On the Verwey charge ordering in magnetite

Abstract

Magnetite, a famous permanent natural magnet, is probably the oldest magnetic material known to humans [1]. Fe3O4 is a ferrimagnet with an anomalously high Curie temperature of ∼850 K which arises from the inverse spinel type crystal structure. This is formally written as AB2O4 where A and B are two crystallographically distinct tetrahedrally and octahedrally coordinated Fe sites. Both have different oxidation states due to the large difference of the corresponding average Fe–O distances, 1.876 and 2.066 A (at 300 K), respectively. Thus the former (the A site) is occupied by Fe3+ ion, whereas octahedral B sites are occupied by an equal number of randomly distributed 2+ and 3+ Fe ions, which results in an average valence value of 2.5+ per FeB ion. The B sublattice is highly frustrated and can be viewed as a pyrochlore lattice of corner-sharing FeB tetrahedra. Therefore, mixed-valent magnetite is considered to be a charge frustrated system with a highly degenerate ground state. Fe3O4 is a poor metal with an electrical resistivity of 4 m cm (at 300 K) which is remarkably higher then the resistivity of simple metals, e.g. 0.96, 1.59 or 1.7 μ cm for Hg, Ag or Cu, respectively. Upon cooling below TV ∼ 120 K magnetite undergoes a sharp first-order metal–insulator transition (the so-called Verwey transition) at which the conductivity abruptly decreases by two orders of magnitude and the symmetry of the structure lowers from the cubic one [2]. To explain this anomaly Verwey proposed a theoretical model according to which the transition is caused by the ordering of Fe2+ ions on the B sublattice with formation of charged (001) planes alternately occupied by 2+ and 3 + FeB ions. This was the first report on a charge ordering as well as an orthorhombic (Verwey) superstructure model. The Verwey charge ordering (CO) model obeys the so-called Anderson criterion for minimal electrostatic repulsion which requires the occupation of each FeB tetrahedron by an equal number of 2+ and 3+ ions, thus leading to a short-range CO pattern [3]. At first confirmed by x-ray and neutron diffraction studies [4], the Verwey CO model was disproved by further experiments. The half-integer satellite reflections (h, k, l + 2 ) clearly observed below TV indicate a doubling of the cubic unit cell along the c axis and show the symmetry to be monoclinic Cc [5]. On the other hand, observation of the magnetoelectric effect revealed even lower P1 symmetry in the low temperature phase [6]. Nevertheless, no conclusive structural model or charge ordered arrangement was identified because of the complexity of the low temperature structure and the difficulties caused by microtwinning at the