Experimental investigation of properties of low degree partial melts of garnet peridotite and their role in OIB genesis
Abstract
Trace element concentrations in alkalic OIB are consistent with small-degree (<10%) partial melts of garnet lherzolite. However, high melt fraction liquids (>10%) from peridotite partial melting experiments are unlike OIB in several respects including low incompatible element abundances (e.g., TiO2), high Al2O3, and low FeO*. This has led to suggestions that many OIB require contributions from other lithologies such as pyroxenite or metasomatized peridotite enriched in volatiles and/or K2O. Further evidence for pyroxenite domains in OIB sources comes from abundances and ratios of first row transition elements (FRTE), e.g., Ni in OIB olivines (Sobolev et al., 2005), and Zn/Fe ratios in OIBs (Le Roux et al., 2010). These studies rely heavily on partitioning data from 1-atm experiments and from natural mineral/melt pairs, and would benefit from high-precision partition coefficients determined for the phase compositions and conditions appropriate to the genesis of OIB. A series of modified iterative sandwich experiments (MISE) determined the composition of a low degree (F=0%) partial melt of garnet lherzolite with composition similar to KLB-1 at 3 GPa. To test the hypothesis that this melt could be parental to OIB, we corrected alkalic OIB compositions (GEOROC) for olivine fractionation by adding or subtracting olivine with KDFe-Mg=0.32 in 0.001% increments until the melt reached the Mg# of our experimental melt (Mg#=89.8). TiO2 (2.5%) in the F=0% melt is greater than was predicted from earlier peridotite partial melting experiments (<2%) but is low compared to many OIB (3-5%), requiring that such basalts come from TiO2-enriched sources. Al2O3 content of melt in equilibrium with garnet (12.7 wt.%) is lower than in experimental melts where garnet has been exhausted from the residue (e.g., Walter, 1998), but is still greater than most OIB, and FeO* (9.7%) in the F=0% melt is lower than virtually all OIB. Together, these results suggest a complex petrogenesis for OIB, involving a combination of exotic metasomatic or pyroxenitic sources. We use this 3 GPa near-solidus composition as a starting point for further experiments to examine metasomatic components, volatiles, and trace elements in peridotite partial melts. Further MISE experiments explore the effects of K2O on peridotite partial melts. Because it is strongly incompatible (Dbulk<0.005), modest variations in source K2O result in large changes of K2O in small-degree partial melts. Preliminary results show that increased K2O in partial melts of garnet lherzolite drives liquid compositions to higher SiO2 and toward the compositions of basalts derived from EM sources. Also, we have conducted crystallization experiments with the F=0% melt doped with FRTE to better constrain partitioning behavior of these elements in putative OIB source regions. These formed large (>100 µm) crystals of olivine, two pyroxenes, and garnet from the near-solidus melt. Partition coefficients analyzed by LA-ICP-MS will be available soon.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2010
- Bibcode:
- 2010AGUFM.V11B2278D
- Keywords:
-
- 3615 MINERALOGY AND PETROLOGY / Intra-plate processes;
- 3619 MINERALOGY AND PETROLOGY / Magma genesis and partial melting;
- 3621 MINERALOGY AND PETROLOGY / Mantle processes;
- 3630 MINERALOGY AND PETROLOGY / Experimental mineralogy and petrology