Origin and age of the Eisenkappel gabbro to granite suite (Carinthia, SE Austrian Alps).

The northern part of the Karawanken plutonic belt is a gabbro-granite complex located just north of the Periadriatic lineament near the Slovenian-Austrian border. Petrographic and geochemical studies of the Eisenkappel intrusive complex indicate that this multiphase plutonic suite developed by a combination of crystal accumulation, fractional crystallization and assimilation processes, magma mixing and mingling. The mafic rocks are alkaline and have within-plate geochemical characteristics, indicating anorogenic magmatism in an extensional setting and derivation from an enriched mantle source. The mafic melts triggered partial melting of the crust and the formation of granite. The granitic rocks are alkalic, metaluminous and have the high Fe/Fe + Mg characteristics of within-plate plutons. Temperature and pressure conditions, derived from amphibole-plagioclase and different amphibole thermobarometers, suggest that the analysed Eisenkappel gabbros crystallized at around 1000 ± 20 °C and 380-470 MPa, whereas the granitic rock crystallized at T ≤ 800 ± 20 °C and ≤ 350 MPa. Mineral-whole rock Sm-Nd analyses of two cumulate gabbros yielded 249 ± 8.4 Ma and 250 ± 26 Ma (εNd: + 3.6), garnet-whole rock Sm-Nd analyses of two silicic samples yielded well-constrained ages of 238.4 ± 1.9 Ma and 242.1 ± 2.1 Ma (εNd: - 2.6).

Introduction

The Variscan belt in Europe resulted from complex interactions between Laurussia- and Gondwana-derived microcontinents, ending with the final collision of Gondwana and Laurussia in Late Carboniferous times (Stampfli and Borel, 2002). Subduction of the Palaeotethys continued with slab rollback inducing post-collisional orogenic collapse and the opening of back-arc basins along the Eurasian margin, beginning in the Early Permian (Stampfli and Borel, 2002). In the Adria microplate, Permian lithospheric thinning resulted in high temperature/low pressure metamorphism (e.g. Schuster et al., 2001, Zingg et al., 1990) and magmatic activity in the Adria-derived western Austroalpine (e.g. Dal Piaz, 1993, Monjoie et al., 2007, Tribuzio et al., 1999), Southalpine (e.g. Bellieni et al., 2010, Cassinis et al., 2008, Macera et al., 1994, Marocchi et al., 2008, Rottura et al., 1998) and eastern Austroalpine domains (e.g. Bellieni et al., 2010, Miller and Thöni, 1997, Thöni and Miller, 2000). Discrete Permian-Triassic magmatic pulses, controlled by a transtensional geodynamic regime, produced tholeiitic to alkaline mantle melts in addition to a variety of felsic intrusive and extrusive rocks that often document mantle–crust interactions.

The Eisenkappel intrusive complex in southeastern Austria and northern Slovenija forms part of this post-Variscan magmatism that affected the Adria microplate. This multiphase, ultramafic to silicic plutonic suite was emplaced in the vicinity of the modern Periadriatic Lineament prior to 224 ± 9 Ma (biotite Rb/Sr cooling ages; Scharbert, 1975). It includes minor ultrabasic cumulates, gabbros, intermediate and silicic plutonic rocks. Previous studies of this complex (Bole et al., 2001, Exner, 1972, Faninger and Štrucl, 1978, Visonà and Zanferrari, 2000) have focused on the mineralogy and geochemistry of intermediate and granitic rocks, documenting extensive interaction of mafic and felsic magmas. Bole et al. (2001) argued that mafic magma provided the heat necessary to trigger partial melting of the crust in this terrane. In contrast, based on geochemical variations and initial Sr isotope ratios of the NKPB, Visonà and Zanferrari (2000) suggested partial melting of a mantle source enriched both in LILE and HFSE, with mafic melts evolving by AFC (assimilation-fractional crystallization) processes, followed by FC processes when residual liquids became felsic. However, this assumption is critically dependent on the calculated 87Sr/86Sr intial ratios, but the precise timing of this magmatic event has remained poorly understood. In addition, little is known about the origin and igneous evolution of the mafic magmas.

In this paper we investigate the geochemistry and Sr–Nd isotope systematics of six gabbro samples and of two rare garnet-bearing granitoids with the aim of identifying magmatic processes and to better constrain the timing of the plutonic activity by providing magmatic ages for mafic and felsic lithologies, based on remarkably fresh primary mineral assemblages. New information about the trace element and isotopic composition of the gabbroic rocks is used to evaluate the nature of the mantle source. In addition, regional-scale information on the origin of the Permo-Triassic magmatism in the eastern Austroalpine domain is discussed.

Geological background and field relationships

The Karawanken plutons just north of the Periadriatic Lineament (Fig. 1) consist of two E–W trending elongated belts that are separated by a narrow zone of metamorphic rocks (Exner, 1972, Faninger and Štrucl, 1978). The southern belt is tonalitic and Oligocene in age (Scharbert, 1975). In contrast, the older northern part of the Karawanken plutonic belt (NKPB) is composed of gabbro, monzonite, syenite and syenogranite, showing mingling phenomena and relations of mutual intrusion or cross-cutting dikes. It is about 46 km long and intrudes a presumably Early Palaeozoic diabase/shale complex (Eisenkappler Diabaszug) and Triassic dolomite to the north and amphibolite-facies rocks (“Altkristallin”) to the south (Exner, 1972, Faninger and Štrucl, 1978) (see Fig. 1b). Contact metamorphism implies an emplacement pressure of ≤ 350 MPa (Exner, 1972, Exner, 1976). Lippolt and Pidgeon (1974) dated biotite (K/Ar), hornblende (K/Ar) and titanite (U/Pb) from a NKPB diorite at 227 ± 7 Ma, 244 ± 8 Ma and 230 ± 5 Ma, respectively. In addition, Scharbert (1975) published biotite Rb–Sr ages of 224 ± 9 and 216 ± 9 Ma from granodiorite and Cliff et al. (1975) measured a single hornblende K–Ar age from a hornblende-bearing pegmatite of 244 ± 9 Ma. Recent U/Pb LA–ICP-MS dating of zircon and titanite from granitic samples show “a spread of data along the concordia”, with some pre-Variscan ages, but mainly clusters of ages between 300–280 and 250–240 Ma (Genser and Liu, 2010). These variations have been interpreted as indicating (i) preservation of an Ordovician magmatic event (U/Pb zircon age of 450–500 Ma), (ii) intrusion of the main Eisenkappel granitoids between 280 and 300 Ma and (iii) cooling below ca. 550 °C at about 245 Ma (Genser and Liu, 2010).

Fig. 1

(a) Geological-tectonic sketch of the southeastern Alps, showing the E Karawanken/Eisenkappel study area. From: Miller et al. (2007), strongly modified. (b) Lithostratigraphic-tectonic sketch map of the study area near Eisenkappel, with sampling locations (a, b, and c). From: Bauer et al. (1981), modified.

Although mafic rocks are widespread in the NKPB, major plutons are not exposed. Gabbro outcrops are small and discontinuous, making spatial correlations difficult. In the study area, gabbro–granite contacts and contacts of the gabbros with the country rocks are not exposed. However, felsic rocks in the NKPB frequently contain dioritic to gabbroic enclaves (Bole et al., 2001, Exner, 1972, Visonà and Zanferrari, 2000). In these composite zones, the mafic enclaves are rounded, lobate and variable in size. Enclave margins are sharp, fine-grained, and frequently serrate, indicating an origin as quenched pillows of mafic magma (Fig. 2). Similar enclave characteristics are known from many plutons and indicate mingling of mafic and felsic magmas (e.g. Blundy and Sparks, 1992, Didier and Barbarin, 1991). In addition, porphyritic syenite dikes are common, containing alkali feldspar mantled by plagioclase. Rapakivi-textured K-feldspars, quartz ocelli, plagioclase mantled by K-feldspar and intrusive breccias also suggest some interaction between mafic and felsic magmas (Bole et al., 2001, Visonà and Zanferrari, 2000).

Fig. 2

Photographs illustrating microtextures in Eisenkappel (a) cumulate and (b) isotropic gabbros. (c) Plagioclase-phyric mafic magmatic enclave in Eisenkappel granite documenting mingling processes. (d) BSE image of porphyritic mafic enclave showing plagioclase phenocrysts in a fine-grained matrix of magnesio-hornblende (light gray), plagioclase and ilmenite (white).

Analytical techniques

Whole rock and mineral chemistry

Whole rock major elements, Ba, Ga, Hf, Rb, Sr, V, Y, Zn and Zr were determined by EDXRFA (Spectro-Xepos) at the University of Innsbruck, based on calibrations with certified rock standard reference materials using glass discs for major components and powder pellets for trace elements.

Samples (c. 100 mg of whole rock powder) for rare earth elements and other trace elements were dissolved in ultrapure HF-HClO4, subsequently in 6 NHCl, and finally in 10−% HNO3. Trace elements were determined at the University of Graz, using an Agilent 7500ce (Agilent, Waldbronn, Germany) inductively coupled plasma mass spectrometer (ICP-MS), equipped with a Cetac ASX500 autosampler. Solutions were introduced to a Burgener HP ARI Mist nebuliser with the ISIS (integrated sample introduction system). The trueness of the calibration was checked with the certified reference material NIST 1643e “Trace elements in water” (NIST, Gaithersburg, USA). The obtained results agreed well with the certified values.

Mineral compositions were determined by wavelength-dispersive X-ray analysis using the JEOL JXA-8100 Superprobe at the University of Innsbruck. Analytical conditions were 15 kV accelerating potential and a beam current of 20 nA. Natural or synthetic silicates and oxides were used as standards and a PhiRhoZ routine for matrix corrections. Accessory phases were identified using back-scattered electron (BSE) imaging, and energy-dispersive spectroscopy. Mineral abbreviations are after Siivola and Schmidt (www.bgs.ac.uk/scmr/home.html).

Sm–Nd and Rb–Sr isotope analysis

The Sm–Nd analytical work was performed at the Laboratory of Geochronology, Center for Earth Sciences, University of Vienna.

Pure mineral separates of plagioclase, pyroxene and garnet were obtained by careful handpicking of optically inclusion-free grain fragments from the 0.1–0.4 mm sieve or/and magnetic fractions. Pure biotite concentrates (≫99%) were obtained by repeated grinding of the mica concentrate in an agate mill using alcohol, drying and sieving, and final magnetic purification.

Sample weights used for Sm–Nd mineral analysis weighed between 30 and 100 mg. Before decomposition, or leaching, the fractions were washed for 30 min in warm (70 °C) 0.2 N (for Pl and Cpx), or 2.5 N HCl (for Grt). Leaching procedures applied to garnet to eliminate phosphate inclusions followed Anczkiewicz and Thirlwall (2003; using concentrated H. Chemical sample digestion, isotope dilution, and element separation for Sm–Nd and Rb–Sr analysis follow those given in Thöni et al., 2008a, Thöni et al., 2008b. Total procedural blanks were < 1 ng for Rb and Sr, and < 50 pg for Nd and Sm. Sm and Nd ID samples were run as metals from a Re double filament, while Rb was evaporated using a Ta filament, and a Finnigan® MAT262. Sr (ID and IC) and Nd IC samples were analysed using a ThermoFinnigan® Triton TI TIMS machine. A 143Nd/144Nd ratio of 0.511844 ± 0.000006 (2σ; n = 24) and a 87Sr/86Sr ratio of 0.710251 ± 0.000018 (2σ; n = 30) were determined for the La Jolla (Nd) and the NBS987 (Sr) international standards, respectively, during the c. 1 year period of investigation. Within-run mass fractionation for Nd and Sr isotope compositions (IC) was corrected for relative to 146Nd/144Nd = 0.7219, and 86Sr/88Sr = 0.1194, respectively. Uncertainties on the 143Nd/144Nd and the 87Sr/86Sr isotope ratios are quoted as 2σm. For the 147Sm/144Nd and the 87Rb/86Sr ratios, a mean error of ± 1% is applied (representing maximum errors), including blank contribution, uncertainties on spike composition, and machine drift; regression calculation is based on these uncertainties and the isochron calculations follow Ludwig (2003). Age calculations are based on decay constants of 6.54 × 10− 12 a− 1 for 147Sm (Lugmair and Marti, 1978) and 1.42 × 10− 11 a− 1 for 87Rb (Steiger and Jäger, 1977), respectively; age errors are given at the 2σ level. For Nd, a continuous depletion of the upper mantle is assumed throughout geological time, and the following Depleted Mantle (DM) model parameters were used: 147Sm/144Nd = 0.222, 143Nd/144Nd = 0.513114 (Michard et al., 1985).

Results

The gabbroic and granitic samples used for whole rock geochemical and mineral analysis, and for geochronological studies are described in detail below.

Sample locations A, B, C for sample names listed in text, tables and Fig. 1b are as follows (see Bauer et al., 1981):

Dell immediately SW Eisenkappel, near tennis court: gabbros 94T31EK, 94T33EK, 08EK06, 08EK07, 08EK10. GPS coordinates: 46° 28′ 31.76″ N/14° 35′ 22.5″ E

Crest north of Remschenig Graben, ESE Eisenkappel, c. 800 m A.S.L.: syenite, syenogranite 08EKR12, 08EKR13. GPS coordinates: 46° 28′ 21.18″ N/14° 37′ 10.23″ E

Leppen Graben, c. 4 km ESE Eisenkappel, 950 m A.S.L.: gabbro 08EKL18. GPS coordinates: 46° 28′ 42.28″ N/14° 38′ 38.82″ E

Igneous petrography and mineral chemistry

Gabbroic cumulates

The gabbroic cumulates are medium-grained without discernible preferred orientation of crystals (Fig. 2a). Minerals interpreted to be post-cumulus overgrow cumulus phases, filling the intercumulus pore volume and producing poikilitic textures. The cumulates are composed of approximately 40–45 modal % olivine, 10–15 modal % plagioclase, 5–10 modal % clinopyroxene, 35–46 modal % brown amphibole and phlogopite. Olivine is the most abundant mafic phase, and is included in poikilitic plagioclase or clinopyroxene (Fig. 3a–c). The grains are typically euhedral with an average grain size of 0.4–1.5 mm. Olivine is often rimmed by orthopyroxene and commonly contains inclusions of Cr-spinel in addition to micrometer-sized composite exsolution lamellae (Fig. 3d). Partial alteration to a mixture of serpentine and magnetite is sometimes observed along fractures. Anhedral plagioclase oikocrysts range from 3 to 7 mm and enclose olivine ± clinopyroxene. Clinopyroxene is pale pink-brownish, forming small euhedral grains (0.2–0.4 mm) if enclosed in plagioclase (Fig. 3b) or anhedral oikocrysts (2–4 mm; Fig. 3c). Exsolution lamellae of orthopyroxene are absent in clinopyroxene. Minerals that crystallized from, or by reaction with, intercumulus melt include othopyroxene (after Ol), amphibole (after Cpx) and phlogopite. Orthopyroxene is usually present as < 0.1 mm rims on olivine or as 0.1–0.3 mm interstitial grains (Fig. 3e). Late-magmatic brown amphibole forms poikilitic grains and overgrowths surrounding clinopyroxene and, together with phlogopite, dominates the intercumulus assemblage (Fig. 3e). Fe–Ti oxides comprise 1–4% of the cumulate samples. Magnetite is the most common oxide and forms irregular interstitial grains or inclusions in brown amphibole. Ilmenite also forms anhedral interstitial grains or exsolutions in magnetite. Pyrrhotite, pentlandite, and chalcopyrite are minor accessory minerals. Interstitial apatite grains are acicular, with their longest dimension being 1.4 mm and an aspect ratio of 1.10. The acicular apatites are often intergrown with late interstitial albite and have an elongate central cavity occupied by albite (Fig. 3f). The cumulate gabbros are modified to variable degrees by subsolidus reactions and minor development of secondary minerals. Plagioclase is locally altered into fine-grained aggregates of sodic plagioclase and Al-rich pumpellyite, and spinel is replaced by magnetite, indicating lower greenschist-facies conditions.

Fig. 3

Backscattered electron (BSE) images illustrating micro-textures of Eisenkappel gabbros. (a) sample 94T31: euhedral olivine enclosed in plagioclase; (b) sample 08EK06: euhedral olivine and clinopyroxene enclosed in plagioclase; (c) sample 94T33: olivine enclosed in clinopyroxene; (d) sample 94T33: olivine containing composite lamellae consisting of Cr-magnetite + Di ± Mg-Hs; (e) sample 94T31: Opx rim at Ol/Pl interface. (f) sample 94T33: acicular apatite intergrown with albite.

Olivine is euhedral, homogeneous or slightly zoned, with core and rim compositions of Fo78–76 and Fo77–75, respectively. MnO contents are 0.26–0.57 wt.%, NiO is 0.10–0.19 wt.%, CaO < 0.05 wt.% and P is below detection limit (Table 1). Olivine contains rounded inclusions of pyrrhotite and chromian spinel and is commonly clouded by lamellar inclusions that are aligned, regularly spaced and between 4 and 30 μm long. The width of these lamellae is generally < 1 μm. These micrometer-sized composite lamellae consist of intergrown diopsidic clinopyroxene or magnesiohastingsite and Cr-bearing magnetite (Fig. 3d).

Table 1

EPMA analyses of olivine and pyroxene from Eisenkappel cumulate gabbros.

Sample	94T31	94T33	94T33	08EK06	08EK06	94T31	94T31	94T33	94T33	08EK06	08EK06	08EK06	08EK06	08EK07
Spot	1.3c	1.1c	1.2r	1c	3r	2.3	1.15	1.9	1.12	4	13c	19	22	7
Mineral	Ol	Ol	Ol	Ol	Ol	Opx	Cpx	Opx	Cpx	Opx	Cpx	Cpx	exsln in Ol	Cpx
SiO2	38.87	38.32	38.33	38.06	38.02	54.97	46.68	55.33	48.48	55.44	48.74	47.67	53.03	51.70
TiO2	0.00	0.02	0.06	0.10	0.05	0.00	2.53	0.03	2.00	0.00	1.95	2.30	0.00	0.49
Al2O3	0.00	0.00	0.00	0.00	0.00	0.44	7.48	0.30	6.56	0.70	6.26	6.87	1.16	1.16
Cr2O3	0.00	0.00	0.00	0.01	0.00	0.02	0.85	0.02	0.83	0.00	0.52	0.69	0.10	0.04
FeOTOT	19.94	20.06	20.54	21.36	21.37	14.26	5.72	15.25	5.59	13.34	5.71	6.12	5.20	9.09
MnO	0.29	0.36	0.37	0.39	0.57	0.49	0.14	0.47	0.06	0.44	0.11	0.20	0.00	0.42
MgO	40.73	40.61	40.16	39.61	39.35	28.36	13.09	27.91	14.17	29.71	14.17	13.8	16.94	12.97
CaO	0.01	0.00	0.03	0.00	0.02	0.95	23.05	0.81	22.57	0.31	22.49	22.25	22.97	22.94
Na2O	0.00	0.00	0.00	0.00	0.00	0.00	0.45	0.04	0.40	0.00	0.43	0.34	0.37	0.64
NiO	0.12	0.22	0.16	0.10	0.17	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.
Total	99.96	99.59	99.65	99.63	99.55	99.49	99.99	100.16	100.66	99.94	100.38	100.24	99.77	99.45

Clinopyroxene is an early crystallizing phase that is partially replaced by igneous amphibole. It has a limited compositional range with high Al2O3 (6.3–7.9 wt.%), 2.0–2.9 wt.% TiO2 and 0.31–0.58 wt.% NaO. Mg-values (Mg/Mg + Fe2+tot) range from 0.79 to 0.82, Cr2O3 contents are 0.39–0.85 wt.%. Aluminum preferentially occupies the tetrahedrally coordinated site. Zonation when present is towards decreasing Cr and increasing Fe and Na at constant or decreasing Al contents (Table 1). Clinopyroxene in the intra-olivine exsolution lamellae is Ti- and Cr-poor, and contains < 1.2 wt.% Al2O3. Orthopyroxene forms thin rims (< 250 μm) between olivine and plagioclase in cumulate samples. It is bronzite (En75–80) containing 0.2–1.2 wt.% CaO, < 0.04 Cr2O3 and 0.2–0.7 wt.% Al2O3 (Table 1).

Amphibole is mostly kaersutite or Ti-bearing pargasite, with Mg-values of 0.72–0.77, 3.2–5.7 wt.% TiO2 and 0.12–0.17 wt.% F. Structural formulae show extensive substitution of Si (5.88–6.37 apfu) by AlIV, coupled with low AlVI (0.03–0.33 apfu). Kaersutitic amphibole is a late magmatic phase, forming patchy replacements of clinopyroxene as well as anhedral interstitial to poikiloblastic grains enclosing olivine, ilmenite, apatite and sulfides. It is well preserved and frequently rimmed by green (ferri-) magnesio-hornblende. The magnesio-hornblende could either be a late-magmatic or a metamorphic phase. A magmatic origin seems more likely because there is no evidence of a medium-grade metamorphic overprint having affected the area. In addition, Ti-poor Mg-hastingsite is present, forming irregular veins within kaersutite, narrow (< 20 μm) rims at olivine/plagioclase interfaces or part of the composite micro-lamellae in olivine (Fig. 3d, Table 2).

Table 2

EPMA analyses of amphibole, phlogopite and plagioclase from Eisenkappel cumulate gabbros.

Sample	94T31	94T31	94T31	94T33	94T33	94T33	94T33	08EK06	08EK06	08EK07	08EK07	94T31	08EK06	94T31	94T33	08EK06
Spot	1.18	1.17	1.21	1.5	1.4	1.6	3.2	15	5	4	5	1.24	16	1.12	1.29	7
Mineral	Krs	Mg-Hs	Mg-Hbl	Krs	Mg-Hs	Mg-Hbl	Ol-exsln	Krs	Mg-Hs	Ti-Prg	Mg-Hs	Phl	Phl	Pl	Pl	Pl
SiO2	41.40	41.88	48.79	40.81	44.02	49.88	44.16	41.10	44.11	41.25	42.26	39.87	38.01	50.88	51.60	52.14
TiO2	5.51	0.81	1.39	5.23	2.61	0.74	.82	4.77	0.10	2.41	1.95	3.30	4.17	0.08	.10	0.00
Al2O3	12.38	15.50	7.71	12.27	10.62	5.97	12.18	11.65	13.86	12.45	10.74	14.14	14.69	31.57	30.62	30.37
Cr2O3	0.00	0.04	0.00	0.20	0.04	0.01	.54	0.23	0.01	0.02	0.09	0.06	0.20	0.00	0.00	0.00
FeOTOT	8.86	7.37	7.12	9.21	10.46	8.03	8.26	9.92	8.24	15.87	16.43	7.35	7.95	0.05	0.03	0.00
MnO	0.09	0.11	0.20	0.15	0.23	0.22	0.05	0.03	0.10	0.24	0.17	0.00	0.05	0.00	0.01	0.00
MgO	13.31	16.23	18.76	13.97	15.29	19.54	17.39	14.70	16.59	10.36	11.08	20.29	19.95	0.00	0.00	0.00
CaO	12.37	12.25	10.93	11.71	10.60	10.23	11.59	11.57	11.43	11.93	11.93	0.00	0.00	14.04	13.25	13.17
Na2O	2.95	3.43	2.71	2.89	3.28	2.51	3.15	2.91	3.23	2.34	1.93	1.44	1.53	3.56	4.08	4.12
K2O	1.17	0.31	0.34	1.10	0.32	0.27	0.06	1.04	0.33	1.09	1.04	8.04	8.38	0.05	0.01	0.04
Total	98.04	97.93	97.95	97.54	97.47	97.41	98.20	97.92	98.00	97.96	97.62	94.49	94.93	100.23	99.70	99.84

Plagioclase compositions show a bimodal distribution that correlates with microtextures. Compositions of the large oikocrysts enclosing olivine and clinopyroxene are limited between An62 and An70. FeO contents are low (< 0.2 wt.%), systematic core-rim variations were not detected. In contrast, rare interstitial plagioclase is distinctly more sodic with An7–16. Locally, plagioclase shows incipient alteration into fine-grained aggregates of Al-rich pumpellyite ± clinozoisite ± prehnite ± amphibole + Na-rich plagioclase (An30–34) (Table 2).

Phlogopite formed late in the crystallization sequence and exhibits a restricted range in composition with 2.1 to 4.2 wt.% TiO2, 7.1–8.8 wt.% FeO, 14.0–15.2 wt.% Al2O3, 1.4–1.9 Na2O, less than 0.2 wt.% Cr2O3 and 0.18–0.24 wt.% F. Its high Na2O and Mg# (0.80–0.84) confirm a magmatic origin (Table 2). Apatite contains < 1.9 wt.% fluorine and < 0.7 wt.% chlorine, resulting in calculated H2O contents of 0.7–1.3 wt.%. Oxides of minor components (Sr, Fe, Mn, Mg, Na, Ce, and Nd) are generally well below 0.75 wt.%. Cr-spinel compositions are variable and characterized by Cr2O3 contents ranging from 18.8 to 25.0 wt.%, 7.3–13.8 wt.% Al2O3, 1.9–3.4 wt.% MgO and high TiO2 contents (1.7–4.7 wt.%). Anhedral grains of Mg-ilmenite (7–13% geikielite component) and Cr-magnetite commonly form inclusions in kaersutite. Cr-magnetite contains two sets of narrow exsolution lamellae. One set is ilmenite (< 5 μm), the other Cr-rich spinel (Cr# = 0.12; TiO2 = 0.15 wt.%). Pyrrhotite, pentlandite and chalcopyrite form small (< 120 μm) disseminated rounded grains that are included in plagioclase, clinopyroxene or kaersutite.

Isotropic gabbros

In contrast to the cumulate gabbros the fine-grained (0.5–2 mm) isotropic gabbros (Fig. 2b) are subhedral granular and consist of plagioclase and clinopyroxene mantled by brown amphibole (Ti-pargasite) that in turn is mantled by green amphibole (Mg-hastingsite). Biotite forms anhedral interstitial grains. Acicular apatite, anhedral to subhedral titanite and ilmenite are accessory minerals. Plagioclase is altered to saussurite, late actinolite replaces Mg-hastingsite and biotite is replaced by chlorite.

Clinopyroxene has lower Mg# (0.72–0.78) and lower Cr2O3 contents (< 0.24 wt.%) when compared to cumulate clinopyroxene. Amphiboles: Ti-pargasite with Mg-values of 0.53–0.71 and 2.4–3.8 wt.% TiO2 forms rims around clinopyroxene or separate grains some of which contain cores riddled with ilmenite. Ti-pargasite is rimmed by Mg-hastingsite that, in turn, is frequently surrounded by fine-grained aggregates of Mg-hornblende and/or actinolite. Plagioclase is clearly altered. Phlogopite has Mg# ranging from 0.71 to 0.74, TiO2 contents of 3.2–3.4 wt.% and 0.74–0.76 wt.% Na2O. Cr2O3 and F contents range from 0.07 to 0.18 wt.% and from 0.14 to 0.16 wt.%, respectively.

Mafic enclaves

In the composite zones, mafic enclaves are either very fine-grained plagioclase-hornblende ± biotite-ilmenite assemblages or porphyritic rocks containing plagioclase (Ab40–48) and clinopyroxene phenocrysts in a fine-grained matrix of magnesio-hornblende, plagioclase (Ab70–73) and dendritic acicular grains of ilmenite (Fig. 2c, d). Clinopyroxene phenocrysts are Al- and Ti-rich, but always mantled and partially replaced by magnesio-hornblende.

Granitic rocks

The granite unit consists mainly of syenogranite but includes granodiorite, syenite and monzonite. Granitic rocks form either massive outcrops or the felsic matrix surrounding mafic enclaves in composite zones (Fig. 2c). Detailed petrographic descriptions of the different granitic rocks are given by Exner, 1972, Faninger and Štrucl, 1978, Visonà and Zanferrari, 2000, Bole et al., 2001. Garnet-bearing samples, however, are rare and have not been described previously.

Sample 08EKR12

The magmatic assemblage is composed of K-feldspar, plagioclase, quartz, biotite, rare garnet, with zircon, F-apatite, monazite and ilmenite as accessories. Quartz forms anhedral crystals in the framework of the rock. Plagioclase commonly shows patchy or reverse zoning with core and rim compositions of Ab96.4–93.5An2.8–4.0Or0.8–2.4 and Ab87.5–84.8An11.4–14.4Or0.8–0.9, respectively. Some grains also exhibit a discontinuous outer albite-rich rim (Ab97.9An1.5Or0.6). Subhedral porphyritic K-feldspar crystals (Ab5.2–5.5Or94.5–94.8) with perthitic microtextures are abundant, containing inclusions of quartz, plagioclase and zircon. Garnet forms highly irregular grain aggregates that are complexly intergrown with quartz, K-feldspar and plagioclase (Fig. 4a–b). In addition to quartz, garnet may contain a few inclusions of F-apatite and zircon. These zircon inclusions are similar to matrix zircon, being euhedral, growth zoned, and including F-apatite, quartz and plagioclase (Fig. 4c–d). In garnet, almandine and spessartine constitute 91–93% of the total molecular composition, grossular is always < 2.5 mol%. EPM studies revealed weak zoning characterized by a Sps inverse bell-shaped profile with central garnet domains containing 6–10 mol% Sps, and rim zones containing 10–12 mol% Sps.

Fig. 4

BSE images illustrating micro-textures of the analysed garnet-bearing Eisenkappel granitoids. Sample 08EKR12: complex intergrowths of (a) garnet-quartz and (b) garnet-quartz-perthitic K-feldspar-plagioclase. Cathodoluminescence images of zircon containing F-apatite and quartz inclusions: (c) zircon included in garnet; (d) matrix zircon; low-T Alpine brittle deformation could account for the broken matrix zircon whereas zircon included in garnet was preserved intact; (e) BSE image of rim domain of large euhedral garnet in sample 08EKR13; (f) compositional profile across garnet in sample 08EKR13.

Sample 08EKR13

The sample is a slightly megacrystic but predominantly medium- to coarse-grained inequigranular rock, with subhedral perthitic K-feldspar (Ab6.1–6.5Or93.5–93.9) in addition to plagioclase (turbid), anhedral quartz, biotite, zircon, fluorapatite, and trace amounts of garnet. Plagioclase is usually euhedral and complexly zoned, with alternating An-rich and An-poor domains. Thus, cores of Ab81.0An18.5Or0.5 are overgrown by a zone of Ab89.9An9.0Or1.1, which in turn is surrounded by a zone of Ab83.1An16.4Or0.5 and an outermost rim of Ab96.8An2.7Or0.5. Biotite is Fe-rich, with Fe/(Fe + Mg) = 0.746–0.757, Ti = 0.165–0.182 apfu, MnO < 0.33 wt.%, and incipient alteration to chlorite and/or pumpellyite. Garnet grains are euhedral, up to 1.5 cm in diameter and contain a few inclusions of quartz, albite, F-apatite and zircon (Fig. 4e). Matrix zircon and zircon inclusions in garnet are characterized by growth zoning, quartz and F-apatite inclusions. Garnet crystals are Fe–Mn-rich. A microprobe traverse across the largest garnet showed a symmetrical pattern of zoning characterized by a near homogeneous core domain of Alm71.2–72.4Prp5.6–6.1Grs4.7–4.9Sps16.8–17.7, a spessartine-rich intermediate zone (Alm64.0Prp4.6Grs5.5Sps25.5) and an outer rim of Alm77.7Prp7.2Grs2.1Sps13.0 (Fig. 4f).

Whole rock geochemistry

Gabbroic rocks

Major and trace element compositions of six Eisenkappel gabbro samples are listed in Table 3. All analysed rocks are characterized by low SiO2 (42.1–46.7 wt.%) contents. Al2O3 contents range from 8 to 15 wt.%, CaO ranges from 5.6 to 10.9 wt.%. The gabbros contain between 9.8 and 13.2 wt.% FeO and 1.4–2.9 wt.% TiO2. Mg-values [Mg# = Mg/(Mg + Fe2+) and Fe2+ = 0.85 Fetot] reveal two distinct groups of rocks: cumulates with Mg# of 0.77–0.80 and isotropic gabbros with Mg# of 0.59–0.68 (Fig. 5a). The high Mg# of the cumulates is coupled with high SiO2/Al2O3 ratios (5.0–5.4) and low CaO (5.6–5.8 wt.%), indicating that their whole-rock chemistry is mainly controlled by a process of olivine ± clinopyroxene ± amphibole accumulation. Based on the total alkali content (2.5–4.7 wt.%), all samples can be classified as alkaline gabbros (Fig. 5b).

Table 3

Major and trace elements of Eisenkappel gabbros and granitoids.

	94T31EK	94T33EK	08EK06	08EK10	08EK07	08EK18	08EKR12	08EKR13
	cumulate	cumulate	cumulate	cumulate	isotropic	isotropic	s-granite	syenite
SiO2	42.27	42.43	43.38	42.14	46.72	43.57	70.67	61.44
TiO2	1.62	1.66	1.64	1.40	2.55	2.87	0.11	0.32
Al2O3	8.37	8.54	8.12	7.83	14.84	12.65	14.54	18.83
Fe2O3	14.67	14.4	14.48	13.96	10.92	12.72	2.14	3.28
MnO	0.20	0.20	0.19	0.19	0.16	0.17	0.13	0.04
MgO	22.96	22.03	21.02	23.82	6.75	11.72	0.35	0.66
CaO	5.59	5.82	7.41	5.71	10.89	9.68	0.83	1.30
Na2O	1.66	1.89	1.56	1.46	3.24	2.57	5.46	6.53
K2O	0.99	1.07	0.93	1.02	1.46	1.34	4.44	5.93
P2O5	0.28	0.31	0.24	0.27	0.46	0.42	0.10	0.26
LOI	0.70	0.96	0.64	1.41	1.23	1.63	0.50	1.09
Total	99.31	99.31	99.61	99.21	99.22	99.34	99.27	99.68
Sc	18	18	26	18	32	30	n.a.	n.a.
V	202	208	250	183	289	353	8	27
Cr	1250	1150	1260	1470	265	455	< 1	< 1
Co	96	90	88	94	36	55	< 2	< 2
Ni	751	698	703	847	99	246	2	3
Cu	56.6	58.6	80.1	55.5	56.2	62.1	3	10
Zn	91	90	83	82	77	79	17	43
Ga	12	12	12	9	23	20	15	20
Rb	26.3	27.3	25.5	32.1	45.7	34.6	91	134
Sr	317	318	274	286	597	558	59.4	332
Y	14	14	14	12	24	22	15	18
Zr	109	117	104	106	141	156	254	360
Nb	21.9	24.3	20.1	20.3	36.8	39.5	9	15
Ba	131	138	122	148	189	206	128	574
Hf	2.86	3.2	2.98	2.8	4.31	4.31	0.24	0.31
Ta	1.84	2.2	2.03	2.19	2.99	3.34	0.69	0.75
Pb	2.21	2.4	2.22	2.63	3.24	3.67	19.9	27.8
Th	4.13	4.2	4.16	5.49	4.5	3.98	19.5	9.82
U	1.02	1.1	0.97	1.29	1.27	1.14	0.91	1.34
La	13.6	15.1	12.4	13.4	22.5	24.3	53.5	93.2
Ce	30.9	34.1	28.7	30.6	50.4	51.9	89.2	155
Pr	4.1	4.4	3.9	4.1	6.5	6.6	8.2	12.7
Nd	17.2	18.5	16.9	17.1	27.5	27.1	26.1	39.1
Sm	3.82	4	3.8	3.57	6.22	5.85	3.73	4.47
Eu	1.18	1.2	1.17	1.09	1.96	1.87	0.38	1.44
Gd	3.75	3.8	3.72	3.34	6.14	5.72	4.19	5.12
Tb	0.53	0.6	0.54	0.47	0.9	0.82	0.47	0.48
Dy	2.84	3	2.88	2.54	4.85	4.4	2.57	2.41
Ho	0.52	0.5	0.53	0.47	0.89	0.8	0.50	0.39
Er	1.37	1.4	1.39	1.21	2.32	2.09	1.38	0.86
Tm	0.18	0.2	0.18	0.16	0.31	0.28	0.19	0.09
Yb	1.1	1.2	1.11	0.99	1.87	1.68	1.2	0.47
Lu	0.16	0.2	0.15	0.14	0.26	0.24	0.14	0.04

n.a. = not analysed.

Fig. 5

(a) Eisenkappel gabbros plotted on the TiO2 vs. Al2O3 discrimination diagram of Pearce (1983). Note that only isotropic gabbros plot in the basalt (= liquid) field. V&Z = Visonà and Zanferrari (2000). (b) Total alkali vs. Silica diagram of Le Maitre et al. (1989) with subalkaline-alkaline fields after Miyashiro (1978) showing the compositional range of the analysed Eisenkappel gabbros and granites and other non-hybrid felsic rocks from the Karavanke pluton (Bole et al., 2001, Visonà and Zanferrari, 2000). (c) Chondrite-normalized (Boynton, 1984) REE patterns and (d) primitive-mantle normalized incompatible trace element patterns of the Eisenkappel gabbros compared with average ocean island (OIB) and mid-ocean ridge (MORB) basalts (PM-normalizing values, OIB and MORB: McDonough and Sun, 1995). Note the absence of Eu and negative Nb–Ta–Ti anomalies.

The cumulate gabbros are rich in both Ni and Cr, and poor in Sr when compared to the isotropic gabbros, in agreement with the higher modal proportions of Ol + Sp, and the low modal proportion of plagioclase. There are broadly positive correlations between Zr, Nb, P2O5, Ba and total REE. The REE patterns are LREE enriched (CeN/YbN = 6.7–8.0) and parallel but at different levels, with HREE about 6 times C1 for cumulates and about 9 times C1 for isotropic gabbros (Fig. 5c). The absence of a positive Eu anomaly (Eu/Eu* = 0.95–0.99) indicates that plagioclase was not a major accumulating mineral. Fig. 5d shows that LREE and other highly incompatible trace elements are strongly enriched compared to N-MORB, whereas Y and the HREE are depleted. Negative Nb–Ta–Ti-anomalies that would indicate subduction-related magmatism and/or significant crustal contamination are absent from the primitive-mantle (PM) normalized trace element patterns of both cumulate and isotropic gabbros.

The Sr isotope compositions of six gabbro whole rocks are uniformly low, showing little variation between 0.70284 and 0.70336, when calculated for t = 250 Ma, based on the Sm–Nd age of sample 94T31EK. At similar 147Sm/144Nd ratios (0.1263–0.1363), the ε(t)Nd values are somewhat variable, ranging between + 5.4 (sample 08EK18) and + 2.5 (sample 08EK10) (Tab. 4).

In the TAS diagram (Fig. 5b) the analysed rocks plot in the syenogranite (08EKR12) and syenite (08EKR13) fields, as does the majority of felsic rocks from the Karavanke pluton (Bole et al., 2001, Visonà and Zanferrari, 2000). Using the geochemical classification of granitic rocks of Frost et al. (2001), the granites are ferroan (FeOtot/( FeOtot + MgO) = 0.82–0.85), alkalic (MALI = 9.1–11.1) and metaluminous (ASI = 0.95–0.97; Na + K < Al). They have relatively high sodium (> 5.4 wt.% NaO) and contain 0.2–1.2 wt.% normative Di. These parameters are typical of within-plate plutons (Frost et al., 2001). The Eisenkappel felsic rocks are more enriched in LREE (Table 3) and have larger LREE/HREE ratios than the mafic rocks (CeN/YbN = 19–85). HREE patterns are strongly fractionated in sample 08EKR13 (TbN/YbN = 4.59) and relatively flat in sample 08EKR12 (TbN/YbN = 1.7), suggesting equilibration with garnet-bearing and garnet-free sources, respectively. The large negative Eu− (Eu/Eu* = 0.29) and Sr-anomalies seen in sample 08EKR12 indicate that plagioclase control was important (Fig. 6a). Both samples show negative Ta–Nb, Sr, P and Ti anomalies in a chondrite-normalized plot (Fig. 6b). These trace element characteristics are similar to trends seen in other granitoid rocks from the Karavanke pluton. Initial 87Sr/86Sr ratios are in the range 0.7067–0.7082 and coupled with ε(t)Nd values of − 2.6 (Table 4).

Fig. 6

(a) Chondrite-normalized (Boynton, 1984) REE patterns and (b) chondrite-normalized (Anders and Grevesse, 1989) incompatible trace element diagram of Eisenkappel syenogranite 08EKR12 and syenite 08EKR13. The gray field represents felsic NKPB samples analysed by Visonà and Zanferrari, 2000, Bole et al., 2001. Note the strongly fractioned pattern of syenite 08EKR13 that suggests equilibration with a garnet-bearing source.

Table 4

Rb–Sr and Sm–Nd data for whole rocks and minerals of gabbros and syenites, Eisenkappel meta-igneous suite.

Sample	Rb, ppm	Sr, ppm	87Rb/86Sr	87Sr/86Sr	± 2σm	wr-mineral age	Sri (250 Ma)
94T33EK wr	30.15	338.3	0.2578	0.703851	0.00005		0.70294
94T33EK Bt (1)	264.7	27.1	28.51	0.797139	0.00002	232.2 ± 2.3
94T33EK Bt (2)	250.5	23.2	31.50	0.797105	0.00004	209.9 ± 2.1
08EK06 wr	27.55	292.2	0.2727	0.704153	0.000003		0.70318
08EK07 wr	43.28	543.2	0.2305	0.703655	0.000004		0.70284
08EK10 wr	34.00	289.8	0.3395	0.704564	0.000005		0.70336
08EK18 wr	34.04	542.2	0.1816	0.703523	0.000003		0.70288
94T31EK wr	27.80	327.5	0.2456	0.704094	0.000003		0.70322
							Sri (240 Ma)
08EKR12 wr	93.2	59.4	4.5458	0.724367	0.000004		0.70887
08EKR13 wr	134.5	315.1	1.2351	0.711068	0.000004		0.70686
Sample	Sm, ppm	Nd, ppm	147Sm/144Nd	143Nd/144Nd	± 2sm	age / initial	ε Nd (t)
94T33EK wr	4.27	19.56	0.1320	0.512722	0.000006
Pl	0.097	1.125	0.0521	0.512577	0.000013
Cpx 1	9.89	40.09	0.1492	0.512739	0.000009
Cpx 2	8.86	35.56	0.1507	0.512745	0.000005
wr, repeat	4.15	19.19	0.1309	0.512712	0.000002	250 ± 26 (n = 7)	ε (250) = + 3.6
Ap	135.5	818.1	0.1001	0.512668	0.000002	ε (t) = + 3.6
Cpx 4MF	11.41	44.97	0.1534	0.512751	0.000002	MSWD = 3.2
08EK06 wr	3.94	17.48	0.1363	0.512719	0.000002		ε (250) = + 3.5
08EK07 wr	6.33	28.30	0.1354	0.512769	0.000002		ε (250) = + 4.5
08EK10 wr	3.65	17.50	0.1263	0.512652	0.000002		ε (250) = + 2.5
08EK18 wr	6.09	28.43	0.1294	0.512804	0.000002		ε (250) = + 5.4
94T31EK wr	3.91	17.88	0.1323	0.512714	0.000002
Cpx 1MF	7.55	29.70	0.1537	0.512750	0.000004	249.0 ± 8.4 (n = 6)
Cpx b 2MF	9.74	37.87	0.1555	0.512753	0.000004	ε (t) = + 3.6	ε (249) = + 3.6
“Cpx” g 2MF	3.39	13.36	0.1528	0.512748	0.000004	MSWD = 0.11
Pl	0.43	3.32	0.0784	0.512628	0.000004
Ap	27.35	162.6	0.1017	0.512665	0.000003
08EKR12 wr	3.84	26.8	0.08649	0.512327	0.000004
Grt	2.82	2.88	0.59083	0.513129	0.000005	242.1 ± 2.1 (n = 4)
Grt 2 R	2.60	1.10	1.42229	0.514440	0.000004	ε (t) = − 2.6	ε (242) = − 2.6
“ L	15.2	78.9	0.11625	0.512374	0.000004	MSWD = 0.15
08EKR13 wr	4.68	40.99	0.06903	0.512304	0.000001
Grt (1) R	5.50	6.50	0.51158	0.512995	0.000002	238.4 ± 1.9 (n = 3)	ε (238) = − 2.6
Grt (2) R	4.83	2.97	0.98471	0.513732	0.000004	ε (t) = − 2.6
Grt < 0.073 (SF)	11.59	63.43	0.11049	0.512372	0.000002	MSWD = 0.023

Geochronological results: Sm–Nd and Rb–Sr mineral–wr dating

Rb–Sr and Sm–Nd ID-TIMS analytical data of whole rock splits (wr) and of pure, handpicked mineral concentrates (plagioclase, clinopyroxene, apatite, garnet, and biotite) from six gabbros and two felsic lithologies are listed on Table 4, results are shown in Fig. 7.

Fig. 7

Sm–Nd mineral-whole rock isochron diagrams of mafic (a, b: gabbros 94T31EK and 94T33EK) and felsic members (c: syenogranite 08EKR12; d: syenite 08EKR13) of the Eisenkappel igneous suite. See text for discussion.

Abbreviations used in Table 4 and Fig. 7: Ap = apatite, Bt = biotite, Cpx = clinopyroxene, Grt = garnet, Pl = plagioclase, wr = whole rock; hp = handpicked, R = residual, L = leachate from H2SO4 leaching experiment.

Sm–Nd data

Gabbroic cumulate 94T31EK

The data points of five mineral concentrates (Ap, Pl, and 3 Cpx fractions) and the whole rock are aligned along a best-fit regression line corresponding to an age of 249.0 ± 8.4 Ma and an initial εNd value of + 3.6 (Fig. 7a). Despite the overall good fit of the data points (MSWD = 0.11; n = 6), internal age precision remains moderate; this is due to the limited spread in 147Sm/144Nd (0.0784 to 0.1555), which encompasses not even 50% of the total spread in Sm/Nd (see Discussion section, below).

Gabbroic cumulate 94T33EK

Five mineral fractions (Pl, Ap, and 3 Cpx fractions) and two whole rock splits show a moderate fit (MSWD = 3.2) along a regression line defining an age of 250 ± 26 Ma and an initial 143Nd/144Nd isotopic ratio of 0.51250 ± 0.000023 (εNd250 =+ 3.6) (Fig. 7b). Though with greater uncertainty, this result is identical to that obtained from sample 94T31EK.

Syenogranite 08EKR12

Two garnet separates (one of which was leached with concentrated H2SO4) and the whole rock define a best-fit (MSWD = 0.26) isochron age of 242.0 ± 2.1 Ma, and an initial 143Nd/144Nd of 0.512191 ± 0.000004 (εNd = − 2.6) (Fig. 7c). If the leachate of the second garnet fraction (Grt 2 L) is included in the regression, the result is identical (n = 4: t = 242.1 ± 2.1 Ma; MSWD = 0.15), demonstrating that LREE-rich inclusions are in equilibrium both with the garnet host and the whole rock and have, therefore, no falsifying effect on the “true” garnet age.

Syenite 08EKR13

The crushed garnet material from this sample has been prepared by isolating 2 half crystals, each ca. 1 cm in diameter, using a microsaw. From this sieved crushate (0.16–0.4 mm sieve fraction), two handpicked, leached garnet separates: Grt (1) R and Grt (2) R, an impure sieve fraction extracted from the garnet crushate (Grt < 0.073, SF, containing all the inclusions) and the whole rock have been analysed. The two garnet residues (R) and the whole rock define a best-fit (MSWD = 0.023) isochron age of 238.4 ± 1.9 Ma, and an initial 143Nd/144Nd of 0.512196 ± 0.000006 (εNd = − 2.6) (Fig. 7d). The impure fine sieve fraction does not perfectly fit the isochronous relationship, indicating slight isotopic disequilibrium (MSWD for n = 4 is 2.3), though the age result is almost identical (t = 238.5±1.9).

Rb–Sr data

Two different magnetic biotite fractions of gabbro sample 94T33EK were analysed. They yielded significantly different Triassic wr-Bt age results of 232.2 ± 2.3 and 209.9 ± 2.1 Ma (Table 4).

Discussion

Magma source and differentiation of the mafic rocks

The low initial Sr isotope ratios (0.7028–0.7033), the positive εNd(t) values (Fig. 8a; Table 4) and geochemical constraints clearly indicate a mantle origin for the Eisenkappel gabbros. The isotropic gabbro sample 08EK18 could represent a near-primary magma: 11.7 wt.% MgO, Mg# = 69, 455 ppm Cr, 246 ppm Ni, and a calculated equilibrium olivine composition of Fo87–89, assuming KD = 0.27–0.33 (Roeder and Emslie, 1970). It also has a low 87Sr/86Sr initial ratio (0.70288) and the highest εNd(t) (= + 5.4) within the suite. These isotopic characteristics plot in the depleted mantle quadrant (Fig. 8a), and may best approximate the isotopic composition of the mantle source.

Fig. 8

143Nd/144Nd vs. 87Sr/86Sr diagrams for the Eisenkappel gabbros and granites. (a) Note the narrow range in the initial Sr composition (at 250 Ma) of the gabbroic rocks and the relatively wide range in their initial Nd isotopic ratios, expressed as εNd. DMM = depleted MORB mantle at t = 250 Ma. (b) a model AFC curve links the most primitive mafic rock and syenogranite 08EKR12. The assumption that Sr behaves as incompatible element is in agreement with subordinate modal plagioclase in the mafic cumulates. See text for discussion.

Ratios between incompatible elements with similar geochemical characteristics can also be used to monitor mantle source compositions because they are not significantly affected by fractionation processes during mafic melt genesis. The Zr/Nb ratios of the Eisenkappel gabbros vary between 3.8 and 5.2 with an average of 4.6, comparable to OIB and rift-related lavas (e.g. Pearce and Norry, 1979). Th/U ratios varying between 3.5 and 4.3 with an average of 3.9 are also comparable to OIB. Thus, the Eisenkappel gabbros appear to have been derived from a slightly enriched mantle source. Low YbN and LuN values (< 9) at high LaN/LuN (8.6–10.5) and high GdN/YbN (2.7–2.8) suggest the presence of residual garnet in the mantle source. This garnet signature implies that melting started at a depth of ≥ 60 km (Klemme and O'Neill, 2000).

The high Mg/(Mg + Fe*) ratios (0.74–0.77), high Cr (1150–1470 ppm) and Ni (712–858 ppm) contents of samples 94T31EK, 94T33EK, 08EK06 and 08EK10 clearly identify these as cumulates. Petrographic and geochemical data suggest that they formed by low-pressure crystallization that began with olivine ± Cr-spinel, followed by clinopyroxene and plagioclase, and finished with brown amphibole, phlogopite, ilmenite and titanomagnetite. The presence of appreciable amphibole, phlogopite and magnetite indicates that the magma was volatile rich and oxidizing. The appearance of minor orthopyroxene could have resulted from late-stage fractionation of pore liquids. Alternatively, oxidation of the system could have stabilized orthopyroxene and magnetite by a reaction of the type 3 Ol + ½ O2 = 3 Opx + Mt. Oxidation could also explain the exsolution lamellae in olivine: as discussed by Markl et al. (2001), these composite exsolution lamellae formed during rapid cooling of originally Ca-rich olivine due to an overstep of the reaction Fa + Kst (kirschsteinite = Ca-rich component in olivine) + 1/3 O2 = Hd + 2/3 Mt. The oxygen necessary for this reaction could be supplied from olivine with an excess Fe3+ component or from a late-stage residual melt or fluid. As there is no indication of an Fe3+ component in the EK olivine, the latter possibility seems more likely, especially since the formation of calcic amphibole in some of the composite exsolution lamellae and at olivine/plagioclase interfaces in addition to late-stage kaersutite and phlogopite also require the presence of an H2O-rich fluid. Heterogeneous exsolutions of clinopyroxene + spinel ± orthopyroxene have been reported by Moseley (1984) in olivine from the Rhum, Skye, Bushveld and Skaergaard intrusive complexes and by Markl et al. (2001) in olivine from the Ilimaussaq intrusion.

The relatively low Fo content (Fo76-78) of olivine suggests that the EK cumulates crystallized from already evolved parental liquids. The somewhat higher 87Sr/86Sr(i) and lower εNd(t) of the cumulate gabbros suggests contamination, the amount of which remains unconstrained (Fig. 8a). The absence of negative Nb–Ta–Ti anomalies argues (Fig. 6d) against significant crustal contamination. However, Nb/U (15.7–20.7 in cumulates, 28.9–34.6 in isotropic gabbros) is low compared to the range of MORB and OIB sources (47 ± 10; Hofmann, 1997). In conjunction with low Ce/Pb (11.6–14.2 in cumulates, 14.1–15.6 in isotropic gabbros, 25 in MORB and OIB) and slightly elevated Th/Nb (0.10–0.27, highest in cumulates) this suggests some influence from continental crust and could be explained as a result of mixture between an OIB-type mantle source and continental crust.

The negative correlation between Sr and Nd isotopic compositions can be interpreted as result of AFC (assimilation fractional crystallization) processes or mixing involving a depleted mantle melt and crust. The analysed Eisenkappel samples define a curved trend in Sr–Nd space that is close to an AFC model curve (DePaolo, 1981), assuming constant partition coefficients D, a constant assimilation/fractionation ratio (r) of 0.5 and a silicic melt similar in composition to 08EKR12 as contaminant (Fig. 8b). However, for the model curve to be close to the data points, Sr must be an incompatible element and Nd should have a D value close to 1. This is unusual in basaltic systems, but could apply to cumulates where olivine, clinopyroxene and amphibole, and not plagioclase, are the dominant fractionating phases since DSr is a function of the amount of plagioclase fractionation (DePaolo, 1981). An AFC process was also postulated by Visonà and Zanferrari (2000) to explain the shift in initial 87Sr/86Sr ratios from a gabbro enclave (sample EK11b: 0.70313) to diorite (sample EK12: 0.70334) and monzonite (sample EK2: 0.70525), followed by fractional crystallization to granite (sample EKX: 0.70473). However, this remains controversial because their Sr iotope initial ratios were based on a biotite Rb–Sr cooling age of 221 Ma. Assuming a more realistic timing of 250 Ma, their samples would have had different initial 87Sr/86Sr ratios, precluding a purely FC relationship of granite and monzonite.

Granitic rocks

Although major element parameters of the NKPB granites are typical of within-plate plutons (Frost et al., 2001), most of them plot in the VAG (volcanic arc granite) field of Pearce et al. (1984). The position of the samples in the VAG field (Fig. 9), however, does not necessarily reflect their tectonic setting: it could reflect plagioclase accumulation. However, since plagioclase accumulation should have resulted in positive Eu anomalies, it is more likely that the trace element concentrations were inherited from the source rocks. Because the geochemical composition of granitic magmas is determined by the composition of their source rocks it is not possible to distinguish granitoids formed in a subduction setting and granitoids generated from source rocks that had been formed by earlier subduction events. In any case, Sr–Nd isotope data (Table 4) and depleted mantle model ages in the range of 0.8–0.9 Ga suggest derivation of the Eisenkappel samples by melting of older meta-igneous rocks in the middle to lower crust where mantle-derived melts provided the heat source for partial melting. Involvement of crustal melting in the generation of other NKPB granitoids was also postulated by Bole et al. (2001), whereas the low 87Sr/86Sr initial ratios of the samples analysed by Visonà and Zanferrari (2000) do not exclude the possibility that at least some of the silicic rocks were FC and/or AFC products of mantle melts. However, in the absence of pertinent Nd isotopic data their model is not well constrained.

Fig. 9

Rb vs. (Y + Nb) discrimination diagram for granites (Pearce et al., 1984) showing that the analysed syenite and syenogranite from Eisenkappel plot in the field of volcanic arc granites (VAG) as does the majority of other A-type granitic rocks of the Northern Karawanken pluton (data from Visonà and Zanferrari, 2000, Bole et al., 2001), suggesting an input by melting of older meta-igneous rocks. ORG = ocean-ridge granite, WPB = within-plate granite, syn-COLG = syn-collisional granite.

Interestingly, although metaluminous, the Eisenkappel granitoids contain garnet. Almandine garnets in granites can be (1) primary phenocrysts, (2) xenocrysts derived from disaggregated country rocks or (3) restite phases. The fact that garnets in samples 08EKR12 and 08EKR13 are intergrown with K-feldspar phenocrysts and contain inclusions of quartz, K-feldspar, apatite and zircon compositionally akin to matrix minerals suggests that they are primary igneous phases. Primary almandine-rich garnet is an uncommon mineral in calc-alkaline igneous rocks but petrologically significant. The garnets are zoned, showing an enrichment of Mn at the expense of Fe and Mg towards the rims. Sps-reverse zoning of primary garnet has been documented in several granitic plutons (e.g. Dahlquist et al., 2007, Du Bray, 1988, Manning, 1983, Stone, 1988). Experimental studies of garnet in silicic igneous rocks (Clemens and Wall, 1981, Green, 1977) have shown that increasing MnO contents will stabilize garnet to lower pressure and that magmatic crystallization at different crustal levels could explain the reverse zoning patterns of garnet. Igneous garnet zoning is mainly controlled by the chemical potentials of the MnO component in garnet and in the melt, and by the rate at which Fe and Mn can diffuse to, and within, the growing crystal (e.g. Manning, 1983). Diffusion rates for Mn in garnet are negligible below 640 ± 30 °C, whereas self-diffusion of Mn is sufficiently rapid above ca. 700 °C to eliminate any compositional growth zoning (Manning, 1983, Yardley, 1977). Thus the homogeneous Mn-poor (< 10 mol% Sps) garnet cores in sample 08EKR12 suggest crystallization at temperatures above 700 °C and at pressures of 0.5–0.7 GPa, whereas the Mn-rich rims (> 10 mol% Sps) document that crystallization ceased at temperatures below 640 ± 30 °C at 0.5 GPa or less. The garnet zoning observed in sample 08EKR13 is more complex. It shows that equilibrium was achieved only between magma and the exterior of grains and could reflect changes in the composition of the magma. Sps contents of > 16 mol% within the core domain suggest garnet crystallization at pressures of ≤ 0.5 GPa.

Geothermobarometry and pluton emplacement

Nimis and Ulmer (1998) performed crystal structure modeling of Ca-rich clinopyroxene coexisting with ultrabasic and basic melts, calibrating a geobarometer based on structural parameters. The temperature-independent clinopyroxene geobarometer (BA) of Nimis, 1999, Nimis and Ulmer, 1998 can be used for anhydrous basic Qtz- to Ne-normative melts but underestimates pressures by ca. 100 MPa per 1% H2O in the melt. This calibration therefore yields only minimum pressures (220–250 MPa), and the standard error is large (Nimis, 1999). The expanded version of the geobarometer (BH) that can be applied to hydrous basaltic systems is strongly temperature-dependent and requires a precise temperature input. Thus, underestimating T by 20 °C would increase the calculated pressure by 100 MPa.

The semiquantitative Ca-amphibole thermobarometer of Ernst and Liu (1998) suggests T ≥ 1000 °C and P ≤ 600 MPa for the kaersutite crystallization. Crystallization temperatures for magmatic amphiboles calculated with the amphibole-plagioclase thermometer of Holland and Blundy (1994) range from 980 to 1000 °C, assuming pressures between 0.5 and 1 GPa. The Ti-in-amphibole thermometer of Otten (1984) yields 1035–1050 °C for magmatic kaersutite in the cumulate rocks and 978–998 °C for magmatic Ti-pargasite in the isotropic gabbros. The thermobarometer of Ridolfi et al. (2010), based on different compositional components of amphibole, suggests kaersutite crystallization at 1020–1036 °C and 407–445 MPa from melts containing 3.4–4.0 wt.% H2O.

For the analysed granitic rocks, zircon saturation temperatures based on total Zr abundances (Watson and Harrison, 1983) are 812 and 829 °C. This compares well with the temperature range of 755–838 °C for monzonitic to granitic NKPB rocks, calculated with the data of Visonà and Zanferrari (2000), and of 739–791 °C, based on the data of Bole et al. (2001). Discrete grains of sodic and potassic feldspars require pressures of at least 500 MPa if PH2O equaled lithostatic pressure. Because this is unlikely, 500 MPa may be the limiting pressure, suggesting depths of emplacement of less than 15 km. According to Exner, 1972, Exner, 1976, Visonà and Zanferrari, 2000, the NKPB granites were emplaced at shallow levels in the crust, at a depth of ca. 4–8 km.

The fact that the initial Sr–Nd isotopic compositions clearly indicate different sources for the analysed EK gabbros and granitoids rule out a direct parent–daughter relationship, suggesting that these mafic and felsic rocks are not genetically related through fractionation. Mafic magmatic enclaves common throughout the NK plutonic complex document that mafic and felsic magmas interacted during ascent and emplacement, involving processes such as mingling (Fig. 2c–d), entrainment of crystals from previously consolidated magma batches and magma mixing (Bole et al., 2001, Exner, 1972, Visonà and Zanferrari, 2000). The association of mantle-derived mafic rocks and evolved igneous rocks strongly suggests that fractionating mafic magmas provided the heat source for partial melting of the crust and the formation of granite (e.g., Hildreth, 1981).

Sm–Nd ages: Timing of crystallization and magma emplacement

Since diffusivity of Nd in clinopyroxene is slow (Davidson et al., 2007, Sneeringer et al., 1984) and the assemblages show only minor signs of post-magmatic alteration, the Sm–Nd data of the gabbros are interpreted to date igneous crystallization of the cumulate assemblages. The Sm–Nd mineral-whole rock ages in Fig. 7a–b therefore indicate that the Eisenkappel gabbros crystallized in the time window ca. 260–240 Ma ago.

Though clearly overlapping within analytical errors, the Sm–Nd ages (Table 4, Fig. 7) suggest that the gabbros crystallized somewhat earlier than the syenitic-granitic rocks. If the closure temperature for the Sm–Nd system in garnet is taken as ca. 700 °C (e.g., Tirone et al., 2005), the Sm–Nd garnet–wr ages of 242.1 ± 2.1 (syenogranite) and 238.4 ± 1.9 Ma (syenite) (Fig. 8c, d) for the more acidic members of the NKPB would date an initial stage of cooling, rather than igneous crystallization s. str., since their magmatic crystallization temperature is estimated at ≥ 700–800 °C. Therefore, the coincidence of all ages in Fig. 7 proves within analytical uncertainties that mafic and more acidic members of the NKPB are broadly contemporaneous, as also clearly documented by the petrographic/microstructural and geochemical data.

A potential problem for dating major phases in these gabbros using the Sm–Nd mineral dating technique is the relatively high modal content of apatite, because apatite is a typically L-MREE-enriched accessory phase (Thöni et al., 2008b) and may control both the Nd and the Sr budget of the assemblage. If present as undetected inclusions or intergrowths in the handpicked separates, e.g. in low-LREE plagioclase, even small quantities of this Nd-rich tracer phase would influence (i.e., shrink) the relative spread, pushing the data point for plagioclase towards higher values, since apatite has generally Sm/Nd ratios intermediate between pure Pl and Cpx. To a certain extent, this may apply to the present case, since (based on Sm and Nd partition behavior in the Pl-Cpx pair) pure Pl is expected to show somewhat lower Sm/Nd ratios than those actually analysed. However, the good internal fit of the data (e.g. sample 94T31EK, MSWD = 0.11 for n = 6, Fig. 7a) at a narrow relative internal spread (< 50%) indicates rather perfect Nd isotope equilibrium in the system; hence, apatite inclusions in major phases could lower age precision, but have, in the present case, no falsifying effect on the slope of the regression line (i.e., on the age).

Regional geologic implications

The geochronological data presented in this study and those published earlier (Cliff et al., 1975, Lippolt and Pidgeon, 1974, Scharbert, 1975), clearly document a Late Permian to Early Triassic (ca. 260–230 Ma) magmatic emplacement age for both mafic and granitic members of the NKPB. Older U–Pb ages in zircons could reflect age inheritance from pre-Permian or early magmatic stages (Genser and Liu, 2010). Permian to Lower Triassic ages (mainly Rb–Sr whole rock and mineral ages and U–Pb zircon ages) have been determined for a number of intrusive and extrusive rocks from the Southern Alps and the southern Eastern Alps, e.g., the Brixen-Iffinger-Monte Croce intrusives (Borsi et al., 1972, Thöny et al., 2008), the Monte Sabion/Cima d'Asta (Borsi et al., 1966), the Predazzo and Monzoni intrusive complexes (Borsi and Ferrara, 1967, Borsi et al., 1968), the “Athesian Volcanic Group” (Klötzli et al., 2003, Marocchi et al., 2008), but also from more westerly segments of the Alpine chain, e.g., the Ivrea-Verbano Zone (Mayer et al., 2000, Voshage et al., 1990).

The majority of these Permian-Triassic magmatic rocks are aligned along the Periadriatic Lineament (PAL, Fig. 1) (Exner, 1976), a fault system that is suggested to trace a significant, and at least Late Palaeozoic paleogeographic border between Northern and Southern Alps. The fact that important Permian to Triassic igneous bodies are present both north and south of this lineament could indicate a relative proximity of Austroalpine and Southalpine units since Late Palaeozoic time.

In the Karawanken area, the Permian-Triassic igneous suite and the Oligocene tonalite form conspicuously thin, steep “lamellae”, oriented parallel to the E–W trending PAL (Fig. 1). These elongated forms resulted from “adaptation” during large-scale, flexural deformation of the lithosphere and Cenozoic transpressive strike-slip movements along the active PAL (Nemes et al., 1997). This could have occurred syn-intrusive in the case of the Oligocene Karawanken tonalite, allowing ductile deformation of the hot tonalitic magma during emplacement in the higher crust. In contrast, the “ribbon-like” appearance of the Permo-Triassic Eisenkappel intrusives is largely a product of brittle Alpine deformation along discrete E−W trending transpressive/strike-slip faults (Exner, 1972, Nemes et al., 1997, Polinski and Eisbacher, 1992, Wölfler et al., 2010).

The magmatic crystallization age of the NKPB fits a scenario characterized by widespread igneous activity and metamorphism at sillimanite and andalusite grade that affected large areas of the southern Austroalpine basement nappes, now exposed NE of the PAL, in Late Palaeozoic/Early Mesozoic time (Habler et al., 2007, Schuster et al., 2001, Thöni and Miller, 2009, Thöni et al., 2008b). Schuster and Stüwe (2008) suggested that this Permian thermal event reflects orogen-scale “lithospheric thinning accompanied by magmatic underplating”. The Eisenkappel igneous suite was generated within an Austroalpine crustal segment near the southern margin of Meliata that experienced large-scale N–S extension due to westward propagation of the Neotethys ocean in the Permian (Schuster and Stüwe, 2008). It presently forms part of the Drauzug-Gurktal nappe system (Schmid et al., 2004, Schuster, 2004), characterized by Alpine low to very low-grade conditions (Fig. 1). The geochemical affinities of the Eisenkappel gabbros indicate parental OIB-like mafic mantle melts generated in an intra-plate rifting environment. Our data suggest that these mantle-derived melts have interacted with continental crustal material during ascent and induced partial melting of continental crust: the depleted mantle model ages of the analysed granitic rocks clearly document derivation of some of the felsic NKPB melts from old crustal sources. In contrast, the Permian gabbroic protoliths of the Cretaceous Saualpe-Koralpe-Pohorje (SKP) eclogites of the adjacent Austroalpine basement nappes (Koralpe-Wölz HP-nappe system after Schmid et al., 2004) are derived from a N-MORB type mantle source along a Permian rift zone that was probably connected with the westernmost Tethyan oceanic domains (e.g., Handy et al., 2010, Miller and Thöni, 1997, Miller et al., 1988, Miller et al., 2007). Reworking of older, differentiated sources is not documented in the SKP eclogite protoliths.

Conclusions

– Gabbros from the Northern Karawanken Plutonic Belt (NKPB) near Eisenkappel (S Carinthia), close to the Periadriatic Fault, are essentially composed of 40–45 modal % olivine, 10–15 modal % plagioclase, 5–10 modal % clinopyroxene, 35–46 modal % brown amphibole and phlogopite.

– Based on major and trace element data and Sr–Nd isotopic compositions, the gabbros are alkaline and show within-plate geochemical characteristics. They suggest anorogenic magmatism in an extensional setting and derivation from an enriched mantle source.

–The volumetrically dominant intermediate and felsic rocks (syenites and syenogranites) of this multiphase plutonic suite developed by a combination of processes, including assimilation, fractional crystallization, crustal melting, magma mixing and mingling.

–Temperature and pressure conditions for igneous crystallization are around 1000 ± 20 °C and 380–470 MPa for the gabbros, whereas the granitic rocks crystallized at T ≤ 800 ± 20 °C and ≤ 350 MPa.

–Sm–Nd mineral-whole rock dating of two cumulate gabbros yielded 249 ± 8.4 Ma and 250 ± 26 Ma (εNd: + 3.6), whereas garnet-whole rock Sm–Nd analyses of two silicic samples yielded ages of 238.4 ± 1.9 Ma and 242.1 ± 2.1 Ma (εNd: − 2.6). Within error limits, these ages are identical, suggesting an intrusion age of ca. 250 Ma.