Specifically, the HSiO3 (yellow curve below) and the Si-SiO3 (light blue curve below) assignments directly overlap with the spectral regions normally assigned soley to SiO2 or Si+1 respectively.

The cluster-based model interfaces caused us to seriously question the assignment scheme grounded upon the formal oxidation state of the silicon species. Following the lead of previous workers in both gas phase and solid state photoemission studies, we adopted group electronegavity as a better method of approximating the valence electron distribution. Shown below is a plot of binding energy shift vs group electronegativity for a series of well-defined structural fragment in the Si/SiO2 interface region (red dots). The triangle, square, and diamond represent hypothetical species assigned to the binding energy shifts observed for the Si(100) Si/SiO2 interface shown above. Note fragments at the same position in the interface region, basically all species between the diamond and triangle, fall on a straight line. Group electronegativity is primarily an indicator of initial state effects on the binding energy shift. Fragments well-screened by the bulk silicon fall off the line to lower binding energies because the final core-hole state is stablized. Fragments in the oxide feel less of this screening effect, and therefore fall off the line to higher binding energies.

Shown below is a model structure proposed for the Si/SiO2 interface based on a combination of experimentally verifiable Si 2p core-level assignments and group electronegavity calculations. The model is striking in that it contains for silicon in the formal +1 oxidation. This work was published in Applied Physics Letters.