Subduction of a rifted passive continental margin: the Pohorje case of Eastern Alps–constraints from geochronology and geochemistry

This study presents geochronological and geochemical data from newly dated Permian granitic orthogneisses associated with the Eclogite-Gneiss unit (EGU) from the southernmost part of the Austroalpine nappe stack, exposed within the Pohorje Mountains (Slovenia). LA-ICP-MS zircon U–Pb ages of two samples of the augen-gneisses are 255 ± 2.2 Ma and 260 ± 0.81 Ma, which are interpreted as the age of magmatic crystallization of zircon. In contrast, all round zircons from leucogneisses give Cretaceous ages (89.3 ± 0.7 Ma and 90.8 ± 1.2 Ma), considered as the age of UHP/HP metamorphism. The round zircons overgrew older euhedral zircons of Permian and rare older ages tentatively indicating that these rocks are of latest Permian age, too. Zircon εHf(t) values of the four orthogneiss samples are between − 13.7 and − 1.7 with an initial 176Hf/177Hf ratio ranging from 0.282201 to 0.282562; TDMC is Proterozoic. The augen-gneisses show geochemical features, e.g. high (La/Lu)N ratios and strong negative Eu anomalies, of an evolved granitic magma derived from continental crust. The leucogneisses are more heterogeneously composed and are granitic to granodioritic in composition and associated with eclogites and ultramafic cumulates of oceanic affinity. We argue that the Permian granitic orthogneisses might be derived from partial melting of lower crust in a rift zone. We consider, therefore, that segment of the EGU is part of the distal Late Permian rift zone, which finally led to the opening of the Meliata Ocean during Middle Triassic times. If true, the new data also imply that the Permian stretched continental crust was potentially not much wider than ca. 100 km, was subducted and then rapidly exhumed during early Late Cretaceous times.


Introduction
A-type subduction is considered to occur at the initial stage of continent-continent collision after consumption of an intervening oceanic lithosphere and subduction of the stretched passive margin. In many cases, the UHP/ HP metamorphic conditions are well known but data on the type of subducted continental crust is lacking. In terms of end members, the type of subducted continental crust is either (1) normal thick continental crust or (2) the crust of a rift zone later transformed to a passive margin, which is influenced by strong extension, hightemperature metamorphism due to thinning of even the subcontinental mantle lithosphere and intense bimodal magmatism. To resolve these alternative scenarios, we investigated the southernmost part of the Eclogite-Gneiss Unit (EGU) and associated units of Cretaceous metamorphic age (Fig. 1a), which is part of the Austroalpine nappe stack and which is exposed in the Pohorje Mountains in the southeasternmost part of the Eastern Alps (Fig. 1b). R. Chang et al. There, UHP eclogites and ultramafic mantle rocks are exposed in a matrix of paragneiss and associated with hitherto undated granitic orthogneisses (Kirst et al. 2010;Janák et al. 2015).
The Alps are the type continent-continent collisional orogenic belt (Schmid et al. 2004). The Eastern Alps in particular, are the result of the convergence of two independent Alpidic collisional orogenic belts (Neubauer et al. 2000;Froitzheim et al. 2008). The main tectonic events of the Austroalpine nappe stack, which stretches from the Eastern Alps to the Western Carpathians, are (1) rifting and lithospheric thinning in the Permian following the Variscan orogeny and deposition of Late Carboniferous Variscan molasse; (2) Middle Triassic opening of the oceanic Meliata basin; (3) Jurassic convergence and subduction of the Meliata basin; (4) internal nappe stacking within the Austroalpine nappe complex after closure of the Meliata oceanic basin during Early Cretaceous times; and (5) Paleogene collision of the southern margin of the stable European continent and the Austroalpine nappe complex after the subduction of the South Penninic ocean under the Austroalpine microplate (Neubauer et al. 2000;Schuster et al. 2008;Froitzheim et al. 2008;Thöni et al. 2008;Janák et al. 2004Janák et al. , 2006Janák et al. , 2009Sassi et al. 2004;Miller et al. 2005aMiller et al. , 2005bMiller et al. , 2007Bruand et al. 2010;Kirst et al. 2010).
This paper presents the first evidence for Permian granites within or associated with EGU exposed in the Pohorje Mountains of Slovenia. We describe here the petrography, whole-rock geochemistry, zircon U-Pb and Hf isotopic data of four samples, which represent two distinct groups of orthogneisses. We use these data to constrain the petrogenesis of the two types of orthogneisses and their tectonic setting. Integrating collected data with results of previous studies of eclogite, gabbro and ultramafic rocks, we discuss the significance of Permian magmatic and tectonic events, interpret these as rifting followed by Middle-Late Triassic passive margin formation and discuss then the subduction processes of this rifted passive continental margin.

Geological setting
The Pohorje Mountains in NE Slovenia are located at the southeastern part of the Eastern Alps just north of the Periadriatic Line (Fig. 1). Together with the Saualpe and Koralpe areas exposed in the north and some small remnants in the Sieggraben at the eastern termination of Eastern Alps, the Pohorje Mts. belong to the above mentioned Eclogite-Gneiss unit, recently considered as a part of the Koralpe-Wölz nappe system, which is part of the Lower Central Austroalpine units with respect to the Eo-Alpine UHP/HP subduction zone (Janák et al. 2004. Kirst et al. 2010;Sandmann et al. 2016;Fig. 2) and was labeled as Middle Austroalpine in the previous literature (e.g., Neubauer et al. 2000). EGU comprises a succession of metamorphic rocks with continental affinity and includes various paragneisses, rare marble and quartzites and intercalated eclogites, rare metagabbro and pegmatites (e.g. Miller et al. 2005aMiller et al. , 2005b. At its southernmost exposures, this succession is in direct contact with a km-sized ultramafic body termed as Slovenska Bistrica Ultramafic Complex (Kirst et al. 2010) (Fig. 3). Based on widespread andalusite paramorphs in paragneisses, EGU is affected by Permian low-pressure metamorphism, whereas the presence of Variscan metamorphism is uncertain. During early Late Cretaceous, these rocks were subducted towards the SE beneath the units of the Upper Central Austroalpine. The subduction polarity is indicated by a south-eastward increase in peak pressures along the Koralpe/Pohorje traverse (Janák et al. 2004;Bruand et al. 2010;Sandmann et al. 2016).
In the Pohorje Mts., the uppermost nappe is the southeastern extension of the Gurktal nappe system and comprises low-grade metamorphic Paleozoic metasediments overlain by non-metamorphic Permo-Triassic sedimentary rocks (Fig. 2). The metamorphic series of the underlying Eclogite-Gneiss unit (here locally termed Pohorje nappe) mainly comprises gneisses and schists with minor marbles and quartzites and embedded bodies and lenses of eclogite and amphibolite and few orthogneisses and augen gneisses (Janák et al. 2004;Sandmann et al. 2016). At the south-eastern margin of the Pohorje Mts., the Slovenska Bistrica Ultramafic Complex (SBUC) extends over 8 km in E-W direction. This body consists of serpentinized ultramafites, eclogites and amphibolites as well as minor garnet-bearing ultramafic rocks (Sassi et al. 2004;Janák et al. 2006;Kirst et al. 2010;De Hoog et al. 2009, 2011. The Pohorje Mts. are intruded by the Miocene Pohorje granodiorite to tonalite (Fodor et al. 2008).
Two stages of metamorphism affected the EGU. The dominant HP-UHP metamorphism is preceded by a HT-LP metamorphic event, expressed by andalusite pseudomorphs. It is interpreted as Permian rifting event (Schuster and Stüwe 2008) causing metamorphism of Paleozoic and older sediments of continental and oceanic affinity. Rifting also led to underplating and emplacement of gabbros in the Koralpe region into the thinned continental crust (Thöni and Jagoutz 1992;Thöni et al. 2008). Sm-Nd ages of these metagabbros are 275 ± 18 Ma for protolith crystallization and 93 ± 15 Ma for eclogite metamorphism (Thöni and Jagoutz 1992). Furthermore, numerous seemingly rootless pegmatite bodies of Permian age are preserved in the EGU and in underlying and overlying units (Knoll et al. 2018 and references therein; Fig. 1b).
The timing of UHP/HP metamorphism in the Pohorje Mts. is generally accepted to be Eo-Alpine (i.e. Cretaceous). U-Pb zircon ages of eclogites and metapelites are c. 93-90 Ma, Miller et al. 2005b). Sm-Nd garnet ages of eclogites from the Saualpe, Koralpe and Pohorje areas are in the range of 108 to 87 Ma (Thöni 2002). K-Ar and Ar-Ar muscovite ages are 85-90 Ma in the northern part of the nappe system (as well as in the underlying Gleinalpe-Stubalpe region; Neubauer et al. 1995) and about 75-80 Ma in the central part (Saualpe-Koralpe region) (Miller et al. 2005b;Wiesinger et al. 2006;Thöni et al. 2008), Lu-Hf ages for Grt-Omp and Grt-Cpx of eclogites and ultramafic rocks, respectively, from the Pohorje EGU are between 97 and 92 Ma (Sandmann et al. 2016). These geochronological data indicate a maximum burial of the Eclogite-Gneiss unit in the Cenomanian and Turonian (Thöni et al. 2008), followed by rapid exhumation in the Coniacian and Santonian (Wiesinger et al. 2006). Wiesinger et al. (2006) documented a N-directed thrust fault at the base of EGU, whereas the upper contact is a ductile low-angle normal fault (Ratschbacher et al. 1989;Wiesinger et al. 2006), later called Plankogel detachment (Schorn and Stüwe 2016). Neogene fault systems have a major influence on the recent morphology and are responsible for the final exhumation of the Saualpe-Koralpe Complex in the Pohorje region (Ratschbacher et al. 1989(Ratschbacher et al. , 1991. The Miocene Pohorje granodiorite reveals that the region was still buried during Early Miocene, whereas in the surrounding Pannonian basin sediment deposition was already ongoing (Fodor et al. 2008).

Samples and petrography
In this study, we focus on two augengneiss, and two leucogneiss samples. All rocks are deformed and metamorphic granitic gneisses, but show quite different characteristics. The augengneisses are Bt-gneisses which still show some magmatic features, the leucogneiss samples are Grt-Ms-gneisses. The protolith of the augengneisses intruded a sequence of gneisses with intercalations of micaschists, amphibolite and marbles (Kirst et al. 2010), the leucogneiss samples are associated with eclogites, Btgneisses and UHP-paragneisses and micaschists (Janák et al. 2015).
The two leucogneiss samples (SK10 and SK12) were collected ~ 1 km east and 1.5 km west of Veliko Tinje village, respectively (Fig. 3, Table 1). According to the map of Kirst et al. (2010) sample SK10 is from a leucogneiss body, sample SK12 from a biotite-gneiss unit. Both leucogneiss samples consist of quartz, plagioclase, K-feldspar, white mica, garnet, and some late-stage biotite (Fig. 4c, d, e, f ). Zircon, apatite, and magnetite occur as Microstructures of augengneisses and leucogneiss from the Pohorje area. For explanation, see text. Images combined from microphotographs with both crossedand parallel (greyish photomicrographs) polarizers. Sample numbers in the lower left, scale in the lower right corner. Qz, quartz; Pl, plagioclase; Kfs, K-feldspar; Bt, biotite; Ms, white mica; Grt, garnet minor constituents. Simple inclusion-free garnet porphyroblasts are common (Fig. 4c, f ). Large K-feldspar phenocryst show graphic intergrows with quartz. Myrmekites are often developed in K-feldpar in contact with plagioclase ( Fig. 4d, e). Quartz grains with an average size from 0.1 to 0.8 mm indicate dynamic recrystallization by sub-grain rotation and grain boundary migration (Fig. 4g).
Zircon grains were extracted from samples using conventional density and magnetic separation techniques at the Yuneng Mineral Separation Company, Hebei Province. Over 500 zircons were handpicked under a binocular microscope, then mounted in epoxy resin and polished until the grain centers were exposed. To remove any lead contamination, the surface was cleaned using 3% HNO 3 prior to analysis. Cathodoluminescence (CL) images were obtained using a Mono CL3 + microprobe, in order to characterize internal structures and to choose potential target sites for U-Pb dating.
Measurements of U, Th, and Pb isotope data and trace element compositions were conducted using a high-precision laser ablation-inductively coupled plasma-mass spectrometer (LA-MC-ICP-MS) at the Beijing Createch Testing Technology Co., Ltd. Detailed operating conditions for the laser ablation system and the MC-ICP-MS instrument and data reduction are the same as described by Hou et al. (2009). Laser sampling was performed using a Resolution 193 nm laser ablation system. A 24 μm spot size was adopted in this study with a laser repetition rate of 6 Hz and energy density up to 6 J/cm 2 . Helium was applied as a carrier gas from the sample chamber to the analyzer. An Agilent 7500 ICP-MS instrument was used to acquire ion-signal intensities. Each analysis incorporated background acquisition of approximately 15-20 s (gas blank) followed by 45 s data acquisition from the sample. Off-line raw data selection and integration of background and analytical signals, and time-drift correction and quantitative calibration for U-Pb dating was performed by ICPMSDataCal (Liu et al. 2010).
Zircon GJ-1 (Jackson et al. 2004) was used as external standard for U-Pb dating, and was analyzed twice every 5-10 analyses. Time-dependent drifts of U-Th-Pb isotopic ratios were corrected using a linear interpolation (with time) for every 5-10 analyses according to the variations of GJ-1 (i.e., 2 zircon GJ-1 + 5-10 samples + 2 zircon GJ-1) (Liu et al. 2010). Uncertainty of preferred values for the external standard GJ-1 (with an age of 599.6 ± 2.9 Ma) was propagated to the ultimate results of the samples. No common Pb correction was necessary due to the low signal of common 204 Pb and high 206 Pb/ 204 Pb ratios. U, Th and Pb concentrations were calibrated by NIST 610. Concordia diagrams and weighted mean calculations were made using Isoplot/Ex_ver3 (Ludwig 2003).

Zircon Hf isotope analysis
In situ zircon Hf isotope analyses were performed on the same zircon grains that were subjected to U-Pb dating at the Beijing Createch Testing Technology Co. Ltd., China, using a Resolution SE 193 nm laser-ablation system attached to a Thermo Fisher Scientific Neptune Plus ICP. Instrumental conditions and data acquisition protocols were described by Hou et al. (2007). A stationary spot with a beam diameter of ~ 38 μm was used for ablation. As carrier gas helium was used to transport the ablated sample aerosol mixed with argon from the laser-ablation cell to the MC-ICP-MS torch. 176 Lu/ 175 Lu = 0.02658 and 176 Yb/ 173 Yb = 0.796218 ratios were applied to correct for the isobaric interferences of 176 Lu and 176 Yb on 176 Hf. For instrumental mass bias correction Yb isotope ratios were normalized to 172 Yb/ 173 Yb = 1.35274 and Hf isotope ratios to 179 Hf/ 177 Hf = 0.7325 using an exponential law. The mass bias behavior of Lu was assumed to follow that of Yb, the mass bias correction protocol was described by Hou et al. (2007). Further external adjustment was not applied to the unknowns because our determined 176 Hf/ 177 Hf ratios of 0.282004 ± 38 (2SD, n = 107) for zircon standard GJ-1 are in good agreement with the reported values (0.282000 ± 0.000005) (Morel et al. 2008). Hafnium isotopic data are age-corrected using the 176 Lu decay constant of 1.867 × 10 −11 a −1 (Soderlund et al. 2004). εHf(t) values and Hf model ages were calculated using the methods of Bouvier et al. (2008) and Griffin et al. (2002), respectively.

Whole-rock major and trace element determinations
In addition to the four dated samples, we collected in each outcrop several additional samples for geochemical analysis. These are similar in petrography varying slightly in the amount of main minerals and microfabrics. The major and trace element compositions of 13 multiple samples from the four magmatic rocks from the studied area were determined by X-ray fluorescence (XRF-1800; Shimadzu) on fused glasses and inductively coupled plasma mass spectrometry (Agilent 7500ce ICP-MS) at Beijing Createch Test Technology Co. Ltd. Prior to analysis, all samples were trimmed to remove weathered surfaces before being cleaned with deionized water and crushed to 200 mesh in an agate mill. Sample powders (~ 40 mg) were digested using HNO 3 and HF acids in Teflon bombs. Loss-of-ignition (LOI) values were measured after heating 1 g of sample in a furnace at 1000 °C for several hours in a muffle furnace. The precision of the XRF analyses is within ± 2% for the oxides greater than 0.5 wt% and within ± 5% for the oxides greater than 0.1 wt%.
Sample powders (about 50 mg) were dissolved in Teflon bombs using a HF + HNO 3 mixture for 48 h at about 190 °C. The solution was evaporated to incipient dryness, dissolved by concentrated HNO 3 and evaporated at 150 °C to dispel the fluorides. The samples were diluted to about 100 g for analysis after redissolving in 30% HNO 3 overnight. An internal standard solution containing the element Rh was used to monitor signal drift during analysis. Analytical results for USGS standards for ICP-MS (Jochum et al. 2005) indicate that the accuracy for most elements were within 5%.

U-Pb zircon dating
Zircons in CL images show that igneous zircons with oscillatory zoning in Fig. 5 but metamorphic ones with relic cores in Fig. 6. The zircon U-Pb isotopic data are given in Additional file 1: Table S1 and are graphically shown in Figs. 7 and 8. The ages given in the text are concordia ages except for the few inherited ages of more than 1000 Ma, for which the 206 Pb/ 238 Pb ages are given.
The zircon grains of the two augengneiss samples SK1 and SK2 are mostly subhedral-euhedral, generally elongated and prismatic and range in length from 100 to 250 μm and in width from 60-100 μm. Most zircon grains exhibit a well-developed oscillatory zoning in CL images (Fig. 5); only a minority of grains contain inherited cores. These characteristics, along with their 67-3631 ppm U and 98-641 ppm Th contents resulting in relatively high Th/U ratios of 0.12-1.30 (except for two zircon rims with 0.08-0.09; Additional file 1: Table S1, Fig. 9a), are typical of magmatic zircons (Hoskin and Schaltegger 2003;Corfu et al. 2003). The chondrite-normalized REE patterns of these zircon grains were dated by the same analytical techniques with U-Pb age on the same spot. Results are given in Additional file 2: Table S2 and are graphically shown in Fig. 9a, reveal a strong depletion in light REEs relative to heavy REEs, and are characterized by strongly negative Eu anomalies, which indicate the presence of plagioclase during the zircon growth (Hoskin and Schaltegger 2003;Rubatto, 2002;. Moreover, relatively steep HREE patterns suggest absence of garnet and low-pressure conditions during the formation of these old zircon domains ) while positive Ce anomalies and negative Eu anomalies are nearly ubiquitous in crustal magmatic zircon (Fig. 9b), Lu/Tb ratios are higher than 5 and up to 10 (Fig. 9a). These geochemical characteristics further support an igneous origin for these zircons (Grimes et al. 2007). Twenty-seven of the 30 analyses of magmatic grains of sample SK1 yield a concordant age of 260.5 ± 0.81 Ma (MSWD of concordance = 4.7; Fig. 7), twenty-nine from thirty analyses of . We interpret these age data as the time of zircon crystallization in the magma and thus the formation age of the porphyric granites. A few zircon cores have concordia ages that range from 640 ± 15 Ma to 2074 ± 30.4 (Additional file 1: Table S1).
The zircon grains of the two leucogneisses SK10 and SK12 are rounded with only a slightly discernible oscillatory zoning and Th/U ratios of 0.005 to 0.009, except several higher ratios of distinct cores (Additional file 1: Table S1, Fig. 9a). Some grains, however, have inherited cores, with a more clearly discernible oscillatory zoning. The chondrite-normalized REE patterns of these zircon grains show markedly lower REE contents than zircons from the augengneisses (Fig. 9b), There is a clear distinction between REE patterns of old and young domains in the investigated zircons ( Fig. 9). They show a relatively flat HREE pattern compared to the HREE enrichment in magmatic zircon, which indicate that these are of metamorphic origin and formed in the presence of a HREEbearing phase; garnet is commonly a main host of HREE in medium to high-grade mafic to pelitic metamorphic rocks (Rubatto 2017). The absence of a strong negative Eu anomaly suggests growth of zircons in the absence of plagioclase. There are a few magmatic zircons with Eu anomalies in cores. Except sample SK1, all samples have a pronounced positive Ce anomaly typical for crustal magmatic zircons (Hoskin and Schaltegger. 2003). The Lu/Tb composition of zircon can distinguish between magmatic zircons which show increasing normalized HREE-patterns and metamorphic zircons that grew in equiblibrium with garnet which show very flat HREE-distributions. In our samples, the normalized Lu/Tb ratios of metamorphic zircons is generally < 5, just several cores in zircon show high Lu/Tb ratios. Especially in sample SK10, these zircons reveal the presence of strongly luminescent cores and darker rims consistent with radiation damage suppression of CL in rims with high U contents (Fig. 6). The size of seven zircon cores are large enough to enable U-Pb isotopic dating. The discussion focuses on data obtained from analyzed zircons that record a primary magmatic growth. 31 of 40 analyses of sample SK10 are concordant and 17 of them yield a concordant weighted mean average 206 Pb/ 238 U age of 89.6 ± 0.6 Ma (MSWD of concordance = 0.71) (Fig. 8). Inherited cores of zircon grains have ages ranging between 255 ± 6 Ma and 281 ± 5 Ma and a mode at about 265 Ma; only one analyzed core is much older at 2103 ± 29 Ma (Fig. 6). In sample SK12, 29 of 40 analyses are concordant and 24 spots yielded a 206 Pb/ 238 U weighted mean age of 90.5 ± 0.5 Ma (MSWD of concordance = 0.0057) (Fig. 8). The ages of zircon cores range between 260 ± 3 Ma and 1489 ± 32 Ma (Additional file 1: Table S1, Fig. 8), but there is no grouping of ages unlike for sample SK10. Only Based on the CL-attributes, the REE contents and patterns, Lu/Tb ratios and the ages the zircons must represent metamorphic growth or recrystallization, thus dating the HP-event. The age group at about 265 Ma from oscillatory zoned zircon cores could be interpreted as magmatic formation ages.

Crystallization temperatures
We calculated the formation temperatures of the orthogneiss samples from the Pohorje area with the whole-rock zirconium saturation thermometer (Boehnke et al. 2013;Watson and Harrison 1983). The zirconium saturation temperatures can provide an estimate of the initial magma temperatures of granites based on the Zr content and a correction term for the major element chemistry of the whole rock (Boehnke et al. 2013;King et al. 1997;Miller et al. 2003;Watson and Harrison 1983). The input parameters for the thermometer and the calculated zirconium saturation temperatures of the orthogneiss samples are listed in Additional file 3: Table S3.
Based on the very different Zr contents and M factors and the resulting magmatic crystallization temperatures, the orthogneiss samples can be divided into two types: a low and a high temperature type, consistent with the previous grouping. The augengneisses give lower temperatures of ~ 500-550 °C, whereas the leucogneisses yield higher formation temperatures of ~ 700-770 °C. Though the calculated temperatures for the augeneisses are unrealistic, this temperature-based subdivision corresponds to the zircon-based subdivision presented above ( Table 2). Subduction of a rifted passive continental margin

Zircon Hf isotopic data
Thirty-five zircon grains of samples SK1, SK2 and SK12 were analyzed for Hf isotopes (Table 2; Fig. 10). The 176 Hf/ 177 Hf(t) values of 20 analyses of magmatic zircon grains from SK1 and SK2 range between 0.282703 and 0.282868 (t = 255 Ma), which correspond to ɛHf(t) values between − 6.4 and − 1.7 and crustal model ages (T C DM ) of 1392 to 1617 Ma (Table 2). An inherited zircon core exhibits an ɛHf(t) value of − 6.6 (t = 640 Ma, 176 Hf/ 177 Hfi = 0.282201), which corresponds to a crustal model ages (T C DM ) of 1972 Ma. Eight analyses of zircons from the leucogneiss sample SK12 gave 176 Hf/ 177 Hf values of 0.282385-0.282562. These analyses correspond to εHf(t) values of − 13.7 to − 7.9 and yield Hf crustal model ages (T C DM ) of 969-1195 Ma. These data suggest a predominantly Proterozoic crustal source for both the augengneisses and the leucogneiss with both orthogneiss groups on the same Hf isotope evolution line.

Geochemistry
In order to chemically characterize and constrain geodynamic settings of the plutonic rocks we analyzed whole-rock major and trace element data which are listed in Table 3. We can use, with due caution regarding the strong Alpine metamorphic overprint, major elements (e.g., Na 2 O, K 2 O and MgO) combined with high-fieldstrength elements (HFSE; e.g., Zr, Hf, Nb, Ta, and Ti), transitional elements (e.g., Ni, Cr, V, and Sc), and rare   The augengneiss samples (SK1 and SK2) have relatively high SiO 2 (66.0-76.3 wt%) and medium to high K 2 O contents of 1.58-4.18 wt%. They are characterized by moderate contents of Al 2 O 3 (14.4-16.1 wt%), Fe 2 O 3 T (1.58-4.18 wt%), and low TiO 2 (0.53-1.30 wt%), MnO (0.03-0.08 wt%), which are typical of metaluminous rocks according to the classification of Villaseca et al. (1998). All samples have Na 2 O/K 2 O ratios less than 1 (0.58-0.70), and fall into the high-K calc-alkaline and calc-alkaline fields of the SiO 2 -K 2 O diagram (Peccerillo and Taylor 1976) and the granodiorite field in the SiO 2 -(Na 2 O + K 2 O) diagram (Middlemost 1994). According to the A-B classification after Debon and Le Fort (1983) the augengneisses are metaluminous. But as K, Na, and Ca, are mobile elements, migration of these elements may occur during metamorphism and thus affect the classification result, we also use the Th vs. Co diagram (Hastie et al. 2007) to classify the samples (Fig. 11), This diagram yields results that are consistent with major elements classification, suggesting no major alteration of whole-rock compositions.
In contrast, the leucogneiss samples SK12 and SK10 exhibit moderate Al 2 O 3 (15.0-15.6 wt%) and low TiO 2 (0.07-0.11 wt%), MnO (0.05-0.10 wt%) and Fe 2 O 3 T (0.54-1.08 wt%) contents. All samples have a high Na 2 O/ K 2 O ratio, with sample SK10 of c. 1.9, sample SK12 of 1.2-0.85 with the former plotting in the calc-alkaline, the latter in the high-K calc-alkaline field of the SiO 2 -K 2 O diagram (Peccerillo and Taylor 1976). The leucogneisses are peraluminous according to Villaseca et al. (1998) and the A-B classification of Debon and Le Fort (1983), and plot into the granite field in the SiO 2 -(Na 2 O + K 2 O) diagram (Middlemost et al. 1994). Also, according to the Co-Th diagram (Hastie et al. 2007) the leucogneisses are high-K calc-alkaline rocks.
The trace element contents of the two groups show quite different patterns with the augengneisses in general more enriched than the leucogneisses. In primitive-mantle-normalized multi-element diagrams (McDonough and Sun 1995), both groups of gneisses show enrichment in large-ion lithophile elements (LILEs; e.g., Cs, Rb, and K) with pronounced negative anomalies of Nb and P, (Fig. 12a). The augengneisses display also relative negative anomalies for Ba, Sr and Eu. Compared with augengneiss samples with positive anomalies for Ti and high La N /Yb N ratios (15.4-33.6), the leucogneiss samples display lower normalized light REE abundances and pronounced negative anomalies for Ti as well as lower La N / Yb N ratios (0. 42-8.27). In addition, they display positive anomalies for U, K, and Sr. Positive and negative Ti anomalies in analyzed augengneisses and leucogneisses, respectively, are interpreted to be due to the extraction of Ti in biotite crystallizing in augengneiss during the melt evolution.
Chondrite-normalized REE patterns are plotted in Fig. 11b, and display for all augengneiss samples high total contents (182.7-275.5 ppm; Table 3), whereas leucogneisses have much lower REE contents (18.0-57.9 ppm; Table 3). Augengneisses have especially much more enriched LREE contents than leucogneisses and show a strong decrease in enrichments from light to heavy REEs. Leucogneisses show also a decrease of enrichments of LREE and almost flat HREE patterns. These patterns are    Peccerillo and Taylor (1976). Classification plots of Na 2 O + K 2 O versus SiO 2 is from Middlemost (1994). The A-B diagram is from Debon and Le Fort (1983), the Co-Th classification diagram from Hastie et al. (2007) also expressed in the (La/Yb) N ratios (Table 3) with about 15-20 for the augengneisses (one sample 34) and about 2-8 for the leucogneisses (one sample 0.4). Sr/Y ratios show also large differences between the two groups with 7-12 (one sample 25) for augengneisses and 16-49 (one sample 5) for the leucogneisses. The augengneisses have strong negative Eu anomalies with a narrow range of Eu/ Eu* ratios of 0.44-0.51 with the exception of one sample with 0.69 (Table 3). Leucogneisses have smaller negative Eu anomalies with Eu/Eu* ratios of 0.63-0.95. However, one analysis of the leucogneiss sample SK10 displays a very flat REE pattern and the most pronounced negative Eu anomaly with a Eu/Eu* of 0.23. Specifically, the augengneisses plot mostly in the field of A-type granite in the Zr + Nb + Ce + Y vs. FeOt/MgO and (Na 2 O + K 2 O)/CaO diagrams of Whalen et al. (1987) (Fig. 13). In the Nb-Y-3 Ga diagram of Eby (1992), which allows a further division of A-type granites, four samples of augengneiss plot in the A1 field, another sample in the A2 field. Samples SK1 plot only just in the field A2, the field of post-collisional to post-orogenic setting, sample SK2 in the field A1, the field for rift-to plume-related granites (Eby 1992).

Discussion
The study demonstrates that the two orthogneiss groups show very different characteristics in geochemistry, properties of zircons and their U-Pb ages and also in the type and metamorphism of their country rocks. We will therefore discuss first the differences of the two groups, then the significance of the zircon ages and the petrogenesis of the two groups separately and then their relationship and the consequence for the geodynamics interpretation of the Austroalpine unit and the Pohorje area in particular. There are strong indications that the two orthogneiss groups belong to two different Alpine tectonic units, not to a coherent unit as assumed by most previous researchers (e.g. Kirst et al. 2010, Janák et al. 2015, which would of course have implications for the interpretation of the genesis and setting of the two plutonic suites, the main topic of this paper. A detailed mapping and the discussion of the Alpine tectonic evolution of the two supposed units is subject of further investigations.

Differences between augengneiss and leucogneiss
The augengneisses and leucogneisses were overprinted during the Alpine metamorphism with variable intensity leading to various structural and chemical changes to the magmatic features. The microfabrics of augengneisses with recrystallization of feldspars indicate deformation at at least upper greenschist facies conditions, but do not allow tight constraints on deformation temperatures. But the stability and recrystallization of biotite and plagioclase, and the absence of HP to UHP minerals in augengneisses, as white mica, garnet or rutile, indicate medium-pressure conditions during the Alpine metamorphism. This metamorphic facies is also indicated by the assemblage Pl-Hbl in amphibolites and especially the paragenesis of St-Grt-Ms-Pl-Ky-Sil-Bt in micaschists (own observations). The well preserved magmatic zircons of the augengneiss, in contrast to the metamorphic (re)crystallized zircons in the investigated leucogneiss samples is a further strong hint that the two orthogneiss groups experienced quite different PT conditions during the Cretaceous metamorphic overprint. We consider, therefore, that two tectonic units are exposed in the southernmost Pohorje: (1) EGU with UHP/HP metamorphism hosting the leucogneisses and (2) the augengneisbearing unit, which is associated with micaschists, with amphibolite facies-grade metamorphism. Fig. 12 Primitive-mantle-normalized multi-element variation diagram (a) from Sun and McDonough (1989) and chondrite-normalized REE patterns (b) from Boynton (1984) of augengneisses and leucogneisses from the Pohorje area

Age and petrogenesis of augengneiss
The augengneiss body is dated at approximately 255-260 Ma on zircon by U-Pb LA-ICP-MS ( Fig. 7; Additional file 1: Table S1), which is Late Permian according to the International Chronostratigraphic Chart v2020/01 of the ICS (Cohen et al. 2013). The mostly porphyric granodiorites intruded into a sequence of gneisses, micaschists, marbles and amphibolites.
The magmatic protolith of the augengneisses is a metaluminous granodiorite, which shows several characteristics of A-type granites as a strong enrichment in incompatible trace elements, including LILE and HFSE, but a relative depletion of Ba, Sr, and Eu that are compatible with fractionation of feldspars, as well as a strong depletion of P by fractionation of apatite (Eby 1990;Bonin 2007). Due to the high contents of the HFSE Zr + Nb + Ce + Y this orthogneiss plots in the field of A-type granites in the classification of Whalen et al. (1987) and belongs to the ferroan A-type group according to Frost et al. (2001). In the subdivision of A-type granites based on Y/Nb ratios by Eby (1992) the augengneisses straddle the field boundary between type A2 and A1 or plot into field A1. Most petrogenetic models for A-type granites postulate either a postcollisional setting with melting of continental crust that has experienced a preceding episode of melt extraction during continent-continent collision or the differentiation of oceanic-island like mantle melts in a rift setting (Whalen et al. 1987;Eby 1990Eby , 1992Bonin 2007). The former should give rise to A2-type, the latter to A1-type granites (Eby 1992). The Y/Nb ratios of the augengneisses on the boundary of A1 and A2 fields fits very well a model with melting of a continental crust, which has undergone a cycle of continent-continent collision during the Variscan orogeny about 60 Ma earlier. The initiation of rifting and the probable involvement of mantle melts should shift the granitoids from the field A2 towards the field A1.
Hf isotopic data show εHf(t) values of − 13.7 to − 7.9, and Hf crustal model ages (T C DM ) of 969-1195 Ma suggesting the derivation of melts from Proterozoic continental crustal source. The continental source is also demonstrated by inherited Proterozoic cores of magmatic zircons.

Fig. 13
Granite discrimination diagrams of Whalen et al. (1987). A = A type granites; OGT = unfractionated I-, S-and M-type granitoids. FG = fractionated I-type granitoids. Nb-Y-3 Ga diagram after Eby (1992) for subdivision of A-type granites into A1 and A2 types (see text for discussion) The Permian intrusion age of augengneiss is related to Permian continental extension, which follows the Variscan orogeny within a few tens of million years (Kozur 1991;Neubauer et al. 2000;Putiš et al. 2018) (Fig. 14).
At present, it is unknown whether the host rocks of augengneisses also experienced the Permian-Triassic low-P/high-T metamorphism as known from the EGU (Habler and Thöni 1998;Thöni and Miller 2000) ( Table 3). As already discussed above, the excellent preservation of Permian zircon ages within an amphibolite-facies grade area indicates that the augengneiss body escaped high temperatures of 800-850 °C typical for the UHP/HP metamorphic conditions in the southwest (Janák et al. 2015; Table 4).

Age and metamorphism of leucogneiss
The leucogneisses are heterogeneously composed and, as part of the EGU, are associated with eclogites, paragneisses, micaschists and ultramafic cumulates of oceanic affinity (SBUC). Previous studies indicate Cretaceous eclogite facies metamorphic overprint for the EGU in the south-eastern Pohorje Mountains, with P-T conditions in the range of 3.0-3.7 GPa and 700-940 °C (Janák et al. 2004(Janák et al. , 2015Vrabec et al. 2012; Table 4 for further data). These conditions are also substantiated by findings of microdiamonds in metasedimentary rocks, which give UHP P-T conditions of ≥ 3.5 GPa and 800-850 °C (Janák et al. 2015). These P-T conditions make it essential to consider this young metamorphic event when interpreting isotopic data from the "pre-Alpine basement". The dated leucogneiss samples were collected nearby to localities were diamond was found by Janák et al. (2015). Consequently, we argue that these high temperatures of 800-850 °C were responsible for new crystallization of broad zircon rims obliterating the older record of magmatic and inherited ages in most zircon grains. Zircons from the leucogneisses have also similar features such as rounded shapes and a faint internal zonation with some preserved cores as zircons from eclogites (Miller et al. 2005b) and metapelites ). The Cretaceous ages of the zircons are also consistent with ages for the UHP metamorphism obtained with Sm-Nd (Miller et al. 2005b) and Lu-Hf (Thöni et al. 2008;Sandmann et al. 2016) dating.
The fabric of leucogneisses shows dynamic recrystallization of quartz and plagioclase. Two types of recrystallization mechanisms can be inferred from the microstructures of quartz (Passchier and Trouw 2005). Highly irregular grain boundaries, pinning structures, and leftover grains in quartz indicate grain boundary migration (Fig. 3). Additionally, small, uniform sized, slightly elongated quartz and feldspar grains argue for subgrain rotation. These observations are consistent with deformation during retrogression within amphibolite and greenschist conditions during retrogression subsequent to Cretaceous eclogite facies metamorphism, and nearly no record of UHP/HP metamorphic fabrics can be observed within the studied leucogneisses. Similar retrogressive fabrics overprinting eclogite-facies fabrics were also observed in more northern areas of EGU (Thöni and Miller 2000;Habler et al. 2007;Thöni and Miller 2010).
The dated leucogneisses give metamorphic ages of 89-90 Ma by zircon U-Pb LA-ICP-MS dating ( Fig. 8; Additional file 1: Table S1), the same age as all eclogites in the EGU. The ages of inherited zircon grains range between 260 ± 3.3 Ma and 1489 ± 31.9 Ma, the minimum inherited ages are similar as those of augengneiss samples. The mode of ages for core analyses of ca. 265 Ma of sample SK10 and the single youngest core age of 260 Ma of sample SK12 are a tentative formation age of these meta-granites. A more definite age would require analysis of more cores and probably samples, which zircons are less overprinted by the Alpine UHP conditions. However, a Permian age for the leucogranites is quite reasonable based on the Permian low pressure, high temperature metamorphism within EGU (Habler and Thöni 2001;Thöni and Miller 2009;Schuster et al. 2015;Schulz 2017;Schorn 2018), magmatic activity as gabbro intrusions of proven Permian age (Thöni and Jagoutz, 1992) and pegmatites (Knoll et al. 2018 and references therein) in the EGU.
Geochemically, the leucogneisses are very different from the augengneisses. Based on major element compositions, they are peraluminous, high-K calc-alkaline granites, supported also by more robust Co-Th contents. Trace element patterns show much lower enrichments than the augengneisses, with the exception of U, K and Sr and also negative anomalies of Nb and P. Ti is strongly depleted in the leucogneisses, in contrast to the augengneisses which are strongly enriched in Ti. REEs show much less enrichment in LREE and almost flat HREE patterns and also only weak negative Eu anomalies (with the exception of one analysis). We consider that this segment of the EGU is part of the distal Late Permian rift zone (Fig. 14). The zircon ages indicate that the UHP/HP metamorphism event occurred in the early Late Cretaceous (ca. 90 Ma).

Relationship of augengneisses and leucogneisses
As discussed before, the two orthogneiss groups display major differences in their petrology and geochemistry, their metamorphic overprint and also their country rocks. However, all previous investigators consider the south-eastern part of the Pohorje area as a coherent tectonic unit that underwent Alpine UHP metamorphism. We argue for the presence of two nappe units that experienced quite different Alpine metamorphic conditions but also show differences in their internal composition. The main difference is the UHP conditions with very high temperatures of the EGU and amphibolite conditions with no hints for high or even ultrahigh pressure conditions of the unit containing the augengneisses. These differences are also expressed in the orthogneisses themselves, with the paragenesis Bt + Plag in the augengneiss vs. Ms + Grt in the leucogneiss and also the very different appearance of zircon with well-preserved magmatic zircons in augengneiss and strongly overprinted/regrown zircon in the leucogneiss. The country rocks of the two gneiss groups show the same differences in metamorphic conditions. Rocks of the EGU show parageneses, which give UHP conditions as described by numerous investigators before, the country rocks of the augengneisses show paragenesis of amphibolite conditions without hints for HP conditions.
The boundary between the two tectonic units should roughly coincide with the tail of the Miocene granodiorite intrusion (Fig. 3). All described samples which gave UHP conditions are from the area south-west of the Miocene intrusion (Radkovec-Visole) with the exception of some eclogites south of Sp. Prebukovie, related with the ultramafic body and the eclogites and their retrogressed amphibolites. The augengneisses east of the Miocene granite are associated with paragneises, micaschists, amphibolites and marbles, which show amphibolite metamorphic conditions. Amphibolites there show no relics of eclogite conditions but well-equilibrated equigranular microstructures, whereas amphibolites retrogressed from eclogite display symplectitic microstructures with a wide range of amphibole grain sizes on the cm-scale. Marbles seem to be typical for the amphibolite facies unit (Vrabec et al. 2018 and geological maps of Mioč and Žnidarčič 1977;and Žnidarčič and Mioč 1989) and are not described from the unequivocal UHP unit. For a definite delineation of the boundary and the characterization of its nature clearly more field work is required. If the interpretation of the different metamorphic evolution holds, it must be a tectonic boundary, either a thrust or detachment fault, that brought two tectonic units with very different Alpine pressure conditions together, thus different subductions depths, in contact at probably amphibolite facies conditions.
According to the mapping and structural analysis by Kirst et al. (2010) two tectonic relations are possible. They mapped two antiforms, the overturned to recumbent Slovenska Bistrica Antiform with the SBUC as its core and the upright Pohorje Antiform to the east (see Fig. 2). The investigated area is on the western limb of the young Pohorje Antiform, with a general dip of foliations to the south (see Figs. 3 and 8 in Kirst et al. 2010). Therefore, the augengneiss unit is now in the footwall of the UHP-unit. This could be the primary relationship, with the UHP unit thrust onto the augengneiss unit. The second possibility is that the augengneiss unit is entangled in the Slovenska Bistrica Antiform and is part of the overturned limb. It would thus be in a hangingwall position before the recumbent folding. These alternatives have of course implications for the parallelization with other Austroalpine units further north. In the area of the Korand Saualpe there are marble-bearing amphibolite facies units without HP conditions both over-and underlying the EGU (Kurz et al. 2002;Thöni 2006;Wiesinger et al. 2006;Schorn and Stüwe 2016).

Tectonic setting of Permian rift-related magmatic rocks
We consider the Pohorje segment of the EGU and of the augengneiss unit are part of the Permian rift zone.
The Eastern Alps and their extension in south-eastern and eastern Europe comprise the Austroalpine domain of continental affinity, which is characterized by a Middle Triassic passive continental margin succession, which opened towards the Meliata Ocean (Channell and Kozur 1997;Lein 1987;Neubauer et al. 2000;Schmid et al. 2004;Kirst et al. 2010;Plašienka 2018). However, there is not much data on the Early Permian to Early Triassic rift development, which finally formed the Meliata Ocean. Many researchers proposed that Alpine tectonic evolution started with onset of Early Permian rifting immediately followed by carbonate platform in the Southalpine unit (Schaffhauser et al. 2015) formed after deposition of the post-Variscan, Upper Carboniferous (Pennsylvanian) molasse. Much evidence for Permian rift magmatism and high-temperature metamorphism is also found in Western and Southern Alps (Schaltegger and Brack 2007;Kunz et al. 2018;Manzotti et al. 2018). For EGU, further evidence of divergence and extension of the lithosphere was the emplacement of tholeiitic gabbros (Thöni and Jagoutz 1992;Schuster and Stüwe 2008;Kirst et al. 2010;Herg and Stüwe 2018), a few leucogranites (Morauf 1980;Tropper et al. 2007;Knoll et al. 2018 and low-pressure, high-temperature metamorphism within EGU (Table 4 for references). Knoll et al. (2018) argue for a derivation of widespread, often Lithium bearing pegmatites from small leucogneiss bodies within and in the footwall of EGU. In this study, we found evidence of A-type granite formation from our augengneiss samples, which could be also the source of pegmatites. The εHf(t) values of zircons are negative (− 6.4 to − 1.7) suggesting that our samples formed by partial melting of old continental crust. So, we suggest that the rifting and lithospheric thinning in the Permian occurred by north-south extension (in presentday coordinates) in response to westward propagation of the Neotethys ocean. Crustal melting caused granite and pegmatite intrusions at mid-crustal levels within the future EGU. Our orthogneisses could be considered as the source magma of seemingly rootless Late Permian (to Triassic) pegmatites widespread within the EGU in the surroundings of the study area (Kirst et al. 2010) and further to the north. We argue that EGU (and underlying and overlying units) represent middle-lower parts of a boudin-like stretched continental crust, within which older fabrics are partly well preserved similar as observed in recent passive margins (Clerc et al. 2015). In that case, our data imply that the stretched continental crust of EGU as part of the Permian rift with magmatic rocks (and the subsequent) Triassic passive margin was potentially not much wider than ca. 100 km, was then subducted and subsequently rapidly exhumed during early Late Cretaceous times.

Continental subduction
The age data summarized in Table 4 and the ages of the leucogneiss samples suggest that the UHP/HP metamorphism event occurred in 89.34 ± 0.69 Ma and 90.8 ± 1.2 Ma, as has been widely reported (Miller et al. 2005b;De Hoog et al. 2011;Knoll et al. 2018). Janák et al. (2015) considered that the Pohorje nappe reached maximum depth at c. 95-92 Ma. The subduction zone formed during Early Cretaceous in the northwestern foreland of the Meliata suture after Late Jurassic closure of the Meliata Ocean and the resulting collision, by progressive subduction of the Triassic passive margin with the associated Permian rift lithologies, when southeastward, intraoceanic subduction of the Meliata Ocean terminated in collision of the Austroalpine continental margin with the accretionary prism associated with this subduction. During the Early Cretaceous, intracontinental subduction zone continued within the Austroalpine continental crust at the site of a Permian-aged rift (Janák et al. 2004(Janák et al. , 2015Stüwe and Schuster 2010).
Previous research also demonstrates that the Lower Central Austroalpine Units were subducted in an intracontinental subduction zone after complete subduction of the Meliata Ocean and the following continent collision, increasing towards SE (in present-day coordinates) and reaching more than 100 km depth in the Pohorje Mts., and exhumation of the Pohorje nappe at ca. 90 Ma (Froitzheim et al. 2008). For the geodynamics of exhumation, several models were proposed: (1) slab break-off (Neubauer et al. 2000), and (2) downward extraction of the Meliata oceanic slab (Froitzheim et al. 2008), which carried the Upper Central Austroalpine continental units in its northwestern part (Froitzheim et al. 2008). In this model, the latter units breakoff caused a reduction in horizontal stress along its trailing edge and thereby promoted exhumation of deeply subducted and directly emplaced on the UHP rocks, the SBUC was part of the down-going plate from the beginning (Kirst et al. 2010). On the other hand, footwall accretion of Austroalpine nappes during Cretaceous (between ca. 120 and 75 Ma) was already proposed by Dallmeyer et al. (1998). E.g., the Lower Austroalpine nappes were incorporated into the thrust belt during late Cretaceous (Neubauer 1994) and units from the Penninic ocean during Eocene (Liu et al. 2001). Consequently, a continuous slab could have caused subduction and the Austroalpine passive margin is entirely consumed, and no slab-break-off is needed to explain the Austroalpine nappe structure. This model is shown in Fig. 14 for the Late Cretaceous. There is a significant difference to previous models of, e.g., Froitzheim et al. (2003) and Janák et al. (2009), which assumes several distinct slabs and slab extraction during Late Cretaceous. As Dallmeyer et al. (1998) have shown, the Austroalpine nappe complex accreted by footwall propagation of thrusting continuously between ca. 110 Ma and 78 Ma. This observation implies that always the same heavy mantle slab is the driving force for thrusting and frontal nappe accretion and no interruption by slab extraction is needed to explain accretation of Meliata remnants and Austroalpine continental units.

Conclusions
(1) The Pohorje area contains two groups of distinct orthogneisses which represent part of Permian granitic to granodioritic bodies which are dated at 255-260 Ma and are derived from partial melting of lower continental crust in a rift zone.