Journal of Petrology Advance Access originally published online on January 4, 2006
Journal of Petrology 2006 47(4):705-744; doi:10.1093/petrology/egi091
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Low-pressure Granulites of the Li
ov Massif, Southern Bohemia: Viséan Metamorphism of Late Devonian Plutonic Arc Rocks
CH JANOU
EK1,2,*CH ERBAN2
1 DIVISION OF MINERALOGY & MATERIAL SCIENCE, UNIVERSITY OF SALZBURG, HELLBRUNNERSTRASSE 34, A-5020 SALZBURG, AUSTRIA
2 CZECH GEOLOGICAL SURVEY, KLÁROV 3/131, 118 21 PRAGUE 1, CZECH REPUBLIC
3 INSTITUTE OF MINERALOGY, J. W. GOETHE-UNIVERSITY, SENCKENBERGANLAGE 28, D-60054 FRANKFURT, GERMANY
4 DIVISION OF GENERAL GEOLOGY & GEODYNAMICS, UNIVERSITY OF SALZBURG, HELLBRUNNERSTRASSE 34, A-5020 SALZBURG, AUSTRIA
5 DIVISION OF EARTH SCIENCES, UNIVERSITY OF GLASGOW, GLASGOW G12 8QQ, UK
RECEIVED DECEMBER 20, 2004; ACCEPTED NOVEMBER 11, 2005
 |
ABSTRACT |
|---|
The Li
ov Granulite Massif differs from neighbouring granulite bodies in the Moldanubian Zone of southern Bohemia (Czech Republic) in including a higher proportion of intermediate–mafic and orthopyroxene-bearing rocks, associated with spinel peridotites but lacking eclogites. In addition to dominantly felsic garnet granulites, other major rock types include quartz dioritic two-pyroxene granulites, tonalitic granulites and charnockites. Minor bodies of high-pressure layered gabbroic garnet granulites and spinel peridotites represent tectonically incorporated foreign elements. The protoliths of the mafic–intermediate granulites (quartz-dioritic and tonalitic) crystallized 
360–370 Ma ago, as indicated by laser ablation inductively coupled plasma mass spectrometry U–Pb ages of abundant zircons with well-preserved magmatic zoning. Strongly metamorphically recrystallized zircons give ages of 330–340 Ma, similar to those of other Moldanubian granulites. For the overwhelming majority of the Liov granulites peak metamorphic conditions probably did not exceed 800–900°C at 4–5 kbar; the equilibration temperature of the pyroxene granulites was 670–770°C. This is in sharp contrast to conditions of adjacent contemporaneous Moldanubian granulites, which are characterized by a distinct HP–HT signature. The mafic–intermediate Liov granulites are thought to have originated during Viséan metamorphic overprinting of metaluminous, medium-K calc-alkaline plutonic rocks that formed the mid-crustal root of a Late Devonian magmatic arc. The protolith resembled contemporaneous calc-alkaline intrusions in the European Variscan Belt. KEY WORDS: low-pressure granulites; geothermobarometry; laser-ablation ICP-MS zircon dating; whole-rock geochemistry; Sr–Nd isotopes; Moldanubian Zone
|
INTRODUCTION |
|---|
Granulite-facies metamorphic rocks provide crucial insights into the conditions and timing of continental collision, crustal thickening and orogenic collapse. Ultimately granulite-facies metamorphism can result in the production of granitic to tonalitic melts, and is, thus, a major process leading to crustal differentiation (e.g. Arculus & Ruff, 1990
; Clemens, 1990). Hence careful and comprehensive study of these rocks provides a key to understanding the nature and development of ancient orogens, including the Variscan collisional belt in Central Europe (Pin & Vielzeuf, 1983; O'Brien & Rötzler, 2003).
The Variscan granulites of the Bohemian Massif have received much attention because many display a clear high-P–high-T metamorphic signature [see O'Brien & Rötzler (2003) and O'Brien (2006) for reviews]. The high pressures have commonly been attributed to subduction at an active Variscan continental margin; the reasons for the high temperatures are still a matter of debate. In particular, the Gföhl Assemblage of the Moldanubian Zone (Lower Austria, SW Moravia and southern Bohemia: Fig. 1), following what was probably the very first scientific account of granulites (von Justi, 1754), has become a classic and much investigated granulite-bearing terrain (e.g. Fiala et al., 1987b; Vellmer, 1992; O'Brien & Carswell, 1993; Wendt et al., 1994; Cooke, 2000; Kröner et al., 2000; O'Brien, 2000).
|
A distinctive feature of the Moldanubian granulites is the dominance of felsic types with igneous (leucogranitic) protoliths (Fiala et al., 1987b
; Vellmer, 1992; Janouek et al., 2004b); by contrast, intermediate–mafic igneous and metasedimentary granulites are relatively rare. The reasons for this distinction are not yet fully understood [see review by Janouek et al. (2004b)]. Unlike several larger granulite outcrops in southern Bohemia and Lower Austria (especially Blansk
les, St. Leonhard and Dunkelsteiner Wald) that have been the subject of numerous detailed studies (Carswell & O'Brien, 1993; Fiala et al., 1995; Vrána et al., 1995; Petrakakis, 1997; O'Brien & Rötzler, 2003; O'Brien, 2006, and references therein), the Liov Granulite Massif (LGM, Fig. 1b) has so far received little attention. This is unfortunate, as both petrological and geochemical data set it apart from other Moldanubian occurrences (see also Vrána & rámek, 1999). The most striking features of the LGM are:
- a predominantly low-pressure, medium- to high-temperature metamorphic record;
- an abundance of orthopyroxene-bearing rock types (mafic granulites–felsic charnockites);
- an absence of Al2SiO5 phases, corundum and cordierite in felsic granulites;
- a presence of spinel peridotites (ultramafic rocks in other Moldanubian granulite terranes are mostly garnet-bearing);
- an absence of eclogites.
The aim of this study is to characterize the granulites of the Liov Massif in terms of petrology and metamorphic evolution, together with the age, nature and petrogenesis of their igneous protoliths. The mechanism of LP granulite-facies metamorphism, following soon (20 Ma) after the formation of the protoliths, and the incorporation of the resulting LP granulites into HP–HT crustal units at the climactic stage of the Viséan continental collision, is of particular significance. Thus, the Liov granulites have important geotectonic implications for the interpretation of the development of the European Variscides, as most low-pressure granulite rocks documented so far formed much later in the orogen's history (at 300 Ma, Pin & Vielzeuf, 1983).
Here we also provide an explanation for the relative abundance of mafic to intermediate granulite types at Liov in comparison with other Moldanubian granulite massifs, and examine the sources and characteristics of the continental-arc magmatism that seems to have been responsible for the generation of their protoliths.
An additional general point is the ability of zircon to preserve growth zoning through subsequent high-grade metamorphism. The mechanism of metamorphic recrystallization of the zircon, ultimately leading to resetting of the U–Pb system, is critically assessed. The discussion demonstrates the utility of laser-ablation inductively coupled plasma mass spectrometry (LA–ICP–MS), in conjunction with optical cathodoluminescence (CL) and back-scattered electron (BSE) imaging, in resolving igneous and metamorphic ages separated by intervals of only a few tens of millions of years.
|
GEOLOGICAL SETTING |
|---|
South Bohemian granulites
The Moldanubian Zone of the Bohemian Massif (Fig. 1) is a tectonic assemblage of medium- to high-grade metamorphic rocks, intruded by numerous granitoid masses. It comprises several crustal segments of contrasting age, ranging from
2·1 Ga to Palaeozoic, but overall relationships are obscured by a complex polyphase deformational history (Dallmeyer et al., 1995, and references therein). The Moldanubian Zone has been subdivided into a structurally lower and mainly metasedimentary Drosendorf Assemblage and above this, the higher grade allochthonous Gföhl Assemblage (Fuchs & Matura, 1976; Franke, 2000).
The lower part of the Drosendorf Assemblage is referred to as the Monotonous Unit (Ostrong Unit) and consists mainly of partly migmatitic garnet–biotite–sillimanite paragneisses with minor orthogneisses and amphibolites. The upper part, the Variegated Unit (the Drosendorf Unit sensu stricto), comprises paragneisses with intercalations of amphibolite, calc-silicate gneiss, marble, quartzite and graphite schist. The Gföhl Assemblage includes abundant granulite- and eclogite-facies metamorphic rocks: anatectic gneisses with granulites, garnet and spinel peridotites, pyroxenites, dunites and eclogites (Fuchs & Matura, 1976; Fiala et al., 1995).
Liov Granulite Massif
A number of granulite bodies are present in southern Bohemia. The Blansk les Massif is the largest, followed by the K
it'anov, Prachatice, Liov and Krasejovka masses and bodies too small to be shown in Fig. 1b. The Liov Massif (c. 40 km2) lies c. 5 km east of 
eské Bud
jovice. Much of its western part consists of felsic granulites (Suk et al., 1978, 1981), but two small intermediate–mafic granulite masses crop out near Zvíkov to the east (Vajner, 1964) and there are also several spinel peridotite bodies, up to 0·7 km long (Fig. 2).
|
The Li
ov Granulite Massif (LGM) is bordered to the south and west by migmatitic biotite (± sillimanite and cordierite) paragneisses of the Monotonous and Variegated units. The southern limits of the massif are modified by faulting but the eastern and northeastern boundaries are covered by Upper Cretaceous sediments. The structural setting has been described by Vrána & rámek (1999).
A single multigrain U–Pb zircon age of 345 ± 5 Ma is available for the rocks of the LGM (from a quartz dioritic two-pyroxene granulite at Rudolfov; van Breemen et al., 1982). This age corresponds well to the best age estimate for the granulite-facies metamorphism in the Gföhl Assemblage [see Kröner et al. (2000) and O'Brien & Rötzler (2003) for reviews].
|
ANALYTICAL TECHNIQUES |
|---|
Cathodoluminescence
The petrographic studies have benefited from optical CL observations (Marshall, 1988
; Pagel et al., 2000) using a CITL Technosyn 8200 Mk 4 apparatus at the University of Glasgow. Typical operating conditions were 26 kV, with a gun current of 210 µA, and a vacuum of 1·2 mbar. Images were recorded on a Nikon DN100 digital camera. Selected zircon grains were mounted in epoxy and polished prior to imaging and ICP-MS analysis. The CL investigations of the zircons were carried out on a LEICA Stereoscan 430 scanning electron microscope, equipped with an OXFORD MiniCL detector, at the Division of General Geology and Geodynamics, University of Salzburg (operating conditions: 15 kV, 500 pA probe current).
Electron microprobe
Preliminary mineral analyses were carried out using an ARL-SEMQ microprobe and CamScan 4-90DV microprobe equipped with Link-ISIS Oxford Instruments EDX, both at the Czech Geological Survey (CGS).
Final analyses included in this study were obtained with a Cameca SX-100 electron microprobe at the Joint Laboratory of Electron Microscopy and Microanalysis, Masaryk University, Brno (analysts P. Sulovsk and R. opjaková). The accelerating voltage was 15 kV, the probe current 10 and 20 nA, beam diameter 1 µm (5 and 15 µm for feldspars), and counting times were 10–20 s for major elements and 30–60 s for minor elements. The following standards were used: sanidine for Si, Al and K; albite and jadeite for Na; almandine for Fe; olivine for Mg; andradite for Ca; rhodonite for Mn; hornblende for Ti; chromite for Cr; topaz for F; apatite for P; barite for Ba; a modified set of standards was used for spinels and ilmenite. Data were reduced using the X–
routine.
Whole-rock geochemistry
Most of the major-element whole-rock analyses were performed by wet chemistry in the CGS. The relative 2
uncertainties for the given concentrations were better than 1% (SiO2), 2% (FeO), 5% (Al2O3, K2O, Na2O), 7% (TiO2, MnO, CaO), 10% (Fe2O3) and 15% (but mainly <6%) MgO. A few of the major-element analyses and most trace elements were determined by X-ray fluorescence (XRF; Bruker AXS) at the Division of Mineralogy and Material Science, University of Salzburg, on lithium tetraborate glass beads and pressed pellets, respectively. The tube conditions were 4 kW and 60 kV; counting time was optimized automatically up to 400 s to obtain a detection limit of c. 1 ppm. Typical errors from the counting statistics were ±1 ppm at low concentrations (<10 ppm) and better than ±5% (relative) for the rest.
The rare earth elements (REE) were obtained mainly by ICP using an OES Perkin-Elmer Plasma II in the CGS. Some Co and Zn data were determined by atomic absorption spectroscopy (AAS) in the same laboratory.
Several samples were analysed in the Acme Analytical Laboratories, Vancouver, Canada. The majority of the trace elements were determined by LiBO2 fusion and ICP-MS/OES, except for precious and base metals analysed by aqua regia digestion followed by ICP-MS.
Data management, recalculation, plotting and statistical evaluation of the geochemical data were facilitated using the R software package GCDkit (Janouek et al., 2003).
Laser-ablation ICP-MS dating of zircon
The U–Pb isotopic analyses were performed using a Thermo-Finnigan Element II sector field ICP-MS system coupled to a Merkantek/New Wave 213 nm UV laser system at Frankfurt University. The masses of 202Hg, 204Hg + Pb, 206Pb, 207Pb, 208Pb, 232Th, 235U and 238U were measured in peak jumping mode during 30–90 s sample ablation. A spot size of 25–30 µm was used as the best compromise between spatial resolution and acceptable internal precision of the 207Pb/206Pb and 206Pb/238U ratios. At this spot size a typical mean 238U signal of 5 to 9 x 105 counts per seconds (c.p.s.) on the GJ-1 reference zircon (240 ppm U) was achieved using 44% laser power and 10 Hz repetition rate. During the 90 s ablation the laser drilled 40–50 µm deep and the signal dropped to about 1·5 to 2·5 x 105 c.p.s. Data were collected at four points per peak by integrating 35 mass scans in 2·5 s. A teardrop-shaped, low-volume laser cell (NIGL, Nottingham) with a washout time below 1 s (Horstwood et al., 2003) allowed precise time-resolved data acquisition; this was important when the laser drilled more than one growth zone during analysis.
Raw data were processed offline in an Excel spreadsheet. The inter-element fractionation was corrected by linear regression, whereby the intercept gives the ‘true’ ratio at the initial stage of ablation. The typical internal precision of 206Pb/238U was 0·6% (1 standard error, SE). The ratio obtained was normalized to the reference zircon GJ-1 by multiplying by the difference between the inter-element fractionation corrected ratio of GJ-1 and its reference value (obtained by isotope dilution thermal ionization mass spectrometry; ID-TIMS). Frequent analyses of zircon with known (GJ-1, 91500, Temora, SL13 and AS3) and unknown U–Pb ages in the Frankfurt laboratory, using 91500 or GJ-1 zircon as references, show that an external precision and an accuracy of about 1% can be achieved routinely for crack-free grains with a homogeneous U–Pb composition (Gerdes, 2005). All errors (1 confidence level on ratios, 2 on ages) are propagated with the reproducibility (standard deviation, SD) of the standard over the session or day taken into account. For the U/Pb and Th/Pb ratios this is usually twice the internal precision of the single analyses. For the calculations of concordia ages and weighted averages, as well as plotting of concordia diagrams, we used Isoplot/Ex 2.49 (Ludwig, 2001).
The typical background, measured for 30 s prior to each analysis, was below 50 and 5 c.p.s. for Pb and U isotopes, respectively. During the two analytical sessions the mean 204Hg gas blanks, calculated from the measured 202Hg signal, were below 300 and 200 c.p.s., respectively. A common lead correction was applied based on the interference-corrected 204Pb and the two-stage model of Stacey & Kramers (1975). However, the 204Pb signal was commonly low or below the limit of detection. Nevertheless, in cases of relatively low radiogenic Pb signals the applied correction caused a scatter in 207Pb/206Pb ratios and significantly increased the 207Pb/206Pb errors. For some analyses there was a noticeable tendency to overcorrect the common-Pb contribution (see, e.g. Table 6). The interpretations in this paper are mainly based on the 206Pb/238U age as this is the more precise and accurate, because of better counting statistics and general robustness against a common-Pb correction. The 206Pb/238U analyses usually gave the most stable time-resolved isotope ratio (1 SE = 0·5–1·3%), indicating ablation of only homogeneous material. The 207Pb/206Pb ages suffer from a low 207Pb signal resulting from the small laser spot size used and the uncertainty arising from the detection of common Pb. Within-run precision of the common-Pb corrected ratio (1 SE = 1·2–7·7%) is generally worse than the external reproducibility of the GJ-1 standard, as precision and reproducibility depend mainly on counting statistics. Because of the low 207Pb abundance and the dependence of the 207Pb/206Pb error on the slope of the concordia, the 207Pb/206Pb ages of Palaeozoic grains are generally less accurate than U–Pb ages. The signal intensity of 208Pb in the sample grains was similar to that of the 207Pb. Thus the 208Pb/232Th ratio is very sensitive to common Pb corrections and to errors derived from it. With 208Pb signals of less than 1000 c.p.s. the 208Pb/232Th has the poorest reproducibility (c. 6·2%, 2) of all isotope ratios determined on the GJ-1 standard.
Radiogenic isotopes
For the isotopic study, samples were dissolved using a HF–HCl–HNO3 mixture. Strontium and bulk REE were isolated by cation-exchange chromatography on quartz columns with BioRad AG-W X8 resin; Nd was further separated on quartz columns with Biobeads S-X8 coated with HDEHP (Richard et al., 1976). Isotopic analyses were performed using a Finnigan MAT 262 TIMS system in static mode with a double Re filament assembly (at CGS). The 143Nd/144Nd ratios were corrected for mass fractionation to 146Nd/144Nd = 0·7219, and 87Sr/86Sr ratios assuming 86Sr/88Sr = 0·1194. External reproducibility is given by the results of repeat analyses of the La Jolla [143Nd/144Nd = 0·511858 ± 20 (2) n = 74] and NBS 987 (87Sr/86Sr = 0·710262 ± 86, n = 136) isotopic standards.
The Rb and Sr concentrations, and Rb/Sr ratios, were determined by an approach similar to that of Harvey & Atkin (1981, and references therein) using the XRF devices in the CGS and at the University of Salzburg. The Sm and Nd determinations were carried out on ICP-OES Perkin-Elmer Plasma II system (at CGS) or by isotope dilution on a Thermo-Finnigan Neptune multi-collector ICP-MS system (at the University of Frankfurt).
The decay constants are from Steiger & Jäger (1977: Sr) and Lugmair & Marti (1978: Nd). The 
values and single-stage CHUR Nd model ages were obtained using the Bulk Earth parameters of Jacobsen & Wasserburg (1980); the two-stage Depleted Mantle Nd model ages (
) were calculated after Liew & Hofmann (1988).
|
PETROLOGY AND MINERAL CHEMISTRY |
|---|
 TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGANALYTICAL TECHNIQUES PETROLOGY AND MINERAL CHEMISTRY MINERAL CHEMISTRYTHERMOBAROMETRYWHOLE-ROCK GEOCHEMISTRYDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The main rock types and their field relations
Throughout this paper the Li
ov granulites are classified into two main suites: (1) LF (‘felsic’): garnet–orthopyroxene charnockites (LF ch) and leucocratic garnet granulites (LF gr); (2) LM (‘mafic’): two-pyroxene quartz dioritic granulites (LM qtzD) enclosing pyroxenite xenoliths and cut by picritic dykes, and two-pyroxene tonalitic (LM to) granulites.
As noted, the LGM includes a significantly higher proportion of intermediate and mafic rocks than other granulite massifs in the Moldanubian Zone; felsic garnet (± biotite) granulites (LF gr) nevertheless predominate. In Liov, locally preserved, unfoliated and pristine garnet-bearing hypersolvus leucogranulites are considered to be the protolith to widespread foliated and mylonitic felsic granulites that are characterized by the disintegration of mesoperthite to fine-grained mosaics of K-feldspar, sodic plagioclase and platy quartz (Vrána & Jake, 1982). In the northern part of the LGM, felsic garnet–biotite granulite gneisses contain what may be relics of a migmatitic texture.
Two-pyroxene, garnet-free quartz dioritic granulites (LM qtzD) form two larger (c. 4 km2) bodies near Zvíkov (Fig. 2). These contain small enclaves of pyroxenites, <0·5 m across, with either high-Ca clinopyroxene or enstatite predominating. Fine-grained basic dykes several centimetres wide cut the foliation of the quartz dioritic granulite (Fig. 3a) and also bear the imprint of granulite-facies metamorphism.
|
There are no units of clinopyroxene (± biotite) tonalitic granulites (LF to) large enough to be shown in Fig. 2, but sporadic outcrops occur at several places in the LGM (e.g. at Vlkovice and in borehole LV-2, south of Li
ov; Kotková, 1998b; Fig. 2). In the quarry near Rudolfov, alternating layers 0·1–2 m thick range from quartz diorite through tonalite to minor charnockite. The original relationships of the juxtaposed rock types have been obscured by tectonic transposition, which has obliterated many of the primary contacts. Garnet–orthopyroxene charnockites, including fine-grained, relatively massive and foliated types, are known from four localities (LF ch). They form tabular and podiform bodies, up to 300 m long, and also c. 0·2 m thick layers in banded sequences of quartz dioritic and tonalitic granulites (at Rudolfov).
‘Mafic’ suite (LM)
Two-pyroxene quartz dioritic granulites (LM qtzD). Two-pyroxene granulites (± biotite, hornblende) are fine-grained, foliated rocks with orthopyroxene + clinopyroxene + biotite + plagioclase + quartz (± K-feldspar, hornblende) and accessory minerals, mostly ilmenite, apatite and zircon (Table 1). Pyroxenes with plagioclase (± biotite, K-feldspar and quartz) form an equilibrated granular mosaic (Figs 3b and c, and 4a and b). Exceptionally, sample Li-26 contains scattered rare garnet porphyroblasts 10 mm across.
|
A dark green–grey massive dyke intruding the granulites LM qtzD (Li-25) has a granular structure with a grain size of 0·1–0·6 mm. The rock is dominated by brown pargasitic amphibole. Minor orthopyroxene occurs scattered in an equilibrium amphibole–plagioclase mosaic. Another dyke, sample BR-462 shown in Fig. 3a, has a mineral assemblage and mineral composition closely comparable with the surrounding granulite LM qtzD.
Clinopyroxene tonalitic granulites (LM to). Foliated fine-grained pyroxene–biotite granulites are characterized by a mineral assemblage including clinopyroxene + biotite + plagioclase + quartz (± K-feldspar, hornblende). In sample Li-3, CL studies revealed large bright greenish-yellow (more calcic) subhedral centres of plagioclase augen, surrounded by dull violet-brown plagioclase rims resting in a recrystallized plagioclase–K-feldspar matrix (Fig. 4c and d). Accessory minerals are ilmenite, apatite and zircon. A few of the mafic granulites LM to are mylonitic and are retrogressed to foliated fine-grained hornblende–biotite gneiss (Li-7). In these rocks, the hornblende is newly formed after pyroxene, whereas biotite and quartz are recrystallized; some plagioclase grains remain from the granulite stage.
|
‘Felsic’ suite (LF)
Garnet–orthopyroxene charnockites (LF ch). The fine-grained garnet–orthopyroxene charnockites consist of quartz, plagioclase, mesoperthite, orthopyroxene and garnet. Accessory minerals are apatite, zircon and ilmenite. The granulites LF ch appear massive in hand specimen, but under the microscope they reveal a foliation defined by elongated quartz grains, with aggregates of orthopyroxene and feldspars (Figs 3e, and 4e and f). The granular mosaic texture indicates an equilibrated assemblage (Fig. 4f).
Feldspars are typically perthitic K-feldspar, with mesoperthite in some samples, but calcic oligoclase is also relatively abundant. Quartz typically forms nearly equant grains, although they are sometimes elongated in the foliation planes (e.g. Fig. 4e and f).
Leucocratic garnet granulites (LF gr). These fine-grained massive rocks were described as ‘hypersolvus garnet leucogranites’ by Vrána & Jake (1982) (e.g. samples Li-6, LV-11, 12, 17). Information on their early history is derived from relict domains, accounting for less than 5 vol. % of the outcrops, that are weakly foliated and preserve their primary high-temperature mineralogy (Figs 3f and g, and 4g and h). However, typically, the rocks are largely mylonitized, recrystallized and foliated with platy quartz and fine-grained feldspar mosaic (Fig. 3h). Accessory minerals are magnetite, apatite, zircon and fine-grained acicular rutile.
Garnet-bearing two-pyroxene granulites of gabbroic composition (LGb)
A modally layered sequence of gabbroic garnet–two-pyroxene (± hornblende, biotite) granulites is confined to a small area 0·5 km SW of Liov (borehole LV-1: Kotková, 1998b) (Fig. 2). They are fine-grained, foliated rocks with compositional banding on a millimetre to decimetre scale. Dark bands are enriched in ferromagnesian minerals (orthopyroxene, hornblende, clinopyroxene, garnet and biotite), whereas light bands contain dominantly plagioclase.
According to Kotková (1998b), garnet porphyroblasts up to 5 mm diameter are surrounded by, or completely transformed to, radiating coronas consisting of symplectites of plagioclase + orthopyroxene ± hornblende (± spinel). Early high-Ca clinopyroxene, which originally coexisted with garnet, is sometimes preserved in these relics. Within a short distance from the symplectite, the texture changes to become an equilibrated granular mosaic. Planar fabrics commonly reflect mylonitic banding, overprinted by high-T recrystallization. This rock type, with an early high-pressure imprint is, like the spinel peridotite bodies, considered to be a foreign component in the massif.
|
MINERAL CHEMISTRY |
|---|
Selected mineral analyses for individual minerals are given in Table 2. The complete dataset can be downloaded from the Journal of Petrology website at http://www.petrology.oxfordjournals.org.
|
Orthopyroxene
Orthopyroxene in the quartz dioritic granulites forms an equilibrium mosaic with plagioclase, high-Ca clinopyroxene and biotite. The composition is Fs46–50Wo1·4–1·6 (Table 2). High-magnification BSE images demonstrate that it is free of lamellar unmixed phases. The mineral is pleochroic (X, light red-brown; Y, beige; Z, light grey-green). In a few samples it is altered to a low-temperature assemblage of actinolite + anthophyllite/cummingtonite (Leake et al., 1997
) and chlorite. The composition of orthopyroxene in the pyroxenites is Fs24Wo1. Orthopyroxene in charnockites is ferrosilite (Fs65–71Wo0·7–1·1; Morimoto, 1988) but the more magnesian sample LV-24 contains Mg-rich ferrosilite Fs54Wo0·3. The mineral is strongly pleochroic (X, red brown; Y, beige; Z, bluish green).
Clinopyroxene
In the quartz dioritic and tonalitic granulites clinopyroxene occurs as equant grains in equilibrated mosaics of other major minerals. It is light grey green and Ca-rich (Fs14–20Wo44–45), whereas in pyroxenites its composition is Fs8Wo47–48. A few samples contain relict grains up to 1·5 mm across of primary magmatic pyroxene, with inclusions of plagioclase, magnetite and ilmenite, together with minute (<1·5 µm) lamellae of unmixed orthopyroxene, forming <3–5 vol. % of the host grain (Fig. 3d). Thin orthopyroxene lamellae also occur in some relatively large grains of recrystallized clinopyroxene in the granulites LM qtzD. In the pyroxenites, the unmixed lamellae are considerably thicker.
Amphibole
Amphiboles occur as an accessory or minor phase in the primary granulite assemblages. In the quartz dioritic samples they are Ti-rich pargasites (2 wt % TiO2) with a greenish-brown pleochroism. Both fluorine and chlorine are below detection limits. The amphibole in the pyroxenite (Li-22) is a bright light brown magnesiohornblende. Secondary, green amphibole in tonalitic sample Li-7 falls at the boundary between magnesiohornblende and ferrohornblende, with 0·3–0·4 wt % TiO2.
Garnet
Garnet forms anhedral grains <1 mm across but scattered euhedral crystals are also present, particularly in mesoperthite. Garnet Alm75–78Prp10–13Grs6–8 is confined to garnet–orthopyroxene charnockites; the relatively magnesium-rich sample LV-24 contains garnet Alm71–72Prp22–23Grs3. Garnets in leucocratic garnet granulites have the composition Alm68–77Prp18–26Grs1–3. Although there is some compositional variation among samples with rather similar whole-rock compositions, individual garnet crystals do not vary by more than 2 mol % of end-members. An insignificant decline in Ca towards crystal rims is compensated by minor increases in Fe and Mg. The spessartine content is uniformly low, 2·6–3·5 mol %. It is notable that crystals lack reaction rims, and plagioclase moats in particular, probably indicating a negligible loss of calcium during the evolution of these rocks.
The yttrium content is high (0·1–0·2 wt % Y2O3) in leucocratic garnet granulites and higher still (0·2–0·7 wt % Y2O3) in charnockite garnet. It seems that in the charnockites the garnet Y concentrations are inversely proportional to the modal proportion of garnet in any given sample, reaching the highest value in LV-24 with <1 vol. % garnet. By contrast, the garnets from leucocratic Moldanubian granulites in western Moravia contain three- to five-fold lower yttrium abundances (opjaková et al., 2005).
Exceptionally, garnet also occurs in quartz dioritic granulite (Li-26). In this sample, characterized by the presence of magnetite, the garnet contains 3·5–7·4 mol % andradite that may reflect stabilization by a locally increased oxygen fugacity.
Biotite
The biotite in quartz dioritic granulites is rich in TiO2 (5·3–6·1 wt %), with an Mg/Fe ratio near unity. It shows equilibrium relationships to pyroxenes in the metamorphic assemblage. The dark mica in the pyroxenite is phlogopite with 4·4 wt % TiO2 and 0·4 wt % F.
Feldspars
The compositions of plagioclases in the main granulite groups are shown in Table 2. Subhedral tabular plagioclase crystals with more calcic cores occur in a few quartz dioritic and tonalitic granulites (e.g. BR-365: core An48–51, rim An45; Li-3: core An47–51, rim An35–37). On textural grounds (Fig. 4c and d), and with regard to comparison with the recrystallized plagioclase in the matrix, the tabular crystals are interpreted as relics from the plutonic protolith. Potassic feldspars and sodic plagioclases in leucocratic garnet granulites occur predominantly as mesoperthite and antiperthite intergrowths, owing to unmixing of the original ternary feldspars. The pristine leucogranulite samples containing nearly equant quartz grains were little affected by mylonitization and the related disintegration of mesoperthite to fine-grained mosaics of secondary feldspars. They preserve a simple assemblage of quartz + unmixed ternary feldspars (Fig. 3g) (mesoperthite Ab/Or 1 and An4–6, locally also antiperthite) with a small amount of primary plagioclase (<5 vol. %). Estimates of the temperature required to form the original ternary compositions are typically above 900°C (Fuhrman & Lindsley, 1988; O'Brien & Rötzler, 2003). However, a more specialized study is required to rigorously apply feldspar thermometry to both ternary feldspar types.
|
THERMOBAROMETRY |
|---|
Two-pyroxene thermometry (Brey & Köhler, 1990
) on five quartz dioritic granulite samples gave T = 710–720°C and on two pyroxenite samples T = 770–840°C (Table 3). The equilibration conditions of four charnockites have been constrained by garnet–orthopyroxene–plagioclase–quartz geothermobarometry (Lal, 1993; implemented by Reche & Martinez, 1996) to P = 3·5–5·0 kbar and T = 625–745°C (Table 3). Core compositions were used in all calculations. However, the differences in Grs and Prp content between cores and rims of garnet crystals are typically <2 mol %, Fs in orthopyroxene varies by <1 mol %, and An in plagioclase varies by 1–3 mol %. This fact points, in agreement with textural relations of the phases, to a prolonged annealing and equilibration of the mineral assemblages. Samples selected for thermobarometry were devoid of amphibolite-facies retrogressive imprint.
|
The somewhat higher temperatures calculated for pyroxenites can be interpreted as reflecting metamorphic modification of relict magmatic clino- and orthopyroxene. Some large clinopyroxene crystals in pyroxenites and several quartz dioritic granulites (e.g. Li-4, BR-365) contain thin unmixed lamellae of orthopyroxene, accounting for <3–5 vol. % of the host clinopyroxene. Numerical simulation, adding 3 % of orthopyroxene back to clinopyroxene composition, showed that the temperatures obtained by the two-pyroxene thermometry (Table 3) may be underestimated by up to c. 40°C for quartz dioritic granulites (Li-4) or c. 100°C for pyroxenites (Li-22).
The samples contain no evidence for a previous higher grade metamorphic history, such as mineral inclusions indicative of an older HP stage. On the contrary, several samples still preserve large (Fig. 4c and d) subhedral plagioclases (LM qtzD) and poikilitic clinopyroxenes (Fig. 3d), both of which are interpreted as having survived from the original magmatic protolith (see Table 1 and previous section)
|
WHOLE-ROCK GEOCHEMISTRY |
|---|
Major elements
The whole-rock geochemical variations in the LGM have been studied using over 20 new analyses (Tables 4 and 5)supplemented by data from the literature (Jake
, 1967; Slab, 1983; Fiala et al., 1987a; Kotková, 1998b). Major-element based cluster analysis was employed to classify a few samples from the literature for which accurate petrographic descriptions were lacking.
|
The granulites are subalkaline (Fig. 5a), defining a calc-alkaline trend in the AFM plot (Fig. 5b). In the multicationic P–Q diagram of Debon & Le Fort (1983)
(Fig. 5c) many of the felsic Liov granulites (LF gr) fall into adamellite–granite domains, close to the frequency maximum for most other Moldanubian granulite massifs (see Janouek et al., 2004b, fig. 4). In contrast, the charnockitic types (LF ch) correspond to granodiorite–adamellite and the more mafic varieties fall mainly into the fields of gabbro–quartz diorite–tonalite, albeit with small overlaps into adjacent monzonitic domains.
|
The mafic Li
ov granulites are exclusively metaluminous, as shown by the mean values of Shand's index A/CNK [molar Al2O3/(CaO + Na2O + K2O), Fig. 6]: LM qtzD = 0·80 (0·73–0·89) and LM to = 0·89 (0·78–0·98). The granitic granulites LF gr and the charnockites (LF ch) are mostly slightly peraluminous, A/CNK = 0·95–1·20 (mean 1·07) and A/CNK = 0·95–1·10 (1·05), respectively. Most of the Liov granulites can be characterized as medium-K to high-K calc-alkaline rocks on the basis of the SiO2–K2O diagram (Peccerillo & Taylor, 1976) (Fig. 6).
|
There are strong negative correlations between whole-rock SiO2 and TiO2, Al2O3, FeOt, MgO and CaO for most of the petrographic groups (Fig. 6). The graphs with K2O and mg number [mg number = molar Mg/(Mg + Fet)] are characterized by inflections or discontinuities at SiO2
60 and/or 70 wt %. The pyroxenites with relatively low Al2O3, TiO2 and alkalis, coupled with high MgO, CaO and mg number, show trends that are commonly oblique to those defined by the LM qtzD granulites.
Trace elements
The Liov samples resemble other Moldanubian granulites in being practically undepleted in large ion lithophile elements (LILE) (Fiala et al., 1987a). However, binary plots of SiO2 vs trace elements (Fig. 7) display conspicuous differences for the various granulite groups. Two types of trace-element distribution can be discerned: (1) elements whose concentrations generally increase or decrease in the succession LM qtzD–LM to–LF ch–LF gr (Rb, Y), and (2) those forming highly scattered convex-upward trends, with an inflection or discontinuity at SiO2 60–65% (La, Zr). The pyroxenites again have a special position (characterized by low La coupled with high Ni).
|
The compositions of the granulites LM qtzD were normalized to average normal mid-ocean ridge basalt (N-MORB; Sun & McDonough, 1989
) (Fig. 8a). The patterns are all enriched in LILE, starting at >40 x N-MORB (Cs, Rb) and falling to slightly less than 0·8 x N-MORB (heavy REE; HREE). All show superimposed troughs in Nb and peaks in Ba and Pb; P and Zr are depleted in most samples. Interestingly, the pattern of the picritic dyke (Li-25) reveals contents of high field strength elements (HFSE) and REE only slightly higher than those of MORB (Fig. 8b). Elements mobile in hydrous fluids (Cs, Rb, Ba, Th, U, K, Pb and Sr) are, however, strongly enriched. Similar LILE/HFSE enrichments are typical of K-rich basalts from continental margin arc settings (e.g. Saunders et al., 1991; Pearce & Parkinson, 1993; Tatsumi & Eggins, 1995).
|
The quartz dioritic granulites (the shaded fields in Fig. 8a–d) differ from other Moldanubian massifs in having elevated divalent LILE concentrations (Sr, Pb ± Ba), whereas K and HREE tend to be somewhat lower (Fig. 8c). These differences can be accounted for by a higher degree of fractionation in the mafic–intermediate granulites from bodies other than that at Li
ov, which nearly all have fairly high silica contents and mg number <60.
The dyke Li-25 is geochemically rather primitive, as shown by the flat, slightly convex-upward REE pattern (CeN/YbN = 1·2: Table 5) at c. 15–20 x chondrite and lack of any Eu anomaly (Eu/Eu* = 1·04) (Fig. 9a). The pyroxenite Li-22A is somewhat more fractionated, as indicated by a weak LREE/HREE enrichment (CeN/YbN = 3·1) and a distinct negative Eu anomaly (Eu/Eu* = 0·60) (Fig. 9a). The LM qtzD granulites are characterized by moderately light REE (LREE)-enriched patterns (CeN/YbN = 5·8–6·4) with slightly negative Eu anomalies (Eu/Eu* = 0·80–0·86) (Fig. 9a). The tonalitic granulites (LM to) contain the highest 
REE in the dataset (Fig. 9b and c), with moderately LREE-enriched patterns (CeN/YbN = 3·7–7·9) and pronounced negative Eu anomalies (Eu/Eu* = 0·50–0·62). A similar pattern is also seen in sample JK-1 of Kotková (1998b), which has the highest total, and in particular HREE content.
|
|
The charnockite (LF ch) has a relatively flat REE pattern (CeN/YbN = 2·9) with a strong negative Eu anomaly (Eu/Eu* = 0·39) (Fig. 9c). The leucogranulite (LF gr) has the lowest LREE/HREE ratio (CeN/YbN = 2·1) and a deep Eu anomaly (Eu/Eu* <0·16, with the Eu below the detection limit of 0·18 ppm).
U–Pb zircon geochronology
A combined study of internal zoning (CL, BSE) and LA–ICP–MS dating of zircons in quartz dioritic (Li-4, Zvíkov) and tonalitic (Li-3, Vlkovice) granulites aimed to constrain the nature and age of the protolith to the ‘mafic’ suite from Liov. Representative CL images of zircons are presented in Figs 10 and 12, analytical data in Table 6, and concordia diagrams in Fig. 11. The majority of the LA–ICP–MS spot analyses yielded concordant Late Devonian to Early Carboniferous ages.
|
|
|
Quartz diorite Li-4 (LM qtzD)
This sample contains numerous small prismatic zircons, typically slightly rounded (Fig. 10a, b and e). Although many grains preserve an oscillatory zoning, suggesting a magmatic origin (Fig. 10a), others commonly seem to be recrystallized with the primary zones blurred and thickened (Pidgeon, 1992
; Pidgeon et al., 1998; Corfu et al., 2003) (Fig. 10c). The complex texture of some grains is illustrated in Fig. 10e. A rounded, xenocrystic core in the centre is surrounded by a mantle that exhibits an original oscillatory zoning. However, this is progressively blurred by recrystallization and is obliterated, most markedly to the left and upper side of the outer zone of the grain.
At a more advanced stage of the solid-state recrystallization, a convoluted zoning appears, which is usually interpreted as a result of a complex elemental diffusion (Hoskin & Black, 2000; Hoskin & Schaltegger, 2003). The process can result in almost complete obliteration of the magmatic zoning, with only rare relics preserved (Fig. 10c). About 15 % of the grains, mostly prismatic and small (120 µm x 60 µm), show well-preserved oscillatory zoning, but this is usually truncated by a relatively small nearly featureless metamorphic rim (Fig. 10a and b). Figure 10f shows an example with a larger overgrowth around a small residual prismatic core.
In addition to small and generally unbroken prismatic zircons, the mineral concentrates contain numerous larger angular fragments. A BSE study of the thin section reveals that these are crushed interstitial grains (Fig. 10d), whose shape is mostly determined by adjacent existing phases. Similar morphologies with only partly developed crystal faces are typical of zircons in many basic rocks, where they commonly reflect relatively late crystallization (Corfu et al., 2003, and references therein). Delayed nucleation is due to the low original Zr content of the magma, when compared with the typically high Zr solubility in basic melts (Watson & Harrison, 1983). Indeed, the parental magma of the protolith of the granulites LM qtzD may initially have been zircon-undersaturated, as indicated by a positive trend in the silica-poor part of the SiO2–Zr plot (Fig. 7) (Hoskin et al., 2000) and the calculated zircon saturation temperatures (c. 600–800°C; Li-4: 700°C), too low for a magma of comparable SiO2 content.
The interstitial grains show a continuous spectrum of variously modified primary zoning, starting from oscillatory, through more complex (convoluted or fir-tree sector zoning: Fig. 10d) to nearly unzoned grains or fragments thereof.
Thus, based on textural observations we can distinguish at least three domains of different ages in zircons in this sample: (a) rounded xenocrystic cores; (b) oscillatory-zoned domains of probably magmatic origin; (c) recrystallized and blurred zones