Library

Notes: A theoretical investigation of the electrical characteristics of GaN/AlxGa1−xN (x is the Al mole fraction in AlGaN) modulation doped field-effect transistors (MODFETs) is carried out. Using a self-consistent solution of Schrödinger's equation and Poisson's equation, relations between the concentration of two-dimensional electron gas (2DEG), the Fermi level in GaN, and the average distance of the electrons from the heterointerface are calculated. A relation between the gate bias and the 2DEG concentration is obtained for a flat quasi Fermi level in AlxGa1−xN. Based on the relation between the 2DEG concentration and the applied gate bias, a model for the drain current and the transconductance of the device is developed. The effects of the n-AlxGa1−xN layer thickness, the spacer i-AlxGa1−xN layer thickness, the n-AlxGa1−xN doping level, the aluminum mole fraction x, and the channel length L on the electrical characteristics of MODFETs are presented. Theoretical results are compared with the recent experimental data, which show striking agreement. The 2DEG concentration is found to be as high as 1013 cm−2, and transconductance as high as 1000 mS/mm. Finally, the effect of the difference in the properties of AlN and GaN, and a device structure that would take advantage of the bending of the quasi Fermi level for electrons in n-AlxGa1−xN, are discussed. © 1996 American Institute of Physics.

Notes: We report a novel metal–insulator–semiconductor (MIS) structure exhibiting a pseudomorphic In0.05Ga0.95As layer on GaAs with interface state densities in the low 1011 eV−1 cm−2. The structure was grown by a combination of molecular beam epitaxy and chemical vapor deposition methods. The hysteresis and frequency dispersion of the MIS capacitor were lower than 100 mV, some of them as low as 30 mV under a field swing of about ±1.3 MV/cm. The 150-Å-thick In0.05Ga0.95As channel between Si and GaAs is found to bring about a change in the minority carrier recombination behavior of the GaAs channel, in the same way as done by In0.53Ga0.47As channel MIS structures. Self-aligned gate depletion mode In0.05Ga0.95As metal–insulator–semiconductor field-effect transistors having 3 μm gate lengths exhibited field-effect bulk mobility of 1400 cm2/V s and transconductances of about 170 mS/mm. © 1997 American Institute of Physics.

Notes: A theoretical method for potential distribution in abrupt heterojunctions (HJs) made of uniformly doped degenerate semiconductors has been developed. The method reduces automatically to that in HJs from nondegenerate semiconductors in the limits of low carrier concentrations. For the development of the method the rigid band approximation of degenerate semiconductors has been considered to be valid. The transport equations of Marshak and Van Vliet [Solid-State Electron. 21, 417 (1978)] and an analytical approximation for the Fermi–Diracintegral of order half by the present author [Solid-State Electron. 30, 713 (1987)] have been employed for the formulation. The average of the scattered experimental data for band-gap narrowing of n-Si, n-Ge, p-GaAs, and n-InP have been fitted to the same form as that for the Fermi–Dirac integral of order 1/2 to ease this formulation. Local electrostatic field and local electrostatic potentials obtained from the formulation reduce to those of Chatterjee and Marshak [Solid-State Electron. 24, 1111 (1981)], Cserveny [Int. J. Electron. 25, 65 (1968)], and Kroemer [J. Appl. Phys. 52, 873 (1981)] under special conditions. It is noted that band-gap narrowing and consideration of Fermi–Dirac statistics represent opposite effects for effective intrinsic carrier concentration and local electrostatic field. At some critical concentration belonging to the degenerate limit of a semiconductor, these two effects cancel the influence of each other on effective intrinsic carrier concentration of the semiconductor and on transition region properties of an HJ. Below this critical concentration, band-gap narrowing rather than a consideration of Fermi–Dirac statistics dominantly influences the device properties. However, above this critical concentration, consideration of Fermi–Dirac statistics dominates over the other. Applications of electrostatic field and electrostatic potential to isotype and anisotype HJs have been discussed. On the basis of present formulas a general form for potential distribution in Schottky barriers has been derived. The relation reduces to that of Gummel and Scarfetter [J. Appl. Phys. 38, 2148 (1967)] under special conditions. Theoretical reasons underlying the lack of rectification in various n-N HJs have been analyzed. In light of this analysis, a theoretical model in terms of many-body electron-electron and electron-donor interactions, and in terms of lowering of band edge in the vicinity of transition region, has been proposed. Numerical results obtained from this model for n-GaAs agree remarkably with observations from experimental measurements.

Notes: An analytical investigation of the space-charge region junction properties of heterojunction semiconductor devices from heavily doped and degenerate semiconductors has been carried out. On the basis of a new formula for Fermi–Dirac integral of order (1)/(2) theoretical formulas for junction boundary conditions, minority-carrier concentrations at the edges of space-charge region and excess minority-carrier concentrations at the edges of space-charge region have been derived. All of these formulas take the spatial dependence of band structures, carrier degeneracy, and band-gap narrowing into account. Under special conditions the formulas reduce to the well-known standard formulas for homojunction devices from both degenerate and nondegenerate semiconductors. The new relation for Fermi–Dirac integral is very highly accurate. Numerical calculations performed on an n-AlzGa1−zAs/p-GaAs (z=0.1) diode indicate that all these parameters significantly influence the junction properties of heterojunction semiconductor devices, and without which theoretical modeling of heterojunction devices with spatially dependent and heavily doped semiconductor regions are likely to involve errors.

Notes: A theory of saturation photocurrent and photovoltage has been developed for p-n junction solar cells. The theory is based on ambipolar transport equations for electrons and holes near the junction, and on empirical models for band-gap narrowing and Fermi–Dirac integrals. It is applicable to solar cells made of nondegenerate or lowly degenerate semiconductors with position dependent band structures. Interestingly, it includes provision of both short-circuit and open-circuit configurations and involves the use of boundary conditions valid at the junction for all levels of injection. The boundary conditions automatically reduce to those of Dhariwal et al. for nondegenerate semiconductors with uniform doping. The empirical models for band-gap narrowing and Fermi-Dirac integrals are found to be significantly accurate when compared with available experiments or with exact results. Numerical calculations have been carried out for a number of silicon solar cells possessing varied doping levels, and the results have been found to be in good agreement with available experiments. The analysis shows that the saturation photovoltage developed by a p+n solar cell is higher than that developed by an equivalent n+p solar cell, and that the photovoltage is lower than the diffusion potential of the corresponding solar cell.

Notes: A three-dimensional analysis of the validity of emitter-collector reciprocity of double heterostructure bipolar transistors with heavily doped and degenerate base regions has been carried out. The analysis is based on current-density relations of Marshak and Van Vliet [Solid-State Electron. 21, 429 (1978)], an effective intrinsic carrier concentration derived in terms of band-gap narrowing and the effect of Fermi–Dirac statistics, and p-n product derived by considering potential barriers and related properties of heterostructures. The analysis indicates that the bipolar reciprocity holds under low-level injection even when parameters such as band-gap narrowing of the band, and mobility, diffusivity, lifetime, and recombination velocity, etc., of the carriers depend on spatial coordinates, and the carrier concentrations are described by Fermi–Dirac statistics. The analysis is general enough to be applicable to both n-p-n and p-n-p heterostructure and homostructure bipolar transistors.

Notes: An approximate formula for electron-hole generation rate has been proposed. The proposed approximation is found to give good representation of the actual generation rate provided by Dunbar and Hauser and by Fossum. In order to determine the applicability of this approximation attempts have been made to calculate short-circuit current, open-circuit voltage, and solar cell efficiency of an n+p silicon solar cell. The calculations involve consideration of nonuniform doping, spatial dependence of the band, and drift field for the n+ region, and band gap narrowing, Shockley–Read–Hall recombination, and Auger recombination for both n+ and p regions. The variations of short-circuit current, open-circuit voltage, and solar cell efficiency with doping concentration of the substrate p region are found to be in good agreement with results obtained from experiments. The trend of the results indicates that proper optimization of the width and impurity concentration of the diffused n region, of the width and impurity concentration of the substrate p region, and of the front-surface recombination velocity is necessary for achieving the highest efficiency of a solar cell.

Notes: A relationship among the resonance energies in superlattices of different numbers of well and barrier layers has been established. It has been shown that the resonance wave vectors can be related to each other in such systems. The calculated dependence of resonance energies and resonance wave vectors on the mole fraction, and of the thickness of barrier layers of superlattices, throws new light on the quantum effects of electron transport through the superlattices. © 2002 American Institute of Physics.

Notes: A metallization scheme has been developed for obtaining low ohmic contacts to n-GaN with a low contact resistance. The metal contact is a Ti/Al/Ti/Au composite with layers that are respectively 30, 100, 30, and 30 nm thick. Contacts with a specific contact resistivity ρs, as low as 6.0×10−7 Ω cm2 for a doping level of 1.40×1020 cm−3 were obtained after annealing the sample for 30 s at 750 °C in a rapid thermal annealer. The Ti placed on top of the traditional Ti/Al contact appears to have the advantage of tying up the excess Al; therefore it does not form a mottled contact. Some of the additional Ti–Al intermetallic alloys that are formed also have beneficial effects. The Ti–Au layer forms a robust upper portion of the composite, which enables the contacts to have high-temperature applications. © 2001 American Institute of Physics.