Emil S. Koteles

In situ measurements of critical layer thickness and optical studies of InGaAs quantum wells grown on GaAs substrates

Reflection high‐energy electron diffraction (RHEED) intensity oscillations have been used during molecular beam epitaxy (MBE) to accurately determine threshold layer thicknesses for two‐dimensional (2D) growth of InxGa1−xAs on GaAs for a wide range of substrate temperatures and indium compositions. InxGa1−xAs/GaAs single quantum wells were also grown by MBE and studied using low‐temperature photoluminescence (PL) spectroscopy. PL peak energy, intensity, and linewidth measurements provided information on the critical layer thicknesses for the formation of dislocations which, under our experimental conditions, were the same as the threshold layer thicknesses for 2D growth measured from the damping behavior of RHEED intensity oscillations.

GaAs/AlGaAs quantum‐well intermixing using shallow ion implantation and rapid thermal annealing

Low‐energy As+‐ion implantation followed by rapid thermal annealing (RTA) was utilized to modify exciton transition energies of GaAs/AlGaAs quantum wells (QW). A variety of structures were irradiated at an energy low enough that the disordered region was spatially separated from the QWs. After RTA, exciton energies showed large increases which were dependent on QW widths and the implantation fluence with no significant increases in peak linewidths. The observed energy shifts were interpreted as resulting from the modification of the shapes of the as‐grown QWs due to enhanced Ga and Al interdiffusion at heterointerfaces in irradiated areas. These results are consistent with the model of enhanced intermixing of Al and Ga atoms in depth of the material due to diffusion of vacancies generated near the surface.

Low substrate temperature molecular beam epitaxial growth and the critical layer thickness of InGaAs grown on GaAs

We report on the critical layer thickness of InxGa1−xAs on GaAs grown at low substrate temperatures in a wide range of indium compositions. Compared with ordinary growth conditions, the transition between pseudomorphic and relaxed regions (in the epilayer thickness versus x plane) occurred at higher indium compositions when the growth temperature was lowered. An increase in critical thicknesses for pseudomorphic growth by at least a factor of seven for alloy compositions with less than 45% indium was observed. This was determined by low temperature photoluminescence spectroscopy and transmission electron microscopy measurements on single quantum wells.

Effect of heat treatment on InGaAs/GaAs quantum wells

We report on the effect of furnace annealing on 60‐A‐wide InxGa1−xAs/GaAs single quantum wells (SQWs) in the range of indium composition 0.1≤x≤0.5. Excitonic energy shifts of up to 120 meV were observed after annealing of the samples at 825 °C for 30 min. The fact that these energy shifts were strongly dependent of the indium composition in the well material was consistent with enhancement on the indium diffusion out of the wells associated with the presence of dislocations. The most dramatic changes, as a result of annealing, manifested by strain recovery were observed from the SQW with x=0.3 which as‐grown had a low dislocation density (quantum well thickness slightly exceeding the critical layer thickness for formation of dislocations).

Exciton photoluminescence linewidths in very narrow AlGaAs/GaAs and GaAs/InGaAs quantum wells

A study of the low‐temperature photoluminescence characteristics of very narrow one‐dimensional quantum‐well structures, grown by atmospheric pressure organometallic chemical vapor deposition, is presented. Theoretically predicted narrowing of photoluminescence peaks as quantum‐well widths approach zero was experimentally observed in both AlGaAs/GaAs and strained GaAs/InGaAs samples. The role of such data in determining interface microstructure is discussed.

Reversal of light- and heavy-hole valence bands in strained GaAsP/AlGaAs quantum wells

Abstract We have experimentally determined the magnitude of the light-hole-heavy-hole exciton energy difference as a function of biaxial tensile strain in GaAsP/AlGaAs quantum wells using 5 K photoluminescence excitation spectroscopy. The strain is induced by the addition of phosphorus into the GaAs well layer which decreases its lattice constant so that it is less than that of the AlGaAs barrier material. Under certain conditions, the strain resulting from the lattice mismatch is large enough to reverse the order of the light- and heavy-hole valence bands. We found good overall agreement between the experimentally determined dependence of the light-hole-heavy-hole energy difference on the phosphorus concentration in the well layer and a simple calculation which included the effects of spatial confinement and biaxial tensile strain on quantum well exciton energies.

Determining energy‐band offsets in quantum wells using only spectroscopic data

We have developed an experimental technique for accurately determining energy‐band offsets in semiconductor quantum wells (QW) based on the fact that the magnitude of the ground‐state light‐hole (LH) energy is more sensitive to the depth of the valence‐band well than is the ground‐state heavy‐hole (HH) energy. In a lattice‐matched, unstrained QW system, this behavior causes the energy difference between the LH and HH excitons to go through a maximum as the well width, Lz, increases from zero. Calculations show that the position, and more importantly, the magnitude of this maximum is a sensitive function of the valence‐band offset, Qv, the parameter which determines the depth of the valence‐band well. By using Qv, or alternatively Qc=1−Qv, as an adjustable parameter and fitting experimentally measured LH‐HH splittings as a function of Lz, an accurate determination of band offsets can be derived. However, we further reduce the experimental uncertainty by plotting LH−HH as a function of HH energy (which is, ...

Metal-semiconductor-metal demultiplexing waveguide photodetectors in InGaAs/GaAs quantum well structures by selective bandgap tuning

A two-wavelength demultiplexing metal-semiconductor-metal (MSM) waveguide photodetector has been fabricated using impurity-free vacancy diffusion and partial intermixing of an InGaAs/GaAs strained layer quantum well structure. The importance of growth and process parameters, such as aluminium composition in the cladding layer and the oxygen plasma treatment of the sample during processing, on the related device performance is discussed. This photodetector is a potential candidate for monolithic integration with other optoelectronic devices.<<ETX>>

Magneto‐optical transitions in GaAs‐AlGaAs coupled double quantum wells

Magnetophotoluminescence excitation spectroscopy has been employed to study GaAs‐AlGaAs coupled double quantum wells. The investigations were undertaken at liquid helium temperatures and in magnetic fields up to 15 T applied parallel to the growth direction. In zero magnetic field, under flat‐band conditions, the spectrum is dominated by ground‐state excitonic transitions between electron and hole states with envelope functions of the same symmetry. In a magnetic field, the oscillator strengths of excited excitonic states are enhanced and four series of transitions are observed. The spectral features corresponding to the ground state and the excited states have been analyzed and these results were utilized to determine the exciton binding energies for several excitonic transitions of different subbands in this coupled double quantum‐well structure. A number of additional features are visible at higher magnetic fields; some of these peaks are believed to be associated with 2p exciton transitions between ho...

Time‐resolved photoluminescence study of molecular beam epitaxial growth induced defect lines in GaAs

A time‐resolved photoluminescence study of the ‘‘damage’’ lines often observed in molecular beam epitaxial GaAs shows two recent models of the origin of these lines to be incorrect. Our results show that the p through v lines have identical transient behavior which is significantly different from that of the g line. These measurements support the general conclusion of a recent excitation spectroscopy study that the g line is due to an acceptor bound exciton, but call for a reinterpretation of several of the lines related to g which leads to a new value for the g‐acceptor binding energy.