R. J. Nelson

Minority‐carrier lifetimes and internal quantum efficiency of surface‐free GaAs

Minority‐carrier lifetimes, internal quantum efficiencies, and values of the radiative recombination coefficient B are determined from photoluminescence time‐decay and external quantum efficiency data taken for LPE GaAs samples (germanium doped or undoped) in the doping range 1.9×1015?p0 ?1×1019 cm−3 at 300 °K. Measurements are made on isotype double heterostructure samples where the optically exicted GaAs layer is bounded by wider‐band‐gap Ga0.5Al0.5As layers which provide for confinement of minority carriers and also minimize the importance of surface or interfacial recombination on the measured lifetimes. Comparison with time‐decay data for samples without the ternary cladding layers shows the dominant effect of surface and substrate recombination on the decay time if confinement layers are not provided. Luminescence decay times are observed from 1.2 nsec for heavily doped samples up to 1.3 μsec for lightly doped samples. Values of the bulk minority‐carrier lifetimes, radiative lifetimes, and internal ...

Long‐lifetime photoconductivity effect in n‐type GaAlAs

A long‐lifetime (τ∼hours, T≲60 °K) photoconductivity effect is observed in Te‐doped Ga1−xAlxAs (0.25≲x≲0.7). Analysis of Hall‐effect data showing a pronounced decrease in the electron mobility upon photoexcitation shows that a donor level is involved. Similar effects are observed in Se‐ and Sn‐doped Ga1−xAlxAs (x=0.3). The magnitude of the effect which is typically of the order of the room‐temperature electron concentration seems to correlate linearly with the concentration of Te, showing that a constant concentration background impurity is not responsible for this effect. A large lattice relaxation is indicated by the large difference between the thermal (0.12 eV) and optical (1.1 eV) ionization energy of the donor level. The potential barrier to electron capture by the donor level is estimated to be 180 meV (x=0.36) from time decay measurements of the photoexcited electron population at low temperatures. Extrapolation to room temperature gives a characteristic decay time of ∼0.5 nsec for the electron co...

The case for Auger recombination in In1−xGaxAsyP1−y

The possible Auger recombination mechanisms in direct‐gap semiconductors are investigated. These include band‐to‐band processes, phonon‐assisted processes, and Auger recombination via shallow traps. The band‐to‐band Auger rates are calculated including Fermi statistics, nonparabolic bands, and screening effects both for n‐type and p‐type semiconductors. The nonparabolicity is calculated using the Kane‐band model. The band‐to‐band Auger processes are characterized by a strong temperature dependence, the Auger rate decreasing rapidly with decreasing temperature. The phonon‐assisted and the trap processes do not exhibit such a strong temperature dependence. This is because the additional momentum conservation for the four‐particle states in band‐to‐band processes gives rise to a ’’threshold energy’’ for the process. For the same reason, the band‐to‐band Auger rate decreases rapidly with increasing band gap. In large‐band‐gap semiconductors the weakly temperature‐dependent phonon‐assisted processes are expect...

Interfacial recombination velocity in GaAlAs/GaAs heterostructures

Photoluminescence time‐decay measurements on Ga0.5Al0.5As/GaAs double heterostructures were made over a wide range of GaAs active layer thickness and doping levels at room temperature. Observed decay times τ in variously doped GaAs samples range from 10 to 450 nsec. Effects of self‐absorption of luminescence and doping level are demonstrated for GaAs layer thickness d≳1 μm. For d<1 μm, the observed decay times are nearly independent of doping level and vary almost linearly with d. The data are interpreted in terms of a small interfacial recombination velocity (Si=450±100 cm/sec) at the Ga0.5Al0.5As/GaAs interface. The value of Si determined here is an average over the doping levels examined.

Reduction of GaAs surface recombination velocity by chemical treatment

Chemisorbed ruthenium ions on the surface of n‐GaAs decrease the surface recombination velocity of electrons and holes from 5×105 to 3.5×104 cm/sec. It is shown that the ions, in a one‐third monolayer thickness, are confined to the surface and do not form a new junction by diffusing into the GaAs. This use of Ru appears to be the first observation of the reduction of the surface recombination velocity for GaAs by the simple chemisorption of ions.

Near‐equilibrium LPE growth of low threshold current density In1−xGaxAsyP1−y(λ=1.35 μm) DH lasers

The growth of In1−xGaxAsyP1−y double heterostructure (DH) laser material by liquid phase epitaxy (LPE) under near‐equilibrium growth conditions which produce small growth rates is described. Broad‐area threshold current densities for this material are as low as 670 A/cm2 for 0.1‐μm active layers which is the lowest value yet reported for this material system. This value is comparable with the best reported value for LPE Ga1−xAlxAs with a similar refractive index step. For comparison, material grown at higher growth rates using the commonly employed two‐phase and supercooling techniques are found to give consistently higher threshold current densities than those grown under near‐equilibrium conditions in the same LPE system. The spontaneous luminescence observed in window lasers grown by the near‐equilibrium method appears uniform with no evidence of dark absorbing regions which could cause self‐pulsations in the laser output during initial operation.

Cyclotron resonance in n‐type In1−xGaxAsyP1−y

First experimental data are presented on the nonparabolicity of the conduction band of In1−xGaxAsyP1−y lattice matched to InP. The values for the effective mass, determined from our energy‐dependent measurements, are smaller than all previously reported values. The nonparabolicity of the conduction band is much stronger than previously assumed.

Frequency chirp under current modulaton in InGaAsP injection lasers

We have measured the frequency chirp in gain‐guided, weakly index‐guided, and strongly index‐guided InGaAsP lasers under direct‐current modulation. The measured chirp width is largest for gain‐guided and smallest for weakly index‐guided ridge waveguide lasers. The chirp width for 1.5‐μm InGaAsP lasers is about a factor of 2 larger than that for 1.3‐μm InGaAsP lasers of the same structure. The frequency chirping results from a modulaton of the carrier density which modulates the effective refractive index of the guided mode. The frequency chirping can introduce a limitation on the performance of single‐frequency injection laser sources for high bit‐rate digital transmission in fiber communication systems at 1.55 μm.

Enhanced indium phosphide substrate protection for liquid phase epitaxy growth of indium‐gallium‐arsenide‐phosphide double heterostructure lasers

Thermal degradation of indium phosphide (InP) single crystal substrates prior to liquid phase epitaxy growth has been virtually eliminated by using an improved protection technique. The phosphorus partial pressure provided by a Sn‐In‐P solution localized inside an external chamber surrounding the InP substrate prior to growth prevents thermal damage to the surface. Nomarski contrast photomicrographs, as well as photoluminescence and x‐ray diffractometric measurements indicate that InP substrates protected by this method suffer a negligible deterioration, in contrast to the results of the more commonly used cover‐wafer method.

Nonradiative recombination in GaAlAs proton‐bombarded stripe‐geometry lasers

Nonradiative recombination processes in GaAlAs double‐heterostructure lasers are examined using luminescence time‐response measurements. Photoluminescence time‐decay data for proton‐bombarded (PB) isotype (p‐type) heterostructures give an effective recombination velocity for the PB region of (7±1) ×105 cm/sec. Electroluminescence decay times of laser devices indicate that the interfacial recombination velocity of the anisotype interface (p‐GaAs/N‐GaAlAs) is 1500±500 cm/sec compared with 400±100 cm/sec for isotype (p‐type) interfaces. The contribution of each important recombination process to the laser threshold current is calculated. For devices in which the PB region penetrates into the active layer, recombination in the PB region and interfacial recombination are the major nonradiative paths.