Kwyro Lee

Current—Voltage and capacitance—Voltage characteristics of modulation-doped field-effect transistors

A model describingI-VandC-Vcharacteristics of modulation doped FETs is developed and used to predict the performance of AlxGa1-xAs/GaAs FETs in good agreement with our experimental results. It is shown that the change in the Fermi energy with the gate voltage changes the effective separation between the gate and the two-dimensional electron gas by about 80 Å. Current-voltage characteristics were calculated using a two piece as well as a three piece linear approximation for the electron velocity and compared with experimental results. At 300 K, the two piece model overestimates the current predicted by the three piece model only by approximately 10-20 percent. At 77 K, however, the three piece linear approximation for the velocity field characteristic should be used since the electron mobility decreases very abruptly at about 200 V/cm. The effect of the nonlinear source resistance is also discussed along with the gate-to-source and gate-to-drain capacitances, parameters of paramount importance in determining device performance. These capacitances are calculated as functions of gate-to-source and drain-to-source voltages below saturation.

Electron density of the two‐dimensional electron gas in modulation doped layers

The electron density of the two‐dimensional electron gas in modulation doped structures is calculated as a function of the doping density in (Al,Ga)As, the thickness of the undoped (Al,Ga)As layer, the lattice temperature, and other device parameters. The results of the calculation show that the depletion approximation is not accurate enough and that the Fermi–Dirac statistics (rather than the Boltzmann statistics) should be used in the calculation. A simple analytical model which takes these factors into account is shown to be in good agreement with our computer calculations and experimental data. The obtained results may be used to evaluate the maximum intrinsic transconductance and the maximum gate voltage swing for the modulation doped field effect transistors.

Low field mobility of 2‐d electron gas in modulation doped AlxGa1−xAs/GaAs layers

We derive a simple analytical formula for the low field electron mobility which uses the 2‐d degenerate statistics of the 2‐d electron gas. This takes into account the finite width of the depletion layer in (Al,Ga)As for the scattering by remote donors, scattering by the interface charge, and the polar‐optical and acoustic deformation potential and piezoelectric scattering. The largest measured value of mobility is determined by scattering due to interface charge in some cases. The ultimate value of the mobility which may be achieved is limited by the acoustic deformation potential and piezoelectric scattering at about 6.5×106 cm2/V s for an interface carrier density of the 2‐d electron gas ns0 ≂4×1011 cm−2. Our results agree very well with experimental data obtained in our laboratory as well as other laboratories.

Parasitic MESFET in (Al, Ga) As/GaAs modulation doped FET's and MODFET characterization

A charge-control model for n-channel modulation doped FETs (MODFETs) is extended to include the drain-to-source current through the doped (Al, Ga)As layer which becomes important for large positive gate voltages. This parasitic conduction leads to decreased device transconductances at high gate voltages. A unified and complete characterization technique for deducing the parameters of our model is introduced and used for the device characterization. Parameters, e.g., the saturation velocity, two-dimensional gas concentration at equilibrium, thickness of the doped (Al, Ga)As layer, etc., deduced using the model, are in good agreement with the independent calculations and measurements. However, the deduced values of the room-temperature low field mobility of the two-dimensional electron gas are considerably smaller than those measured by Hall effect and in long-gate MODFETs. This model is in good agreement with the characteristics of high-current normally on MODFETs. The maximum measured current swing of 300 mA/mm gate is reported.

Temperature dependence of minority‐carrier mobility and recombination time in p‐type GaAs

The electron mobility in p‐type GaAs, μpe, has been determined as a function of temperature by measuring the common‐emitter cutoff frequency, fT, of an AlGaAs/GaAs n‐p‐n heterojunction bipolar transistor. The base was 0.6 μm thick and it was doped with 4×1018 cm−3 Be. The 300 K value of 1055 cm2/V s and 79 K value of 5000 cm2/V s for μpe are comparable to the previously measured values. The discrepancy with the calculated values is pointed out. The recombination lifetime is also measured as a function of temperature for minority carriers. The results agree reasonably well with the calculated radiative recominbation time.

Persistent photoconductivity in (Al,Ga)As/GaAs modulation doped structures: Dependence on structure and growth temperature

When modulation doped (Al,Ga)As/GaAs heterostructures are exposed to light, the interface electron concentration and in most cases the mobility increases substantially for lattice temperatures below ∼100 K. A substantial part of the increase is persistent and attributed to the excitation of electrons from a donor–vacancy complex in the (Al,Ga)As. The change in the free electron concentration in the (Al,Ga)As is calculated using a recently developed model for electron transfer across a modulation doped heterojunction. The persistent increase in the free electron concentration is used as a measure of the trap concentration. The magnitude of the photoconductivity response decreased with increasing growth temperature in the range 580–750 °C. The dependence on Al mole fraction was studied for 0.16≤x≤1.0 and found to be a maximum for x∼0.3.

Characteristics of modulation‐doped AlxGa1−xAl/GaAs field‐effect transistors: Effect of donor‐electron separation

The dc characteristics of modulation‐doped AlxGa1−xAs/GaAs field‐effect transistors have been studied experimentally and theoretically to determine the effect of the thickness of the undoped AlxGa1−xAs spacer layer commonly left at the heterointerface. Increasing the thickness of the spacer layer decreases charge transfer and increases mobility. Current transport in short channel transistors, however, is limited by the electron saturation velocity which is independent of the spacer thickness. Due to increased charge transfer, decreasing the spacer thickness from 100 to 20 A doubled the maximum saturation current and transconductance. This should allow faster switching speeds to be obtained. A maximum current of 24 mA was obtained for a gate width of 145 μm with a 40‐A spacer and a transconductance of 250–275 mS/mm was obtained for a device with a 20‐A spacer. Theoretical results indicate that intrinsic transconductances greater than 900 mS/mm are possible. Preliminary small‐signal rf measurements indicate...

Modeling the transition region of a symmetric step junction

The authors are indebted to Dr.B.L.Grung for several helpful suggestions concerning this paper.

The kinetics of the oxide charge trapping and breakdown in ultrathin silicon dioxide

The kinetics of carrier trapping and breakdown in oxides of less than 5 nm was studied. It was found that electron trapping was negligible, but hole trapping was relatively high. An effective oxide trap density due to tunnel annealing was proposed. The rate equation of carrier trapping in the bulk oxide was presented in connection with the generated hole injection by anode surface plasmons. The voltage variation during a constant current test was analyzed using the hole trapping model and capacitance‐voltage measured interface trap generation, and approximate values for the capture cross section and hole generation rate were extracted. The gate voltage shift rebounded after 30–100 C/cm2 electron fluence due to interface trap generation. In ultrathin oxides hole trapping causes breakdown, and that trapping is mainly developed in localized weak areas. Using the weak area breakdown model we found that the ratio of weak to robust area is about 5%.

The influences of traps on the generation-recombination current in silicon diodes

Abstract The recombination current in the space charge region of an abrupt Si P - N junction is calculated under the assumption that the electrochemical potentials or quasi-Fermi levels for the majority carriers are constant throughout the bulk and space-charge regions, and that the Hall-Shockley-Read centers are distributed uniformly with respect to position in the space-charge region, but can be either a single, discrete level or a continuous and uniform distribution with respect to energy. The results for the single level show that the slope can have constant value over a wide range of applied forward bias. These values are mainly dependent on the trap level position and are nearly independent of the minority carrier lifetime at a given doping concentration. They are also dependent on the lower doping concentration. Furthermore, very good linearity in the characteristic curve can be obtained if we assume that the recombination centers are uniformly distributed in energy throughout the energy gap.