ALBERT

Description: Highlights • Parental melts of sulfide-bearing KM rocks have near primary MORB-like composition. • Crystallization of these S-saturated melts occurred in near-surface conditions. • Extensive fractionation and crustal assimilation are not the causes of S-saturation. • S content in melts can be restored by accounting for daughter sulfide globules. Abstract Sulfide liquids that immiscibly separate from silicate melts in different magmatic processes accumulate chalcophile metals and may represent important sources of the metals in Earth's crust for the formation of ore deposits. Sulfide phases commonly found in some primitive mid-ocean ridge basalts (MORB) may support the occurrence of sulfide immiscibility in the crust without requiring magma contamination and/or extensive fractionation. However, the records of incipient sulfide melts in equilibrium with primitive high-Mg olivine and Cr-spinel are scarce. Sulfide globules in olivine phenocrysts in picritic rocks of MORB-affinity at Kamchatsky Mys (Eastern Kamchatka, Russia) represent a well-documented example of natural immiscibility in primitive oceanic magmas. Our study examines the conditions of silicate-sulfide immiscibility in these magmas by reporting high precision data on the compositions of Cr-spinel and silicate melt inclusions, hosted in Mg-rich olivine (86.9–90 mol% Fo), which also contain globules of magmatic sulfide melt. Major and trace element contents of reconstructed parental silicate melts, redox conditions (ΔQFM = +0.1 ± 0.16 (1σ) log. units) and crystallization temperature (1200–1285 °C), as well as mantle potential temperatures (~1350 °C), correspond to typical MORB values. We show that nearly 50% of sulfur could be captured in daughter sulfide globules even in reheated melt inclusions, which could lead to a significant underestimation of sulfur content in reconstructed silicate melts. The saturation of these melts in sulfur appears to be unrelated to the effects of melt crystallization and crustal assimilation, so we discuss the reasons for the S variations in reconstructed melts and the influence of pressure and other parameters on the SCSS (Sulfur Content at Sulfide Saturation).

Description: Assessing the N content of arc magmas and their mantle source remains a challenge because the volatile element composition of melts and gases can be modified during magma ascent, storage, and eruption. Given that melt inclusions (MIs) in Mg-rich olivine represent the best proxies for primary arc melts, we applied, for the first time, an in situ high-resolution secondary ion mass spectrometry (SIMS) method to determine the N concentration in olivine-hosted MIs from Klyuchevskoy volcano in Kamchatka. To reverse the effects of post-entrapment modification processes (i.e., exsolution of volatiles into a fluid bubble), the MIs were partially to completely homogenized at high temperatures (1150–1400 °C) and pressures ranging from 0.1 to 500 MPa under dry to H2O-saturated conditions at variable oxygen fugacities (CCO to QFM + 3.3). After the experiments, N concentrations in water-rich MI glasses correlate positively with H2O and CO2 contents as well as with N/CO2 ratios, and negatively with the volume of the remaining fluid bubble. Glasses of completely homogenized (fluid bubble-free) MIs contain up to 25.7 ± 0.5 ppm N, whereas glasses of three unheated (natural, bubble-bearing) MIs have significantly lower N concentrations of ~1 ± 0.3 ppm. The N-CO2-Nb characteristics of completely homogenized MIs indicate that melts feeding Klyuchevskoy volcano have high absolute concentrations of both N and CO2, as well as large excess of these volatiles relative to Nb, compared to primary mid-ocean ridge melts. This implies that large amounts of N and CO2 in Klyuchevskoy melts and their mantle source are derived from the subducting slab, and that these subducted volatiles are (partially) returned to the crust and atmosphere by arc-related magmatism.

Description: Here we present a confocal Fe K-edge μ-XANES method (where XANES stands for X-ray absorption near-edge spectroscopy) for the analysis of Fe oxidation state in heterogeneous and one-side-polished samples. The new technique allows for an analysis of small volumes with high spatial 3D resolution of 〈100 µm3. The probed volume is restricted to that just beneath the surface of the exposed object. This protocol avoids contamination of the signal by the host material and minimizes self-absorption effects. This technique has been tested on a set of experimental glasses with a wide range of Fe3+  ΣFe ratios. The method was applied to the analysis of natural melt inclusions trapped in forsteritic to fayalitic olivine crystals of the Hekla volcano, Iceland. Our measurements reveal changes in Fe3+  ΣFe from 0.17 in basaltic up to 0.45 in dacitic melts, whereas the magnetite–ilmenite equilibrium shows redox conditions with Fe3+  ΣFe ≤0.20 (close to FMQ, fayalite–magnetite–quartz redox equilibrium) along the entire range of Hekla melt compositions. This discrepancy indicates that the oxidized nature of glasses in the melt inclusions could be related to the post-entrapment process of diffusive hydrogen loss from inclusions and associated oxidation of Fe in the melt. The Fe3+  ΣFe ratio in silicic melts is particularly susceptible to this process due to their low FeO content, and it should be critically evaluated before petrological interpretation.