Although one third of the mass of our planet resides in its metallic core (divided into a molten outer part and a solid inner part), fundamental properties such as its chemical composition and internal structure remain poorly known. Although it is well established that the inner core consists of iron with some alloying lighter element(s), the crystal structure of the iron and the nature and concentrations of the light element(s) involved remain controversial. Seismologists, by studying the propagation characteristics of primary earthquake waves (P-waves), have shown that the inner core is anisotropic and layered, but the origins of these properties are not understood. Seismically observed shear waves (S-waves) add to the complexity as they show unexpectedly low propagation velocities through the inner core.
Interpretation of these seismic observations is hampered by our lack of knowledge of the physical properties of core phases at core conditions. In addition, the accuracy of derived inner core seismic properties is limited by the need to de-convolve inner core observations from seismic structure elsewhere in the Earth. This is particularly relevant in the case of shear waves where detection is far from straightforward. A combination of well-constrained seismological data and accurate high-pressure, high-temperature elastic properties of candidate core materials would allow for a full determination of the structure and composition of the inner core - an essential prerequisite to understanding Earth’s differentiation and evolution.
Unfortunately, the extreme conditions of pressure (up to 360 GPa or 3.6 million times atmospheric pressure) and temperature (up to 6000 K) required make results from laboratory experiments unavoidably inconclusive at present. An alternative and complementary approach, that has only recently become available, is computational mineral physics, which uses computer simulations of materials at inner core conditions. Ab initio molecular dynamics simulations have been used to determine the stable phase(s) of iron in the Earth’s core and to calculate the elasticity of iron and iron alloys at core conditions. Calculated S-wave velocities are significantly higher than those inferred from seismology. If the seismological observations are robust, a possible explanation for this discrepancy is that the inner core contains a significant amount of melt (possibly >10%). The observed anisotropy can only be explained by almost total alignment of inner core crystals.