The petrogenesis of the eastern Pyrenean peridotites: an integrated study of their whole-rock geochemistry and Re-Os isotope composition

Abstract

The Re-Os isotopic compositions of sixteen peridotites from the eastern Pyrenean ultramafic massifs have been determined in conjunction with their whole-rock major, trace, and chalcophile element contents. The data suggest that the peridotites represent residues after the extraction of up to 25% partial melt from a fertile mantle source at relatively high pressures (>20 kbar). Linear correlations between the incompatible-element contents of the peridotites and indices of major element depletion require that the initial melt extraction must have been followed by metasomatism by an incompatible-element enriched component. The S contents of the majority of the peridotites show a linear correlation with major element depletion, consistent with the extraction of a S-saturated basaltic melt from a fertile peridotite source. Low Cu/S ratios in a number of samples, however, suggest that parts of some of the massifs may have been infiltrated by an exotic sulphide component.

The range of Os abundances in the peridotites is relatively limited (3.13–4.74 ppb) and unrelated to either the major or trace element contents of the peridotites. In contrast, the Re abundances vary widely (65–497 ppt) and correlate with their degree of major element depletion. Rhenium, sulfur, and Al₂O₃ are all positively correlated, suggesting all are controled by melt extraction. The ¹⁸⁷Os/¹⁸⁸Os ratios of the peridotites range from 0.1157 to 0.1294 and show good correlations with whole-rock Re, S, and Al₂O₃ abundances. From the y-intercept of the ¹⁸⁷Os/¹⁸⁸Os-S trend an initial ¹⁸⁷Os/¹⁸⁸Os ratio of ∼0.1159 is calculated for the peridotites, yielding a model age of 1.9 ± 0.3 Ga for the isotopic homogenisation of the mantle lithosphere that underlies the massifs. This model age is similar to the age of major crustal growth in Europe and implies a broadly synchronous stabilisation of both crust and lithospheric mantle at this time.

The linear decrease of Re with degree of melt extraction is consistent with nonmodal melting of a Re-bearing base-metal sulphide phase. A sulphide-silicate melt partition coefficient (D_sul/sil^Re) of ∼325 is calculated for Re, indicating that the chalcophile behaviour of Re during melting may be similar to that of the moderately chalcophile elements, such as Cu rather than the highly chalcophile Pt-group elements.

Introduction

As a consequence of the chalcophile and siderophile geochemistry of Re and Os, the Re-Os isotope system has been shown to provide a view of mantle processes that is unique and contrasts markedly with that provided by the Rb-Sr, Sm-Nd, and U-Pb isotope systems Walker et al 1989, Reisberg et al 1991, Pearson et al 1995. During mantle melting, Os is highly compatible and remains in the residue, whereas Re is incompatible and partitions into the melt. Consequently, residues after melt extraction have Re/Os ratios that are much lower than those existing prior to melting. The high concentration of Os in the residue renders the Os isotopic compositions of melt residues more resistant to the effects of subsequent metasomatic processes, allowing the most highly depleted residues to retain the Os isotopic composition of the mantle at the time of melt extraction.

Although the unusual behaviour of Re and Os during silicate melt generation may be accounted for by their recognised chalcophile and/or siderophile properties (depending on the presence of metallic iron), the processes responsible for Re/Os fractionation remain only poorly understood. Recent studies of orogenic peridotites from the Ronda ultramafic complex (Reisberg et al., 1991) and the eastern Pyrenean ultramafic massifs (Reisberg and Lorand, 1995) have demonstrated correlations between the ¹⁸⁷Os/¹⁸⁸Os ratios of the peridotites and their major element compositions (e.g., Al₂O₃). Although these correlations confirmed that fractionation of the Re-Os isotope system in the massive peridotites is associated with major element depletion, either coincident with, or prior to the incorporation of the peridotites into the mantle lithosphere, the relationship between the Re-Os isotopic compositions and the abundances of the chalcophile elements was poorly defined. In order to investigate further the fractionation of Re from Os during mantle melting, the Re-Os isotopic compositions of peridotites from the eastern Pyrenean ultramafic massifs were determined in conjunction with both their lithophile and chalcophile element contents. The results lead to a re-evaluation of the processes responsible for major-element variations in the massifs and a better model for the behaviour of Re and Os during mantle melting.

Previous Re-Os isotope studies of the continental mantle lithosphere underlying and adjacent to Archaean cratons have demonstrated that the stabilisation of their lithospheric keels coincided with a period of crustal generation, indicating a potentially causative link between the generation of mantle lithosphere and crust early in Earth history Pearson et al 1994, Pearson et al 1995. In contrast to the material analysed by Pearson and coworkers, the eastern Pyrenean ultramafic massifs were tectonically emplaced into relatively young crust (during Cretaceous movements of Iberia relative to France), and so offer an opportunity to investigate whether the crust and mantle portions of the continental lithosphere remained coupled during the stabilisation of younger continental regions and to examine crust-mantle coupling during orogenesis and plate movement.

The Pyrenean orogenic ultramafic massifs comprise over forty separate ultramafic bodies which crop out along the narrow North Pyrenean Metamorphic Zone (NPMZ) on the Franco-Spanish border and range in diameter from a few metres to over a kilometre. Although the ultramafic bodies occur in groups along over 350 km of the North Pyrenean Fault (which bounds the NPMZ) more than half are located in the Ariège Department, France, at the eastern end of the NPMZ (Fig. 1 ).

The eastern Pyrenean ultramafic bodies are composed predominantly of layered spinel lherzolites, in which the layering is defined by the intercalation of bands of harzburgite and clinopyroxene-poor lherzolite and 3–20 cm thick veins of spinel websterite. Isoclinal folding of the pyroxenite veins indicates that the websterites were introduced while the peridotites were still hot and plastically deforming and probably pre-date the accretion of the peridotites onto the base of the lithosphere (Fabriès et al., 1991). Rare garnet pyroxenite layers (0.3–1 m thick) are also present in some of the bodies (e.g., Lherz and Freychinède). Their orientation is commonly parallel to the peridotite layering, but also cuts across it at high angles Conquere 1978, Bodinier et al 1987b. In addition to these anhydrous pyroxenites, the peridotites of Lherz and Freychinède are cross-cut by a later generation of Ti-rich amphibole pyroxenite dykes up to 30 cm thick and thin phlogopite-bearing hornblendite veins Lacroix 1917, Conquere 1971, Bodinier et al 1987a.

Away from the brecciation that is present around the margins of the bodies and along internal faults, the peridotite massifs are moderately fresh, with large internal areas apparently free from alteration. At least two distinct phases of deformation and recrystallisation have been identified in the unaltered peridotites of the massifs (Conquéré and Fabriès, 1984), creating a range of protogranular to porphyroclastic textures. Although estimates of the temperature and pressure during the earlier of these deformation-recrystallisation events (900–950°C and 12–16 kbar; Sautter and Fabriès, 1990) indicate that it occurred while the peridotites were located within the continental mantle lithosphere, the P-T conditions during the second phase of deformation (700–750°C and ≤8 kbar; Conquere 1977, Conquere and Fabries 1984) are consistent with shearing of the upper mantle prior to the emplacement of the ultramafic bodies into the base of the crust. No absolute date is available for the earlier phase of deformation and recrystallisation, but the age of the later phase may be estimated by its coincidence with the introduction of the amphibole-rich pyroxenite veins, the composition and radiometric ages of which are consistent with their emplacement during Cretaceous crustal thinning and extension in the region Montigny et al 1986, Bodinier et al 1987a. Throughout the sequence of deformation and recrystallisation recorded by the peridotites of the eastern Pyrénées, they appear to have remained at temperatures below the dry peridotite solidus.