The fate of subducted oceanic crust: a mineral segregation model

Abstract

Plate subduction and mantle plumes are two of the most important material transport processes of the silicate Earth. Currently, a debate exists over whether the subducted oceanic crust is recycled back to the Earth's surface through mantle plumes, and can explain their derivation and major characteristics. It is also puzzling as to why plume heads have huge melting capacities and differ dramatically from plume tails both in size and chemical composition. We present data showing that both ocean island basalt and mid-ocean ridge basalt have identical supra-primitive mantle mean Nb/U values of ∼46.7, significantly larger than that of the primitive mantle value. From a mass balance calculation based on Nb/U we have determined that nearly the whole mantle has evolved by plate subduction-induced crustal recycling during formation of the continental crust. This mixing back of subducted oceanic crust, however, is not straightforward, because it generally would be denser than the surrounding mantle, both in solid and liquid states. A mineral segregation model is proposed here to reconcile different lines of observation. First of all, subducted oceanic crustal sections are denser than the surrounding mantle, such that they can stay in the lower mantle, for billions of years as implied by isotopic data. Parts of subducted oceanic crust may eventually lose a large proportion of their heavy minerals, magnesian-silicate-perovskite and calcium-silicate-perovskite, through density segregation in ultra-low-velocity zones as well as in very-low-velocity provinces at the core-mantle boundary due to low viscosity. The remaining minerals would thus become lighter than the surrounding mantle, and could rise, trapping mantle materials, and forming mantle plumes. Mineral segregation progressively increases the SiO2 content of the ascending oceanic crust, which enhances flux melting, and results in giant Si-enriched plume heads followed by dramatically abridged plume tails. Therefore, ancient mineral-segregated subducted oceanic crust is likely to be a major trigger and driving force for the formation of mantle plumes.