HR: 1340h
AN: V13B-1473    [Abstracts]
TI: Experimental Phase Equilibria and Trace Element Partitioning in Differentiated Hydrous Calc-Alkaline Melts at High Pressures
AU: * Alonso-Perez, R
EM: raquel.alonso@erdw.ethz.ch
AF: Department of Earth Sciences, ETH-Zentrum, Zurich, CH-8092 Switzerland
AU: Ulmer, P
EM: peter.ulmer@erdw.ethz.ch
AF: Department of Earth Sciences, ETH-Zentrum, Zurich, CH-8092 Switzerland
AU: Muntener, O
EM: othmar.muntener@geo.unibe.ch
AF: Institute of Geological Sciences, University of Bern Baltzerstr. 1, Bern, CH-3012 Switzerland
AU: Thompson, A B
EM: alan.thompson@erdw.ethz.ch
AF: Department of Earth Sciences, ETH-Zentrum, Zurich, CH-8092 Switzerland
AB: Garnet and amphibole fractionation from high-pressure, mantle-derived hydrous basaltic liquids exert an important control on the genesis of lower arc crust. Here we present experimental phase equilibria and partitioning of trace elements between garnet, amphibole and differentiated hydrous calc-alkaline melts at elevated pressures to evaluate the potential role of these minerals in generating specific trace element characteristics that have been attributed to their presence either during partial melting or during fractional crystallization. We carried out piston cylinder experiments at conditions of P=0.8-1.2 GPa and T=800-1000$\deg$C, using a synthetic H$_{2}$O under-saturated andesite composition that was derived by fractional crystallization experiments from a primary hydrous picrobasalt. Once the phase equilibria were established, the andesite starting material was doped with 20 trace elements. Run products, analysed by electron microprobe and ion microprobe, include amphibole, plagioclase, Fe-Ti-oxides, garnet, clinopyroxene and quenched liquid. Amphibole could successfully be analyzed (grain-size $>$10$\mu$m), whereas the grain-size of plagioclase and Fe-Ti-oxides was always too small for SIMS analysis ($\leq$5 $\mu$m). Large (up to 100 $\mu$m in diameter) and homogeneous garnets were produced, but clinopyroxene grain size was always smaller than 10 $\mu$m, just at the limit of successful analysis. The solid/liquid partition coefficients for garnet exhibit very steep variations, with LILE being highly incompatible (D$_{Ba}$=0.003) and heavy REE having very high D-values (D$_{Lu}$=163.8). The quantitative model proposed by van Westrenen et al., (2000, GCA 64) to calculate partition coefficient for garnets predicts much high values then we observe, indicating that models based on partition coefficients determined in anhydrous MORBs cannot adequately been used to predict garnet partition coefficients in hydrous, siliceous melts of calc-alkaline affinity. Amphiboles show basically constant HREE partition coefficients and higher but variable D's for HFSE (Nb=0.98, Ta=0.61) than published data for amphiboles in equilibrium with andesitic liquid compositions. Similarly to Tiepolo et al. (2000, EPSL 176), we obtain D$_{Nb}$ $>$ D$_{Ta}$. Our results indicate that amphibole fractionation from andesitic magma at elevated pressure producing amphibole-rich cumulates at the base of the arc crust is an additional scenario to explain sub-chondritic Nb/Ta ratio in differentiated island-arc magmas. The peraluminous to metaluminous restitic liquids are silica-rich andesites to dacites. High-pressure melts in equilibrium with garnet and cpx exhibit very low Y and HREE, very high Sr contents and lack Eu anomalies. High Sr/Y ratios of intermediate to acidic island-arc magmas are often regarded as indicative for so called adakites. The present data show that 10% fractionation of garnet and cpx at the base of a growing island arc results in a change from normal Sr/Y ratio of 20 for a basaltic andesite to values exceeding 200. Therefore, high Sr/Y ratios cannot be used as an unequivocal characteristic identifying "slab-melts"; identical or similar geochemical patterns can be produced through high-pressure fractionation of mantle-derived "normal" calc-alkaline basaltic primary magmas at the base of a growing island-arc.
DE: 3620 Crystal chemistry
DE: 3630 Experimental mineralogy and petrology
DE: 3640 Igneous petrology
DE: 1065 Trace elements (3670)
SC: Volcanology, Geochemistry, Petrology [V]
MN: 2004 AGU Fall Meeting