Photoemission satellites and electronic structure of<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"><mml:mrow><mml:msub><mml:mrow><mml:mi mathvariant="normal">Fe</mml:mi></mml:mrow><mml:mrow><mml:mn>2</mml:mn></mml:mrow></mml:msub></mml:mrow></mml:math><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"><mml:mrow><mml:msub><mml:mrow><mml:mi mathvariant="normal">O</mml:mi></mml:mrow><mml:mrow><mml:mn>3</mml:mn></mml:mrow></mml:msub></mml:mrow></mml:math>

The electronic structure of \ensuremath{\alpha}-${\mathrm{Fe}}_{2}$${\mathrm{O}}_{3}$ with the high-spin ${d}^{5}$ ground-state configuration has been studied by ultraviolet photoemission spectroscopy with use of synchrotron radiation as well as by x-ray photoemission and Auger-electron spectroscopy. The results are interpreted in terms of the configuration-interaction theory based on a ${\mathrm{FeO}}_{6}$-cluster model. The main lines of the valence-band photoemission spectra are identified with Fe?d sup 5 ndash---ligand-hole final states produced by ligand-to-3d charge-transfer screening of 3d holes (3${d}^{4}$ states), whereas the satellite at higher binding energies is assigned to unscreened (or poorly screened) 3${d}^{4}$ final states. The Fe 3d versus O 2p partial density of states and symmetry characteristics of 3d-derived peaks are found to be quite different from assignments based on ligand-field theory or band theory. These results indicate that ${\mathrm{Fe}}_{2}$${\mathrm{O}}_{3}$ cannot be considered as a Mott-Hubbard insulator in its original sense but is classified as a charge-transfer-type insulator according to a theory of Zaanen, Sawatzky, and Allen. A possibility is suggested that the lowest unoccupied state is not Fe?d ndash---like but is the bottom of the Fe 4s band. The large exchange energy of the high-spin 3${d}^{5}$ configuration is shown to greatly stabilize the localized 3d states relative to the itinerant state.