ALBERT

Notes: Capacitance-voltage (C-V) analysis of high quality MBE grown quantum well samples shows that carrier distributions are averaged over the scale of the Debye length. This averaging process results in strongly temperature dependent C-V-deduced doping distributions that can be very different from the actual ones. The doping distribution of the structure is obtained by fitting numerically simulated curves to the measured C-V curves and doping profiles, respectively. Although the calculations do not take quantum size effects into account they show good agreement with the measured data. © 1996 American Institute of Physics.

Notes: In a recent paper we measured the lateral hole diffusion in a GaAs/AlGaAs single quantum well (SQW) by a novel method. At helium temperature, we estimate a lateral hole diffusion length in the QW of 1.5 μm. However, the assumption that diffusion takes place mainly in the SQW needs to be checked, as the measured diffusion length is the result of two competing processes: (i) hole diffusion in the SQW plane itself and (ii) hole diffusion in the barrier followed by recombination in the SQW. We present here a comparison between the lateral hole distribution in the SQW and in the AlGaAs barrier. First, we estimate the hole diffusion length in the barrier fitting experimental cathodoluminescence linescans on simulated ones. Second, using the measured diffusion lengths in the QW plane and in the bulk barrier and modeling the carrier transport, we deduce the lateral hole distribution in both layers. It is found that even for very large barriers (1.2 μm), the hole diffusion in the barrier contributes less than 0.1% of the total lateral hole diffusion. The lateral transport is mainly carried by holes in the QW (2D diffusion) due to their confinement in the well.

Notes: We report a clear evidence of bistability in the current–voltage characteristics of p-i-n heterostructures containing InGaAs V-shaped quantum wires. The observed phenomenon is explained in the framework of a single carrier transport model in which the quantum wires act like traps for the vertical current. The charge trapping phenomenon is indeed demonstrated by temperature-dependent photocurrent experiments. © 1996 American Institute of Physics.

Notes: Identical GaAs/Al0.2Ga0.8As multiple-quantum-well (MQW) structures uniformly doped with Si at various concentrations ranging from 1×1017 to 1×1019 cm−3 are grown by molecular-beam epitaxy to study the effects of the background Si-doping level on the Zn diffusion-induced disordering process. After Zn diffusions at 575 °C for 4 and 16 h, the structures are investigated by secondary-ion-mass spectrometry, and by transmission electron microscopy on cleaved wedges of the sample. The results show that the totally and partially disordered regions are always behind the Zn diffusion front. A dependence of the effective Zn diffusivity and of the disordering rate of the structures on the background Si-doping level is observed. The effective Zn diffusivity and the disordering rate are significantly reduced with increasing background Si concentration. Before Zn diffusion, photoluminescence spectra of the Si-doped MQW structures exhibit an increase in intensity of the Si donor–column-III vacancy complex emission band with increasing Si-doping level. This indicates that the concentration of column-III vacancies in the MQW structures increases as the background Si concentration increases. After Zn diffusion, an important decrease in intensity of the column-III vacancy related emission band is observed on the photoluminescence spectra taken in the Zn-diffused regions.The systematical analysis of the photoluminescence spectra of the Zn-diffused MQW structures as a function of diffusion time and as a function of etching depth below the sample surface makes it possible to describe the physical processes occurring during Zn diffusion. A model based on the "kick-out'' mechanism of Zn diffusion is proposed to explain the effect of the background Si-doping level on the effective Zn diffusivity. The model shows that the effective Zn diffusivity is controlled by the concentration of column-III interstitials behind the Zn diffusion front and by the donor concentration in the sample. During the incorporation of Zn into the crystal lattice, column-III interstitials are generated. The supersaturation of these interstitials behind the Zn diffusion front is responsible for the enhancement of Al–Ga interdiffusion. Since column-III interstitials and column-III vacancies can mutually annihilate, the concentration of column-III interstitial and column-III vacancy in the Zn-diffused region is reduced with increasing Si-doping level, leading to a retardation of Zn diffusion into the MQW structure. On the other hand, a decrease of the effective Zn diffusivity caused by an increase in donor concentration in the samples is also demonstrated.Our results give evidence for the Fermi-level effect and the interactions between different point defects during Zn diffusion-induced disordering of GaAs/AlGaAs multilayered structures. © 1996 American Institute of Physics.

Notes: This work aims at a systematic study of phase modulation in GaAs/AlGaAs double-heterostructure waveguides with different doping profiles. Both theoretical (part I) and experimental (part II) aspects are investigated, leading to interesting new results. Phase modulation is the sum of a linear electro-optic term, a quadratic electro-optic term, and a free-carrier term. The carrier term is shown to be the sum of a plasma term, an intervalence-band term (for holes only), a band-filling term, and a band-shrinkage term, the latter being due to many-body effects. A new analytic expression for the band-filling term is derived which shows that the band-filling effect does not depend on the carrier effective mass. We prove that the band-shrinkage term is approximately half of the band-filling term, but has opposite sign. The phase modulation is computed using an overlap integral between the optical intensity and the local refractive-index difference. We also report an analytic expression for the modulation efficiency of a p-i-n junction double-heterostructure modulator. This expression is very accurate and only requires knowledge of the depletion width and the guiding parameters.

Notes: Phase modulation in GaAs/AlxGa1−xAs double heterostructures with different doping profiles is systematically investigated. Very good agreement between the experimental measurements and the theory developed in Part I of this paper is reported. By measuring the phase modulation for both transverse electric and transverse magnetic modes along the [110] and [11¯0] crystallographic directions, we are able to deduce accurate values for the linear electro-optic coefficient. Values of r41=−1.68×10−10 cm/V at λ=1.15 μm and r41=−1.72×10−10 at λ=1.09 μm are obtained with an estimated accuracy of ±5%. An accurate estimation of the carrier effect permits us to deduce the quadratic electro-optic coefficient for GaAs. The values are R11=−2.0×10−16 cm2/V2, R12=−1.7×10−16 cm2/V2 at λ=1.15 μm, and R11=−2.9×10−16 cm2/V2, R12=−2.4×10−16 cm2/V2 at λ=1.09 μm with an estimated uncertainty of ±25% for all values. Contrast measurement of Fabry–Perot fringes enables us to evaluate the modulator losses. A value of 4.8 cm−1 is reported for a P-n doped modulator (n=6×1017 cm−3). Free-carrier absorption is shown to be the dominant loss process in high-quality structures.

Notes: Careful investigation of the reflectivity of two very high finesse integrated Fabry–Perot interferometers is reported. These two structures, made of GaAs active layer (1.7 μm thick) surrounded by two superlattice/AlAs Bragg reflectors, exhibit vertical cw lasing action at and above room temperature when photopumped with thresholds of 16 mW at 300 K and 56 mW at 380 K. Reflectivity measurements together with theoretical calculations show that layer regularity, accurate thickness control, and low interface roughness are key parameters for high-performance structures. Transmission electron microscopy on cleaved wedges and reflection electron microscopy are shown to be unique tools for measuring and characterizing these layers. Electron microscopy, optical reflection, and laser linewidth measurements are correlated and show that the layer flatness is dramatically increased by the introduction of six (2.5 A(ring)) GaAs wells in the AlAs growth of the integrated dielectric reflectors. Reflectives of 97%, Fabry–Perot finesse as high as 100 (61 for direct measurements), and laser linewidths as small as 1.2 A(ring) are reported.

Notes: The effects of background n- and p-type doping on Zn diffusion in GaAs/AlGaAs multilayered structures are investigated by secondary-ion-mass spectrometry and photoluminescence measurements. Zn diffusions are performed at 575 °C into Si-doped, Be-doped, and Si/Be-codoped identical GaAs/Al0.2Ga0.8As multiple-quantum-well structures. The results obtained by secondary-ion-mass spectrometry show that the Zn diffusion induces an enhancement of Be out-diffusion and the disordering of all structures. The effective Zn diffusivity and the disordering rate are increased by Be doping and reduced by Si doping. Photoluminescence measurements give information about the reactions of different point defects during the diffusion process. Before Zn diffusion, the Si-doped structures contain a high concentration of column-III vacancies, whereas As vacancies are the dominant defects in the Be-doped structures. After Zn diffusion, we observe a reduction of column-III vacancy concentration in Si-doped structures and an increase of column-III interstitial concentration in Be-doped structures. A model based on the "kick-out" mechanism of Zn diffusion is proposed to explain our observations. The supersaturation of column-III interstitials behind the Zn diffusion front is responsible for the enhancements of Al–Ga interdiffusion and Be out-diffusion. The effective Zn diffusivity is controlled by the background donor or acceptor concentration ahead of the Zn diffusion front and by the concentration of column-III interstitials behind the Zn diffusion front. For Be-doped structures, the increase in the background acceptor concentration and the supersaturation of column-III interstitials in the Zn-diffused region results in an enhancement of the Zn diffusivity. For Si-doped structures, the effective Zn diffusivity decreases with increasing background donor concentration. Moreover, the concentrations of column-III interstitials and column-III vacancies in the Zn-diffused region are reduced due to their mutual annihilation, leading to a retardation of Zn diffusion. © 1999 American Institute of Physics.

Notes: The linear (r41) and quadratic (R11 and R12) electro-optic coefficients of GaAs have been measured at the wavelengths 1.32 and 1.52 μm using single-mode AlGaAs/GaAs strip-loaded waveguides. The results are r41=−(1.54±0.08)×10−10 cm/V, R11=−(9.3±2.8)×10−17 cm2/V2 and R12=−(5.1±l.9)×10−17 cm2/V2 at 1.32 μm and r41=−(1.50±0.08)×10−10 cm/V, R11=−(3.2±2.3)×10−17 cm2/V2 and R12=−(5.1±2.6)×10−17 cm2/V2 at 1.52 μm. These results, together with those previously measured at different wavelengths, are compared with published theoretical models. The linear electro-optic coefficient r41 is slightly dispersive near the band gap and roughly constant for longer wavelengths, in good agreement with the models. The quadratic electro-optic coefficients drop rapidly with increasing wavelength, in good agreement with a model based on electroabsorption.

Notes: The mechanism of silicon diffusion in GaAs, Al0.3Ga0.7As, and the silicon diffusion-induced layer disordering of multiquantum wells have been studied by photoluminescence, secondary-ion-mass spectroscopy, and transmission electron microscopy across a corner of a wedge-shaped sample. The diffusion source was a grown in highly Si-doped layer. The main photoluminescence properties of point defects in GaAs and Al0.3Ga0.7As are reviewed to interpret the experimental data. The depth profile of the photoluminescence allows the spatial correlation between the luminescence spectra and the Si concentration profile obtained from secondary-ion-mass-spectroscopy measurements. On the basis of the photoluminescence results, the physical processes occurring during the Si diffusion are discussed. Frenkel defects (pairs of element-III vacancies and interstitials) are generated in the highly Si-doped region. The element-III interstitials rapidly diffuse towards the surface where they react with the element-III vacancies generated at the surface when annealing is performed in an external As pressure. This induces a supersaturation of element-III vacancies in the Si-doped region which drives the Si diffusion. Annealing in vacuum reduces the oversaturation of element-III vacancies and, hence, reduces the Si diffusion. A domination of the Si donor–element-III vacancy complex emission band was found in the spectra taken in the Si-diffused region. This gives evidence for the vacancy-assisted mechanism in the Si diffusion and in the impurity-induced disordering.