Mixed valence transition metal oxides exhibit many interesting properties such as superconductivity, colossal magnetoresistance and metal insulator phase transitions. Generally, the description of the electronic state of these compounds is made on the basis of the ionic model and it implies the spatial or temporal charge localisation on the transition metal atoms.

Magnetite was the first material in which a charge ordering transition was proposed to explain the metal insulator phase transition which occurs at 125 K [1]. The classical mechanism to describe this transition states that the distribution of octahedral Fe3+ and Fe2+ ions changes from dynamical disorder to long range order by lowering the temperature. Nevertheless, this point has not yet been demonstrated.

We report the observation of the d-glide plane 002 and 006 forbidden reflections, below and above the Verwey transition (125 K), by means of X-ray resonant scattering at the iron K-edge [2]. Experiments were performed at beamline BM2 (D2AM French CRG). The energy dependence of the intensity of the 002 and 006 reflections, with the X-ray beam polarisation along the [110] direction, at 20 K and at room temperature is shown in Figure 62 together with the fluorescence spectrum. As it can be observed, the energy dependence of the scattered intensity is the same for both temperatures. No detectable intensity below the absorption edge was observed. Three main features can be distinguished on the spectra as a function of the energy [3]: a) a resonance at the pre-peak energy of the fluorescence spectrum corresponding to a virtual excitation and de-excitation through dipolar-quadrupolar channels at the tetrahedral iron atom b) a strong resonance at the 1s-4p energy transition, due to the anisotropy of the dipolar scattering factor of octahedral iron atoms at the trigonal (-3m) site. iii) the extended part above the edge that has the same origin as the main resonance and shows an oscillatory behaviour as a function of the energy.

The oscillatory signal is a sum of cosine terms whose frequencies are proportional to the differences in distance of the coordination shells. This is the first time that anisotropy of the atomic scattering factor in the extended part is reported and a coherent explanation of its origin is discussed. The dependence of the integrated reflection intensity, at 7124 eV, as a function of the azimuthal angle , is shown in Figures 63a and 63b for the two reflections. The theoretical azimuthal dependence (solid lines) is sketched as total, [( - ) + ( - )] and ( - ) channel, intensities. The analysis of the experimental data show that the anomalous scattering factor is anisotropic but identical for all the octahedral iron atoms at room temperature and remains unaltered across the Verwey transition. As a matter of fact, it is well established that the main difference of the anomalous scattering factors between different oxidation states is given by the chemical shift (edge splitting). In the case of Fe3+ and Fe2+, the chemical shift is between 3 and 5 eV, appreciably larger than the calculated splitting due to the anisotropy (1 eV), so charge fluctuation at room temperature with a time duration longer than the interaction time for the virtual process involved (10­16 sec), would destroy the structural coherence for diffraction. The identical behaviour (energy and azimuthal dependence) for both, low and room temperature phases, discard a charge localisation in a periodic arrangement. As a general conclusion we can say that octahedral iron atoms in magnetite can not be described as pure ionic Fe3+ or Fe2+ ions, neither spatially or temporally. As a consequence, the metal-insulator transition is likely produced by crystallographic distortion rather than by charge localisation.

References
[1] E.J.W. Verwey, Nature, 144, 327 (1939).
[2] D.H. Templeton, L.K. Templeton, Acta Cryst. A36, 237 (1980).
[3] J. Garcia, G. Subias, M.G. Proietti, H. Renevier, Y. Joly, J.L. Hodeau, J. Blasco, M.C. Sanchez, J.F. Berar, Phys. Rev. Lett., in press.

Authors
J. Garcia (a), G. Subias (a), M.G. Proietti (a), J. Blasco (a), H. Renevier (b), J.L. Hodeau (b), Y. Joly (b).

(a) Instituto de Ciencia de Materiales de Aragón, Consejo Superior de Investigaciones Científicas y Universidad de Zaragoza (Spain)
(b) Laboratoire de Crystallographie, CNRS, Grenoble (France)