Charge ordering on the surface ofFe3O4(001)

Abstract

We have studied the (001) surface of a ${\mathrm{Fe}}_{3}{\mathrm{O}}_{4}$ single crystal using scanning tunneling microscopy (STM), low-energy electron diffraction (LEED), and Auger electron spectroscopy. A clean surface was obtained by a combination of ${\mathrm{Ar}}^{+}$ ion sputtering, in situ annealing in ${\mathrm{O}}_{2}$ atmosphere, and further annealing in ultra high vacuum conditions. A sharp $(\sqrt{2}\ifmmode\times\else\texttimes\fi{}\sqrt{2})R45\ifmmode^\circ\else\textdegree\fi{}$ reconstruction was observed by LEED and STM analysis at room temperature. The STM measurements were performed using a tip of antiferromagnetic MnNi alloy. Atomically resolved STM images provide evidence of a surface terminated at the octahedral plane, with rows of Fe cations running along the [110] and $[11\ifmmode\bar\else\textasciimacron\fi{}0]$ crystallographic axes. The 3 \AA{} periodicity of the Fe rows expected for a bulk-terminated B plane is not observed. Instead, two different kinds of Fe cations with a separation of 6 \AA{} are imaged. The periodicity between Fe cations of the same kind is about 12 \AA{}. We propose a model to explain the observed symmetry in terms of charge ordering of the Fe cations on the octahedral sites of the inverse spinel lattice. We also explain the $(\sqrt{2}\ifmmode\times\else\texttimes\fi{}\sqrt{2})R45\ifmmode^\circ\else\textdegree\fi{}$ LEED pattern in terms of charge ordering, as opposed to a structural rearrangement of the atoms on the surface. We further suggest the possibility of a spin polarized effect, in view of the anomalous corrugation observed along the [110] Fe rows on the octahedral plane, in agreement with the different spin configuration of ${\mathrm{Fe}}^{2+}$ and ${\mathrm{Fe}}^{3+}$ ions.