Hybrid quasiatomic models

The implicit assumption in the previous model is that atomic structure has little influence on the degree of / electron localisation in the solid. There are difficulties in reconciling this view with any persistence of atomic multiplet structure, and with the fact that changes of localise tion are clearly associated with certain regions of the Periodic Table. For this reason, various attempts have been made to explore what properties of atoms might survive in solids and, perhaps, provide a quasiatomic mechanism to drive changes of valence. 

Provided it can be interpreted unambiguously and separated from the localisation effects (which, unfortunately, are not unambiguous), quasiatomic multiplet structure is useful it yields the electronic configuration, and hence provides the signature of a specific valence state in a case of uncertain valence. Thus, there is a powerful motivation to compare 

In this sense, atomic multiplet theory provides complementary information about the valence state. One can take the view that this information should be used, and then blended in some way with the conceptual framework of the Anderson single-impurity model, so that the matrix elements coupling the / electrons to the conduction band can continue to play the decisive role in determining the extent of / electron localisation. 

A good example of this hybrid approach is the work by Thole et al. [619] on the 3d subshell X-ray absorption spectra of the lanthanides. While a substantial measure of agreement is achieved for a whole sequence of elements, there are areas of uncertainty in the atomic structure calculations themselves, which turn out to be most troublesome precisely for the elements which exhibit valence fluctuations.  

In a specific instance, that of Sm (further discussed below), the discrepancy between theory and observations for the condensed phase could not be resolved even by using empirically scaled atomic structure calculations. 

---