Heterointerface dipoles: Applications to (a) Si–SiO₂, (b) nitrided Si–N–SiO₂, and (c) SiC–SiO₂ interfaces

This article identifies procedures to calculate charge-transfer dipoles at semiconductor–dielectric interfaces, focusing primarily on the Si–SiO₂ system. Since SiO₂ is more polar than Si, there is a transfer of electrons from Si to SiO₂ to balance the difference in chemical potentials creating a dipole localized at the semiconductor–dielectric interface. This dipole increases the conduction-band offset energy difference between Si and SiO₂, and therefore, has important effects on interface electrical and optical properties. Empirical chemistry and ab initio methods have been applied to molecular clusters which emulate the interface bonding, and have been used to calculate interfacial charge transfer at (i) abrupt Si–SiO₂ interfaces and (ii) nitrided Si–N–SiO₂ interfaces. Additional calculations have applied to determine the average dipoles at Si–SiO₂ interfaces with suboxide bonding in excess of the monolayer level required to form an ideal interface. The calculations support experimental data that indicate that the effective conduction-band offset energies at nitrided Si–N–SiO₂ interfaces and at interfaces with minimized suboxide bonding are essentially the same. Finally, the calculations have been extended to SiC–SiO₂ interfaces to illustrate the effects of changing from a homopolar to a more ionic or heteropolar semiconductor. © 1998 American Vacuum- - Society.