Metal-silicate Partitioning of Cesium – Implications for Planetary Core Formation

Introduction: We present here preliminary results on partitioning of cesium between metal and silicate at high pressure and temperature. The purpose of these experiments is to develop a data set for the metalsilicate partitioning of cesium, a lithophile element with apparent mantle depletion, to determine whether, and under what conditions, cesium may take on more siderophile behavior. These data will be used to assess the validity of current magma-ocean core formation models [e.g. 1,2]. If the Earth experienced a widespread melting event after accretion, the stage would have been set for an equilibrium differentiation of the core. It may be possible to constrain potential equilibrium conditions of core segregation by relating an element’s measured abundance in the mantle to its presumed abundance in the bulk Earth, inferred from the planetary volatility trend. The difference in concentrations would be due to the partitioning of an element into a descending metal phase with the same major element composition as the core [3]. Thus, if an equilibrium differentiation scenario is correct, we should be able to explain the observed mantle elemental abundances by experimentally reproducing the appropriate D for each element. A coherent set of conditions that generate the correct D-values for all elements would define the parameters of equilibrium segregation [4]. Because both the metal and silicate phases are completely molten in these experiments, the behavior of cesium may help to determine those parameters. Experimental Methods: The initial composition for these experiments is a mixture of FeS powder and finely ground pollucite, a cesium feldspar. The mixture was placed in graphite capsules and loaded into a Walker multi-anvil press, located at the University of New Mexico, at pressures from 3.5 to 6.0 GPa. A complete treatment of the Walker multi-anvil is provided in [5]. Run products generally have metal blobs distributed evenly throughout a glassy silicate quench. All run products were analyzed on the JEOL 8600 electron microprobe, also at University of New Mexico, at 15kV with a defocused beam of 15 microns to prevent volatilization of cesium under the beam. Fig. 1 is a backscattered electron image of a typical run product. Initial experiments were prepared with a wet polishing procedure, but since alkali elements in a sulfide phase are known to be soluble in water [6]; therefore, all contact with liquids are now avoided. Polishing was done on carborundum sheets with boron nitride as a lubricant. In addition, the samples, once exposed, were not re-impregnated with epoxy to fill vacancies caused by decompression. Finally, samples were not polished until as soon as practical before analysis, to prevent the cesium from reacting with water in the air. In all cases, samples were carbon coated immediately and stored in an evacuated desiccation chamber until analysis. Currently we are undertaking additional experiments with lower sulfur in the metallic phase as well as the concentration of cesium in the silicate phase. Also, the applied pressure range will be extended to approximately 20 GPa.