北极Gakkel超慢速扩张脊大洋橄榄岩的地球化学特征及其意义

Gakkel Ridge in Arctic Ocean is the ulstraslow spreading ridge in the world with a full spreading rate decreasing from 14 mm/yr in the western end to 7mm/yr in the eastern end. To study the histories of partial melting and melt referilization occurred in the oceanic mantle beneath Gakkel Ridge, both extremely fresh and altered abyssal peridotites from two dredge hauls (PS66-238 and HLY0102-D70) have been selected in this research. Major and trace element data of the residual minerals suggest that all samples have been refertilized by late enriched melts after low to moderate degrees (3-12%) of partial melting in the stability field of spinel, whereas some samples also inherited signatures of partial melting in stability field of garnet. Os isotopic compositions of Gakkel samples have not been significantly affected by late processes, e.g., seawater alteration and melt refertilzaiton. Samples from both dredge hauls have similar range of 187Os/188Os, from strongly unradiogenic (~0.114) in the harzburgites to approximating the inferred values of PUM (primitive upper mantle) in some lherzolites (~0.129). Inherited ancient depletion events in the harzburgites with Re-depletion age up to 2 billion years are unrelevant to the recent genesis of MORB (mid-ocean ridge basalts) beneath Gakkel Ridge. Comparisons of highly siderophile elements (HSEs) between the fresh and altered samples suggested both Pd and Re were affected and thus are mobile during seawater alteration, whereas the other HSEs (i.e., Os, Ir, Ru an Pt) are stable. The fractionated HSEs patterns in the harzburgites suggest both PPGEs (Pt and Pd) and Re can be fractionated from IPGEs (Os, Ir and Ru) at low degree of partial melting, which might be due to physical dredging of sulfide melts by silicate melts rather than equilibrium partitioning between residues and silicate melts. Inferred HSEs budget of the PUM confirm the previous study that both Ru/Ir and Pd/Ir are suprachondritic in the PUM. Some modifications of late-veneer hypothesis are required in light of the unique PUM composition. HSEs and Os isotopic compositions of Gakkel abyssal peridotites indicate the oceanic mantle is highly heterogeneous within a scale of one dredge haul (<5 km). Both depleted and fertile mantle domains are likely to be mechanically juxtaposed in the asthenosphere in a state of ‘plum pudding’. Widely distribution of ancient depleted components in the asthenosphere suggests that DMM (depleted MORB mantle) should not be synonymous with the MORB source. The later is just the fertile part of the former, i.e., the depleted components in the DMM do not or contribute little to the genesis of MORB.