PARTITION COEFFICIENTS FOR ZIRCON FROM HIGH-TEMPERATURE ICELANDIC RHYOLITES, DETERMINED BY IN SITU ANALYSIS OF GLASS AND CRYSTAL RIMS

Accurate trace element partition coefficients (Kds) for minerals are essential for modeling magma genesis and evolution and for using zoned crystals to interpret compositional fluctuations in melt. However, determining reliable Kds is challenging, given the inherent difficulty of determining true concentrations of crystals and melts in equilibrium; sluggish equilibration hampers experimental analyses, while natural materials typically yield heterogeneous crystals and most lack analyzable melt. Recent efforts involving in situ analysis of crystal rims and coexisting glass have yielded valuable results (Bachmann et al. 2005; Colombini et al. 2011; Padilla and Gualda 2014), but these studies span a relatively narrow range of conditions.

This study focuses on zircon rim/glass partitioning in high-temperature Icelandic rhyolites. We performed elemental analysis of glass and zircon rims from six samples from the Pleistocene to modern volcanic centers Krafla, Kerlingarfjöll, Torfajökull (2 samples), Hekla, and Öraefajökull. Compositions of glasses were determined by SEM-EDS (major elements) and LA-ICP-MS (trace elements) at Vanderbilt. Trace element compositions of zircon rims were determined by SHRIMP-RG (Stanford-USGS).

Rare earth element Kds in all samples are very high for HREE (Lu 30-140) and very low for LREE (La 0.0002-0.002); Kd vs. ionic radii patterns are identical in shape to those presented in references cited above, but Kds are lower by a factor of ~2-5. Likewise, Kds for U (6-50), Th (1.5-14), and Hf (430-800) are consistent with, but substantially lower than, those determined in previous studies employing a similar approach. The lower Iceland Kds are likely at least in part a consequence of much higher temperatures: zircon saturation temperatures estimated from glass compositions range from 880-950°C (Watson and Harrison 1983; temperatures from Boehnke et al. 2013 are slightly lower), in contrast with saturation temperatures of 740-850°C for samples used in previous studies. Our Icelandic data will be of use for future studies involving processes involving high-T silicic magmas, and they point to the need for additional future work to better characterize sensitivity of partitioning to temperature and other factors.