Tracing subduction zone processes with magnesium isotopes

Abstract

Subduction and recycling of oceanic plates change the chemical composition of mantle and affect its physical properties, thereby modulating Earth’s dynamics. Stable Mg isotopes (δ26Mg) can trace this recycling process as crustal materials are highly fractionated compared to the average mantle composition. This dissertation focuses on Mg isotope fractionation during subduction-related processes and the consequent mantle heterogeneity. The dissertation first compares inconsistent δ26Mg values of San Carlos peridotitic olivines that are published by several laboratories. We analyzed mineral grains from two San Carlos peridotites with disparate lithologies and determined that all mineral phases have indistinguishable δ26Mg values to within 0.07‰. With analytical precision and accuracy being confirmed, the significance of anomalous δ26Mg values can be understood in the context of mantle heterogeneity. The next paper investigates the Mg isotopic heterogeneity in mantle pyroxenites that have formed by multi-stage interactions between peridotites and melts with diverse origins. Pyroxenites formed by reaction with melts derived from subducted oceanic crust and carbonate sediments are shown to have variable δ26Mg values (−1.51‰ to −0.10‰). In contrast, pyroxenites that are formed by reaction with silicate melts from deep mantle has δ26Mg values similar to common mantle peridotites. Therefore, subducted oceanic slab can indeed produce local Mg isotopic heterogeneities. The third paper aims to characterize the Mg isotopic compositions of subducting sediments, which account for a substantial Mg input to global subduction zones. We analyzed 77 bulk sediments that were recovered from drill cores worldwide; these cores contain diverse lithologies with a wide compositional variation. The sediments display a large variation in δ26Mg, ranging from −1.34‰ (carbonate-rich sediments) to +0.46‰ (clay-rich sediments), indicating that sediment recycling is a feasible process to alter the Mg isotopic composition of local mantle. The fourth paper assesses the Mg isotopic distribution in a typical mantle wedge overlying an active subduction zone, where fluids released from subducting slab acting as a transfer medium for elements between slab and mantle wedge. A suite of sub-arc peridotites from Avacha, Russia, are chosen to represent western Pacific mantle wedge. The isotopic compositions of these peridotites are similar to normal mantle peridotites, suggesting that fluids produced by slab dehydration at relatively shallow depth (pressure of 2-3 GPa) are too depleted in Mg to leave a measurable fingerprint on the Mg-rich mantle wedge. Therefore, large-scale dehydration of isotopically distinct minerals at higher pressures, such as serpentine, are likely to be required to produce mantle domains with atypical δ26Mg. Collectively, this dissertation confirms the existence of locally heterogeneous mantle domains caused by subducted slabs, highlights the potential of recycling sediments for introducing heterogeneous δ26Mg to the mantle, and emphasizes the importance of subduction zone thermal structure, which controls the dehydration path of a subducting slab, in generating atypical δ26Mg in mantle wedge. The results from this dissertation contribute to the overall development of using Mg isotopic system to trace crustal recycling and associated mantle heterogeneities.