Chemical Effects on Core-Electron Binding Energies in Iodine and Europium

A theoretical and experimental study was made of the shift in atomic core‐electron binding energies caused by the chemical environment. Two models are presented to account for these “chemical shifts.” The first uses an energy cycle to break the core‐electron binding energies into a free‐ion contribution and a classical Madelung energy contribution. The Madelung energy contributes a significant part of the binding‐energy shift. It can, in principle, be evaluated rigorously although there is some ambiguity as to a surface correction. The reference level for binding energies must also be considered in comparing theory with experiment (or in comparing experimental shifts with one another). Electronic relaxation could also introduce errors of ∼1 eV in shift measurements. The second, more approximate, model consists of a “charged‐shell” approximation for bonding electrons in atomic complexes. It gives semiquantitative estimates of shifts and demonstrates the relationship between bond polarity and core‐electron binding energy shifts. These models indicate that several features of the free‐ion state will be reflected also in chemical shifts. Free‐ion Hartree–Fock calculations were made on F, Cl, Br, I, and Eu in several oxidation states. These indicate that the removal of a valence electron shifts the binding energies of all core levels by nearly equal amounts (10–20 eV). This shift decreases with increasing atomic number in a given chemical family. The removal of an inner “valence” electron (e.g., $\text{4f}$ in europium) gives rise to relatively large shifts (∼20 eV). These features were also found in the experimental chemical shifts. Chemical shifts were measured for iodine in KI, KIO₃, and KIO₄ and for europium in EuAl₂ and Eu₂O₃, using the technique of photoelectron spectroscopy. Binding energies in iodine were found to increase by ∼0.8 eV per unit increase in oxidation number. A Madelung energy calculation indicates that this corresponds to a loss of ∼0.5 electronic charges from the valence shell per unit increase in oxidation state, and this value agrees approximately with previous results obtained from Mössbauer measurements. A shift of 10 eV was found between Eu^2 + and Eu^3 +. This very large shift is due largely to the loss of a full $\text{4f}$ electron in this change of oxidation state. With some refinements, the above technique could produce very useful information about bonding in ionic solids, in particular, allowing the determination of the charge on each atom. Their application to such problems as a determination of the oxidation states of metals in biological molecules seems even more promising.