Theory of the peroxy-radical defect ina-SiO2

We have investigated the peroxy-radical defect in glassy Si${\mathrm{O}}_2$ by means of MOPN, a semiempirical molecular-orbital program, applied to a cluster of atoms chosen to simulate the defect. Our calculations are consistent with important features of Griscom's model of the defect as a perturbed ${\mathrm{O}}_2^{-}$ ion substituted for a single-bridging ${\mathrm{O}}^{2-}$ ion in Si${\mathrm{O}}_2$ and attached to a single silicon, and they place geometrical constraints on the defect structure for this model to be valid. We predict the existence of a related defect (the small peroxy radical, or SPR) wherein the peroxy radical is strongly bonded to two silicons. We have also investigated the formation of the peroxy radical. Griscom and co-workers envision peroxy linkages substituting for single-bridging oxygens during the growth process. They suggest that upon annealing these linkages readily give up an electron to form the observed radical. Our calculations lead us to argue against this process; rather, capture of a free hole seems more likely. We suggest that the SPR could form via a process in which neutral oxygen molecules diffuse through the solid and combine with $E_1^{\prime }$ centers to form peroxy radicals.