A. Kavner, D. Walker
Lamont Doherty Earth Observatory, Palisades NY 10964
The Earth's core-mantle boundary is subject to a temporally- and spatially varying electrical field generated by the geodynamo. Here we report new results from a suite of piston-cylinder experiments designed to elucidate the physical, mineralogical, and trace-element geochemical behavior of a metal/silicate interface subjected to an externally applied voltage. All experiments were performed on silicate, silicate/iron, or silicate/sulfide/iron systems at $1350\deg$C and 20 kbar in an electrically insulated sintered MgO capsule. Conductivity measurements are made and voltage is delivered via Pt and Pt-Rh electrical leads in contact with either end of the sample. Behavior at the anode (positively charged electrode) is significantly different from the behavior at the cathode (negatively charged electrode) and a non-charged boundary. The reactions also depend on the chemical nature of the boundary being charged. At a Pt-Rh (20%)/silicate anode (+1 V), Fe-bearing spinel-like oxide phases are formed at the rate of ~100µm/hr. A corresponding drop in probe-to-probe electrical conductivity is also measured. Under these conditions, Rh preferentially segregates into the oxide phase, with a strong concentration gradient in the silicate phase and a corresponding Rh concentration depletion near the surface of the anode electrode. When the sulfide phase is the anode (100 mv-1 V), increased sulfide wetting of silicate grain boundaries is observed, and the probe-to-probe electrical conductivity increases significantly, indicating sulfide percolation to the cathode probe. In recovered samples, the platinum cathode is plated with a ~500 µm thick sulfide-poor iron layer. In the absence of sulfide, the surface of a platinum cathode shows evidence of significant corrosion, with corresponding blebs of Pt-Fe metal alloy within the silicate surrounding the cathode. If these observations can be applied to the core-mantle boundary, they provide a means to a) generate significant geochemical mixing between core and mantle phases, including platinum-group element partitioning; b) generate different mineralogical assemblages at different areas of the core-mantle boundary; and c) locally flux sulfide into the silicate, creating pockets of melt within the lowermost mantle. Each of these phenomena may be testable through a combination of geochemical analysis of plume material, seismic observations of the core-mantle boundary, and geophysical estimates of the topological behavior at the core-mantle boundary.