Table of Contents:
This Work is giving an overview of recent developments in the field of metastable III-V semiconductors where a fraction of the anions of the host is replaced by an isoelectronic impurity. The large differences in size and electronegativity between the isoelectronic impurity and the host anions are responsible for the dramatic band-structure changes observed already at very low concentrations of the isoelectronic impurity. The first part of the work dealt with the electronic band structure of bulk GaNxAs1-x. The band formation from the very dilute N-regime (i.e. doping levels) to N-contents of several percent was discussed. The intriguing coexistence between N-localized states and the extended band states of the host is addressed. The N-related localized states persist up to N-concentrations of almost a percent whereas the N-induced shift of the band gap of the GaNxAs1-x alloy sets in at much lower N-contents. A new N-induced band is detected for x=0.21% which shifts to higher energies with increasing x. All the higher conduction bands are almost unaffected by the N-incorporation and the corresponding transitions are to a first approximation the same as in GaAs, but considerably broadened. The red shift of the alloy band-gap and the blue shift of the new N-induced band can be qualitatively described as a level repulsion between the unperturbed band edge of the host state (e.g. GaAs) and the localized isoelectronic impurity level (e.g. N) within the conduction band.
The strong correlation between the local environment of the isoelectronic impurity and the global band structure was explored further in the second part of the work. The change of the band structure of GaNxAs1-x due to hydrogenation as well as the band structure changes of quaternary Ga1-yInyNxAs1-x due to thermal annealing were discussed. In the former case, hydrogen almost solely binds to nitrogen forming various N-Hn complexes whereas the host atoms are not affected by the hydrogenation. The complex formation compensates the size difference between N and As and satisfies the desire of N for electrons, thus basically diminishes the perturbation of the host lattice by the N-atom. Consequently the observed electronic band structure of fully hydrogenated GaNxAs1-x resembles again that of GaAs i.e. the E+ has disappeared and the E- band shifts back to the energy of the GaAs band gap. In the case of thermal annealing of Ga1-yInyNxAs1-x, the situation is similar. At moderate annealing conditions, it is possible to cause a rearrangement of the local N-environment from Ga-rich environments to In-rich environments by hopping of N via As-vacancies. In In-rich environments the local perturbation of the lattice due to N is reduced compared to Ga-rich environments because In-atoms are larger than Ga-atoms and the In-N bond is weaker than the Ga-N bond. Under these conditions, this reduction of the perturbation is again manifested in a blue shift of the Ga1-yInyNxAs1-x band gap towards that of the Ga1-yInyAs host on annealing.
The main focus of the third part of this work lay on studies of the electronic states in GaNxAs1-x/GaAs heterostructures and suitable models for describing them. The band-anticrossing model in its simplicity accounts already for a large number of experimental observations in Ga(N,As) and (Ga,In)(N,As). In the context of QWs, these are in particular the strong nonparabolicity of the conduction-band dispersion and the strong dependence of the electron effective mass on N-content. Therefore, it suggests itself to modify existing k·P-models for describing the electronic states of conventional III-V containing QW structures by combining them with the band-anticrossing model. So the existing 8 band k·P-model was extended by including two additional spin degenerate N-related states which couple directly to the G6 conduction-band states. By comparison of the model with experimental data for interband transitions of various Ga(N,As)/GaAs QWs a set of material parameters was derived. This parameter set, in conjunction with the model, allows a prediction of the electronic states of any GaNxAs1-x/GaAs with 1%