Cadomian and Variscan metamorphic events in the Léon domain (Armorican Massif, France): P-T data and EMP monazite dating
Published:January 01, 2007
- PDF LinkChapter PDF
Bernhard Schulz, Erwin Krenn, Fritz Finger, Helene Brätz, Reiner Klemd, 2007. "Cadomian and Variscan metamorphic events in the Léon domain (Armorican Massif, France): P-T data and EMP monazite dating", The Evolution of the Rheic Ocean: From Avalonian-Cadomian Active Margin to Alleghenian-Variscan Collision, Ulf Linnemann, R. Damian Nance, Petr Kraft, Gernold Zulauf
Download citation file:
The Léon domain adjacent to the Cadomian realm in the North Armorican domain appears to be a displaced crustal block, as its metamorphism and rock types bear a resemblance to the South Armorican domain of the internal Variscan belt. The amphibolite-facies Conquet-Penze Micaschist unit overlies the high-grade Lesneven Gneiss unit in the central part of the Léon. Timing and conditions of the metamorphic evolution have been evaluated. At the base of the Lesneven Gneiss unit, a high-pressure eclogite-facies stage (700 °C at >13 kbar) was followed by a high-temperature event (800 °C at 8 kbar), which is characterized by the crystallization of garnet-cor-dierite assemblages in aluminous paragneisses. Maximal temperatures in the upper parts of the Lesneven Gneiss unit were 630 °C at 6 kbar. Zoned garnet in assemblages with staurolite recorded prograde P-T paths from 490–610 °C at 5–8 kbar in the upper and at 6–9 kbar in the lower parts of the Conquet-Penze Micaschist unit. Garnet Y, heavy rare earth elements, and Li are low in high-grade gneisses and display strong zonations in the micaschists. A younger population of monazite with a broad range of Y contents displays Th-U-Pb ages between 340 and 300 Ma. It crystallized subsequent to formation of foliations S1-S2 and Variscan peak metamorphic assemblages. In contrast, an older population of Cadomian monazite at 552–517 Ma is uniformly rich in Y, suggesting an earlier crystallization than garnet, however, at elevated temperatures. The findings do not support a South Armorican provenance of the Léon domain. The Léon units appear as part of a Cadomian crust at the northern margin of the former Armorican microplate. During a Variscan collision, this crust was strongly overprinted by underthrusting toward the southeast or east beneath the Central Armorican domain and by later uplift accompanied by Late Carboniferous dextral shear tectonics. The features are typical of the Variscan Saxo-Thuringian zone, which faced the Rheic Ocean to the north.
The Armorican Massif in western France is assembled out of several crustal domains. To the north, the Neoproterozoic Avalonian-Cadomian orogen was overprinted to a variable degree. To the south, structures and metamorphism of a Paleozoic continental collision are dominant. Within this frame of well-zoned Cadomian and Variscan orogenic belts, the Léon domain to the northwest appears as a strange (“exotic”) unit. Some arguments for displacement of the Léon domain arise from similar rock types, ages, and metamorphic events, as observed in the South Armorican domain, especially the occurrence of eclogites and orthogneisses (Cabanis and Godard, 1987; Le Corre et al., 1989). Tectonic studies revealed dextral shearing in ENE-trending zones along the southern border of the Lèon domain, which were interpreted as major displacement lines (Balé and Brun, 1986). The discussion of whether crustal domains represent displaced and allochthonous terranes is crucial for paleogeographic and plate tectonic models of the Ibero-Armorican segment of the Variscan belt (Franke, 1989; Martinez-Catalan, 1990; Matte, 1991; Dalziel, 1997; Shelley and Bossière, 2000, 2002; Robardet, 2002; Stampfli et al., 2002; Cartier and Faure, 2004). A detailed reconstruction of the magmatic, metamorphic, and structural evolution is essential for this discussion. The present article deals with the P-T evolution, the mineral trace element chemistry, and Th-U-Pb monazite ages from the two major lithotectonic units of the Léon domain. Geothermobarometry on garnet-bearing assemblages in paragneisses and micaschists and on Ca-amphibole in metabasites revealed single prograde-retrograde P-T paths at different temperatures and pressures in the upper and lower parts of the normal crustal pile. The majority of the Th-U-Pb monazite ages confirm that a Barrovian-type metamorphism can be assigned to the Variscan collision. However, an earlier Cadomian thermal event is documented in a distinct population of monazite and provides new details for the zoneography of the Variscan belt in the Armorican Massif.
Regional Geological Setting
Two major west–east-trending late Variscan shear zones separate the South, Central, and North Armorican domains (Fig. 1). Each of the domains is subdivided into distinct Upper Proterozoic to Lower Paleozoic lithotectonic units. The South Armorican domain is part of the internal Variscan belt (Cogné, 1988; Ballèvre et al., 1994). It involves greenschist-, amphibolite-, and blueschist-facies rocks as well as high-grade and eclogitic units, with partly complex P-T evolution during eo-Variscan (463–376 Ma) and Variscan (330–300 Ma) times (Jones and Brown, 1990; Audren and Triboulet, 1993; Ballèvre et al., 1994; Schulz et al., 2001; Lucks et al., 2002). In the Central Armorican domain a Brioverien (Upper Proterozoic) unit with a narrow southern amphibolite-facies zone (Schulz et al., 1998) can be distinguished from an unconformably overlying low-grade to epizonal “classical” Cambrian to Upper Devonian cover sequence (Le Corre et al., 1991; Paris and Robardet, 1994; Rolet, 1994). In the North Armorican domain a unique section across the Cadomian belt is developed (Strachan et al., 1989; Brun and Balé, 1990; Cogné, 1990; Ballèvre et al., 1994; Egal et al., 1996; Brun et al., 2001; Chantraine et al., 2001). The Cadomian Domnonean and Mancellian domains and subunits are separated by northeast-trending major thrusts and shear zones that turn to the northwest in the Trégor province (Fig. 1). Within this framework, the eastern part of the Léon domain is juxtaposed onto the Cadomian realm, whereas its southern part is linked to the Paleozoic of the Central Armorican domain.
Geological Setting in the Léon Domain
In the central part of the Léon domain two main metamorphic sequences, the amphibolite-facies Conquet-Penze Micaschist unit and the high-grade Lesneven Gneiss unit, can be identified (Rolet et al., 1994). At the southern border and in the hangingwall of these units (Fig. 1), the very low-grade phyllitic Proterozoic schists of L'Elorn have been intruded by a granodiorite, the later Gneiss de Brest, which was dated at 466 ± 25 Ma (Deutsch and Chauris, 1965; Michot and Deutsch, 1970; Cabanis et al., 1977; Le Corre et al., 1991). To the east, Silurian to Devonian schists overlie the micaschists of Conquet-Penze (Fig. 1B). The eastern margin of the Léon domain and the transition to the northwest-trending major Cadomian structures in the Trégor region are masked by the Carboniferous basin of Morlaix (Cabanis et al., 1979a). To the northwest, the fault of Porspoder-Guisseny with a sinistral sense of shear separates the Léon metamorphic pile from the migmatic complex of Landunvez-Plouguerneau (Outin et al., 2000). Two series of granites intruded the metamorphic pile: the older granite complex of Saint Renan-Kersaint with ages of 340–330 Ma and the younger granites of Aber-Ildut-Ploudalmézeau-Kernilis with ages of 300–280 Ma (Cogné and Shelley, 1966; Leutwein et al., 1969; Michot and Deutsch, 1970). Metamorphic structures and the southern margin of the Saint Renan-Kersaint granite were mylonitized by dextral shearing along the North Armorican shear zone (Goré and Le Corre, 1987). An offset of ∼15 km along the shear zone has been estimated within the 329 ± 9-Ma Plouaret-Commana granitic complex (Peucat et al., 1984).
The lower part of the central Léon is a high-grade metamorphic sequence and was described as the Lesneven Gneiss unit (Cabanis et al., 1979b; Chantraine et al., 1986). The migmatized ortho-augen-gneisses of Plounevez-Lochrist and Treglonou appear in antiformal structures at its base (Fig. 1C). The orthogneisses provided ages of 400 ± 40 Ma (U-Pb zircon), 392 ± 14 Ma (Pb-Pb zircon; Chauris et al., 1998), and 385 ± 8 Ma (Rb-Sr whole rock), considered as protolith ages by the authors but also interpreted to date a major tectonometamorphic event (Le Corre et al., 1989, 1991). Lenses of eclogites occur within the orthogneisses and overlying paragneisses. The metabasites with a normal mid-oceanic ridge basalt (N-MORB)-type character (Cabanis and Godard, 1987) recorded 650–700 °C at minimum pressures of 13–14 kbar (Paquette et al., 1987; Godard and Mabit, 1998). A 439 ± 13-Ma U-Pb zircon lower intercept age was interpreted to date the high-pressure metamorphism (Paquette et al., 1987). In the region of Plounevez-Lochrist, a distinct horizon of garnet-bearing aluminous paragneisses crops out in the vicinity of the eclogites and related amphibolitized eclogites (Fig. 1B). Mineral assemblages with fibrolitic sillimanite and K-feldspar, and stromatic migmatites prevail in paragneisses of the upper part of the Lesneven Gneiss unit. According to the huge anticlinal structure in the central Léon (Fig. 1C), the amphibolites of Lannilis should be part of the Lesneven Gneiss unit. The upper part of the Lesneven Gneiss unit, the Conquet-Penze Micaschist unit, and the Gneiss de Brest are exposed in an almost continuous coastal cross-section in the region of Le Conquet between the Anse de Porsmoguer and the Pte de St. Mathieu. The granodioritic gneiss of Pointe des Renards (565 ± 40 Ma Rb-Sr WR) occurs in lenses both in the micaschist and the gneiss units (Michot and Deutsch, 1970; Chauris and Hallégouët, 1989). Lenses and layers of amphibolites, partly with garnet, a metagabbro, and a metaporphyroid are intercalated in the Conquet-Penze micaschists (Chauris and Hallégouët, 1989). In the coastal section, the main foliation uniformly strikes ENE. It is dipping steeply in the northern parts and 30–50° to SSE in the southern parts (Fig. 1B). A mineral lineation is dominant in the southern part and plunges 20–30° WSW. The ENE-trending Léon shear zone, where progressive synmetamorphic unroofing of the Léon domain (Jones, 1994) along dextral strike-slip transtensional movement should have been accommodated in Upper Devonian times (Balé and Brun, 1986), is located within the Gneiss de Brest to the south of the micaschist unit. A steeply increasing metamorphic grade toward the north, coinciding with a transition from micaschists to gneisses, was recognized in the coastal section (Jones, 1994). Garnet-bearing assemblages with staurolite prevail in the Conquet-Penze micaschists and display metamorphic conditions of 550–600 °C at 6.3–8.4 kbar (Jones, 1994). In the Kermorvan peninsula, slightly higher temperatures of ∼630 °C were estimated from garnet-bearing gneisses (Jones, 1993, 1994). Although the metamorphic conditions in the Lesneven gneiss and Conquet-Penze micaschist units have already been evaluated by geothermobarometric studies (Paquette et al., 1987; Jones, 1994; Godard and Mabit, 1998), data on metamorphic ages are still sparse. A possible maximum age of the high-pressure metamorphism is provided by the 439 ± 13-Ma U-Pb zircon lower intercept age from the eclogites (Paquette et al., 1987). However, the high-temperature metamorphism could be younger, if the Devonian ages of the orthogneisses are considered as protolith ages. Intrusion of the younger granites at ca. 300 Ma and the development of the North Armorican shear zone give a minimum age limit.
It has been demonstrated in various parts of the Variscan belt and other metamorphic terrains (Finger and Helmy, 1998; Williams et al., 1999; Finger et al., 2002; Dahl et al., 2005) that in situ “chemical” Th-U-Pb dating of monazite by analysis with an electron microprobe (EMP monazite dating; Montel et al., 1994, 1996; Suzuki et al., 1994) provides valuable constraints on the timing of metamorphic events and allows the researcher to resolve and distinguish single thermal events in a polymetamorphic evolution. Metamorphic monazite crystallizes as an accessory phase in metapelites and metagraywackes with a limited range of Ca-poor bulk compositions. The presence of garnet with biotite, muscovite, plagioclase, quartz, and aluminosilicates in these rocks can be used to evaluate the metamorphic conditions and P-T paths by geothermobarometry. We combined the EMP monazite dating method with a detailed geothermobarometric study in garnet-bearing micaschists and gneisses. In addition, trace element analyses of garnet were used to evaluate the relative timing of monazite and garnet crystallization in the samples. Geothermobarometry on garnet-bearing metasediments gives insight into a limited part of the P-T evolution of a metamorphic unit. The study has been completed further by geothermo-barometric data from metabasites. Monazite Th-U-Pb dating and related geothermobarometry is focused on the garnet-bearing aluminous paragneisses to the south of Plounevez-Lochrist in the lower part of the Lesneven Gneiss unit and the coastal section to the west, with the transition of the Conquet-Penze micaschists into the upper parts of the Lesneven gneisses (Fig. 1B).
The main element whole-rock compositions of orthogneisses and of monazite-bearing micaschists and paragneisses were analyzed by X-ray fluorescence spectrometry (XRF). Trace element analyses were performed at Institute of Mineralogy University of Wuerzburg by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) with a single collector quadrupole The laser was adjusted to a scan speed of 5 µm/s at an energy of 0.25 mJ and a repetition rate of 10 Hz. It was traced along 1.6-mm-long and 50-µm-wide extraction lines on LiBO4 fusion glass beads (Brätz and Klemd, 2002). Bulk rock concentrations of SiO2 measured by XRF were used for normalization of LA-ICP-MS analyses. The glass reference material NIST SRM 612 with the values of Pearce et al. (1997) was used for external calibration and calculation of trace elements concentrations by the GLITTER® 3.0 Online Interactive Data Reduction for LA-ICP-MS Program version 2000 by Macquarie Research, Ltd. Mean results from five extraction lines coincide within 1σ error with the data from acid solution ICP-MS (Centre de Recherches Pétrographiques et Géochemiques–Centre National de la Recheche Scientifique, Nancy, France), except for the results for the element La. At measured La concentrations between 16 and 60 ppm, this divergence is not negligible. A constant amount of ∼7 ppm La was introduced with LiBO4 into the glass beads and was empirically corrected by regression through comparison of samples with different La concentrations, as described by Sylvester (2003). From the means calculated from five extraction lines, the error is <5% (based on 1σ standard deviation) for the rare earth elements (REE), Hf, Nb, and Ta, except for Sm, Lu, Hf, and Th (5–13%). Reproducibility, accuracy, and precision of the applied method were controlled by repeated analysis of NIST SRM 614 (data by Horn et al., 1997) and whole-rock geostandards (BE-N Basalt, MAG-1 Marine Mud; Govindaraju, 1994). Based on analysis of NIST SRM 614, the precision for REE, Nb, Ta, and Th is <5% (based on 1σ), except for Gd and Er (5–7.5%).
The mineral chemistry (major elements) has been analyzed by EMP (900 points, CAMECA SX 50, SX 51 at 15 kV, 10 nA, counting time 20 s, with CAMECA routine ZAF and PAP corrections; for representative data see Table 1) in eleven micaschist and garnet gneiss samples and in five amphibole-bearing metabasites. Amphibole was characterized by profiles of five to fifteen points. The cores and rims of chlorites, epidotes, and plagioclases next to the amphibole porphyroblasts were analyzed and further profiles with three to five points were taken. Cations in amphiboles were normalized to 23 oxygens and sum(T1 + T2 + M1 + M2 + M3) = sumMg = 13 (13eCNK), as detailed in Triboulet (1992). Fe3+ has been estimated as maximum.
In micaschists and paragneisses, the composition of zoned garnet, staurolite, cordierite, biotite, muscovite, and plagioclase was determined from cores to rims. The core of mica inclusions in garnet was also analyzed. The chemical evolution of garnets in micaschists was studied in several samples by profiles with a point spacing between 0.05 and 0.2 mm. However, single profiles may not have passed the entire core region; furthermore, porphyroblasts display zonation gaps, and some show only a part of the complete chemical evolution of garnet. These complications are not necessarily evident from the single zonation profiles. Therefore, garnet compositions were investigated using spessartine-grossular-pyrope coordinates. As the Mn component is often controlled by Rayleigh fractionation during crystallization (Hollister, 1966), it allows recognition of the relative temporal chemical growth evolution when garnet is the main Mn fractionating phase.
Trace element analyses of garnet in monazite-bearing samples were carried out by LA-ICP-MS (Longerich et al., 1993) in those thin sections that were used for the EMP analyses. Garnet profiles included five to ten single-shot craters with diameters of 35 µm. The laser was adjusted to a scan speed of 5 µm/s, an energy of 0.23–0.28 mJ (40 J/cm2), and a repetition rate of 5 Hz. After 25 s of background detection, an analysis of 50 s with time-resolved signals and integration times of 10 ms for 29Si and 49Ti; 20 ms for 45Sc, 51V, 53Cr, 89Y; and 30 ms for 7Li, 60Ni, 69Ga, and REE with masses 139–175 was compared to the glass reference material standards NIST SRM 612 and SRM 614 (Jackson et al., 1992; Pearce et al., 1997). SiO2 of garnet rims known from EMP analysis was used as the internal standard; an in-house standard garnet K23 allowed verification of the data. After signal quantification by the GLITTER 3.0, the 1σ errors, based on counting statistics from signal and background, range ∼29% (Gd, 3.2 ± 1.1 ppm) and 14% (Yb, 123 ± 18 ppm), depending on the absolute concentrations. Averaged errors are given in Table 1. Ti in the zoning profiles has been considered, as no high Ti resulting from unintentional analysis of tiny Ti-phase inclusions in garnet occurred.
The in situ “chemical” Th-U-Pb dating of monazite by EMP analysis (Montel et al., 1994, 1996; Suzuki et al., 1994) is based on the observation that concentration of common lead in monazite (light REE [LREE], Th)PO4 is negligible when compared to radiogenic lead resulting from decay of Th and U (Cocherie et al., 1998). As Th concentrations are mostly high (3–14 wt%), a sufficient amount of radiogenic lead to be analyzed by EMP can accumulate in Paleozoic monazite. All monazite analyses of this study were carried out on a JEOL JX 8600 at Salzburg University, using conditions of 15 kV, 250 nA, and a beam diameter of ∼5 µm (Finger and Helmy, 1998). Mα1 lines were chosen for Th, U, and Pb; Lα1 for La, Y, and Ce; Lβ1 for Pr and Nd; and Kα1 for P, Si, and Ca. The counting times for Pb, Th, and U were 240–360, 30, and 50 s, respectively, per analysis on peak and 2 × 120, 2 × 15, and 2 × 25 s, respectively, on background. This procedure results in statistical errors (1σ) of typically 0.012, 0.05, and 0.015 wt% for Pb, Th, and U, respectively (Finger and Helmy, 1998). All other elements were determined with 10 s (2 × 5 s) counting times. The small Y interference on the Pb Mα line was corrected by linear extrapolation after measuring a Pb-free yttrium standard (Montel et al., 1996). A small Th interference on U Mα was also empirically corrected. For each single analysis, a chemical age was calculated using the equations of Montel et al. (1996) plus a respective error resulting from counting statistics, which was mostly between ±20 and ±40 m.y. (1σ). Weighted average ages were calculated after Ludwig (2001). To control the quality of analyses, monazite with a concordant U-Pb age of 341 ± 2 Ma analyzed by thermal ionization mass spectrometry (TIMS) (Friedl, 1997) has been measured together with the specimen. Weighted average model ages of 346 ± 13, 341 ± 9, 340 ± 8, and 344 ± 11 Ma were obtained from the standard during four different sessions and are in agreement with the U-Pb TIMS age.
Bulk Rock Compositions of Metasediments
For evaluation of bulk rock compositions, monazite-bearing metasediments can be subdivided into micaschists with garnet and staurolite, paragneisses without garnet, and aluminous para gneisses with garnet. Metagranitoids, such as the Gneiss de Brest, the Pte des Renards granite gneiss, and the ortho- augengneiss of Lochrist, were analyzed for comparison. In the La-Th-Sc and Th-Sc-Zr/10 discrimination diagrams for the provenance character and geotetonic setting (Bhatia and Crook, 1986), all former clastic sediments (micaschists and paragneisses) as well as the Léon meta-granitoids uniformly plot in the field of a continental magmatic arc (Fig. 2A). In the Ni-TiO2 and SiO2-K2O/Na2O discrimination diagrams of Floyd et al. (1989) and Roser and Korsch (1986), the provenance characteristics are confirmed by indication of a source region dominated by felsic rocks and by the placement of the sequence in an active continental margin setting (Fig. 2B). It should be noted that the analyzed rocks have a Ca-poor composition, which is typical for monazite-bearing metasediments. Nethertheless, the data are regarded as being representative for the recognition of the sedimentary source region, because the lithological units are rather monotonous. Monazite-bearing rock types appear to be abundant within the suite. Similar geo chemical characteristics and geotectonic setting have been described from other Neoproterozoic metasedimentary sequences in the Saxo-Thuringian zone (Linnemann and Romer, 2002).
Mineral Chemistry and Geothermobarometry of Metabasites
Lenses of eclogites and their retrogressed equivalents occur around the contact of paragneisses and the ortho-augengneiss of Plounevez-Lochrist in the lower parts of the Lesneven Gneiss unit. In the samples from location Kao (in the village of Plounevez-Lochrist), with the best preserved eclogites, garnet of a relict eclogitic peak metamorphic assemblage displays a pro-grade zonation, with increasing pyrope contents (20–45 wt%) and decreasing grossular (30–18 wt%), almandine (47–35 wt%), and spessartine (3.3–0.5 wt%) contents from core to rim. As was similarly observed by Paquette et al. (1987) and Godard and Mabit (1998), no clinopyroxene was preserved in the matrix, and symplectites with plagioclase and amphibole prevail. The garnet core composition corresponds to the observations by Paquette et al. (1987), who calculated 700 °C at 13–14 kbar using garnet and enclosed omphacite with Jd 27% by garnet-clinopyroxene equilibria (Ellis and Green, 1979; Krogh, 1988) and related geo-barometers (Holland, 1983). This P-T estimate represents the minimum temperatures as well as corresponding minimum pressures that occurred during the crystallization of garnet cores (see Fig. 5A). Although the increase of Mg in garnet can be attributed to growth during increasing temperature, decrease of Ca toward the garnet rim signalizes decreasing pressure during crystallization of the garnet-bearing assemblage. Godard and Mabit (1998) found further petrological evidence for such a temperature-dominated evolution subsequent to the peak pressure conditions.
Ca-amphibole, which replaced garnet and clinopyroxene in overprinted eclogites, is also abundant in the numerous meter-scale metabasite lenses and in the large Lannilis metabasite complex (Fig. 1B). Ca-amphibole is best described by its IVAl-VIAl and Ti-IVAl composition, but the other chemical parameters, Si (decreases with increase of Altot), Mg, NaM4, and (Na + K)A show the corresponding variations (Triboulet and Audren, 1988). P-T conditions of the amphibole-bearing assemblages were estimated by a least squares approximation of amphibole compositional dependence on temperature and pressure for the assemblage Ca-amphibole + plagioclase (An > 10) + quartz ± zoisite ± chlorite ± calcite, determined by Plyusnina (1982), as used by Gerya et al. (1997) and empirically modified for Fe3+ in amphibole by Zenk and Schulz (2004):
In the retrogressive post-peak assemblage of the kyanite-bearing eclogite (sample Kao), unzoned magnesio-hornblende with IVAl 1.0/VIAl 0.3 shows no preferential orientations. It coexists with andesine, epidote, quartz, and sphene and crystallized at 530 °C and 4 kbar (Fig. 5A).
Preferentially oriented tschermakite in an amphibolite from the Lannilis metabasite complex (sample Lan) in the upper part of the Lesneven Gneiss unit crystallized with andesine, epidote, quartz, and sphene and is unzoned, with IVAl 1.7/VIAl 0.4 and elevated Ti of 0.13 compared to the other samples. These amphiboles indicate metamorphic conditions at 620 °C and 6 kbar, when the thermobarometer reported above is used. An amphibolite with a similar mineral assemblage (sample Kerhorn) from the western part of the Lesneven Gneiss unit displays zonations from magnesio-hornblende (IVAl 1.2/VIAl 0.5) in the cores to actinolite rims (IVAl 0.7/VIAl 0.2). The continuous zonation corresponds to a retrograde P-T path from 540 °C and 5 kbar to 400 °C and 2 kbar (Fig. 5A).
Amphibolite samples from the Conquet-Penze Micaschist unit come from Pors Liogan (sample PLi3 in Fig. 5) and the harbor of Le Conquet (PortCo). Long axes of preferentially oriented green amphiboles define a southwest-plunging mineral lineation, associated with a weak foliation that is parallel to the structural elements in surrounding micaschists. The unzoned tschermakitic hornblendes and tschermakites (Leake et al., 1997) with IVAl 1.8/VIAl 1.0 in sample PLi3 and tschermakitic amphiboles with IVAl 1.5/VIAl 0.45 in sample PortCo coexist with andesine, epidote, quartz, and spene. Garnet occurs in sample PLi3. The amphiboles recorded 610 °C at 8 kbar in sample PLi3 and 590 °C at 6 kbar in sample PortCo and are interpreted to represent the conditions of the thermal maximum in this unit (Fig. 5B).
Mineral Chemistry and Geothermobarometry of Garnet Micaschists and Paragneisses
Lesneven Gneiss Unit
Garnet is rare in the paragneisses of the Lesneven Gneiss unit. The distinct 1.5-km-long northeast-striking horizon with aluminous paragneisses in the vicinity of the eclogite lenses to the south of Plounevez-Lochrist (Chauris et al., 1998) seems to be the only garnet-bearing paragneiss location in the lower part of this unit. In samples from La Garenne (Lage), Traonjulien (Trao), and Keranton (Kerz), a foliated matrix by biotite (muscovite only rarely appears as a late phase), plagioclase, K-feldspar, quartz, and both kyanite and fibrolitic sillimanite surround garnet (up to 4 mm in diameter). Cordierite is abundant in the matrix of samples Trao (Fig. 3A) and Kerz. As kyanite is enclosed, cordierite is interpreted to have crystallized after kyanite. Andalusite appears as a late phase in sample Trao. Cordierite in sample Kerz has higher Na2O (0.55–0.70 wt%) than in sample Trao (0.40–0.50 wt%), whereas the XMg of cordierite is similar in both samples (XMg 0.73–0.78), with slightly higher XMg in the cores. The poikiloblastic garnet encloses biotite, plagioclase, quartz, and kyanite. Chemical analysis along profiles (Fig. 4) indicates that the porphyroblasts have large homogeneous cores with 25% pyrope, 8% grossular, and 5% spessartine contents. In the narrow outermost rim, Mg significantly decreases while Fe increases at almost constant Ca and Mn (Fig. 4A, H, and I), indicating a temperature decrease. The XMg of biotite enclosed in garnet and in the matrix are similar and depend on the XMg in garnet cores. Plagioclase is andesine and zoned with decreasing anorthite contents toward the rim. The P-T conditions for crystallization of garnet cores have been calculated using the garnet-biotite thermometer of Bhattacharya et al. (1992) in combination with the garnet-sillimanite-plagioclase (GASP) barometer of Holland and Powell (1990), involving an internally consistent thermo dynamic data set with the ideal activity models for garnet (Ganguly and Saxena, 1984; Ganguly et al., 1996) and plagioclase (Powell and Holland, 1993). Furthermore, the garnet-cordierite thermobarometer using the thermodynamic data set by Perchuk (1991) was also applied. Tentative calculations by other calibrations of the thermo- and barometers yielded no substantially different results. Maximum temperatures calculated from the Mg-rich garnet cores and enclosed biotite or matrix biotite with a similar XMg range between 740 and 780 °C. Corresponding pressures of 8–10 kbar were calculated by plagioclase enclosed in garnet or from the matrix (Fig. 5A). Temperatures and pressures from the garnet-cordierite equilibria are ∼800 °C at 7 kbar. This difference is in the range of the general error of the thermobarometric estimates and can be attributed to the different thermodynamic data sets involved. According to the P-T data, the lack of muscovite in the assemblage with garnet can be explained by its prograde decomposition to K-feldspar and sillimanite. Godard and Mabit (1998) attributed the appearance of sapphirine after kyanite in quartz-free eclogites nearby to a substantial increase of temperature during the decompression. This increase is further confirmed by the thermobarometric data from the aluminous paragneisses.
In the upper part of the Lesneven Gneiss unit, which is exposed along the coast to the west (Fig. 1B), one sample of garnet gneiss has been gained from a 5-m-thick horizon at Porz Pabu (sample Pabu) in the Kermorvan peninsula, corresponding to the location of sample LC12 in Jones (1994). Layers with relict garnet in a foliated matrix with biotite, plagioclase, quartz, staurolite, and kyanite are embedded in domains with biotite, muscovite, plagioclase, quartz, and fibrolitic sillimanite, the latter occuring with biotite in the S2 foliation. Apparently garnet has been replaced by aggregates of andalusite. Observations and interpretations of phase relations are given in Jones (1993). The garnet porphyroblasts are poorly zoned, with homogeneous cores of pyrope 13.3% and grossular 2.5% (Fig. 4B and G). Increase of spessartine at decreasing pyrope in the outer rim is attributed to retrograde exchange reaction. Temperatures of 635 °C were calculated from Mg-rich garnet zones and enclosed biotite or matrix biotite. They are identical to the results of Jones (1994) and represent the thermal peak. Jones (1994) argued that GASP barometry with low-An plagioclase (An 9) may overestimate pressure, because of poorly understood activity-composition relationships (Ashworth and Evirgen, 1985). However, in regional studies involving comparison of samples with oligoclase and low-An plagioclase, Schulz et al. (1998) showed that the garnet-muscovite-biotite-plagioclase (GMBP) barometers can be alternatively applied to assemblages with low-An plagioclase. Accordingly, pressures of 6.4 kbar at 630 °C (6.8 kbar with GASP) have been calculated for sample Pabu (Fig. 5A).
Conquet-Penze Micaschist Unit
Garnet is abundant in the Conquet-Penze Micaschist unit. Samples for geothermobarometry and monazite dating come from the outcrops at Penzer (Penz), Plage de Pors Liogan (samples labeled with “PLi”: PLi412, PLiN1, PLi1–5-1), Pointe des Renards (RenS3, Bil), Plage de Portez (samples labeled with “Port”: Portez, PortMic). Micaschist horizons in the coastal outcrops to the south of the Le Conquet have garnet of 1.5–5 cm in diameter, and a complex structural evolution is documented in S1–S5 internal and external foliations (Jones, 1994). In samples with garnet <1 cm and without aluminosilicates, an internal foliation, S1i, which is outlined by numerous aligned inclusions of quartz, ilmenite, and rarely plagioclase and mica, is enclosed in garnet (Fig. 3B). Apparently dependent on the orientation of the section, the trails of S1i can be almost straight, slightly curved in an S-shape, but also more complex, with millipede-like structures, as shown by Passchier and Trouw (1996). The S1i is discordant to and does not line up with the surrounding matrix foliation S2. The main foliation S2 (corresponding to S4 in Jones, 1994) curves around the lenticular domains with garnet porphyroblasts, and its pressure shadows with quartz. A composite character of S2 and its polyphase development is obvious from intrafolial planar domains and lenticular microlithons bearing mica oriented at acute angles or perpendicular to the strictly preferentially oriented phyllosilicates, which define the S2 planes (Fig. 3C). Fine-grained staurolite is enclosed in garnet rims. Coarser grained staurolite in the matrix is preferentially oriented with its long axis parallel to the southwest-plunging mineral lineation. It encloses planar trails of quartz and ilmenite (S2i), which line up with the external S2 (Fig. 3B). The microstructures indicate a succession of assemblages (involving muscovite + quartz + plagioclase) with:
garnet + biotite ± chlorite,
garnet + biotite ± staurolite, and
staurolite ± biotite,
The metapelite garnets in the Conquet-Penze unit display prograde zonations (Tracy, 1982) with uniformly decreasing spes-sartine and increasing pyrope and almandine contents from core to rim. Retrogressive overprints in the outermost garnet rims, which show a decreasing pyrope content, are only poorly developed. In samples from Penzer and Porz Liogan to the south, grossular contents (Grs 6.5–9.2%) are slightly higher than in Plage de Portez (3.3–6.4%) to the north (Fig. 4C–F). This difference is matched by different An contents in plagioclase, with higher An contents to the south (An 11–24) than to the north (An 8–15), as has similarly been observed by Jones (1994). Maximum pyrope contents in sample Penz (Prp 12.5%) are slightly lower than in samples from Porz Liogan (13.0–14.6%) and Plage de Portez (13.0–14.3%). The compositions of S2 biotites are always slightly variable and range between XMg 0.45–0.49 (sample Penz) and XMg 0.43–0.48 (sample Portezmic), but a trend toward lower XMg in the north, as described by Jones (1994), is only weakly developed.
Garnet-biotite Fe-Mg exchange thermometry and GMBP barometry in such progressively deformed metapelites was performed by analyses from mica and plagioclase corresponding to early (= core) and successively later (= toward rim) stages of the garnet growth. When compositional zoning of garnet in metapelites is a product of continuous reactions involving garnet, mica, plagioclase, staurolite, aluminosilicates, and quartz of low-variance assemblages (Tracy, 1982), it reflects finite temporal and spatial domains of equilibration. Moreover, when the coexistent minerals (mica and plagioclase) are preserved within a microstructural domain or occur as inclusions that can be correlated with matrix phases (St-Onge, 1987), their chemical compositions in such “local equilibria” allow evaluation of successive P and T or P-T changes for consecutive stages of garnet growth (Triboulet and Audren, 1985; Audren and Triboulet, 1993; Schulz, 1993). The basic assumption is that domain equilibrium was achieved and preserved within a given volume of rock. A minimum error of ±50 °C and ±1.5 kbar has to be considered. Compositional changes of plagioclase enclosed in garnet and in the matrix can be excluded because of slow volume diffusion rates (Spear et al., 1990). From the slight variations in XMg it can be concluded that biotite behaved as a reservoir.
Mica that occurs as inclusion in garnet or in microlithons wrapped by S2 is considered to have grown early and simultaneously with garnet cores. Mica along S2 or in pressure shadows around garnet documents a later stage of crystallization (Fig. 3A–C). Similarly, early stages of mineral chemical evolution are displayed by plagioclase enclosed in garnets, and/or the cores of zoned plagioclase porphyroblasts. The late stage is documented by the rims of zoned porphyroblasts. Mica and plagioclase of S1i in garnet, early mica in microlithons, and cores of zoned plagio clase were related to early garnet core compositions. Syn-S2 mica and the plagioclase rims coexisted with late Mg-rich garnet rims. Accordingly, P and T for crystallization of Mg-rich garnet rims have been calculated using the garnet-biotite thermometer of Bhattacharya et al. (1992) in combination with the GMBP barometer (Holland and Powell, 1990), as reported above. The Holland and Powell (1990) version of the GASP barometer gives values intermediate between the pressure estimates from biotite-Mg and biotite-Fe in GMBP. Temperatures calculated from garnet rims with the maximum Mg content range between 580 °C (Penz) and 610 °C (PLiN1) in the southern part and ∼610 °C in the northern part of the micaschist unit (Fig. 5B); thus, a small difference in temperatures appears. Temperatures calculated from the Mg-poor garnet cores are lower and range between 490 and 570 °C, and apparently are dependent on the exact analysis location in zoned garnets with low Mg contents. Corresponding to decreasing An contents in plagioclase, pressures calculated from garnet rims and cores appear to be systematically ∼1 kbar higher in the northern and structurally deeper parts of the section. The maximum pressures are ∼8 kbar at Penzer and 9 kbar at Plage de Portez for a temperature of 610 °C (Fig. 5B).
The combination of the thermobarometric data from metabasites, paragneisses, and micaschists can be arranged into continuous, single, clockwise P-T paths for each of the lithological units (Fig. 5). A significant increase of metamorphic pressure and temperature is observed toward the lower parts of the crustal pile, which is represented by the Lesneven Gneiss unit. This pattern is in accordance with previous observations (Le Corre et al., 1991; Jones, 1994), suggesting that the central Léon lithotectonic units represent a normal crustal pile, with the highest-grade rocks in the structurally lower part. However, even when hampered by the lack of garnet-bearing rocks in the transition zone, which is best exposed around Le Conquet, the thermobarometric data do not support the existence of a major thrust or metamorphic discontinuity between the gneiss and the micaschist units. As has been stated by Jones (1994), the lithological boundary has a transitional character. It coincides with the occurrence of garnet with flat zoning profiles in Mn. The transitional character of the boundary is further supported by the occurrence of the foliated Pointe des Renards metagranitoid in both units. According to the sparse chronological data (565 ± 40 Ma, Rb-Sr whole rock) and the structural observations (Michot and Deutsch, 1970; Chauris and Hallégouët, 1989), the protolith of the metagranitoid is interpreted to have intruded before the formation of the foliation in the host rocks. However, the present P-T data do not exclude the possibility that the units are both part of a major nappe unit that was underthrusted beneath the Central Armorican Domain, as has been proposed by Rolet et al. (1986, 1994). Numerical thermal modeling (Davy and Gillet, 1986) has shown that underthrusting during a continental collision could result in different shapes and maximum temperatures of P-T paths in the upper and lower parts of a lithotectonic pile, as is observed from the Léon units.
Garnet Trace Element Composition
Y and REE distribution between metamorphic accessory phases (e.g., monazite, xenotime) and major phases (e.g., garnet) is likely to be the result of the whole-rock reaction history. Monazite, xenotime, and garnet appear to be the major fractionating phases for Y in metapelites (Pyle and Spear, 1999; Pyle et al., 2001); therefore, trace element zonation in garnet may provide information concerning the relative appearance and chemical composition of monazite. Maximum Y, Sc (50–250 ppm), and V (20–80 ppm) contents in the garnet zonation profiles are dependent on garnet mode and element abundances in the bulk rock. For Y this bulk rock dependency is obvious in the amphibolite-facies micaschist samples with high maximal Y in garnet, as well as for the low-Y garnets in the high-grade aluminous para-gneisses. The maximum Li contents in garnet are systematically lower (8–74 ppm) in the aluminous paragneisses and high-grade rocks than in the micaschists (Li 26–126 ppm), where the garnet is associated with staurolite rich in Li (500–1000 ppm).
Heavy REE (HREE) and Y display a strictly linear correlation in garnet (Fig. 4A–D). Like Mn, both Y and HREE should decrease from garnet core to rim at increasing temperature when garnet appears as the only Mn, Y, and HREE fractionating phase in equilibrium with xenotime and monazite, which are consumed (Pyle and Spear, 2000). At low temperatures, xenotime is supposed to be a stable phase (Pyle et al., 2001). When the mode or volume of garnet in a xenotime-bearing rock successively increases during metamorphism, garnet, which crystallizes at higher grades, should be progressively lower in Y and HREE. This property could explain the low Y and HREE in garnets of aluminous paragneiss samples where garnet occurs at high modes (10–15 vol%) and bulk rock Y shows a broad variation of 22–35 ppm, as in the other metasediments. Garnet in the Conquet-Penze micaschists, which crystallized at lower temperatures and recorded a prograde P-T path, is rich in both Y (up to 1000 ppm) as well as HREE (up to 500 ppm) when compared to garnet in the aluminous paragneisses. Furthermore, decreases in Y and HREE are observed in the garnet rims of samples Penz, PLiN1, and PortMic (Fig. 4C and D, Table 1). However, in samples Penz and PortMic, the Y and REE are very low in the Mn-rich cores of garnet that crystallized at low temperatures. The Y contents strongly increase toward the garnet inner rims. This trend may indicate that the Y was initially bound in xenotime and/or monazite and then liberated by the breakdown of these phases under prograde metamorphic conditions (Fig. 4C and D).
MINERAL CHEMISTRY AND AGE OF MONAZITE
Bulk rock REE absolute abundances and pattern of monazite-bearing metasediments match the post-Archean average Australian sedimentary rock (PAAS) composition reported in Rollinson (1993). This match indicates that the occurrence of monazite is not related to any specific concentration of REE in the metasediments. The bulk rock REE contents increase, while MgO and Y contents (20–40 ppm) decrease, regardless of metamorphic grade and abundance or absence of garnet. Al2O3 is >20 wt% and K2O ∼1.0 wt% in aluminous paragneisses compared to Al2O3 of 15–17 wt% and K2O of 3–4.5 wt% in the other metasediments. The CaO of monazite-bearing samples is comparably low (0.5–2.7 wt%), supporting the general observation that monazite is not stable when calcsilicate phases occur (Finger et al., 1998).
Monazite grain sizes increase with metamorphic grade and range from 20–100 µm in amphibolite-facies micaschists up to 250 µm in high-grade aluminous paragneisses. Two distinct age populations of monazites, Variscan and Cadomian, as detailed below, have been found in the Léon metasediments (Table 2). The observation of two separated monazite age groups is supported by mineral chemical data. Compositions of natural monazite vary in LREE, Th, Y, Ca, and Si (Franz et al., 1996). Trivalent Y substitutes for LREE, and Th4+ and U4+ can be substituted by Th or U + Si = REE + P (huttonite substitution), with Si replacing P in the tetrahedral site, and by Th + U + Ca = 2REE (brabantite substitution), with Ca replacing in addition REE in the 8-fold site (Spear and Pyle, 2002). In the studied monazites, an excellent linear correlation between Th + U and Ca exists, suggesting that the brabantite exchange is the dominant substitution (Fig. 6A). Furthermore, Th and U display broad variations in both Variscan and Cadomian monazites, as is obvious from the U/Pb-Th/Pb diagram (Fig. 7H), which signalizes an unlimited substitution.
The REE roughly correlate negatively and linearly with Y2O3 (Fig. 6B). Significant variations of REE and Y in monazite with metamorphic grade were reported by Heinrich et al. (1997), Pyle et al. (2001), Finger et al. (2002), and Spear and Pyle (2002). Therefore, it is interesting to compare compositional parameters of Variscan and Cadomian monazites. The Y2O3 in Cadomian monazite is elevated at ∼2.2 wt%. In contrast, considerable variation of Y2O3 is observed from Variscan monazite: high-grade aluminous paragneisses contain monazite low in Y2O3 (∼0.2 wt%), intermediate Y2O3 (1.0–1.8 wt%) contents are observed in monazite from amphibolite-facies garnet-bearing rocks, and high Y2O3 contents of >2.0 wt% occur in monazite in high-grade samples without garnet (Fig. 6C). In the Léon monazites, Y2O3 apparently is not strictly related to metamorphic grade. Therefore, an interpretation of monazite Y2O3 in terms of metamorphic temperatures is difficult. On one hand, increases of Y2O3 in monazite should be correlated with increasing metamorphic grade when xenotime coexists (Heinrich et al., 1997). This trend is not matched by the Variscan monazite, which are lower in Y2O3 in aluminous paragneisses with garnet + sillimanite + cordierite + K-feldspar assemblages when compared to monazite in micaschist with garnet + staurolite assemblages, thereby indicating growth subsequent to the breakdown of xenotime. On the other hand, a low monazite Y content does not automatically indicate a low formation temperature. It may result from unavailability of Y, either because the host rocks are Y-deficient, or because Y is retained and fractionated into other minerals, such as xenotime and prevailing garnet (Pyle et al., 2001). Monazite is low in Y in samples with abundant low-Y garnet, as in aluminous paragneisses, and it is intermediate in samples with few garnets. High Y in monazite occurs in samples without garnet. These observations allow us to conclude that Variscan monazites should have crystallized subsequent to the breakdown of xenotime and during or subsequent to Mg-rich garnet, which represents the thermal peak of the metamorphic evolution. Furthermore, the potential for preservation of older populations of monazite increases with bulk rock Y and when the mode of garnet remains low and xenotime is absent.
All ten studied samples contain monazite with Variscan and older chemical model ages (Fig. 7A–G, Table 2). The weighted averages (Ludwig, 2001) of the Variscan ages range from 304 ± 21 Ma (Port) to 340 ± 16 Ma (PLiN1) in the Conquet-Penze micaschists. In the Lesneven gneisses, the ages are younger and range from 300 ± 10 Ma in Kerz to 324 ± 15 Ma in Pabu (Fig. 7A–G). Three samples with Variscan monazites contain another monazite generation, which yielded ages at 520 ± 28 Ma (Pabu), 517 ± 13 Ma (Port) and 552 ± 22 Ma (Sab1). The “Cadomian” monazites occur both in the Lesneven Gneiss and in the Conquet-Penze Micaschist units. The single ages from individual grains display two distinct groups between 280–350 Ma and 560–510 Ma in the U/Pb-Th/Pb diagram (Fig. 7H) of Cocherie and Albarede (2001). No monazite was found with ages in between or linking these two groups. Thus, the monazites belong to entirely separate events. The existence of clearly separated groups of ages, observed within single samples, indicates that monazite has not been affected by possible lead loss or postcrystalline metasomatism.
CADOMIAN AND VARISCAN METAMORPHIC EVENTS
Thermobarometric data from metasediments and metabasites in the Lesneven Gneiss and the Conquet-Penze Micaschist units signalize a normal crustal pile and increasing maximal metamorphic temperatures and pressures with increasing structural depth. A change from prograde-zoned garnet with increasing XMg toward the rims to Mg-rich garnet with homogeneous cores and zoned outermost rims coincides with the mapped lithological border of the Conquet-Penze Micaschists and the Lesneven Gneisses. The common retrograde P-T evolution in the sillimanite and then andalusite stability fields was mainly recorded by assemblages with Ca-amphibole in the various metabasites. Mineral-chemical, thermobarometric, and Th-U-Pb monazite age data indicate a transitional character and no major metamorphic discontinuity between the Conquet-Penze and Lesneven units.
Monazite Th-U-Pb model ages determined by the EMP range from 340 to 300 Ma, with younger monazite of 310–300 Ma prevailing in lower parts of the Lesneven Gneiss unit (Fig. 7I). In aluminous paragneisses, monazite of 310–300 Ma coexisted with cordierite and serves as a temporal marker for the high-temperature metamorphic stage. The eclogitic high-pressure stage is linked to the high-temperature stage along a single P-T path recorded by the prograde garnet zonation in the eclogites (Fig. 5A). This link provides an argument that the age of the high-pressure stage may be younger than the 439 ± 13-Ma U-Pb zircon lower intercept age given by Paquette et al. (1987). However, the Carboniferous monazite ages coincide with the K-Ar and Rb-Sr mica ages in the granites and orthogneisses and the intrusion of the late granites (l'Aber-Ildut, Ploudalmezeau) at 300–290 Ma. They postdate the early 340–330-Ma granite de Saint Renan-Kersaint, which was already post-tectonic and posterior to an Upper Devonian “phase bretonne” observed in low-grade Paleozoic sequences (Le Corre et al., 1989, 1991). In consequence, the metamorphic monazites support a distinct “late Variscan thermotectonic event” (Le Corre et al., 1989) in the Léon domain that lasted until the Lower Permian during the late generation of granite intrusion. Similar Late Variscan mica ages have been reported from the South Armorican domain (Brown and Dallmeyer, 1996).
The question of whether the amphibolite-facies metamorphism in the Conquet-Penze unit is a Variscan or an older event (Jones, 1993, 1994) can now be discussed in view of new Cadomian monazite ages, the garnet trace element zonations, and the P-T conditions of metamorphism. Cadomian monazites were analyzed in three samples from different locations in both the Conquet-Penze and Lesneven units. Variscan monazite appears in the same samples. Cadomian monazite occurs in a garnet-free paragneiss and in samples with low modes of garnet. All monazites in garnet-rich micaschists and aluminous paragneisses are of Variscan age. Cadomian monazite is rich in Y, regardless of whether it is from samples with Y-rich or Y-poor garnet or without garnet, and contrasting low-Y Variscan monazite observed in the same samples. From their uniform mineral-chemical compositions and their narrow range of ages, a detrital origin of the Cadomian monazites can be excluded. They should record an early thermal event in the units and provide a minimum age of sedimentation. As Y-rich Cadomian and Y-poor Variscan monazite appear in the same sample with garnet, it can be concluded that Cadomian monazite crystallized previous to, and Variscan monazite subsequent to, the garnet; the garnet-bearing assemblages are bracketed by the two populations of monazite. New monazite can crystallize when Y is available through breakdown and retrogressive replacement of nearby garnet and xenotime (Foster et al., 2000, 2002; Pyle and Spear, 2000). This interpretation is supported by the observation of Y-rich Variscan monazites in samples without garnet.
The existence of monazite previous to garnet crystallization is further supported by Y-poor garnet cores, which imply the former presence of another phase (monazite and/or xenotime), which fractionated Y previous to garnet growth. The Y content of a preserved early monazite should be dependent on bulk rock composition, metamorphic temperature, and the presence of xenotime, regardless of whether later garnet growth occurred. This dependence allows some conclusions to be drawn on the nature of the Cadomian thermal event. At increasing temperature, Y in monazite increases at the expense of xenotime, which is consumed and disappears at higher grades (Heinrich et al., 1997; Pyle et al., 2001; Pyle and Spear, 2003). Xenotime was not evident in the studied samples and presumably was consumed; thus, high-Y Cadomian monazite could indicate an elevated temperature of metamorphism. The bulk rock composition was suitable for garnet crystallization in some samples, however, the observed garnet has low Ca and crystallized at medium pressures. Garnet would not be stable at the given Ca-poor bulk rock compositions when pressure is low (Spear, 1993; Pyle and Spear, 2003). Monazite and xenotime are consumed during garnet growth (Pyle et al., 2001; Pyle and Spear, 2003). Thus, the preservation of Cadomian monazite with high-Y content could be explained by the lack of coeval garnet, caused by low pressures during the Cadomian metamorphic event. One could speculate on a Cadomian contact metamorphism in the vicinity of intrusions like the Pointe des Renards metagranitoid. From the available sparse radiochronological data (565 ± 40 Ma, Rb-Sr WR; Michot and Deutsch, 1970) this possibility cannot be excluded. A Cadomian regional low-pressure metamorphism was described from the eastern parts of the North Armorican Cadomian domain (Ballèvre et al., 2001) and appears to be an alternative explanation for the Cadomian monazite ages.
There exists a zoneography of metamorphic cooling ages within the Armorican Cadomian belt, ranging from Neoproterozoic (600 Ma) cooling in the Trégor unit to the northwest to Cambrian (520 Ma) cooling in the Fougères unit to the southeast (Fig. 1A). The northeastern part of the Léon is juxtaposed on the Trégor unit of the internal Cadomian belt along the main Cadomian thrust, which curves to the northwest along the Baie de Morlaix. To the east, the Léon continues into the Saint Brieuc and Saint Malo units of the external Cadomides (Fig. 1A and B). In the St. Brieuc unit and its subordinate Yffinac formation, maximal high-grade conditions at 9 ± 1 kbar and 700 ± 50 °C (Hébert, 1995) were achieved previous to cooling at ca. 570–560 Ma (Dallmeyer et al., 1993). This scenario is in contrast with the later Cadomian deformation and metamorphism in the St. Malo migmatitic unit at ca. 550–540 Ma, which is characterized by low pressures and high temperatures (<4 kbar and 700 °C) and the lack (or rarity) of garnet (Weber et al., 1985; Ballèvre et al., 2001). The boundaries of the Léon are masked and overprinted by the Carboniferous Morlaix basin, the Variscan granites, and the dextral Variscan North Armorican shear zone and their related displacement zones. However, in the light of the Cadomian monazite ages, the Léon domain appears to be a part of the external Cadomian belt, as exposed in the St. Malo or Fougères units and thus is not entirely exotic within the North Armorican domain.
In conclusion, in the Léon domain several features typical of the Saxo-Thuringian zone of the Variscan belt, as defined by Kossmat (1927), can be recognized. One characteristic of the Saxo-Thuringian zone is the occurrence of Cadomian remnants, displayed by the Cadomian monazite ages. According to their bulk rock chemistry, the Léon metasediments were deposited in an active continental margin setting as a part of the Briovérien sequences. Such Cadomian elements still have to be better constrained, for instance, by radiometric dating of the Pointe des Renards metagranitoid and other intrusive rocks. Other features of the Saxo-Thuringian zone observed in the Léon are Variscan granitic plutonism, deformation, and metamorphism to various degrees. The eclogite-facies metamorphism has been taken as an argument for a South Armorican affinity. However, apart from different bulk rock characteristics, the P-T paths and the tectonic relationships of the South Armorican high-pressure metabasites in Baie d'Audierne (Lucks et al., 2002), Ile de Groix (Schulz et al., 2001), and Champtoceaux (Ballèvre et al., 1994) considerably differ from the marked decompression-heating path of the Léon eclogites, regardless of the ages of the high-pressure events, which still need to be made more precise. To the southeast and east, Early Paleozoic to Devonian and Carboniferous sedimentary records are preserved. The central Léon metamorphic units represent the footwall of these sedimentary basins. As regards the Upper Devonian “phase bretonne,” known from the stratigraphic record (Rolet et al., 1986, 1994), the metamorphic monazite ages postdate this event. However, the S1 and S2 deformation structures, with associated crystallization of garnet, aluminosilicates, and staurolite, predate the Carboniferous intrusion of the Saint Renan–Kersaint granite (340–330 Ma). Then the range of Variscan monazite ages (sample PLiN1, 340 ± 16 Ma; sample Port, 304 ± 21 Ma) link this event to a Late Carboniferous stage, with overprinting of the planar structures by dextral shearing. The younger Variscan monazite ages indicate that the Léon units were exhumed no earlier than the Upper Carboniferous, presumably in line with the dextral wrench tectonics.
TheP-T data and P-T paths in combination with the monazite ages emphasize that the central Léon units underwent a common structural and metamorphic evolution during the Variscan orogeny. Inverted metamorphic gradients or superposition of high-pressure rocks on lower-pressure units, which provide arguments for nappe tectonics in other parts of the Variscan belt (e.g., Matte, 1991), are not evident in the Léon. This observation does not exclude the possibility that the units may represent allochthonous or parautochthonous crust. Rolet et al. (1986, 1994) proposed that the Léon units were underthrusted toward the southeast or east beneath the Central Armorican domain during a Variscan collision. In this model, a South Armorican provenance of the nappes is not favorable, despite the possible re-arrangement of tectonic domains during Carboniferous dextral wrench tectonics. As supported by the Cadomian monazite ages, an origin of the Léon units as former parts of a Cadomian realm, situated to the northwest or west appears to be more likely. When interpreted in the light of geodynamic reconstructions of the western part of the Variscan belt (Matte, 2002; Shelley and Bossière, 2002; Stampfli et al., 2002; Cartier and Faure, 2004), our findings do not serve as an additional argument for the existence of an Early Paleozoic oceanic realm (Medio-European Ocean, Massif Central Ocean, or South Armorican Ocean) and suture zone between the southern margin of an Armorican micro-plate and Gondwana. They rather support the notion that the Léon units were parts of a suture zone along the northern boundary of the Armorican microplate, hence related to the margin of a former Rheic Ocean between Armorica and Avalonia.
We commemorate the Breton geologist Claude Audren from Géosciences Rennes, who guided the first author to the Léon in 2001. The constructive reviews by J.F. von Raumer, Fribourg, and J. Rolet, Plouzane, are gratefully acknowledged. We thank U. Kroner, Freiberg, for giving us insight to his manuscript. Microprobe analyses were facilitated by H.-P. Meyer at Mineralogisches Institut Universität Heidelberg and by U. Schüßler, Würzburg. Technical support was provided by R. Baur, P. Spaethe, and K.-P. Kelber at the Institut für Mineralogie in Würzburg. The project was financed through the Deutsche Forschungs-gemeinschaft (SCHU 676/10).
Figures & Tables
The Evolution of the Rheic Ocean: From Avalonian-Cadomian Active Margin to Alleghenian-Variscan Collision
- absolute age
- Armorican Massif
- Cadomian Orogeny
- chemical composition
- crystal chemistry
- eclogite facies
- geologic barometry
- geologic thermometry
- ICP mass spectra
- Ille-et-Vilaine France
- mass spectra
- metamorphic rocks
- P-T conditions
- trace elements
- Variscan Orogeny
- Western Europe
- Fougeres Unit
- Guingamp Unit
- Saint Malo Unit
- Saint Brieuc Unit
- Lesneven Unit
- Conquet-Penze Unit
- Tregor Unit
- Leon Unit