This thesis presents a semi-classical density of states analysis of amorphous semiconductors. This analysis, cast, within an effective mass setting, provides a means of evaluating the density of states by averaging a local density of states over Gaussian distributions of conduction band and valence band potential fluctuations. The results span the transition from the tail states to the band states, and both analytical and numerical results are obtained. Applying this formalism to the case of hydrogenated amorphous silicon, we find that our results are consistent with those of experiment. For hydrogenated amorphous silicon we find that potential fluctuation effects play the dominant role in shaping the distribution of states, the kinetic energy of localization playing a lesser role.

We then use this semi-classical approach to study how the distribution of states is shaped by the distribution of potential fluctuations. We consider a broad class of exponential distributions of potential fluctuations, and find that the tail states are produced at the expense of band states. As each state lost from the band is given to the tail, we find that while the distribution of tail states is relatively heavily influenced by the distribution of potential fluctuations, the distribution of band states is not.

Finally, we use the results of our density of states analysis to determine the functional form of the optical absorption coefficient. The analysis focuses on determining the form of the joint density of states function, whose functional features dominate those of the optical absorption spectrum. To determine the joint density of states, we first determine its form in a local region, and then average over spatially correlated Gaussian distributions of conduction band and valence band potential fluctuations. We find that while the breadth of the absorption tail is a strong function of disorder, the mean energy gap is essentially independent of the amount of disorder.