We review how the tight-binding method provides a particularly useful approach to understand the electronic structure of GaInNAs alloys, and use it to derive a modified k•p model for the electronic structure of GaInNAs heterostructures. Using the tight-binding model, we first confirm that N forms a resonant defect level above the conduction band edge in Ga(In)As. We show that the interaction of the resonant N level with the conduction band edge accounts for the strong bandgap bowing observed in GaInN x As 1−x , in agreement with experimental analysis but contrary to some theoretical interpretations. We then use a Green function model to derive explicitly the two-level band-anti-crossing model describing the interaction between the resonant states and the conduction band edge in ordered Ga(In)N x As 1−x . We extend the Green function model to show that the conventional k•p model must be modified to include two extra spin-degenerate nitrogen states, giving a 10-band k•p model to describe the band structure of GaNAs/GaAs and related heterostructures. We describe how this 10-band model provides excellent quantitative agreement with a wide range of experimental data and finally discuss briefly the effects of disorder on the electronic structure in dilute nitride alloys.

Dilute nitrogen alloys of InSb exhibit strong band gap bowing with increasing nitrogen composition, shifting the absorption edge to longer wavelengths. The conduction band dispersion also has an enhanced nonparabolicity, which suppresses Auger recombination. We have measured Auger lifetimes in alloys with 11 and 15 m absorption edges using a time-resolved pump-probe technique. We find the lifetimes to be longer at room temperature than equivalent band gap Hg 1Ϫy Cd y Te alloys at the same quasi-Fermi level separation. The results are explained using a modified k"p Hamiltonian which explicitly includes interactions between the conduction band and a higher lying nitrogen-related resonant band. © 2001 American Institute of Physics. ͓DOI: 10.1063/1.1355301͔When a small fraction of anion atoms are replaced by nitrogen in GaInAlAs the band gap initially decreases rapidly, by between 0.05 and 0.2 eV at 1% N. [1][2][3][4] This opens the possibility of long wavelength applications in dilute nitride alloys. Despite the decrease in band gap, the conduction band ͑cb͒ edge effective mass has been observed to rise with increasing N content, 5,6 contrary to the prediction of usual k"p theory. The strong band gap bowing and increase in effective mass arise because of a repulsive interaction between the conduction band edge and a higher lying band of nitrogen resonant states. [7][8][9] The band structure of InN x Sb 1Ϫx can be predicted using a modified k"p Hamiltonian previously developed for GaN x As 1Ϫx . 7,8 In this model, the Hamiltonian describing the interaction between the cb edge and nitrogen resonant level is given by ͑1͒The parameters ␣, ␤, ␥, and the energy of the nitrogen resonant level, E N , can be found by analyzing the results of a tight binding model we have developed. 9 In the case of InN x Sb 1Ϫx , the cb energy E C ϭ0.177 eV (ϵ7 m), and we calculate E N ϭ0.647 eV, ␣ϭ␥ϭ0.77 eV, and ␤ϭ2.2 eV. The band gap is therefore predicted to decrease to 110 meV at 1% N, a fractional change of almost 40%. This clearly offers significant potential for long wavelength emission. III-V materials may be made with band gaps smaller than hitherto possible, opening up a longer wavelength region for direct interband transitions.Auger recombination ͑see Fig. 1͒ is the main recombination scattering process in intrinsic and n-type Kane band structure materials with band gap less than about 1 eV and, hence, is the primary factor limiting the maximum operating temperature of emitter and detector devices in this energy range. The Auger lifetime varies as

Picosecond time-resolved far-infrared measurements are presented of the scattering between conductionband states in a doped quasi quantum dot. These states are created by the application of a magnetic field along the growth direction of an InAs/AlSb quantum well. A clear suppression of the cooling rate is seen, from 10 12 s Ϫ1 when the level spacing is equal to the phonon energy, to 10 10 s Ϫ1 away from this resonance, and thus the results provide unambiguous evidence for the phonon bottleneck. Furthermore, the lifetimes had only weak dependence on temperature between 4 and 80 K. ͓S0163-1829͑99͒50612-7͔Electronic lifetimes of two-dimensional ͑2D͒ systems in magnetic fields are of fundamental interest in part because the quantization effect of the magnetic field mimics the effect of a quantum dot potential with an easily variable degree of confinement. 1 The magnetic field perpendicular to the layers forces the electrons into confined orbits and the density of states becomes a ladder of broadened ␦ functions similar to that of a quantum dot. Recently much work has been carried out on the so-called ''phonon bottleneck'' that has been claimed to inhibit the cooling of carriers in quantum dots when the level separation is not equal to the phonon energy. [2][3][4][5][6] However, partly as a result of different groups using different growth techniques for interband photoluminescence samples and partly on fundamental grounds, this is controversial and is the subject of much debate. 7-12 Indeed several mechanisms have been proposed that may bypass the bottleneck, such as multiphonon scattering, 7 Auger processes, 8 excitonic effects, 9 and defect related processes. 10 In the present work we observe clear phonon suppression in n-type quasi 0D dots ͑i.e., Landau quantized rather than spatially quantized͒ by a time-resolved intraband absorption measurement. This provides unambiguous evidence for the phonon bottleneck effect independently of arguments concerning which processes dominate in the interband photoluminescence measurements in dots 3-6,9-12 and quasi dots 13 such as electron-hole scattering. Further, because of the very clean model system ͑much sharper interfaces and no wetting layer, etc.͒, the interpretation is not complicated by detailed questions about different growth techniques and the quality of different dot sample structures. The results should assist in the understanding of which aspect of these processes is fundamental and which is dependent on sample structure.Resonant absorption of longitudinal optic ͑LO͒ phonons causes a variety of transport properties to oscillate with applied magnetic field, such as the magnetoresistance. 14 Resonant phonon scattering occurs whenwhere ប LO is the LO phonon energy, ប c ϭបeB/m* is the cyclotron energy, and ⌬l is an integer. At these resonances the LO phonon absorption/emission probability is strongly enhanced giving rise to large changes in the electron energy relaxation lifetime. 15,16 In the present work we have used the pump-probe technique, with cyclotron resonan...

We report photomodulated reflectance measurements of several intersubband transitions for a series of as-grown In y Ga 1Ϫy As 1Ϫx N x /GaAs multiple quantum well samples as functions of hydrostatic pressure ͑at room temperature͒ and temperature ͑at ambient pressure͒. The experimental results provide support for the effects of disorder due to different nearest-neighbor N-cation configurations. The quantum well transition energies obtained from the photomodulated reflectance spectra are fitted as a function of pressure with a realistic 10 band k•p Hamiltonian, that includes tight-binding-based energies and coupling parameters for the N levels. The quality of match between theory and experiment confirms the theoretical model and predicts some important material parameters for dilute-N InGaAsN alloys.

Cyclotron resonance (CR) and quantum transport measurements are performed on three GaAs/Ga 0.67 Al 0.33 As heterostructures in the quantum Hall regime at T = 2 K. The relaxation time and the 2D electron density N s are determined fitting the CR transmission curves by the Drude-type model. In all samples the density N s exhibits oscillations as a function of magnetic field. The dc quantum transport is also studied on the same structures and, assuming that the Hall resistance determines the electron density at all fields, one obtains density oscillations similar to those measured by CR. This density behaviour is modelled using a reservoir hypothesis.