3D analytical investigation of melting at lower mantle conditions in the laser‐heated diamond anvil cell

Abstract

Diamond anvil cell (DAC) is a unique tool to study materials under static high pressures up to several hundreds of GPa comparable to the pressures in the earth and planets interior. By using laser heating the temperature of the material inside the cell can be raised to several thousand degrees. This allows us to reach to the pressure and temperature conditions of deep mantle in laser heated diamond anvil cell (LHDAC). On the other hand small heated volume of the sample adjacent to the high thermally conductive diamonds results in large temperature and pressure gradients which affect the phase transformation and chemical distribution in LHDAC. To fully understand the phase assemblages and equilibrium inside the LHDAC, it is essential to use three dimensional analytical characterization methods. As a proxy to deep mantle composition, San Carlos olivine has been chosen as a starting material for this study. To observe the effect of pressure and heating time, five samples are prepared. Three samples were melted at ~3000 K and at 45 GPa for durations of 1, 3 and 6 minutes. Other two samples were melted for 3 minutes at 30 GPa and 71 GPa. Each sample was then sliced by focused ion beam (FIB) with slice thickness of 50‐100 nm. A secondary electron image and an energy dispersive x‐ray (EDX) map were acquired from each slice by scanning electron microscope (SEM) in a dual beam FIB instrument. Half of the heated area in each sample was used for 3D FIB tomography and the other half is used to extract a 100 nm thick thin section for subsequent analysis by analytical transmission electron microscope (TEM). TEM is used to obtain accurate EDX maps from the phases. Also, the structure of crystalline phases has been characterized by electron diffraction technique. 3D reconstruction of SEM EDX maps (figure 1) shows that the heated area is roughly spherical and it consists of three main regions in all samples which correspond to ferropericlase (Mg ­­ , Fe)O (Fp), perovskite‐structured bridgmanite (Mg,Fe)SiO 3 (Brg) and iron‐rich core. The bulk of the heated area is surrounded by ferropericlase shell. Then, we find a thick region of bridgmanite phase just inside the Fp shell and in the center lies an iron‐rich core. In addition, in 45 GPa sample heated for 3 minutes we start to see another (Mg, Fe)O phase (Mw) around the core which is more iron‐rich than the Fp shell. In the 45 GPa sample heated for 6 minutes this iron‐rich oxide (Mw) entirely surrounds the iron‐rich core. TEM analysis shows a third and even more iron‐rich (Mg, Fe)O phase forming a thin layer (~70nm) between the Mw and the core. The core is getting richer in iron by increasing the pressure or heating time and its structure varies among the samples. For instance, in 45 GPa sample heated for 1 minute the core has eutectoid structure with iron nanoparticles distributed in it (figure 2) while in the 45 GPa sample heated for 6 minute we have a granular structure with the higher content of iron in the center of grains (figure 3). Moreover, we can see narrow Fp veins connecting the Fp shell to the iron‐rich core in all of the samples, particularly in 71 GPa sample these veins are numerous and thick. In fact, they occupy a substantial part of Brg region in this sample.