Short-range order and Fe clustering in ${\text{Mg}}_{1-x}{\text{Fe}}_x\text{O}$ under high pressure

By combining high-pressure and high-temperature Mössbauer spectroscopic studies of (Mg,Fe)O with results of ab initio simulations, several important properties of this material were established. Under high pressure (Mg,Fe)O shows changes in the short-range order with the tendency to form iron clusters. These changes were found to be irreversible, implying sluggish kinetics of these processes at ambient conditions. The pressure-induced spin crossover is interpreted here as a gradual noncooperative transition. The onset and width of spin crossover depends, therefore, not only on pressure, temperature, and composition, but also on short-range order in the FeO-MgO solid solution.

Figure 1

(Color online) Mössbauer spectrum at ambient conditions of the Fe20 sample (circles), simulated Lorentzian line-shape spectrum (dashed line), and quadrupole splitting distribution (QSD) model (solid line). Residuals for both models are shown above the spectrum. The doppler velocity is given relative to the ${}^{57}\text{C}\text{o}:\text{Rh}$ source. Inset: quadrupole splitting distribution function. The dashed line shows the QS value from the Lorentzian line-shape analysis.

Figure 2

Quadrupole splitting distribution functions (solid line with black circles) of the samples (from left to right) Fe5, Fe13, and Fe20. Histograms show the simulated distribution according to Eq. (3) (for $P_{\text{FeFe}}=0.08$, 0.18 and 0.20, from left to right).

Figure 3

Temperature dependence of the average quadrupole splitting of the Fe5 and Fe20 samples. Solid lines show the fit to the crystal-field model (Ref. 48). Diamonds show the calculated $\Delta$ value for $T=0\text{K}$ for an isolated ${\text{Fe}}^{2+}$ and a 2Fe cluster substituting in a MgO matrix. The square shows the experimental value measured at 14 K for $\left({\text{Mg}}_{0.9996}{\text{Fe}}_{0.0004}\right)\text{O}$ (Ref. 26).

Figure 4

Pressure dependence of the quadrupole splitting of the Fe5 and Fe20 samples (a and b, respectively). Black triangles pointing to the right are compression points, while white triangles pointing to the left are decompression points. Lines are given as a guide for the eye.

Figure 5

Calculated enthalpy difference for the $\left({\text{Fe}}_2{\text{Mg}}_{30}\right){\text{O}}_{32}$ supercell as a function of Fe-Fe distance. Black circles are for $a=4.25\acute{\text{Å}}$ (pressure about 0 GPa) and open circles are for $a=4.00\acute{\text{Å}}$ (pressure about 40 GPa).

Figure 6

Quadrupole splitting distribution of Fe20 samples with different $P$ and $T$ histories. Solid line—starting material, equilibrated at $1200°\text{C}$ and ambient pressure. Dashed line—sample, annealed at $1000°\text{C}$ and 18 GPa in a large-volume multianvil press. Dotted line—sample after compression to about 30 GPa at room temperature. Dashed-dotted line—laser-heated sample (approximately $1800°\text{C}$) at 40 GPa.

Figure 7

$\text{Fe}K$-edge XANES spectra in the MgO-FeO system. (a) Full multiple-scattering simulations of $\left({\text{Mg}}_{0.90}{\text{Fe}}_{0.10}\right)\text{O}$ with $a=4.225\text{Å}$ where we have modified the chemical composition of the NNN shell. Spectra 1 to 3 are for 12, six, and zero Fe atoms in the NNN shell of the absorbing atom, correspondingly. (b) Experimentally measures XANES spectra. Spectrum 4 is the Fe5 sample after compression; 5—Fe13 sample after compression; 6—Fe20 sample after compression; 7—FeO wüstite sample. 8—Fe13 starting material. The arrow indicates the region of major changes in the XANES spectra due to changes in local Fe coordination.

Figure 8

Pressure dependence of (Mg,Fe)O isomer shift (relative to $\alpha \text{-Fe}$). Open and closed symbols are for high- and low-spin components, respectively. Circles, diamonds, and triangles are for Fe5, Fe13, and Fe20 samples, respectively. Solid and dashed lines show the results of ab initio simulations.

Figure 9

(Color online) Low-spin fraction in ferropericlase as a function of pressure. Gray circles, green diamonds, and open triangles are for Fe5, Fe13, and Fe20 samples at room temperature, respectively. Red squares—reanalyzed data from Speziale et al. (Ref. 2), collected at 10 K for the $\left({\text{Mg}}_{0.8}{\text{Fe}}_{0.2}\right)\text{O}$ sample. See text for details.

Figure 10

Calculated energy difference between the high- and low-spin states of Fe in different Fe clusters.

Figure 11

(Color online) Low-spin fraction at different temperatures from ab initio calculations (see text for details). Solid black line is for 10K, dashed (red) line is for 300K and dotted (green) line is for 2400K. Only the temperature contribution to Boltzmann statistics was taken into account.

Figure 12

(Color online) Temperature dependence of spin crossover in Fe5 (a) and Fe20 (b) ferropericlase samples, measured using the Mössbauer spectroscopy in resistively externally heated DAC. Boundaries are drawn at the point where low-spin or high-spin states are no longer detectable.