Identifying Electronic Properties Relevant to Improving the Performance and Stability of Amorphous Silicon-Based Mid-Gap and Low-Gap Cells: Annual Subcontract Report, 16 January 1998 - 15 January 1999

An overriding theme of the work described in this report has been the effect of partial crystallinity, or the approach to partial crystallinity, on the electronic properties of a-Si:H. This includes, of course, how degradation or the relative stability of these films is affected by the approach to, or onset of, microcrystallinity. We first discussed the results on a set of samples produced by dcreactive magnetron sputtering, obtained in collaboration with John Abelson's group at the University of Illinois, for which we demonstrated the existence of a small, but significant, microcrystalline component. For these films, the degradation kinetics was found to be quite unusual; however, it could be well accounted for by a model that postulated two phases of degrading material. One wasa-Si:H host material of good quality and the other was a more defective component associated with boundary regions near the microcrystallites. Our sub-band-gap photocapacitance measurements on these films also indicated the existence of a distinct feature (a 'shoulder' with a threshold near 1.1 eV) that signaled the presence of the microcrystalline phase. The second set of samples investigatedwere produced by Uni-Solar, deposited under conditions of high hydrogen dilution, very close to but just below the microcrystalline phase boundary. Here we found that the defect density following light-induced degradation decreased as the film thickness increased. Corroborating our findings with X-ray diffraction results obtained by Don Williamson on sets of similar films, we concluded that thefilms were becoming more ordered and less defective just prior to the onset of a detectable microcrystalline component. Furthermore, we found that at almost exactly the conditions that Williamson found XRD evidence for the onset of microcrystallinity, we found the appearance of the distinctive 'shoulder' in our sub-band-gap photocapacitance spectra. Third, we investigated two sets of sampleswhere the deposition rate had been varied to include samples grown at moderate to high rates. In one set of samples, produced at ETL, samples deposited under H2 dilution at 10 ?/s were found to exhibit extremely low deep defect densities and narrow Urbach tails, indicating films of exceptional quality. The photocapacitance spectra for these films were found to contain evidence for a small degreeof microcrystallinity. In another set of samples, produced at UniSolar, we found evidence for increasing defect densities plus somewhat larger Urbach energies for the films deposited at higher rates. This is consistent with the fact that the photovoltaic device performance is significantly poorer for the higher deposition rate material. Finally, we discussed the general issue of deep defectdensities in the a-Si,Ge:H alloys. We again demonstrated how well the deep defect densities in such samples from several sources could be fit using the spontaneous bond-breaking model of Martin Stutzmann. This implies that such state-of-the-art alloy films have been optimized in a quantifiable sense. We also found that the increase in deep defect density with small amounts of P and B dopantscould also be reproduced reasonably well by modifying the spontaneous bond-breaking model to include the extra energy terms associated with charged defects.