A critical assessment of different models of the metastability in a-Si:H.

We have studied the light degradation and the thermal annealing of a-Si:H at different light intensities and temperatures monitoring both the defect density and the photoconductivity. This large set of data is used as a benchmark for the models of the metastability in a-Si:H. We show that an entire class of these models (which predict stretched exponential kinetics) is not in good agreement with our data. We discuss the mathematical and physical basis of these models pointing out some weak points that could explain these failures. The data are in better agreement with the “recombination induced bond breaking" models if an accurate description of annealing processes is included. The annealing rate must be computed considering a broad distribution of energy barriers. It must be also taken into account that the energy barrier for the annealing of a metastable defect can be reduced by a monomolecular trapping event at the defect itself. Our calculations show that this reduction amounts to an energy of about half the gap and lasts for a time of the order of 1 ps.