- Investigation of the properties of iron-bearing alloys and silicates and their implications for the Earth’s lower mantle and core (2010)
- (1) (Mg,Fe)(Si,Al)O3 perovskite comprises about 80% of the Earth’s lower mantle by volume, leaving ~ 15% to (Mg,Fe)O ferropericlase and ~ 5% to CaSiO3 perovskite. Therefore characteristics of the lower mantle would be determined mostly by the properties of the silicate perovskite and ferropericlase. While high-spin – low-spin crossover in (Mg,Fe)O is nicely described in literature, the electronic state of iron in silicate perovskite at high pressures and temperatures remains controversial. Conclusions derived from the results of X-ray emission (XES) and nuclear forward scattering (NFS) spectroscopic studies of Fe-bearing silicate perovskite are not in agreement on the pressure and temperature conditions of the transition and on whether Fe2+ or Fe3+ or both iron cations are involved. We undertook an alternative study of (Mg,Fe)(Si,Al)O3 perovskite at a wide pressure and temperature range using a number of different spectroscopic techniques (conventional Mössbauer, X-ray absorption near edge structure (XANES), NFS and X-ray diffraction (XRD) spectroscopies), in order to get a rather complete picture regarding the spin state of iron in this compound. Desirable pressures relevant to those in the Earth’s lower mantle were achieved by means of diamond anvil cells, equipped with miniature external resistive heaters, providing homogeneous heating up to 1000 K, which enables us to estimate the effect of temperature as well. Our Mössbauer and XANES data, collected at pressures to 110 GPa and temperatures to 1000 K for silicate perovskite, revealed a gradual transition involving Fe2+, which at room temperature occurs over a rather wide pressure range, 35-70 GPa, but becomes narrow at high temperatures. This observation coincides with the previously reported drop in spin number revealed by XES. Taking this into account and based on the fact that our XRD measurements, performed at the corresponding pressure-temperature conditions, do not suggest any appreciable structural change in perovskite, we conclude that the origin of the observed transition is fully electronic. Considering the simplified energy diagram of ferrous iron, sitting in 8-12 fold coordinated polyhedra in the perovskite structure, and by analysing the effect of pressure on the distribution of valence electrons over the energy levels, we propose stabilization of intermediate spin Fe2+ in magnesium silicate perovskite at pressures over 35 GPa. The gradual character of high-spin – intermediate-spin crossover in silicate perovskite does not assume any abrupt changes of the lower mantle properties. However due to the negative Clapeyron slope of the transition in some areas in the uppermost lower mantle, located in the vicinity of subducting slabs or hot mantle upwellings, they would have slightly different properties (namely electrical and thermal conductivity, element partitioning) with respect to the surrounding mantle. (2) Although the effect of spin transitions in lower mantle silicate perovskite and ferropericlase on the iron partitioning between these two phases has been widely discussed in the literature, the question remains open, mainly due to the lack of in situ experiments under the relevant pressure and temperature conditions. We approach this problem using combined in situ synchrotron XRD and XANES spectroscopic measurements with laser-heated diamond anvil cell (LH-DAC) technique. The assemblages of perovskite and ferropericlase were synthesized via the breakdown reaction of ringwoodite or a natural San Carlos olivine. Our data, collected from 22 GPa to 115 GPa, after laser heating to 1950 K and 2300 K, confirmed previously reported preferred partitioning of iron in low-spin ferropericlase with respect to high or intermediate spin perovskite. Increase of temperature was shown to slightly increase the amount of Fe in perovskite. (3) The phase diagram of Fe-enriched FeNi alloy at elevated pressures and temperatures comparable to those in the Earth’s core as well as the effect of the potential light element in the core on the system remains under debate. The reason is very simple: the necessary pressure-temperature range (300-350 GPa and 5,000-7,000 K) is hardly achieved in the laboratory. Conditions close to those in the core can be created in LH-DAC experiments; however the main disadvantage of the technique is a lack of control on the system under study. In contrast, the large-volume press (LVP) provides rather equilibrated and controllable experimental conditions, although at significantly lower pressures. Therefore we carried out a “cross-checking” study of phase relations in Fe1-xNix (0.10 < x < 0.22) and Fe0.90Ni(0.10-x)Cx (0.01 < x < 0.05) systems at pressures to 52 GPa and temperatures to 2600 K, using LH-DAC and LVP high pressure techniques. In situ and ex situ sample analyses were done by means of XRD, Mössbauer spectroscopy, and scanning and transmission electron microscopies. We showed that laser heating in the DAC can promote undesired reactions between the sample (FeNi alloy in our case) and carbon, which upon laser heating diffuses from the diamond anvils through the pressure-transmitting layers. We investigated the mechanism of carbon incorporation into the structure of FeNi alloy at high pressures and temperatures, and showed that rapid cooling of fcc-structured carbon-bearing FeNi alloy to room temperature (either in the case of LH-DAC or LVP quenched experiments) results in the formation of the metastable solid solution of bcc-FeNi, bct-Fe-Ni-C (known in metallurgy as martensite) and a certain amount of preserved fcc Fe-Ni-C.