Generation of Calc-Alkaline Magmas during Crystallization at High Oxygen Fugacity: An Experimental and Petrologic Study of Tephras from Buldir Volcano, Western Aleutian Arc, Alaska, USA

Abstract

Abstract Despite agreement that calc-alkaline volcanism occurs at subduction zones and is responsible for the genesis of continental landmasses, there is no consensus on the source of the Fe-depleted signature hallmark to calc-alkaline volcanism. In this study, we utilize mafic tephras collected from Buldir Volcano to address the genesis of strongly calc-alkaline volcanic rocks (those with a low Tholeiitic Index; ≤0·7) in a segment of the western Aleutian Arc to determine if the eruptions are plausibly part of a liquid line of descent, if they are mixtures of crustal melts and parental magmas, or if they are mixtures of melts of the mantle and the subducting slab. We conducted a series of H2O-saturated phase equilibrium experiments (1175–1000°C; 100 MPa) in a rapid-quench cold-seal (MHC) apparatus on the most primitive natural lava from Buldir (9·34 wt % MgO) at oxidizing conditions near the Re–ReO2 buffer. We confirmed that all experiments equilibrated 0·3 ± 0·23 log units above the Re–ReO2 buffer (ΔQFM ∼ +2·8) using X-ray Absorption Near Edge Structure (XANES) spectroscopy. Chromite is the liquidus phase, followed by olivine, then plagioclase, then clinopyroxene, and finally hornblende. Once clinopyroxene saturates, spinel composition shifts to magnetite. We compared our experimental results to the major element geochemistry and petrology of six tephras (51·9–54·8 wt % SiO2) from Buldir collected during the 2015 field season of the GeoPRISMS shared platform. Tephras contain olivine + plagioclase + clinopyroxene + spinel ± hornblende; plagioclase comprises most of the crystalline volume, followed by either olivine or hornblende. Spinel is ubiquitous; with Cr-rich spinel inclusions in olivine and hornblende, and magnetite in the groundmass. Variations in phenocryst assemblages and compositions between samples can be attributed to differences in pre-eruptive temperatures, where hotter samples are devoid of hornblende, and contain Fo-rich olivine and plagioclase with lower An-contents, owing to the position of the mineral-in curves at fluid-saturated conditions. Experimental glasses match the depletion in FeOT observed in the tephra whole rock compositions. The continuous depletion in FeOT is attributable to saturation of spinel as a liquidus phase (initially as chromite) and continuous crystallization through the experimental series (changing to magnetite at colder temperatures). In contrast to the natural samples, the experiments show enrichment in TiO2 with decreasing MgO, suggesting that differentiation did not occur at 100 MPa on Buldir. The TiO2 depletion in volcanic rocks from Buldir can be accounted for if hornblende crystallization occurs close to the liquidus of a parental magma; a condition that is met at higher pressures and hydrous conditions. The emerging picture for Buldir Island is that (1) oxidizing conditions are required to drive the observed depletions in FeOT via crystallization of spinel, and (2) elevated H2O contents and high pressures are required to saturate hornblende close to the liquidus to reproduce the entire suite of major elements. Our study provides a mechanism to generate the calc-alkaline trends observed at Buldir without requiring mixing of slab and mantle melts. We conclude that calc-alkaline volcanic rocks with extremely low Tholeiitic Indices (0·7), like those from Buldir, cannot be generated in absence of high oxygen fugacity, even at high pressure and/or elevated water pressures.