Crystallization Processes in Mercury’s Core Inferred from In-situ High-Pressure Melting Experiments in the Fe-S-Si-C System

Friday, 19 December 2014
Audrey M Martin1, James A Van Orman1, Steven A. Hauck II1, Nan Sun1, Tony Yu2 and Yanbin Wang3, (1)Case Western Reserve University, Cleveland, OH, United States, (2)University of Chicago, Center for Advanced Radiation Sources, Chicago, IL, United States, (3)University of Chicago, Center for Advanced Radiation Sources, Argonne, IL, United States
Based upon the high pressure melting temperatures in the Fe-FeS system, an iron “snow” process has been suggested to occur in Mercury’s core. However, recent results from the MESSENGER mission indicate very reducing conditions in Mercury, under which a substantial amount of silicon should also dissolve into the core. The presence of Si can significantly modify the chemical and physical properties of Mercury’s core (e.g., phase relations, crystallization, density). Moreover, up to 4 wt% C could have been incorporated into the core during the planet formation. In order to test the iron snow hypothesis in a system that is likely to be closer to the actual core composition, we performed in situ high-pressure, high-temperature experiments in the Fe-FeS-Fe2Si-Fe3C system using a multi-anvil press on a synchrotron (Advanced Photon Source, Argonne). To observe low degree eutectic melting, we separated the samples in two parts: (1) an iron rod presaturated with Si and C and (2) a mixture of FeS, Fe2Si and Fe3C. Eutectic melting temperature and phase relations were determined at various pressures between 4.5 and 15.5 GPa using energy dispersive X-ray diffraction and imaging. Temperature was quenched soon after melting in order to preserve the eutectic melt composition. The X-ray images, diffraction spectra and back-scattered electron images of the recovered samples show that eutectic melting occurs in the range of 800 - 900°C in all our experiments. These temperatures are close to the eutectic temperatures in the Fe-FeS-Fe3C system, indicating that Si does not change the eutectic temperatures significantly. Melting therefore occurs at much lower temperature than suggested for the Fe-S-Si system at similar pressures. This difference may be explained by the presence of C and by the higher silicon content in our starting composition. Our experimental setup may also be more suitable for detecting the low degrees of melting in metallic systems. Such low eutectic melting temperatures imply that the iron “snow” process may still be valid even if silicon and carbon are present in Mercury’s core.