Two long standing problems in the origin and evolution of the Earth's core are the identity of the light alloying element(s) and the origin and nature of the seismically anomalous D" layer directly overlying it. It is possible that the key to both these problems lies in high pressure and temperature chemistry. Ringwood and colleagues (1990a and b) showed that high pressure increased the solubility of O and other oxides in Fe-metal. Thus, at high pressures elements O and Si may become soluble

more »

... ugh in Fe-metal for them to contribute significantly to the light element budget in the core. In addition, one proposed origin of the D" layer is a chemical reaction between the liquid metal outer core and the solid silicate mantle overlying it Jeanloz 1989, 1991). In order to investigate the nature and extent of these possible chemical interactions, we have begun a diamond anvil cell study of the high pressure and temperature reactions between metal and silicate. Experimental procedures Our basic sample consisted of a metallic plate in contact with a silicate which was covered with a mineral plate to insulate the sample from the highly conductive diamond anvil. In the majority of runs the sample consisted of San Carlos olivine in contact with Fe-metal with an A1203 cover plate. However, we have also conducted runs with San Carlos enstatite for both the silicate and the cover plate with Fe-metal, and an Fe-free enstatite with an MgSiO3 glass cover plate and Fe-meta|. Also, in order to study very reducing conditions and evaluate the proposal that Si is the major light element in the core, we have performed runs with iron silicides consisting of 17 and 9 wt.% silicon and San Carlos olivine with MgSiO3 glass cover plates.. We have also successfully completed a run with a small amount of FeS added along with the Fe-metal and San Carlos olivine. After loading the sample into the diamond cell, the diamond cell was placed in 100 ~ vacuum oven over night. The oven was repressurized with Ar, and the cell was sealed. The flushing with Ar and immediate sealing ensures that there is no water or O from the atmosphere present so that the oxygen fugacity prevailing during the experiment is set by the sample assemblage. The interface between the metal and silicate was heated with an YLF laser with a hot spot size ranging from 20 to 50 gm. There is an average temperature gradient across the hot spot of 15 to 25 K/gm. However this is not indicative of the true temperature distribution across the hot spot as the temperature gradient is very flat across the central part but increases dramatically as the edges are approached. The metallic portion of the sample was melted during heating, and the initially clear San Carlos olivine and San Carlos enstatite darkened due to the transformation to the high pressure phases of perovskite and magnesiowfistite. The Fe-free enstatite remained clear. The run pressure was determined through the position of fluorescence peaks of rubies distributed throughout the sample. The samples were recovered after the runs and polished down to the heated surface. Results and conclusions The metal and silicate portions of the samples were analysed by electron microprobe. Figures 1 and 2 show the O and Si contents of the metal in typical runs with San Carlos olivine and Fe-Metal versus pressure. The number next to each point indicates the number of degrees above the melting point of Fe the temperature was. When measuring small amounts of O and Si in the metal on such small samples false amounts can be detected though the secondary fluorescence of X-rays from the silicate and the formation of an oxide coating on the Fe-metal. In order to check how much false O and Si we might detect in the metal we simply analysed an unheated sample of San Carlos olivine and Fe-metal. The false 624