BORON PARTITIONING IN THE UPPER-MANTLE - AN EXPERIMENTAL AND ION PROBE STUDY

Piston cylinder experiments were performed on two natural spinel lherzolites, from Lherz (French Pyrenees) and Kilbourne Hole (USA), in order to determine the distribution coefficients of boron between minerals and melt for upper mantle conditions. Additionally, a MORB glass and its phenocrysts were taken as a low pressure analogue. Boron concentration measurements were made by ion probe, the only technique suitable for in situ (approximate to 10 mu m) boron concentration measurements in the range 0.1-1 ppm. The two starting peridotites samples have low bulk boron concentrations of 0.96 and 0.81 ppm. Boron distribution coefficients were determined from (1) direct measurement of melt and crystals in contact, and (2) calculations from profiles of several tens of measurements made using 10 mu m diameter spots containing variable proportions of crystals and melt. Distribution coefficients between minerals and melt decrease in the following order: D-cpx-melt (0.117) > D-sp-melt (0.08) >> D-ol-melt (0.034) > = D-opx-melt (0.027). The limited ranges of composition, pressure, and temperature investigated mimic natural conditions for basalt genesis but sharply limit the present data for determining variations of partition coefficients with P, T, and X. However, boron behavior seems to follow roughly that of aluminum, since the Al-richest minerals are the ones with the highest boron distribution coefficients. The D-ol-melt values increase with pressure, which is similar to what has been observed for the distribution coefficient of aluminum between olivine and melt. Boron is, therefore, strongly incompatible during partial melting in the mantle. Using these distribution coefficients and assuming that MORB have been produced by 5-20% partial melting, the depleted mantle source of MORB is estimated to contain approximate to 0.05-0.3 ppm B. These values correspond to the lower limit for the mean boron content of chondrites (0.55 +/- 0.3 ppm) but are in accordance with a simple budget of boron between mantle, crust, and seawater. In fact, a simple mass balance calculation, where all the boron hosted in the crust and in seawater is presumed to have been extracted from an upper mantle of chondritic original boron content, yields a boron content for the depleted upper mantle of approximate to 0.3 ppm at present. Because in situ boron concentration measurements under the ppm level at the mu m scale are now possible with the ion probe, and because of its strong partitioning in melts and possible partitioning in fluids, boron is a potentially very powerful geochemical tracer for the study of mantle processes such as basalt genesis or metasomatism.