Mg isotope systematics during magmatic processes: Inter-mineral fractionation in mafic to ultramafic Hawaiian xenoliths

Stracke A, Tipper ET, Klemme S, Bizimis M

Research article (journal)

Abstract

Observed differences in Mg isotope ratios between bulk magmatic rocks are small, often on a sub per mill level. Inter–mineraldifferences in the 26Mg/24Mg ratio (expressed as d26Mg) in plutonic rocks are on a similar scale, and have mostly beenattributed to equilibrium isotope fractionation at magmatic temperatures. Here we report Mg isotope data on minerals inspinel peridotite and garnet pyroxenite xenoliths from the rejuvenated stage of volcanism on Oahu and Kauai, Hawaii.The new data are compared to literature data and to theoretical predictions to investigate the processes responsible forinter–mineral Mg isotope fractionation at magmatic temperatures. Theory predicts up to per mill level differences ind26Mg between olivine and spinel at magmatic temperatures and a general decrease in D26Mgolivine-spinel (=d26Mgolivine –d26Mgspinel) with increasing temperature, but also with increasing Cr# in spinel. For peridotites with a simple petrogenetichistory by melt depletion, where increasing depletion relates to increasing melting temperatures, D26Mgolivine-spinel should thussystematically decrease with increasing Cr# in spinel. However, most natural peridotites, including the Hawaiian spinel peridotitesinvestigated in this study, are overprinted by variable extents of melt-rock reaction, which disturb the systematic primarytemperature and compositionally related olivine–spinel Mg isotope systematics. Diffusion, subsolidus re-equilibration,or surface alteration may further affect the observed olivine–spinel Mg isotope fractionation in peridotites, making D26Mgolivine-spinel in peridotites a difficult–to–apply geothermometer. The available Mg isotope data on clinopyroxene and garnet suggestthat this mineral pair is a more promising geothermometer, but its application is restricted to garnet–bearing igneous(garnet pyroxenites) and metamorphic rocks (eclogites). Although the observed d26Mg variation is on a sub per mill rangein bulk magmatic rocks, the clearly resolvable inter–mineral Mg isotope differences imply that crystallization or preferentialmelting of isotopically distinct minerals such garnet, spinel, and clinopyroxene should cause Mg isotope fractionation betweenbulk melt and residue. Calculated Mg isotope variations during partial mantle melting indeed predict differences between meltand residue, but these are analytically resolvable only for melting of mafic lithologies, that is, garnet pyroxenites. Contributionsfrom garnet pyroxenite melts may thus account for some of the isotopically light d26Mg observed in ocean island basalts and trace lithological mantle heterogeneity. Consequently, applications for high-temperature Mg isotope fractionations arepromising and diverse, and recent advances in analytical precision may allow the full petrogenetic potential inherent in thesub per mill variations in d26Mg in magmatic rocks to be exploited.

Details zur Publikation

Release year: 2018
Language in which the publication is written: English