D. J. Dunlavy

INTENSITY-DEPENDENT MINORITY-CARRIER LIFETIME IN III-V SEMICONDUCTORS DUE TO SATURATION OF RECOMBINATION CENTERS

The minority‐carrier lifetime has been measured by time‐resolved photoluminescence in a variety of III‐V epitaxial material including GaAs and AlxGa1−xAs. In cases where Shockley–Read–Hall recombination is dominant, the measured lifetimes are dependent upon the intensity of the excitation source. These lifetime effects can be described by a Shockley–Read–Hall model that includes the injection dependence of the recombination. As the lifetimes increase with the injection level, we describe the effects as the saturation of recombination centers.

Photoluminescence lifetime in heterojunctions

Abstract It is desirable to obtain minority carrier lifetimes on real photovoltaic structures. However, the photoluminescence (PL) lifetime in such devices is less than the real bulk lifetime because of collection of carriers at the p-n junction. A transient analysis of the PL decay produces lifetime as a parameter. By fitting the data to the model, we can estimate the bulk lifetime in window-absorber heterojunction structures. For ITO/InP devices, there was a strong correlation between such lifetime values and the device efficiency. Similar results were found for n-Al0.9Ga0.1As/p-Al0.37Ga0.63As heterostructures.

Measurement of AlGaAs/AlGaAs interface recombination velocities using time‐resolved photoluminescence

Time‐resolved photoluminescence has been used to examine AlxGa1−xAs/AlyGa1−yAs interfaces, focusing on the recombination velocity. For an Al0.08Ga0.92As/Al0.88Ga0.12As interface, important for solar cells, recombination velocities are about 104 cm/s with the growth conditions used in this study. Several types of interface passivation were attempted, but the most successful was the insertion of thin Al0.14Ga0.86As layers between the other two alloys. Using this technique, a 16‐fold increase (to ∼20 ns) of the minority‐carrier lifetime was measured in a 0.8‐μm‐thick Al0.08Ga0.92As layer in which interface recombination would normally have limited the lifetime to about 1–2 ns. Compositional grading was found to be ineffective at passivating the interfaces.

Photoluminescence studies of CuInSe2: Identification of intrinsic defect levels

Photoluminescence (PL) has been used to determine the emission bands in Se-deficient CuInSe2 single crystals that were either Cu- or In-rich. Data are presented for n- and p-type crystals over the temperature range of 7–300°K. The interpretation of these PL data is based mainly on electron microprobe and Auger electron spectroscopy compositional measurements. The origins of these defect levels, that are used in this interpretation are proposed and discussed.

Minority‐carrier lifetime in n‐Al0.38Ga0.62As

The minority-carrier lifetime in n-Al/sub 0.38/Ga/sub 0.62/As has been investigated by laser-induced photoluminescence. A variety of device structures were used to reduce interface recombination effects, including double heterostructures. Bulk lifetimes of about 18 ns were seen at doping levels of 1 x 10/sup 16/ cm/sup -3/ or less. These data suggest that minority-carrier devices are feasible in high aluminum AlGaAs, contrary to the suggestion of earlier work

Minority-carrier lifetime in n-Al/sub 0. 38/Ga/sub 0. 62/As

The minority-carrier lifetime in n-Al/sub 0.38/Ga/sub 0.62/As has been investigated by laser-induced photoluminescence. A variety of device structures were used to reduce interface recombination effects, including double heterostructures. Bulk lifetimes of about 18 ns were seen at doping levels of 1 x 10/sup 16/ cm/sup -3/ or less. These data suggest that minority-carrier devices are feasible in high aluminum AlGaAs, contrary to the suggestion of earlier work

A new method to analyze multiexponential transients for deep‐level transient spectroscopy

A new technique is introduced to analyze digitally recorded capacitive transients in order to determine the properties of deep states. Using a nonlinear double exponential fitting routine, it is shown that a two‐trap model can be applied to the transient data. We determine the individual trap concentrations and produce two Arrhenius plots. The latter yields the thermal activation energies and capture cross sections of closely spaced traps. The excellent agreement between the new technique and the standard rate window technique is shown via a simulation deep‐level transient spectroscopy spectrum. The new method is applied to Se‐doped AlxGa1−xAs (x=0.19 and 0.27) grown by metal‐organic chemical vapor deposition. The measured results for all deep states including the DX centers agree well with the values published in the literature.

Deep-level transient spectroscopy of AlGaAs and CuInSe2

Abstract A new technique is used to obtain the activation energies and capture cross-sections of closely spaced traps. It is applied to single-crystal p-type CuInSe2 prepared by the vertical Bridgeman method and selenium-doped Al x Ga 1−x As (x = 0.27) grown by metal-organic chemical vapor deposition. Deep centers, which are active in the same temperature range, usually yield overlapping peaks in conventional deep-level transient spectroscopy rate-window spectra. The digitally recorded capacitive transients are analyzed directly by fitting a double exponential to the data. This fitting produces two Arrhenius plots which yield the trap depths of the two defects. This technique revealed two hole traps for CuInSe2 at 423 and 489 meV. Likewise, for AlGaAs, electron traps were discovered at 435, 455 and 368 meV. The first two energy levels were obtained by applying the double exponential fit in the temperature range 214–246 K. The energy level of the third state was the result of a single exponential fitting in the temperature range from 184 to 214 K after subtracting the two known traps from the transient. It will also be shown that the agreement between the new technique and the standard technique is excellent, and considerably more information about the traps is obtained.

Non-linear recombination processes in photovoltaic semiconductors

Abstract The single-lifetime model is commonly used to describe recombination in photovoltaic materials. Here we describe two non-linear processes which affect the applicability of that model. Photon recycling is observed in direct band gap materials such as GaAs. This self-absorption and secondary emission of photons makes the effective radiative lifetime a function of device geometry. The saturation of recombination centers by minority carriers produces light intensity dependent lifetimes when the former are present. These effects need to be considered in device design and modeling.

Minority‐carrier diffusion and recombination in CdZnS/CuInSe2 solar cells

Minority‐carrier diffusion has been investigated in as‐grown CdZnS/CuInSe2 solar cells. Capacitance‐voltage (C‐V) and photoconductivity measurements are combined to determine diffusion and interface recombination processes. Diffusion lengths of about 2 μm and a small interface recombination velocity are determined from biased photoconductivity, biased spectral response, and photoluminescence studies. Photoluminescence studies also indicate that a much lower interface recombination velocity occurs at the CdZnS/CuInSe2 interface than that at a bare CuInSe2 surface.