Trace Element Contents of Micas: Defining Signatures of Partial Melting Processes in the Crust

Trace element contents of micas in the protoliths of crustal melts are thought to be low, but are poorly understood. In view of their significant role in generating partial melts in the crust, a better understanding of their trace element chemistry, especially REE abundance patterns, is required. Important dehydration-melting reactions of muscovite and biotite consume mainly mica, quartz and feldspar. Therefore partial melt compositions should reflect the compositions of these phases. An additional complexity is the role of accessory phases because of their potential control on REE behavior. As a complement to natural geochemical studies, we have investigated experimentally the partial melting behavior of two typical crustal rocks, under both static and dynamic (deformation) conditions. We are characterizing the trace element signatures of partial melt in experiments in which crack networks appear and contrasting the signatures with compositions in a dynamic experiments where melt-enhanced deformation occurs. Both static and deformation experiments were conducted on a two-mica pelite and a biotite gneiss at pressures of 0.7-1.0 GPa, and between 725° C and 950° C. Deformation was performed at 10-5/s strain rate. In both sets of experiments, we observe textures associated with the development of permeability, either by cracking associated with reaction or by melt-enhanced deformation. We use the development of these textures to establish melt segregation mechanisms. Trace element data for mica and the glass in cracks and in highly permeable glass-bearing zones have been determined in situ using SIMS and LA-ICPMS. Starting micas in both rock types and the have also been analyzed in situ for REE. Both muscovite and biotite have low abundance REE (<10xCI) with gently sloping pattern from >1xCI for the LREE to <0.1xCI for the HREE. There also is a slight positive Eu anomaly. Compared to published REE data for micas, these abundances are lower. However, previous data were obtained using mica separates and may be compromised by inclusions of REE-rich accessory phases. In situ analyses provides data that are unambiguously mica alone, without inclusions. Complementary LA-ICPMS and SIMS data from low volume, muscovite-derived glass in melt cracks show a flatter REE pattern compared to the muscovite, but have retained the slightly positive Eu anomaly. In longer duration experiments, flat patterns remain but the Eu anomaly becomes negative. The data suggest that mica alone does not produce the melt signature and accessory phases, in addition to minor feldspar, play a significant role in controlling the geochemical signatures of the melts under these experimental conditions.