Core level electron energy-loss spectra of minerals: pre-edge fine structures at the oxygen K-edge: Comment on "Water in minerals detectable by electron energy-loss spectroscopy EELS" by R. Wirth, Phys Chem Minerals (1997) 24:561-568

In a recent paper entitled "Water in minerals detectable by electron energy-loss spectroscopy EELS" by R. Wirth, it has been claimed that OH^-- and H₂O-bearing minerals exhibit a characteristic peak in the ELNES spectra at about 528 eV prior to the onset of the O K-edge at 532 eV, which could be used for (semi-)quantitative determination of water- or OH-contents on a nanometer scale. It is shown here by parallel electron energy-loss spectroscopy (PEELS) recorded in a transmission electron microscope (TEM) that O K-pre-edge peaks with very high intensities may also exist in water-free compounds and minerals, in particular when they contain transition metals. These spectral features arise from covalent mixing of the metal and oxygen states, which introduces oxygen p character in unoccupied states of mainly metal character. The point is illustrated by the comparison of hematite (α-Fe₂O₃) and lepidocrocite (γ-FeOOH) O K-edge PEELS spectra which exhibit similar intensities of the pre-edge peak, despite of their grossly different OH^- contents. As a consequence, the general validity of the method proposed by Wirth is questioned.