Abstract
The previously suggested segregation model for arsenic at Si∕SiO2 interfaces based on a combined trapping∕pairing model [J. Dabrowski, H.-J. Müssig, V. Zavodinsky, R. Baierle, and M. J. Caldas, Phys. Rev. B 65, 245305 (2002)] requires high binding energies for interface vacancies, which our results of ∼0.2eV cannot confirm. As an alternative explanation, we present ab initio results that show that As and hydrogen bond with an energy gain of 1.5–3 eV with their minimum-energy position at the interface, which creates additional trapping sites for As segregation. The inclusion of hydrogen into the modeling might thus be able to explain the differences between the previous model and experiments.

This publication has 23 references indexed in Scilit: