| Two theoretical aspects of high field transport in III-V semiconductors have been studied. First, a new mechanism to obtain negative differential resistance in a GaAs-AlGaAs multilayered structure is described. The mechanism is based on the transfer of electrons in real space from a high mobility GaAs region to an adjacent low mobility AlGaAs region when a high electric field is applied parallel to the interface. It is analogous in many respects to the Gunn effect, except that this mechanism allows greater control of device characteristics. These characteristics can be adjusted by varying the doping densities, the layer thicknesses, and the Al mole fraction in the AlGaAs.; The mechanism is analyzed using the electron temperature model and the Monte Carlo simulation. The electron temperature model is exact in the high carrier density limit, whereas the Monte Carlo method is valid in the low density limit. Both methods clearly illustrate the degree of control possible with this mechanism over device characteristics. Comparisons are made between the two models. Miscellaneous effects which are neglected in the models are discussed. These include two-dimensional effects, band bending, statistical fluctuation, and quantum mechanical transmission at the interface. Switching characteristics have been analyzed, and the switching time is estimated to be approximately 1 x 10('-12) sec. This fast transfer mechanism is attractive for microwave, switching, and memory devices.; The second portion of the present work deals with the band structure dependence of high field transport and impact ionization in GaAs. An understanding of impact ionization is important because of its influence on the performance of avalanche photodetectors and IMPATT diodes. It also determines the ultimate performance limits of small semiconductor devices, such as CCDs and FETs.; A realistic band structure has been included in a Monte Carlo simulation of high field transport in GaAs. The band structure has been calculated using the empirical pseudopotential method. Partly due to the lack of information and partly for simplicity, we have made simplifying assumptions on the phonon scattering rate, the ionization threshold, and the ionization probability. Unlike previous theories of impact ionization, the method requires, in principle, no adjustable parameters as long as the band structure and the scattering mechanisms are known. The calculated drift velocity, mean free path, and impact ionization rate are in fair agreement with the experimental data. It is found that the contribution of ballistic electrons to the impact ionization rate is negligibly small. Within statistical uncertainty we do not observe the anisotropy of the electron ionization rate in contradiction to the recent experimental results. Based on the results of the simulation, a general discussion of impact ionization is given. |
