19~ isotopic systematics of the upper mantle and some plumes

An important factor in determining the physical and chemical evolution of the Earth is the scale of convection in the mantle. If convection is restricted to separate cells within the upper and lower mantle, then little material transport likely occurs across the 670 km discontinuity. If whole mantle convection occurs, which is predicted on the basis of various geophysical observations, then some plumes are likely to have been generated at the core-mantle boundary in the D layer. Highly fluid metal from the outer core is likely transported into the lowermost kilometers of the D layer by capillary flow (Jeanloz, 1990). Low seismic velocities in the D layer have been interpreted in terms of chemical interactions between the mantle and core (Jeanloz and Garnero, 1997). Therefore, if plumes are generated from the D layer, compositional evidence for core-mantle interaction may be present in lavas derived from such plumes. Walker et al. (1995) proposed that some radiogenic lSTOs present in some ocean island basalts could result from addition of outer core material into the D layer source of some plumes. This is because during solidification of the inner core, partitioning of Re and Os between solid and liquid metal may result in enrichment of Re/Os in the outer molten core relative to chondrites. Because of the predicted suprachondritic Re/Os ratio and high abundance of Os (c. 1 ppm) in the outer core, only small additions of outer core liquid (<~ 1%) are required to produce the observed ~SVOs/~SSOs ratios of-0.130 to 0.140 in OIB. For instance, if the inner core formed early, Walker et al. (1995) calculated that the outer core would have a 187Os/lSSOs of ~0.137, assuming Re and Os partitioning during core crystallization similar to that in asteroidal cores. Because radiogenic 1870S/1880S in OIB can be interpreted to result from core-mantle interaction or crustal recycling in their source regions, additional evidence is needed to determine whether core-mantle interaction can be identified. Walker et al. (1995) suggested that inner core crystallization may also enrich the Pt/Os ratio of the outer core relative to chondrites. 19~ decays to lS6Os via the ~ transition ()~ = 1.542 x 10-12a-1). Consequently, the longterm Pt/Os ratio of a geological reservoir can be monitored via examination of the ~86Os/~SSOs ratio of that reservoir. Because of the high Pt/Os ratio that may be present in the outer core resulting from inner core crystallization and using the same inferences for the Re-Os abundances by Walker et al. (1995), the outer core would have an approximately 0.01% more radiogenic 186Os/1SSOs than a chondritic mantle allowing for 4 to 4.5 Ga of in situ growth of lS6Os. Given these constraints, to detect variations in 186Os/lSSOs for plume-derived lavas that might be radiogenic relative to depleted upper mantle, precisions of better than _+ 0.005% must be obtained on 1860s/1880s measurements. As a first step in examining whether 186Os/~S~Os ratios vary between the upper oceanic mantle and plumes, we have determined high-precision 186Os/188Os and lSYOs/188Os ratios for 10 Os-Ir alloys and 3 chromitites from peridotite massifs and ophiolites worldwide, 2 abyssal peridotites from ODP Leg 153 (Site 920). These materials should provide an upper mantle ~86Os/18SOs. Also, 5 picrites from Hawaii were analysed. Hawaiian picrites were chosen because Hawaiian magmas are derived from one of the longest-lived hotspots on Earth. This characteristic points toward an unusually high thermal anomaly beneath the Hawaiian islands, and potentially this is the most favourable present-day case for upwelling of material from the deep mantle.