ELECTRON CORRELATION ENERGIES IN HYDROGENATED AMORPHOUS SILICON.

The existence of defects with a negative correlation energy in amorphous silicon is still a question of controversy. Basically, two classes of defect models exist for this material. The first model states that the bulk of the available experimental data can be explained by considering three major defect states (valence band tail states, dangling bonds, and conduction band tail states), all of which have positive-U character. The second class of models agrees with the existence of some positive-U defect states but, in addition, postulates the existence of some positive-U defect states, but, in addition, postulates the existence of an even larger density of negative-U defects, especially dangling bonds. The authors report some new experimental results which might help resolve part of the discrepancy between these two models.