Master's student: Ingibjörg Andrea Bergþórsdóttir
Title: The role of apatite in Hekla magmas: trace element partitioning between minerals and melt
Faculty: Faculty of Earth Sciences
Advisors: Olgeir Sigmarsson, Research Scientist at the Institute of Earth Sciences
Members of the master committe: Eniko Bali, Associate Professor and Þorvaldur Þórðarson, Professor, both at the Faculty of Earth Sciences
Examiner: Kristján Jónasson, Head of Geology at the Icelandic Institute of Natural History
Pumice from Hekla volcano, Iceland have been studied in this thesis in order to shed light on the importance of fractional crystallization versus crustal melting in the variability of Hekla melts. Presented are major, minor and trace element analysis of glasses and minerals. The pumice range in composition from basalt to rhyolite and contain only a few crystals. Olivine displays an overall large range in Fo contents (Fo: 71-12%) but is relatively homogeneous within each sample. Plagioclase spans a compositional range from An36 to An66 displaying normal zonation. Clinopyroxene range from Mg# 30 to 76 and fluor-apatite is commonly observed as an accessory phase in the basaltic andesite, dacite and rhyolite samples. Other accessory phases include Fe-Ti oxides in all samples and sulphide in rhyolite. Sc and Y have partition coefficients (D) higher than unity in clinopyroxene as well as the REE in more evolved samples. Strontium has D higher than unity in plagioclase and Eu in more silicic samples. The first in-situ analysis of trace elements in fluor-apatites from Hekla and Iceland reveal its great capacity to fractionate trace elements from the melt. Strontium, Y, REE, Th and U have D higher than unity and are thus all compatible in apatite. Apatite has DU/DTh close to unity and does not fractionate U from Th. Therefore, apatite crystallization cannot explain lower U/Th in the silicic magma from Hekla. Partial melting of hydrothermally altered crust is therefore the preferred mechanism producing silicic melt under Hekla. The H2O concentration of the H1158 dacite is estimated by the means of geothermometry to be 4-6 wt%. If the hydrothermally altered crust partially melted 10-20% to produce the dacite, then the crust does not need to contain more than 0.4-1.2 wt% H2O. The low water content of the altered protolith readily explains the elevated δ18O values in Hekla magma compared to silicic formations of the rift-zones.
Ingibjörg Andrea Bergþórsdóttir