Search for: All records

This study evaluates changes in copper (Cu) speciation that occur in sulfate-dominated basaltic and andesitic magmas equilibrated at oxygen fugacities (fO2’s) above the nickel-nickel oxide (NNO) buffer. Cu K-edge microfocused X-ray absorption fine structure spectroscopy (XAFS) data are presented from both natural and synthetic silicate glasses. Natural samples analyzed include olivine-hosted melt inclusions from tephra of mafic cinder cones in the Lassen segment of the Cascade arc (USA) and from the Michoacan-Guanajuato volcanic field (Mexico) as representative samples from melts equilibrated at fO2 > NNO. A comparison with melts equilibrated at fO2 < NNO is provided by analysis of olivine-hosted melt inclusions from Kīlauea Volcano. Data are also presented from copper- and sulfur-bearing synthetic hydrous andesitic glasses synthesized over a range of fO2, from roughly NNO-2 to NNO+2. The Cu spectroscopy data from the natural and synthetic glasses show two dominant Cu species, Cu1+ oxides (referred to here as Cu-O) and Cu1+ sulfides (referred to here broadly as Cu-S, but not precluding Cu-Fe-S species). The relative proportion of each species present correlates with the relative concentration of dissolved sulfide in the melt. Synthetic sulfur-bearing glasses equilibrated at NNO-1.2 were found to contain exclusively Cu-S species. Sulfur-bearing experimental glasses equilibrated at NNO-0.5 give calculated Cu-O/(Cu-O + Cu-S), defined here as the “Cu-O fraction”, of < 0.10, whereas sulfur-bearing glasses synthesized at NNO+0.6 and NNO+1.8 give calculated Cu-O fraction > 0.96. Natural melt inclusions from Lassen and Kīlauea show a bimodal distribution in Cu-O fraction, with overlapping ranges, of 0.14-0.77 for Lassen and 0.18- 0.78 for Kīlauea. Michoacan-Guanajuato inclusions yield Cu-O fractions of 0.68-0.91. The difference in the calculated proportions of Cu-O to Cu-S species appear correlated with available sulfide in the melt. As relative S2- concentrations decrease, the dissolved Cu species in the melt evolves from dominantly Cu-S to Cu-O. This includes melts equilibrated at fO2’s where S6+ is the dominant S species. At intermediate sulfide abundances both species appear to coexist. Thermodynamic modeling of the Cu speciation in these silicate melts suggests that speciation of Cu as a CuFeS2 melt species (akin to chalcopyrite or intermediate solid solution) most accurately predicts the measured Cu species. The modeling suggests that aFeO in the silicate melt, fO2 and melt S2- (expressed as fS2) are the most important parameters controlling the proportions of Cu-O vs. Cu-S species. Our results provide a new perspective for understanding Cu solubility, transport, and partitioning in magmatic systems. 

Abstract Oxygen fugacity is an important but difficult parameter to constrain for primitive arc magmas. In this study, the partitioning behavior of Fe3+/Fe2+ between amphibole and glass synthesized in piston-cylinder and cold-seal apparatus experiments is developed as an oxybarometer, applicable to magmas ranging from basaltic to dacitic composition. The partitioning of Fe2+ is strongly dependent on melt polymerization; the relative compatibility of Fe2+ in amphibole decreases with increasing polymerization. The Fe2+/Mg distribution coefficient between amphibole and melt is a relatively constant value across all compositions and is, on average, 0.27. The amphibole oxybarometer is applied to amphibole in mafic enclaves, cumulates, and basaltic tephra erupted from Shiveluch volcano in Kamchatka with measured Fe3+/FeTotal. An average Fe3+/Fe2+ amphibole-glass distribution coefficient for basalt is used to convert the Fe3+/FeTotal of amphibole in samples from Shiveluch to magmatic oxygen fugacity relative to NNO. The fO2 of primitive melts at the volcano is approximately NNO+2 and is faithfully recorded in amphibole from an amphibole-rich cumulate and the basaltic tephra. Apparently, higher fO2 recorded by amphibole in mafic enclaves likely results from partial dehydrogenation of amphibole during residence in a shallow andesite storage region. We identify three pulses of mafic magma recharge within two weeks of, a month before, and two to three months before the eruption and find that, at each of these times, the host andesite was recharged by at least two magmas at varying stages of differentiation. Application of the amphibole oxybarometer not only gives insight into magmatic fO2 but also potentially details of shallow magmatic processes. 

Abstract Anisotropic absorption in crystals is routinely observed in many spectroscopic methods and is recognized in visible light optics as pleochroism in crystalline materials. As with other spectrosco-pies, anisotropy in Fe K-edge X-ray absorption spectroscopy (XAS) can serve both as an indicator of the general structural arrangement and as a conundrum in quantifying the proportions of absorbers in crystals. In materials containing multiple absorbers, observed anisotropies can typically be represented by a linear relationship between measured spectroscopic peak intensities and relative absorber concentrations. In this study, oriented XAS analysis of pyroxenes demonstrates that the macroscopic theory that describes visible light absorption anisotropy of triaxially anisotropic materials can also be applied to X-ray absorption in pyroxenes, as long as the orientation and magnitude of the characteristic absorption vectors are known for each energy. Oriented single-crystal XAS analysis of pyroxenes also shows that the measured magnitude of characteristic absorption axes at a given orientation is energy-dependent and cannot be reproduced by linear combination of intermediate spectra. Although the macroscopic model describes a majority of the anisotropy, there is distinct discordance between the observed and interpolated spectra in the pre-edge between 7109 and 7115 eV, which is marked by spikes in RMSE/mean intensity ratio. Absorption indicatrices for samples analyzed in the visible and at X-ray wavelengths are modeled with a three-dimensional (3D) pedal surface, which functions as an empirical way of interpolating between the observed absorption data. This surface only requires a maximum of three coefficients, and results from the summation of 3D lemniscates. An absorption indicatrix model can be used to characterize anisotropic absorption in crystals and provides a way of comparing XAS spectra from randomly oriented crystals, such as those from polished sections, to a database of characterized crystals.