Trace element partitioning between high-An plagioclase and basaltic to basaltic andesite melt at 1 atmosphere pressure

We determined plagioclase-melt partition coefficients for 18 elements by performing controlled cooling rate, 1-atmosphere experiments using both natural and synthetic basaltic (51 wt.% SiO(2)) and basaltic andesite (56 wt.% SiO(2)) powders in a vertical quench furnace. The experiments produced An(69) to An(67) composition plagioclase. Three starting powders were Gorda Ridge basalt, synthetic diopside (40%)-albite (28%)- anorthite (32%) mixture, and Arenal volcano (Costa Rica) basaltic andesite. The Gorda and synthetic powders were doped at both low concentrations (20-200 ppm) and high concentration (200-5000 ppm), whereas the Arenal powder was doped only at high concentrations resulting in two doped Gorda powders (low: NP, and high: SDP), two doped diopside/albite/anorthite powders (low: DAD, high: SDD) and one doped Arenal powder (high: AR99-2). Trace elements concentrations in both glass and plagioclase were measured by secondary ion mass spectrometry (SIMS) and/or by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). Results for the partitioning of trace elements between plagioclase and melt at different doping levels demonstrate both adherence to Henry's Law and good agreement between the different analytical techniques. In general, plagioclase-melt partition coefficients determined in the An(69-73) range are similar to other published values, however, some of those occurring for plagioclase compositions >An(75) are distinctly lower than those predicted by current regression formulations. We applied a two-lattice melt model to these data to account for differences in melt composition and temperature and found that there were no aberrations associated with partition coefficients. A new set of regression formulations is determined involving the newly determined dataset for plagioclase with An contents between 75 and 87. (C) 2010 Elsevier B.V. All rights reserved.