NI-CO SYSTEMATICS OF HIGH-NI OLIVINES IN THE CENTRAL MEXICAN VOLCANIC BELT: CONSTRAINTS FOR PYROXENITE FORMATION IN THE MANTLE WEDGE
The reaction model proposes that high-Ni olivines precipitate from Ni-rich initial melts from olivine-free, but Ni-rich pyroxenites. Recently, two alternative mechanisms for increasing melt Ni have been proposed: (i) melt Ni may increase when newly ascending, recharging magmas selectively absorb Ni-rich olivines precipitated from earlier melts, and (ii) Ni-rich initial mantle melts may reflect diffusive uptake of Ni from co-existing sulfides in source (Weber et al. 2011, ESPL). However, both mechanisms fail to satisfy the combined Ni-Co systematics of the high-Ni olivines. As Ni, Co is compatible in olivine and the mineral/melt partition coefficient KdCooliv is similarly dependent on melt composition. However, as KdCooliv < KdNioliv, olivine Co is indifferent or even slightly increases with decreasing Fo during fractional crystallization, while olivine Ni strongly decreases. Model calculations demonstrate that these trends preclude Ni enrichment of melts by recharge magma mixing. The sulfide model is rejected, because Co abundances in Fo-rich (Fo>89) high-Ni olivines are similar to MORB olivines, while olivine Ni is much higher. This argues against a role of sulfides as source for the transition elements, as Ni and Co are both enriched in co-existing mantle sulfides. Moreover, a sulfide source would induce much more random Ni enrichments than those observed which remain consistently within the limits predicted by the reaction model. In summary, Ni-Co systematics further support the reaction model with all consequences for the petrogenesis of arc volcanic rocks, arc recycling and the growth of arc and continental crust.