Magnesium isotope behavior in oceanic magmatic systems: Constraints from mid-ocean ridge lavas from the East Pacific Rise

Abstract

The magnesium (Mg) isotope composition of oceanic basalts has provided useful constraints on the evolution of the upper mantle and crust-mantle interactions. However, the behavior of Mg isotopes during oceanic magma differentiation is still unclear because of the small range in Mg isotope values in typical normal mid-ocean ridge basalts (N-MORB). Here, we present high-precision Mg isotope data on a well-characterized suite of mid-ocean ridge (MOR) lavas from the 9–10°N segment of the East Pacific Rise. These samples range from relatively primitive basalt to evolved dacite with MgO contents decreasing from 8.62 to 0.80 wt.%, and display a resolvable variation in δ26Mg from -0.27 to -0.17‰. The less-evolved samples (MgO > 7.0 wt.%) that have experienced olivine and plagioclase fractional crystallization have δ26Mg values ranging from -0.23 to -0.17‰ that are negatively correlated with MgO content. Samples containing MgO of 3.5–7.0 wt.%, that have experienced significant crystallization of clinopyroxene, together with plagioclase, show a limited variation of δ26Mg (-0.20 to -0.18‰). For those highly evolved samples (MgO < 3.5 wt.%) saturated with Fe–Ti oxides, the δ26Mg vary from -0.27 to -0.20‰ and are positively correlated with MgO content. The variation of δ26Mg in the MOR lavas is consistent with three stages of magma differentiation, beginning with fractional crystallization of olivine and plagioclase followed by increasing amounts of clinopyroxene and finally joined by Fe-Ti oxides. Quantitative modeling of the Mg isotopic variation in EPR samples, shows that the variation of δ26Mg in MOR lavas is primarily a consequence of fractional crystallization of the Mg-bearing minerals when Δ26MgOl-melt ∼ -0.10‰, Δ26MgCpx-melt ∼ 0.00‰, and Δ26MgTi−Mgt-melt ∼ 0.20‰. This study provides evidence that shallow level crystal fractionation can produce significant and predicable variations in Mg isotopic compositions of MOR lavas and that this process should be carefully considered when using evolved lavas to trace mantle source compositions.