The effects of the temperature shift of Fermi level $E_F$ on the conductances of hydrogenated amorphous silicon (a-Si:H) were studied by model calculations, with the assumption that a single density-of-states distribution is valid in a sample.
A numerical method was devised to calculated the temperature-shift of Fermi level at any temperature with any density-of-states distribution. Using this method, the conductivities of a-Si:H were calculate with two types of experimentally determined density-of-states model with the negatively charged dangling-bond band located at 0.9eV below the mobility edge $E_c$ can explain, at least in part, some features of anomalous transport phenomena of n-type samples, such as the kinks or bendings of log conductivity vs. inverse temperature curves and the Meyer-Neldel rule. No matter what type of density-of-states was chosen, the shift of Fermi level has been found to be generally nonlinear, and large in magnitude, since the density-of-states in a-Si:H is a rapidly varying function of energy. So one should always take the shift of Fermi level into account, in the analysis of transport data in a-Si:H.
Related to this point, the validity of the zero-temperature statistics in a-Si:H was discussed. It was explicitly shown that one would get erroneous values of density-of-states, if one used the zero-temperature statistics in obtaining it from experimental data.
Finally, the temperature-dependence of field-effect conductance was simulated. To do this, a numerical method was developed, for the first time, which can calculate the shift of Fermi level and the variation of potential profile with temperature, when bands are bent. The apparent pre-exponential factors were found to satisfy the Meyer-Neldel rule with the slope similar to experimental values, although it was assumed that a single conductivity prefactor is valid in a sample. Thus it is concluded that the Meyer-Neldel rule may not be a local property, as some scientists though...