Ion Probe Studies of Mantle-Melting Run Products

University Of Oregon Eugene, Eugene OR

Investigators

Abstract

Johnston EAR-0105500 This proposal requests funds to support an ion probe study of existing 10 kbar mantle-melting run products which the P.I. proposes to pursue while at Woods Hole Oceanographic Institution during his sabbatical in Fall, 2001. These run products represent 5+ years of NSF-funded work by the P.I. and two recent Ph.D. graduates. Ten peridotitic starting materials, each constructed by recombining hand-picked and acid-washed mineral separates in differing modes, were investigated. Five of these utilized minerals from a fertile lherzolite xenolith from Kilbourne Hole while the other five utilized minerals from a compositionally intermediate lherzolite sample from Mt. Noorat, S.E. Australia. In all, 54 glass- and olivine-bearing run products are available, 36 with cpx, 41 with opx, and 47 with spinel. Calculated melt fractions range from 1-48 wt. %; 22 contain less than 5 wt% melt. All experiments employed a variation of the diamond aggregate technique (e.g., Hirose and Kushiro, 1993; Baker and Stolper, 1994) in which a layer of 70-100 micron diameter vitreous carbon spheres was substituted in lieu of powdered diamond, to provide a melt sink in which some melt could be isolated from the residual crystals, thereby preventing its quench modification upon termination of the experiments. Major element mineral and glass data for all starting materials vary in a systematic and sensible manner with increasing temperature and this, combined with run durations of 2-10 days, supports a close approach to equilibrium in the experiments. The primary goals of the proposed work will be to measure the distribution coefficients between melt and cpx and opx for REE, Ti, V, Cr, Sr, Y, and Sr (Be, B, Rb, Nb, and Ba will also be attempted but will probably be too low in the pyroxenes), and to analyze the water contents of these nominally anhydrous run products. To my knowledge, these will be the first such data germane to mantle melting derived from experiments on peridotitic, as opposed to basaltic, bulk compositions. Earlier work in basaltic systems has revealed complex compositional dependencies of pyroxene/melt Ds demonstrating that the use of single-valued Ds in modeling studies is inappropriate. Our major element data greatly expand the composition space covered by experiment and will allow extension and better definition of these dependencies, thus improving predictive models. We also have 17 cpx-bearing run products with less than 5% melt and some of these melts have the elevated Na2O and SiO2 contents that Baker et. al. (1995) first noted characterize very near-solidus mantle melts. Given the likely importance of fractional melting in the mid-ocean ridge environment, determination of the partitioning behavior of key trace elements between these "exotic" melts and mantle pyroxenes will be critical