Melting of Iron - Light-Element Alloys in the Laser Heated DAC

Seismic data indicate that the Earth's outer core is ~10% less dense than pure iron at the pertinent conditions of pressure and temperature based on the experimentally determined equation of state of iron (Shanker et al., 2004). This core density deficit can be explained by the presence of a light element component such as H, C, N, O, S or Si, or a mixture of these. Constraints on core composition, phase relations and temperature can be derived from knowledge of the melting relations in relevant iron - light-element systems using a thermodynamic approach coupled with observations from seismology (Helffrich & Kaneshima, 2004). Here we make a progress report on our systematic effort to determine melting phase relations in binary Fe-alloy systems at high pressures using laser-heated DAC techniques. Foils of compressed powder or pre-fabricated chips of Fe alloys are loaded into ~100 micron holes in pre- indented stainless steel or rhenium gaskets. We use a variety of pressure media including sapphire, ruby, alumina gel, NaCl, and Argon, which also serve as thermal insulators. Pressures are measured before and after experiments using the fluorescence shift of ruby. Samples are heated using a 60W Nd:YLF laser with a double- sided heating geometry, and temperatures are measured using standard spectro-radiometric techniques (Walter & Koga, 2004). Melting is deduced from sudden, obvious and repeatable discontinuities in the temperature and emissivity vs. laser power functions as expected from invariant melting. In some cases clear visual observation of melt motion is coincident with these discontinuities. Our results to date show good correspondence with previous measurements where data overlap for Fe, Pt, the Fe-S eutectic, the Fe-Fe3C eutectic and Fe3C. Our melting curve for Fe3C up to ~75 GPa is considerably lower in temperature at high pressures than previously predicted (Wood, 1993), yielding an extrapolated temperature of about 4000 K at the core mantle boundary using a Simon fit to the data. We predict a singular point along the Fe3C melting curve at ~ 20 GPa where congruent melting begins, and possibly another singular point at ~ 70 GPa where the Fe-Fe3C eutectic may intersect the Fe3C liquidus indicating that the eutectic composition has risen to become equal to that of Fe3C. We will present these and other new results.