Recycled Crust in the Mantle: Is High-Ni Olivine the Smoking Gun or a Red Herring?

It is widely accepted that small amounts of recycled crustal components are present in some mantle-derived mafic and ultramafic magmas. This concept is supported by many isotopic and trace element studies of basalts, picrites and komatiites in the last 30 years. Recently Sobolev et al. [1,2] used olivine compositions such as Ni content and Mn/Fe ratio to demonstrate that the amounts of the recycled crustal component (i.e. pyroxenite) in these mantle-derived melts are much larger than previously appreciated. Their calculations show that the pyroxenite-derived component varies mostly between 40 and 80% for Hawaiian shield basalts and Siberian flood basalts, and mostly between 10 and 40% for mid-ocean ridge basalts and Archean komatiities. However, a critical test using olivine-liquid Mg-Fe equilibrium that was overlooked by Sobolev et al. [1,2] reveals that mixing of the two end-members (pyroxenite-derived and peridotite-derived melts) that were used in their models cannot generate the parental melts for the above natural samples. Such a discrepancy prompts us to reexamine the conventional view of a peridotite-dominant source for the Hawaiian shield basalts. This hypothesis has been criticized recently by many people because the contents of Ni in olivine phenocrysts in the basalts are significantly higher than mantle olivines in associated peridotite xenoliths and because total pressure has little effect on olivine-liquid Ni partition coefficient (DNi). What has not been generally considered is that the depth of olivine crystallization/equilibration has a negative effect on olivine Ni content because DNi is negatively correlated with melt temperature which decreases during adiabatic ascent. To evaluate such an effect quantitatively we have used all available experimental results of Ni partitioning between olivine and liquid to construct a robust empirical equation for DNi based on melt composition and temperature. The results of our calculations indicate that the contrasting Ni contents between mantle olivines and olivine phenocrysts in the Hawaiian shield basalts can be explained by variation in their crystallization/equilibration temperatures at different depths. [1] Sobolev et al. (2005) Nature 434, 590-597. [2] Sobolev et al. (2007) Science 316, 412-417.