Abstract
The surface‐state energy distribution at the metal‐semiconductor interface of a Schottky barrier has been deduced by combining a general theoretical analysis with one observed fact concerning the anomalous temperature dependence of the forward current on voltage. The fact, empirically found by Saxena, Padovani, and Sumner and by Padovani, is that the exponential terms of the forward‐current characteristic involve the sum of T plus T0. Here T is the absolute temperature and T0 is the empirical parameter which is essentially independent of temperature at constant current. This fact leads to the following new theoretical conclusions: (i) The surface‐state energy distribution near the Fermi level is a simple exponential with a characteristic e‐fold energy increase E0 that can be computed from T0. (ii) A plot of T0−1 vs voltage V yields essentially a straight line whose slope and intercept can be used to compute E0 and the barrier height φB. (iii) In reverse bias, φB varies as the log of the surface electric field F, and E0 can be computed from the slope. (iv) Finally, the log of the reverse current is proportional to (E0/kT)log F. These theoretical predictions have been substantially checked, and E0 deduced from data on the following diverse Schottky barriers: Au/GaAs, Cr/Si, Ag/ZnS, Zu/SrTiO3, and ZrSi2/Si, as measured by Padovani and Sumner, Padovani, Saxena, Wronski, Carnes and Goodman, and Andrews and Lepselter. Furthermore, the forward‐ and reverse‐current temperature anomalies observed by Padovani on the same sample of Au/GaAs lead to the same value of E0, E0=0.045±0.005 eV. Also, φB as obtained from a plot of T0−1 vs V on diverse Schottky barriers agrees closely with φB as obtained from independent methods. These features demonstrate a measure of self‐consistency in the theory. A derivation of the empirical equation involving T0 is given and its limits of applicability are stressed.