EarthRef.org Reference Database (ERR) -- McDermott & Hawkesworth 1991

Destructive plate margin rocks have (Th-230/Th-232) values in the range 0.5-2.5, which is similar to those in MORB and OIB. Approximately 60% of the available data on subduction-related rocks plot within 10% of the (Th-230/Th-232)- (U-238/Th-232) equiline, and most of the remaining analyses are displaced to high (U-238/Th-230). Moreover, despite the large range of Th isotope ratios in sediments and altered MORB in the subducted slab, the majority (approximately 70%) of destructive margin rocks plot within the Th-Sr and Th-Pb mantle arrays defined by MORB-OIB. Th/U ranges from 1.0 to 6.6 in the rocks analysed, and both the measured and the source Th/U inferred from initial (Th-230/Th-232) correlate with Pb-208/Pb-206. Thus, the major (six-fold) difference in Th/U between different arc suites was established several 100 Ma prior to subduction, and it does not appear to reflect element fractionation processes associated with Recent subduction. Oceanic sediments exhibit a wide range in Th/U but relatively restricted Pb-208/Pb-206 suggesting that the average residence age of pelagic sediments is less than 350Ma. Th isotopes sensitive tracers of subducted carbonate, hydrothermally altered oceanic crust and detrial sediments. However, only those from Nicaragua and Tonga exhibit unusually high (Th-230/Th-232) consistent with a significant contribution of Th from altered MORB and/or carbonate. Most of the water in the subducted slab is in sediments and altered MORB, and so they should dominate the slab-derived contribution. Their bulk (Th-230/Th-232) values are likely to be > 4, which suggests that subducted materials constitute < 15% of the Th in arc magmas. Finally, Th/Nb and Ba/Nb ratios are high in subduction-related rocks, but whereas the latter primarily reflects a high Ba/Nb slab flux, the high Th/Nb ratios also reflect partial melting processes in the mantle wedge. If the restricted range in (U-238/Th-230) in many subduction-related rocks (0.9-1.1) simply reflects crystal-liquid trace element partitioning, then the amount of residual amphibole is < 20%. Alternatively, if (U-238/Th-230) in part reflects the dynamics of melt extraction at destructive plate margins then a larger amount of residual amphibole can be accommodated.