A first-principles molecular orbital study was made of the postulate that, in the amorphous material, E’ centers would spontaneously distort so as to minimize the number of second-nearest neighbor Si atoms in back-bonded positions. The results of the present calculations did not support this assertion. It was instead found that back-bonded second-nearest neighbors were preferred. The hyperfine results did not contradict the principal assertion that, at E’ centers, the dihedral angles did not exhibit a flat distribution. The present results indicated that spontaneous puckering was probably not the mechanism. Another possibility was that, since the E’ center was the result of the asymmetrical relaxation of an O vacancy, it could be that the side of an O vacancy which underwent the large relaxation was related to the distribution of dihedral angles. However, it had been shown that the size of a molecular cluster was critical in predicting the relaxation correctly. This implied that the distribution of dihedral angles was not flat, but might depend fundamentally upon sample preparation.

A.H.Edwards, W.B.Fowler, T.S.Elicker: Materials Science Forum, 1997, 239-241, 11-4