The origin of geochemical trends and Eoarchean (ca. 3700 Ma) zircons inMesoarchean (ca. 3075 Ma) ocelli-hosting pillow basalts, Ivisaartoq greenstone belt,SW Greenland: Evidence for crustal contamination versus crustal recycling

摘要

The Mesoarchean Ivisaartoq greenstone belt consists of tholeiitic pillow basalts, picrites, boninites, gabbroicto dioritic dykes and sills, actinolite schists and serpentinites. In addition, the belt includes volumetricallyminor siliceous volcaniclastic sedimentary rocks, cherts, biotite schists, and quartzitic gneisses. Siliceousvolcaniclastic rocks yield an average U-Pb zircon age of 3075 ± 15 Ma, and pillow basalts and gabbros yield aSm-Nd errorchron age of 3080 ± 160 Ma, constraining the maximum age of the belt. The Ivisaartoq rocksunderwent high-temperature seafloor hydrothermal alteration, resulting in extensive epidote-diopside calc-silicate metasomatism. Th-REE-Nb systematics of the least metasomatized volcanic rocks are consistentwith a supra-subduction zone geodynamic setting, displaying a trend projecting from mid-ocean ridge basalt(MORB) field to island arc basalt (IAB) field. This trend is interpreted as reflecting a transition from thedepleted upper mantle to a subarc mantle wedge following the initiation of intra-oceanic subduction and arcmigration. Many pillow basalts in the Ivisaartoq greenstone belt contain numerous millimetre- to centimetre-longwhite polycrystalline ellipsoidal inclusions, called 'ocelli', set in a dark green mafic matrix (basaltic host). Onthe basis of new major and trace element, Nd isotope, and petrographic data, we re-interpret the origin ofthese inclusions and suggest that they were derived from the melting of hydrated mafic to ultramafic rocksin the lower oceanic crust during magma-crust interaction. Some pillow basalts contain rare small inherited magmatic zircons revealing Eoarchean (3706 ± 8 to 3717 ±4 Ma) ~(207)Pb~(206)Pbages. In the least metasomatized pillow basalts, and gabbros diorites there is a strongnegative correlation between initial Nd isotopic composition (‰εNd= 0.76 to + 3.10) and depleted mantlemodel ages (T_(DM)=3100-3800 3100-3800 Ma). Post-magmatic metasomatic alteration can be ruled out as the cause ofthis negative correlation, suggesting the contamination of the Ivisaartoq rocks, or their mantle source, byEoarchean (~3700 Ma) continental crust or sediments derived from such crust. Four possible processes ofcrustal contamination might have taken place during or prior to the emplacement of the Ivisaartoq belt: (1)interaction with continental crust during magma ascent; (2) mixing between the depleted upper mantle andrecycled sediments derived from Eoarchean continental crust; (3) mixing between the depleted uppermantle and delaminated lower continental crust; and (4) assimilation of continentally-derived sedimentsduring eruption. It is unlikely that recycled Eoarchean zircons and their host rocks could have survivedthrough multiple events of partial melting in the Mesoarchean upper mantle. A tectonic slice ofmetasedimentary rocks exposed between the Ivisaartoq belt and the Eoarchean (3600-3870 Ma) Isukasiaterrane contains magmatic zircon grains up to 3800 Ma old. Accordingly, we argue that Eoarchean zirconsare likely to have been derived from the protoliths of these metasedimentary rocks. The negative correlationbetween initial εNd and TOM in pillow basalts, gabbros and diorites, however, can be better explained bycontamination of the mantle source through hydrous melts and/or fluids, derived from Mesoarcheansubducted (recycled) sediments, than by contamination at the surface.