HR: 0800h
AN: V31C-0607    [Abstracts]
TI: Lherzolite-saturated melt compositions of peridotite MM3: results of near-solidus micro- sandwich experiments
AU: * Baker, M B
EM: mikeb@gps.caltech.edu
AF: Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, United States
AU: Stolper, E M
EM: ems@gps.caltech.edu
AF: Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, United States
AB: Results of the 1 GPa diamond-aggregate melting experiments of Baker and Stolper (1) and Baker et al. (2) on the peridotite composition MM3 continue to be controversial (3,4), especially regarding the observation (1,2) of high SiO₂ and low TiO₂ contents of partial melts produced near the solidus. Here we present the results of four 1 GPa micro-sandwich experiments on MM3 at near-solidus temperatures of 1245-1250°C. As in our earlier experiments (1,2), the MM3 mix consists of natural olivine, orthopyroxene, clinopyroxene, and spinel grains that are less than 15 microns in size. Four different synthetic glass compositions were used and all lie off the compositional trends defined by the MM3 partial melts in oxide vs. MgO space (1,2,5). Experimental charges were constructed by placing a small chip of synthetic glass (0.4-0.8 mg) within a much larger mass (13-14 mg) of powdered MM3; each combined glass-MM3 sample was run in a graphite crucible within a sealed Pt capsule for 122 to 144 hr in a piston cylinder apparatus. At P and T, the initial glass chip within a charge generates a small pool of liquid whose composition shifts as it equilibrates with the MM3 solid phase assemblage. If the near- solidus MM3 partial melting trend of (1,2,5) is correct, then experiments with different initial glass compositions will produce final liquids that are consistent with this compositional trend. Except for TiO₂ vs. MgO in one experiment, the final glasses all lie within 1-2 sigma of the oxide vs. MgO trends defined by the partial melts of MM3 (1,2,5). Note, that by minimizing the amount of basaltic glass added to MM3 in each experiment (mass fractions varied from 0.03-0.06), the glasses do not substantially perturb the bulk composition of the starting peridotite, which minimizes the need for iterative sandwich experiments (e.g., 6). We also compared the liquid compositions from both the peridotite melting experiments and traditional sandwich experiments of (4; done using a synthetic oxide mix of MM3) with our data (1,2,5). The two data sets overlap at the 1-2 sigma level when liquid compositions are plotted as a function of liquid MgO content and thus, contrary to the claims of (4), the two data sets are consistent with each other. The divergence of the two data sets when liquid compositions are plotted as a function of temperature reflects the increasing concentration of K₂O, P₂O₅, H₂O, and Cl in the partial melts of (1,2) compared to the liquids of (4), whose synthetic bulk compositions were K₂O, P₂O₅, and Cl-free. Results of our micro-sandwich experiments lend support to the partial melting trend of (1,2,5) on MM3, especially with regard to the high SiO₂ and low TiO₂ contents in liquids at low degrees of partial melting.
(1) Baker and Stolper (1994) GCA 58, 2811-2827; (2) Baker et al. (1995) Nature 375, 308-311; (3) Falloon et al. (1997) EPSL 152, 149-162; (4) Falloon et al. (1999) JPet 40, 1343-1375; (5) Hirschmann et al. GCA 62, 883-902; (6) Robinson et al. (1998) EPSL 155, 97-111
DE: 1065 Major and trace element geochemistry
DE: 3612 Reactions and phase equilibria (1012, 8412)
DE: 3619 Magma genesis and partial melting (1037)
DE: 3630 Experimental mineralogy and petrology
SC: Volcanology, Geochemistry, Petrology [V]
MN: 2007 Fall Meeting