (0.5 vol%).

Figure 7.  P-T path (black) reconstructed for metapelite 14AS6 on the basis of intersections of various isopleths (coloured solid lines, Fig. 6) for garnet (Gt) parameters and Si in potassic white mica (Ms). Light grey ellipses, the size of which might roughly reflect the P and T errors of these intersections, refer to the bulk-rock composition (Table 4, bulk 1), whereas grey ellipses refer to the effective bulk-rock composition (bulk 2). The end of the path for micaschist 14AS6 differs by a solid and a broken line (see text for discussion). For comparison, the P-T path (dark grey) for garnet-bearing micaschist 13F18 (see Fig. 1c) reported by Rahimi and Massonne (2018) is displayed. The P-T limits for chlorite, chloritoid, garnet, and muscovite are shown according to the results presented in Fig. 5a. Abbreviations as in Fig. 5.

To decipher the conditions for the late metamorphic stage of the micaschist, the P-T pseudosection for the effective bulk-rock composition (bulk2 in Table 4) was used (Fig. 5b). We constructed intersections of the isopleths for the garnet outer mantle (X_Ca=0.06, X_Mg=0.05, X_Mn=0.05), innermost rim (X_Ca=0.05, X_Mg=0.06, X_Mn=0.05), inner rim (X_Ca=0.04, X_Mg=0.06, X_Mn=0.05) and outer rim (X_Ca=0.03, X_Mg=0.065, X_Mn=0.04) at about 6.2 kbar and 557 ℃, 5.2 kbar and 590 ℃, 5.0 kbar and 600 ℃, and 4.8 kbar and 615 ℃, respectively (Fig. 7). These isopleths for the innermost and inner rim intersect at about 10 ℃ lower temperatures when bulk1 was used instead of bulk2. The calculated mineral assemblage at 6.2 kbar and 557 ℃ involves chlorite (6 vol%), staurolite (2.5 vol%), and ilmenite (1 vol%).

At the peak-T conditions of 615 ℃, derived by the intersection of isopleths for the garnet outer rim, the calculated modal content of garnet has increased to 2.5 vol%. The Si content of potassic white mica has decreased to 3.05 pfu (Fig. 7). Further minerals in the calculated assemblage at 615 ℃ (Fig. 5b) are staurolite (4.5 vol%), plagioclase (0.5 vol%, X_An=0.26), ilmenite (1 vol%), biotite (8 vol%, Mg^#=0.43), and sillimanite (1 vol%).

Using BSE imaging (Scherrer et al., 2000), monazite was found in different microtextural positions. Petrographic and textural features of monazite are summarized in Fig. 8. Monazite is commonly disseminated within the matrix of phyllosilicates (potassic white mica, chlorite; Figs. 8a–8c) and quartz (Fig. 8d), and rarely occurs in chloritoid and garnet (Gt1, Gt2; Fig. 8f).

For this study, an/subhedral, elongated, or oval shaped monazite grains between 30 and 100 µm large were dated. In BSE images and elemental X-ray maps complex zonations and internal features are sometimes discernible. For instance, a grain within a white mica aggregate was found to show a nearly homogeneous distribution of Y and Th (Fig. 8g), whereas another monazite grain, hosted by chlorite, shows an irregular zoning with enrichment of Th and Y in the core and partly in the rim, respectively (Fig. 8h).

Altogether 128 monazite analyses were conducted with the EMP on 87 grains leading to U-Th-Pb ages (Table 5, Fig. 9). Some of these analyses were discarded because of low oxide sums (< 96.0 wt.%) and relatively high SiO₂ contents (> 0.9 wt.%). According to the remaining 113 EMP analyses (Table 5, Fig. 9a), monazite is characterized by Y₂O₃ contents less than 1.5 wt.%. Most monazite analyses yielded ThO₂ and UO₂ contents in the range of 3.6 wt.%–6.3 wt.% and 0.6 wt.%–0.8 wt.%, respectively (Table 5). These contents are more or less similar to those in igneous/high-grade monazite that shows ThO₂ and UO₂ contents of 3 wt.%–6 wt.% and 0.2 wt.%–0.6 wt.%, respectively (Langone et al., 2011; Rasmussen and Muhling, 2007; Schandl and Gorton, 2004). However, monazite formed at medium-grade metamorphism also shows such contents, frequently with somewhat higher UO₂ contents (Massonne, 2014; Rasmussen and Muhling, 2007; Gibson et al., 2004).

Figure 9.  Compositional data of monazite in sample 14AS6. (a) La/Gd (pfu), Y₂O₃ (wt.%)×10, and ThO₂/UO₂ (wt.%) plotted versus the determined age; (b) probability density diagram and (c) weighted average plots for the obtained U-Th-Pb monazite ages. Ages and their errors (2σ) of eight age populations are given in (c).

The obtained monazite ages range between 315 and 480 Ma without any significant age difference of monazite in distinct textural settings (Figs. 8a–8f, 9b, 9c). Due to the successive growth during a metamorphic cycle (Taylor et al., 2016) and possible mixed analyses of different monazite domains, different age groups are not easily discernable. Histogram analysis with the Isoplot program (Ludwig, 1999) reveals, however, eight populations with age maxima at 325, 346, 357, 368, 394, 423, 443, and 479 Ma. The most prominent maxima and side maxima are at 346.0±1.1 (2σ), 357.3±1.7, and 368.3±1.7 Ma (Figs. 9b, 9c). Among the monazite analyses (Fig. 9a), a group with ages ≤386 Ma can be only vaguely distinguished from a group with ages > 386 Ma by mean values for the La/Gd ratio, Y₂O₃ content, and Th/U ratio of 9.8, 1.02 wt.%, and 5.9, respectively, whereas these mean values for the older monazite population are either somewhat higher (La/Gd=10.05) or slightly lower (Y₂O₃=0.93 wt.%, Th/U=5.4).

The above derived P-T conditions and monazite age data will be used to interpret the metamorphic evolution of the studied micaschist from the Elstergebirge.

The prograde P-T path for the studied chloritoid-bearing micaschist consists of two parts (Fig. 7); the first one is an exhumation path to 15–20 km depths (~5 kbar), whereas the second part indicates an isobaric heating event to about 600 ℃. The P-T path starts at 510 ℃ and 16 kbar. This temperature is defined by the inner-core garnet composition, whereas the pressure is constrained by relatively Si-rich potassic white mica (Si=3.30 pfu), found as relic in the core of mica flakes. Even when these garnet and mica domains would not form an equilibrium pair, the peak pressure would hardly be lower than 16 kbar as the corresponding Si isopleth for phengite is little temperature-dependent above 400 ℃ (Fig. 6b). Subsequently, the rock has experienced significant pressure release to about 5 kbar at moderately rising temperature. This pressure release is mainly evidenced by the formation of low-Si potassic white mica at the expense of Si-rich potassic white mica that was only found in core domains of mica flakes. The temperature increase of about 50 ℃ has been deduced from the growth of garnet from its core to the outer mantle and the consideration of an effective bulk-rock composition for late garnet growth. However, selection of various effective bulk-rock compositions, using subtraction of early-formed garnet domains, has no significant effect on the reconstructed P-T path as already shown in earlier works (Massonne et al., 2018a, b; Massonne, 2014). The isobaric heating path at about 5 kbar was exclusively reconstructed using the chemical composition of the garnet rim (see also below).

Along the derived exhumation path, the calculated amount of garnet in the studied micaschist increases from 0.5% to 2.5%. The heating path at 5 kbar has led to further increase in garnet mode, which is then compatible with the observed quantity of garnet (3%) in the rock. Other minerals such as paragonite and chloritoid, that are predicted to have been present at the peak-pressure conditions, were already decomposed during the exhumation of the rock to 20 km depths (~5 kbar). Thus, no paragonite was found in sample 14AS6, but relics of chloritoid exist. The chloritoid amount of 5%–6% (see above) corresponds to the thermodynamic prediction at P-T conditions of 8 kbar and 550 ℃. Absence of subsequent deformation may have contributed to the preservation of chloritoid. Lack of deformation at metamorphic pressures below 8 kbar is documented by preserved skeletal texture of lately grown garnet around the grain boundaries of pre-existing equigranular quartz grains (Figs. 2, 3). This texture is controlled by reduction of interfacial free energy rather than by dislocation creep processes (Hawkins et al., 2007) and can, thus, only be achieved at very low differential stress, which led Stöckhert et al. (1997) to infer that deformation, if at all, was only localized.

Along the exhumation path to 20 km depth, the calculated amount of chlorite has increased from 1.5 vol% to 6 vol%, in contrast to the mode in the natural sample of slightly more than 10 vol%. This means that during a late retrograde stage after the temperature maximum, additional chlorite has grown compatible with textural observations on micaschist 14AS6.

The nearly isobaric heating path to more than 600 ℃ at about 5 kbar should have led to the formation of significant amounts of staurolite (4.5 vol%) and even some sillimanite (1 vol%) but these minerals were not observed in sample 14AS6. We explain this discrepancy by an overestimation of the peak temperature because the composition of the outer garnet rim is probably the result of local equilibria only and can hardly be approached by selection of an effective bulk-rock composition. Thus, we think that the maximum temperatures have only achieved about 590 ℃ although at such temperatures staurolite should have already started to form at the expense of chloritoid and chlorite (Fig. 7). However, crystallization of significant amounts of biotite (5 vol%) proves at least heating of the rock after exhumation to 20 km depth (see Fig. 6c).

Compared to a garnet-bearing, but chloritoid-free and gneissose micaschist (sample 13F18) occurring 11 km to the WSW from the location of sample 14AS6 (Fig. 1, Rahimi and Massonne, 2018), we noted a similar P-T evolution (Fig. 7). The P-T paths of both rocks show the same shape, but the peak-pressure of 13F18 (10 kbar) is significantly lower than the one derived here (16 kbar). In addition, the temperatures of the exhumation path to 5 kbar are slightly lower for 13F18. Both rocks (14AS6, 13F18) experienced a heating event caused by the nearby intrusions of granitic magma at depths between 15 and 20 km, corresponding to 4–5 kbar lithostatic pressure (Rahimi and Massonne, 2018). Although the peak temperature experienced by this event was somewhat lower for 13F18, the impact of the heating is more obvious in this rock due to the formation of staurolite, biotite, and andalusite, whereas in the micaschist 14AS6 only biotite formed newly. Occurrence of these minerals led earlier workers (e.g., Mielke and Schreyer, 1972) to conclude that the micaschists of the northeastern Fichtelgebirge crystalline complex experienced only low- pressure metamorphism although Mielke et al. (1979) already reported the occurrence of phengite (Si ~3.3 pfu), but these authors were not aware of the geobarometric potential of potassic white mica at that time. Presence of this mineral unequivocally documents an early HP metamorphic stage of the studied micaschist before Barrovian-type metamorphism.

The monazite ages, subdivided into 8 populations, range between 315 and 480 Ma. Such extended age variations are common in medium- to high-grade metamorphic monazite (e.g., Massonne, 2014; Palin et al., 2014; Rubatto et al., 2013; Martins et al., 2009; Foster et al., 2002, 2000). Different mechanism could be invoked to explain such age spreads (see, e.g., Foster et al., 2002). According to the discussion in Rahimi and Massonne (2018) we believe that the presence of different chemical-age domains is the result of continuous and/or discontinuous monazite growth. In addition, we cannot exclude that a few age results are due to mixing of different monazite domains during EMP analysis. On the contrary, Pb loss in monazite due to diffusive processes is unlikely for a medium grade metamorphic rock as lead diffusivity is very slow and comparable with that in zircon (e.g., Gardés et al., 2007; Cherniak et al., 2004; Seydoux-Guillaume et al., 2002; Spear and Pyle, 2002). In addition, observations of systematic correlation of intra-crystalline zoning (for example concerning Y and heavy rare-earth elements) with ages (e.g., Williams and Jercinovic, 2012, 2002; Gibson et al., 2004) make this diffusional process largely unlikely to be responsible for wide age variations.

The determined monazite age populations (Fig. 9) show similarities to those reported from a metasediment (sample 13F18; Rahimi and Massonne, 2018) that occurs 11 km further to the WSW (Fig. 1c). There is a strict coincidence for the age cluster at 325 Ma. The cluster around 362 Ma given by Rahimi and Massonne (2018) might be represented in this work by two clusters at 357 and 368 Ma, whereas our cluster at 394 Ma is rather analogous to two clusters at 386 and 405 Ma in the former study. In general, younger monazite ages were more intensively found in sample 13F18 than in the present micaschist. For instance, about one third of the EMP analyses of monazite in 13F18 yielded ages younger than 315 Ma which are completely lacking in monazite of micaschist 16AS6.

Following Rahimi and Massonne (2018), we assign our age clusters at 394 Ma and older to detrital monazite. The provenance areas of the detrital monazite were Early Paleozoic magmatic arcs located at the northern margin of Gondwana or peri-Gondwanan terranes (cf., Waizenhöfer and Massonne, 2017). The age clusters at 357 and 368 Ma are related to an Upper Devonian HP metamorphism and a subsequent exhumation in Upper Devonian to Lower Carboniferous times. This metamorphism was also reported from the nearby complexes of Münchberg, Erbendorf- Vohenstauß, and Mariánské Lázně (Willner at al., 2000; Fig. 1). Rahimi and Massonne (2018) suggested that the exhumation from HP conditions to 5 kbar (20 km depth) had ended already in the Upper Devonian (362 Ma), but according to our new data, the end of exhumation occurred somewhat later which would be still compatible with Ar-Ar cooling ages of ca. 360 Ma obtained for HP rocks WNW of the Elstergebirge (Faryad and Kachlík, 2013). The age cluster at 325 Ma is assigned to the heating at pressures of ~5 kbar (Fig. 7), compatible with the low-pressure metamorphism in the time interval 330–320 Ma (Okrusch et al., 1990) and the emplacement of granite in the adjacent Fichtelgebirge crystalline complex.

The original clastic sediments of the complex of HP micaschists, over 100 km long between the crystalline complexes of Fichtelgebirge and Erzgebirge, were deposited in the Devonian at the northern margin of Gondwana (Rahimi and Massonne, 2018). These protoliths also apply to metasediments with HP signature in the Münchberg metamorphic complex (Waizenhöfer and Massonne, 2017). Burial of these metasediments started with the closure of the Rheic Ocean in the Late Devonian and proceeded through the subsequent continent-continent collision. In this scenario the upper continental plate should be Avalonia, a peri-Gondwanan terrane being accreted to Laurussia already in the Silurian (e.g., Li et al., 2017; Willner et al., 2013; Zeh and Gerdes, 2010; Zeh and Will, 2010). The HP stage was reached after the micaschists were fully overlain by the Avalonian crust. This crust might have been thickened, eventually by a previously active magmatic arc as in the present-day Andes, in order to explain the depth of 55–60 km reached by the Elstergebirge micaschist at its pressure peak of 16 kbar. The exhumation occurred in an exhumation channel, which had developed between the collided plates (Massonne, 2016b). Typically, a P-T loop as shown in Fig. 7 reflects this exhumation process as was observed for similar geotectonic settings elsewhere (e.g., Massonne, 2016a; Massonne and Toulkeridis, 2012; Massonne and Calderón, 2008).

Other metasediments of the belt of HP micaschist that have experienced peak pressures of 10–13 kbar (Rahimi and Massonne, 2018; Rötzler et al., 1998) were involved in the upwards-directed mass flow of the exhumation channel either earlier, before they were overlain by the thickened portion of the Avalonian crust, or later because of a larger distance to the tip of the downgoing plate. The latter hypothesis could be more plausible when we also consider low-grade metasediments such as those outcropping at the margin of the Münchberg metamorphic complex. There, phyllites occur in the prasinite-phyllite- series, a nappe of this complex, showing K-Ar and Ar-Ar ages of metamorphism between 374 and 366 Ma (Koglin et al., 2018). Thus, the low-grade metamorphism was contemporaneous to the HP metamorphism of the micaschists and suggests that the low-grade metasediments were not overridden by thick Avalonian crust in contrast to the Elstergebirge micaschists.

A further important geotectonic event is the stacking of a crystalline nappe pile that probably occurred in the Visean due to the collision of Gondwana and a northerly situated peri- Gondwanan terrane (northern Avalonia; the southern portion of Avalonia was the overriding plate in the Upper Devonian). This collision followed the closure of a short-lived ocean (Saxothuringian Ocean according to Schulmann et al., 2009) probably around 340 Ma, a common age of metamorphism in the Bohemian Massif (Willner et al., 2000). The Lower Carboniferous crystalline nappe pile was transported to the north or northwest over the autochthonous, low grade metapsammopelites with intercalated acidic metavolcanics occurring to the north and west of our sample location (Fig. 1c) as equivalents of Cambro– Ordovician sediments of the Thuringian lithofacies. Within the pile, a nappe of HP metasediments occurred. However, this nappe does not present a coherent crustal section. By contrast, it is composed of metasedimentary slices that are characterized by similar P-T trajectories but clearly distinct peak-pressure conditions (10–16 kbar). We assume that the mixing of these slices has already occurred in the Upper Devonian exhumation channel as suggested above.

The heating of the Elstergebirge micaschists to about 590 ℃ at 325 Ma was probably caused by the older post-tectonic granitic magmas of the Fichtelgebirge batholith (Rahimi and Massonne, 2018; Siebel et al., 2010). At that time the crystalline nappe pile was already in place. The period of time after 325 Ma is characterized by Late Variscan transpressional tectonics. Associated advective heat transfer to upper crustal levels from exhumed rocks and granitic melts is the likely reason for wide-spread regional high-temperature, low-pressure metamorphism in the late Lower Carboniferous (Kroner et al., 2007).

Petrographic, phase-equilibrium, and geochronological study of a chloritoid-bearing micaschist from the Elstergebirge in the SZ documents the following scenario: (1) HP metamorphism at about 16 kbar and 510 ℃ (Fig. 7) was followed by exhumation to 20 km depth (~5 kbar) and heating to 555 ℃. Subsequently, an isobaric heating to almost 600 ℃ occurred. A similar metamorphic path with lower peak pressure (10 kbar) was already reported from a gneissose micaschist located 11 km WSW of the present location (Rahimi and Massonne, 2018). Both HP micaschists are inconsistent with early hypothesis of low-pressure metamorphism (Okrusch et al., 1990; Mielke et al., 1979). (2) EMP monazite dating suggests that the HP metamorphism occurred in the Upper Devonian (Fig. 9). The subsequent exhumation to 20 km depth ended in the Early Carboniferous. The isobaric heating event is related to the intrusion of granitic magmas.

We envisage the following scenario: The burial of the protoliths of the micaschists started with the closure of the Rheic Ocean in the Late Devonian and evolved into continent- continent collision. HP conditions were reached when the metasediments were overlain by thick Avalonian crust. Exhumation occurred in an exhumation channel. It remains unclear if the stacking of a crystalline nappe pile has already occurred in the exhumation channel or later. A likely al ternative for the stacking in the Visean is a continent-continent collision after a short-lived ocean (Saxothuringian Ocean?) was closed. The nappe pile has already been in place in the Sepukhovian as the intrusion of granites at 325 Ma is clearly post-tectonic.

We interpret the HP micaschists from the Fichtelgebirge- Erzgebirge crystalline complexes as a single nappe unit (cf. Faryad and Kachlík, 2013), although these rocks show a range of peak-pressure conditions between 10 and 16 kbar. This range is taken as hint at different metasedimentary slices of which this nappe unit is composed of.

The authors thank Thomas Theye (Stuttgart) for supporting their EMP work. Two anonymous reviewers and Bin Xia from CUG (Wuhan) contributed significantly to the improvement of an earlier version of this paper. The final publication is available at Springer via https://doi.org/10.1007/s12583-020-1300-3.