3D mapping of chemical distribution from melting at lower mantle conditions in the laser-heated diamond anvil cell

Monday, 15 December 2014: 1:53 PM
Farhang Nabiei1, Marco Cantoni1, James Badro1,2, Susannah M Dorfman1, Richard Gaal1 and Philippe Gillet1, (1)EPFL Swiss Federal Institute of Technology Lausanne, Lausanne, Switzerland, (2)Institut de Physique du Globe de Paris, Paris, France
The laser-heated diamond anvil cell is a unique tool for subjecting materials to pressures over few hundreds of GPa and temperatures of thousands of Kelvins which enables us to experimentally simulate the inaccessible interiors of planets. However, small sample size, laser profile and thermally conductive diamonds cause temperature gradients of 1000s K over a few microns which also affects chemical and structural distribution of phases in the sample.

We have examined samples of San Carlos olivine (Mg,Fe)2SiO3 powder melted in the diamond anvil cell by double-sided and single-sided laser heating for 3-6 minutes to ~3000 K at 35-37 GPa. Moreover, MgO is used as an insulating media in one of the sample. Recovered samples were analyzed by a combination of focused ion beam (FIB) and scanning electron microscope (SEM) equipped with energy dispersive x-ray (EDX) detector. Images and chemical maps were acquired for ~300 slices with ~70 nm depth from each sample, comprising about half of the heated zone. Detailed chemical and structural analysis by transmission electron microscopy (TEM) of lamellas prepared from the remaining section of the samples will also be presented.

In all samples the heated zone included (Mg,Fe)SiO3 perovskite-structured bridgmanite (PV) phase and two (Mg, Fe)O phases, one of which, magnesiowüstite (MW), is richer in iron than the other one, ferropericlase (FP). In double-side heated samples we observe a Fe-rich quenched melt core surrounded by MW phase. Our results show that with increasing heating time, Fe migrates to the molten center of the sample. In the single-side heated sample, the Fe-rich MW phase is concentrated in the center of heated zone. In all samples a FP crust was observed around the heated zone. This crust, however, is broken in the upper part (colder part) of the single-side heated sample due the high asymmetrical temperature gradient within the sample. The results confirm the importance of double-side heating and insulating media for generating homogenous central temperature and chemical distribution.