The Deal fast-surface states are probably deep-level impurities in the semiconductor

A comparison is made between the widely held model that surface states are states that lie at the Si-SiO₂ interface (e.g. dangling bonds) and a model based on deep levels (heavy metals) in the semiconductor. It is shown that calculations based on the two models measure the same quantity but that a model based on dangling bonds is inconsistent with an observed dependence on oxide thickness. Deep-level impurities introduced during or after oxide growth, however, are expected to show a dependence on oxide thickness. Furthermore, the onset of the surface states is in agreement with that known for copper at 0.52 eV as measured from the valence band edge.