Differences between the δ18O values of immiscible Si- and Fe-rich melts in the systems Fe2SiO4–Fe3O4–KAlSi2O6–SiO2, Fe3O4–KAlSi2O6–SiO2, and Fe3O4–Fe2O3–KAlSi2O6–SiO2, with H2O, H2O+P or H2O+S have been determined in isothermal, isobaric experiments at 1100 and 1200°C and 200MPa. The Δ18O values for conjugate Fe2SiO4–Fe3O4–KAlSi2O6–SiO2+H2O and, Fe3O4–KAlSi2O6–SiO2·KAlSi2O6–SiO2+H2O melts are only 0.4–0.6‰ and do not differ significantly from those for anhydrous melts of similar composition. The Δ18O values for melts with added H2O+P or S are more variable, ranging from 0.0 to 0.8‰. Partitioning of 18O between the immiscible melts is 0.6–1‰ less than the partitioning reported for melt–mineral and mineral–mineral pairs. The partitioning of 18O in the network modifier-bearing immiscible melts is not controlled by the relative degree of polymerization in the melts or fO2. The upper limit of the range of Δ18O values (<1‰), and the variation in the δ18O values of conjugate melts that occurs with the inclusion of network modifying constituents, suggest that in some cases, oxygen isotope ratios might be useful to distinguish lithologies evolved from coexisting immiscible silicate melts, from lithologies that have evolved by crystal fractionation only.
Read full abstract