Melting and phase relations of Fe-Ni-Si determined by a multi-technique approach

Abstract

Many studies have suggested silicon as a candidate light element for the cores of Earth and Mercury. However, the effect of silicon on the melting temperatures of core materials and thermal profiles of cores is poorly understood, due to disagreements among melt detection techniques, uncertainties in sample pressure evolution during heating, and sparsity of studies investigating the combined effects of nickel and silicon on the phase diagram of iron. In this study we develop a multi-technique approach for measuring the high-pressure melting and solid phase relations of iron alloys and apply it to Fe_(0.8)Ni_(0.1)Si_(0.1) (Fe-11wt%Ni-5.3wt%Si), a composition compatible with recent estimates for the cores of Earth and Mercury. This approach combines results (20-83 GPa) from two atomic-level techniques: synchrotron Mössbauer spectroscopy (SMS) and synchrotron x-ray diffraction (XRD). Melting is independently detected by the loss of the Mössbauer signal, produced exclusively by solid-bound iron nuclei, and the onset of a liquid diffuse x-ray scattering signal. The use of a burst heating and background updating method for quantifying changes in the reference background during heating facilitates the determination of liquid diffuse signal onsets and leads to strong reproducibility and excellent agreement in melting temperatures determined separately by the two techniques. XRD measurements additionally constrain the hcp-fcc phase boundary and in-situ pressure evolution of the samples during heating. We apply our updated thermal pressure model to published SMS melting data on fcc-Fe and fcc-Fe_(0.9)Ni_(0.1) to precisely evaluate the effect of silicon on melting temperatures. We find that the addition of 10 mol% Si to Fe_(0.9)Ni_(0.1) reduces melting temperatures by ∼250 K at low pressures (<60 GPa) and flattens the hcp-fcc phase boundary. Extrapolating our results, we constrain the location of the hcp-fcc-liquid quasi-triple point at 147±14 GPa and 3140±90 K, which implies a melting temperature reduction of 500 K compared with Fe_(0.9)Ni_(0.1). The results demonstrate the advantages of combining complementary experimental techniques in investigations of melting under extreme conditions.

Files

Additional details