Sr, Nd, AND Pb ISOTOPIC COMPOSITION OF CLINOPYROXENE AND TEMPERATURE AND AGE OF MINERAL EQUILIBRIA: LATE CRETACEOUS MANTLE XENOLITHS FROM THE CENTRAL ANDES

LUCASSEN F.; FRANZ G.; VIRAMONTE JOSÉ G.; ROMER R.

Salvador

Simposio; IV South American Symposium on Isotope Geology; 2004

INTRODUCTION AND PETROLOGYMantle derived magmatism in W Argentina and Bolivia is bound to late Mesozoic extension, e.g. in the Salta Rift System (Fig. 1; Viramonte et al., 1999). The composition of these magmatic rocks indicates two different mantle sources, one comparable to a depleted mantle and a second comparable to old subcontinental mantel of the Brazilian Shield (Fig. 1; Lucassen et al., 2002). This study investigates chemical and isotopic composition of depleted mantle xenoliths from the Salta Rift/ Quebrada las Conchas (Fig.1), 26°S near the town of Cafayate in NW Argentina, which are hosted by smallvolume basanitic dikes and sills within a continental rift basin (e.g. Viramonte et al., 1999; Lucassen et al., 2002). Mantle xenoliths and their host rocks are used toconstrain the composition of the upper mantle before the onset of the Plateau formation in the Central Andes. The occurrence of large xenoliths (ca. 4 – 20 cm), providing sufficient material for bulk methods of investigation as XRF, ICP-MS and TIMS, are restricted to few locations within the extended rift system. Most xenoliths are spinel lherzolite to harzburgite with variable amounts of olivine, clinopyroxene (cpx), orthopyroxene (opx) and spinel. Few samples are dunite. Carbonatization of these xenoliths occurred locally at the intrusion level, preserving clinopyroxene and spinel in a calcite matrix. Many samples show hydration of olivine along grain boundaries. Some large peridotite xenoliths show compositional layering or preferred orientation of the minerals. At the microscopic scale, all samples show well annealed metamorphic fabrics with straight grain boundaries, and 120° triple points are common. Optical zoning of the minerals is absent. Cpx of some samples shows small rims (10 -30 μm) of secondary low Al and Na cpx in palisade textures together with glass, considered to be typical of melting along grain boundaries during fast uplift of the xenoliths. Mineral assemblages of all samples selected for isotopic studieshave been checked for compositional zoning by electron microprobe, which is absent however. The mineral assemblage appears to be in chemical equilibrium. Thermometric studies of cpx – opx indicate rather uniform temperatures between 1000 – 1100°C (12 samples). The peridotite xenoliths represent the uppermost mantle shallower than ca 70 km, outside the garnet stability field.CHEMICAL AND ISOTOPIC COMPOSITIONBulk rock compositions (XRF, 27 samples) are similar to other spinel peridotite xenoliths worldwide (Maaloe & Aoki, 1977) and Mg# varies between 89 – 92 at high Cr (1700 - 4000 ppm) and Ni (1700 – 2500 ppm) content. For Sr, Nd and Pb isotope measurement and REE analyses cpx (21 samples) has been separated (for analytical methods see: Lucassen et al., 2002). Cpx is the principal host of the respective elements in the peridotite because amphibole and phlogopite are absent. Selection and leaching avoids possible effects of alteration and accidental contamination during whole rock sample handling especially on the U-Th-Pb isotope system. Sr,U, Th, Pb, Sm and Nd element concentrations are variable in the cpx, whereas Rb is below detection limit. REE pattern of cpx are highly variable from mildly depleted to essentially flat patterns to prominent light REE enrichment (Fig. 2; La/YbN ratios between 0.4 – 110). Sr, Nd and Pb isotope ratios are corrected for in-situ decay to 100 Ma using the element concentrations from ICP-MS measurements. Sr and especially Nd isotope ratios show considerable variations (Fig. 3). Most samples scatter within the depleted mantle field (the main group, where the discussion is focused on), three samples have elevated Sr isotope ratios, two samples plot into thefield of the crust or enriched mantle. 207Pb/204Pb and 206Pb/204Pb isotope ratios scatter from rather unradiogenic to radiogenic values (Fig. 4a). 208Pb/204Pb ratios (Pb from Th decay) of some xenoliths are high at given 206Pb/204Pb ratios (Fig. 4b). Opx and cpx have been separated from selected samples in order to constrain the age of isotopic equilibration in the Sm-Nd system. Sm/Nd ratios in opx are slightly higher than in cpx. Ages from two point ‘isochrons’ of five samples vary between ca. 70–130 Ma. One sample yielded a late Palaeozoic age (ca. 300 Ma).DISCUSSION – IMPLICATIONS FOR THE EVOLUTION OF THE UPPER MANTLEWell equilibrated fabrics and the lack of chemical zoning in the minerals indicate chemical equilibrium between the minerals of the xenoliths. Temperatures of 1000 – 1100°C). A precise age cannot be derived from the Sm-Nd systematics of opx – cpx pairs, due to the small differences in the Sm/Nd ratios. However, the data allow inferring a late Mesozoic age of the high-T metamorphism with possible Palaeozoic inheritance in the Sm-Nd system of one sample. A late Mesozoic age and high-T from mineral equilibrium are in good accordance with the high-T of metamorphism and Sm-Nd mineral isochron ages in lower crust felsic – mafic granulite xenoliths from the same location (ca. 900°C, 10 kbar; ca. 90 – 100 Ma; Lucassen et al., 1999). Theperidotite xenoliths document the in-situ thermal condition of the upper mantle lithosphere in theCretaceous. Lower crust and mantle lithosphere have been metamorphosed at high-T during the activity of the late Mesozoic rift. Older events of mineral formation i.e. core – rim textures and chemical zoning have not been found in the samples. Primary hydrous phases like phlogopite and amphibole are absent. Major element composition of the peridotite (whole rock) is considered typical for such rocks and varies in a relatively small range. The variation e.g. decreasing Al2O3 and CaO with increasing Mg# could be explained by increasing depletion by melt extraction. Variable REE pattern, trace element content and isotope ratios of the cpx however indicate a more complex evolution of the upper mantle lithosphere. There is no typical ‘LREE depleted mantle’ pattern and most of the variation within the LREE are variable degrees of enrichment. The HREEpatterns are flat in most samples (Fig.2). The La/Yb ratio and Sr (Fig. 5a), REE content and the U/Pb ratios are positively correlated. Pb isotope ratios increase with Sr content (Fig. 5b) and La/Yb. 143Nd/144Nd ratios decrease with increasing Nd content (Fig. 6) and La/Yb ratios.87Sr/86Sr ratios of the main group (0.703 – 0.704; 17 samples) show no correlation with the variable Sr content, but 4 samples with 87Sr/86Sr ratios > 0.705 show positive correlation with Sr content (Fig.7). Radiogenic growth since the Cretaceous is minor and the correlation between the isotope ratios and respective parent elements is weak. Variations in the Pb isotope ratios (207Pb/204Pb: ca. 15.5 – 15.7 and 206Pb/204Pb: ca. 17.5 - 20.5) and Nd isotope ratios (ca. 0.5124 -0.5138) indicate contributions of Pb and Nd from sources with long-term separateisotope evolutions. Principal regional sources are the Pacific MORB-type mantle, old subcontinental mantle of the Brazilian Shield and local continental crust as the contaminant in a subduction system (Figs. 3, 4). However, there is no clear affinity of a sample group to one of the sources or a uniform mixing relation between the sources. Deviations of REE (and other trace elements) from the expected LREE depleted pattern in depleted mantle compositions are usually attributed to LREE enrichment during mantle metasomatism by percolating fluids or melts. In the investigated samples, Nd and Pb isotope ratios are correlated with increasing trace element contents of the cpx, whereas Sr isotope ratios are largely independent. Considering a isotopic composition close to depleted mantle as the starting condition (sample 6-180a; Figs. 3, 4; 206Pb/204Pb: 18.80, 207Pb/204Pb: 15.55, 208Pb/204Pb: 38.50, 143Nd/144Nd: 0.51312, 87Sr/86Sr: 0.70264, La/YbN: 0.9), a melt similar to the average basanite hosting the xenoliths could substitute the second endmember in the Nd – Sr isotope diagram (Fig. 3). However, the Pb isotope composition of the xenoliths does not indicate any systematic relation to the composition of the average basanite or other possible Pb source (Fig. 4). The REE patterns of the xenoliths do notshow a gradual change from depleted to enriched patterns as expected from a reaction with varying volumes of compositional uniform magma under similar physical conditions (Fig. 2). These observations make a single and uniform process causing the observed variations in the peridotite unlikely. Magmatism associated with the Cretaceous Rift is of minor volume in small isolated occurrences (Viramonte et al., 1999). The mainly alkaline rocks are derived from low degrees of melting in the mantle. If the magmatic activity is an expression of the degree of melting in themantle and its regional distribution, most of the mantle volume beneath the Cretaceous Rift was unaffected by melting. Despite the chemical and isotopic equilibrium between the minerals on the handspecimen-scale, a larger scale equilibration of isotope composition and trace element distribution is unlikely without mayor melt extraction/percolation in the upper mantle. Therefore, The western edge of Gondwana has long history as a sporadically active margin at least from the Early Palaeozoic onwards with major reworking of Proterozoic crust in a Palaeozoic mobile belt (Fig. 1). This and the isotopic and chemical composition of the xenoliths and mantle derived magmatic rocks are the basis of our working hypothesis for the evolution of the mantle in this section of the Central Andes. (1) The mantle at the leading edge of the continent has been modified by several timely distant systems of subducting oceanic plate – asthenospheric mantle wedge configurations. This resulted in a generally depleted mantle beneath the Palaeozoic orogens/magmatic arcs in contrast to the old subcontinental mantle beneath the Brazilian Shield orfound in the Cretaceous magmatism between 32- 34°S (Figs. 1, 3, 4). Most Early Palaeozoic to Mesozoic mantle-derived magmas (e.g. Lucassen et al., 2001) andthe mantle xenoliths in the realm of the Palaeozoic mobile belt are from isotopically depleted mantle sources. (2) Compared with MORB mantle, this depleted mantle lithosphere is heterogeneous in isotope and trace element composition. The major heterogeneities occur in the Sm- Nd and U-Th-Pb isotope systems and trace elements (e.g. Sr and REE), whereas Sr isotope ratios are rather uniform (Figs. 3, 4, 7). The measured μ-ratios (238U/204Pb) in the cpx (ca. 1 – 300, the assumed mantle value is ca. 8) and the 147Sm/144Nd ratios (ca. 0.06 -0.25, the assumed mantle value is ca. 0.21) show considerable variations whereas Rb is below the detection limit or negligible compared with Sr content. If metasomatic processes in an ancient active mantle wedge occurred at large scale this could explain the observed differences. Assuming an average μ-ratio of 30 for a metasomatized depleted mantle, 200 Ma of separation (e.g. during lull of subduction) would be sufficient to cause substantial radiogenic growth in the UPb system (e.g. 206Pb/204Pb from 18 to 19 instead to 18.26 for a μ ratio of 8 of the depleted). A147Sm/144Nd ratio of 0.1 would retard the radiogenic growth of 143Nd/144Nd compared with normal depleted mantle evolution (e,g from 0.5127 to 0.51283 instead to 0.51297 for a 147Sm/144Nd ratio of 0.21 of the depleted mantle). The uniform Sr isotope composition of the main group may result from generally anhydrous mantle without Rb bearing phases like phlogopite, whereas the few samples with high Sr-isotope ratios could be related to hydrous parts of the mantle. (3) Such modified mantle realms could be the source of xenoliths sampled by the host basanite or of melts modifying other parts of the mantle during ascent. The increase of La/Yb with increasing Sr(Fig. 5a) and the increase of the 206Pb/204Pb ratios with La/Yb (Fig. 5b) could reflect the varying amount of addition of such melts to an originally depleted peridotite. Small amounts of melt with high trace element content would change the trace element budget of a depleted peridotite dramatically, leaving the major element composition virtually unchanged. (4) Traces of melt/fluidinfiltration such as possible hydrous phases or chemical zoning of minerals have been extinguished by Cretaceous high-T metamorphism.