InGaAs Epitaxial Layers Research Articles

Investigation of Be diffusion in InGaAs using Kick-out mechanism

Be diffusion during post-growth annealing has been investigated in InGaAs epitaxial layers. Kick-out mechanisms considering species charges, build-in electric field and Fermi-level effect have been studied. Several forms of Kick-out mechanism have been implemented in our simulation programs. Experimental concentration profiles obtained by SIMS have been compared systematically with the results of simulations. We have deduced that the Kick-out mechanism Bei0 ⇔ Be− + IIII+ is the dominating diffusion mechanism in InGaAs under our experimental conditions (C0 = 3 × 1019cm−3).

Impurity Gettering Effect in Pr2O3- Added InGaAs Liquid Phase Epitaxy

InGaAs epilayers were grown on semi-insulating (SI) InP substrates by liquid phase epitaxy (LPE) with a rare-earth (RE) compound Pr2O3 added into the growth melt during the epitaxial process for the first time. Without the growth melt baking process, the corresponding Hall measurements indicate that the background concentration of these InGaAs-grown layers will decrease from a value of 1.6×1016 cm-3 to 2.0×1015 cm-3. Their corresponding 77 K mobility also increases significantly from a value of 15321 cm2/ V·s to 32171 cm2/ V·s. The X-ray diffraction measurements, secondary ion mass measurements (SIMS) and photoluminescence (PL) spectra of Pr2O3- added InGaAs epitaxial layers all demonstrate that the grown layers exhibit pure crystalline quality. Finally, InGaAs/InGaAs/InP pin diodes were fabricated, and a smaller leakage current and better response for the Pr2O3- added samples were reported.

Growth and properties of thin InGaAs epitaxial layers on InP

An application of Far Infrared Fourier Transform (FTIR) spectroscopy to estimate the ordering and clustering effects that take place in In 0.53Ga 0.47As layers grown on InP substrate is proposed. The clustering and ordering parameters are evaluated from the temperature phonon behavior, and the difference between the observed optical phonon frequencies and those calculated in terms of the modified random element isodisplacement model, respectively. The influence of “internal” strains, due to the difference in the lattice parameters of the binary compounds InAs and GaAs, has been investigated. Also, the effect of “external” strains has been studied.

Beryllium diffusion in InGaAs compounds grown by chemical beam epitaxy

Be diffusion mechanisms during post-growth annealing have been investigated in p-type InGaAs epitaxial layers. These Be-doped InGaAs layers were grown between two undoped InGaAs layers. Annealing cycles with time durations of 30 s to 3 min and temperatures in the range of 500 - 800 for doping concentrations of , and were performed. It was found that there is no significant Be diffusion during post-growth rapid thermal annealing (RTA) for all doping levels if annealing is performed at 500 and 600 for a time under 1 min. The same observation was made for a Be concentration of with annealing at 700 for 30 s. For a doping level of , significant Be diffusion occurs for temperatures and durations greater than 700 and 1 min respectively. The shapes of the obtained curves represent a `double profile' giving rise to a concave region called a `kink'. It was deduced from the experiments that the Be effective diffusion coefficient is approximately constant for the high region of the concentration profiles and proportional to the square root of concentration for the low part of the experimental curves. Assuming the latter coefficient dependence, an effective diffusivity was substituted into Fick's second law and the resulting differential equation was solved numerically using an explicit finite-difference method. Good agreement was obtained between our depth profiles and the corresponding simulated distributions.

Beryllium diffusion in InGaAs epitaxial layers: a point defect nonequilibrium model

Be diffusion during post-growth annealing has been studied in InGaAs epitaxial layers. To explain the observed concentration profiles, a point defect nonequilibrium model has been proposed. A good agreement has been obtained between experimental and calculated data. The point defect concentration in epitaxial layers during diffusion in InGaAs is also discussed.

A Generalized Model of Beryllium Diffusion in InGaASs Epitaxial Structures Under Point Defect Nonequilibrium Conditions

ABSTRACTBeryllium diffusion during post-growth annealing is investigated in InGaAs epitaxial layers. Indeed, this undesirable diffusion may occur during thermal treatments of InGaAs/lnP Heterojunction Bipolar Transistors (HBT's), which can generate a limitation of frequency performances of these devices. Epitaxial structures have then been grown, one set by Chemical Beam Epitaxy (CBE), and another one by Gas Source Molecular Beam Epitaxy (GSMBE). The post-growth Rapid Thermal Annealing (RTA) was then performed, and Secondary Ion Mass Spectrometry (SIMS) has been used to characterize the Be depht profiles.In parallel with our experimental study, we propose two models of Be diffusion in InGaAs in the case of point defect nonequilibrium. First, a Kick-out Diffusion model considering neutral Be interstitial species and charged point defects has been studied. Then, a Generalized Substitutional-Interstitial Diffusion model based on simultaneous diffusion by Dissociative and Kick-out mechanisms is proposed. Good agreements between experimental depth profiles and simulated curves have been obtained.

P-Type diffusion in InGaAs epitaxial layers using two models: A concentration dependent diffusivity and a point defect nonequilibrium

Be diffusion during post-growth annealing has been studied in InGaAs epitaxial layers. To explain the observed concentration profiles, two models have been proposed. A good agreement has been obtained between experimental and calculated data. The point defect concentration in epitaxial layers during diffusion in InGaAs is also discussed.

Post-growth diffusion of be doped ingaas epitaxial layers: experimental and simulated distributions

Be diffusion during post-growth annealing has been investigated from InGaAs epitaxial layers grown between InGaAs undoped layers. A General Substitutional-Interstitial Diffusion mechanism is proposed to explain the observed concentration profiles and related Be diffusion. The concentration dependent diffusivity has also been covered to perform an improved data fitting of Be diffusion profiles.

Modeling of Be Diffusion in InGaAs Epitaxial Structures

ABSTRACTThe subject of this work is the simulation of Be diffusion during post-growth Rapid Thermal Annealing (RTA) of InGaAs epitaxial layers grown by Chemical Beam Epitaxy (CBE). This diffusion may occur during thermal treatments of InGaAs/InP Heterojunction Bipolar Transistors (HBT's), which contributes to limit the frequency performances of these devices. In order to characterize the Be depth profiles, Secondary Ion Mass Spectrometry (SIMS) has been used. The concentration dependent diffusivity has been covered to perform an improved data fitting of Be diffusion profiles. In a first step, the solid state diffusion mechanisms have been developed, including the Substitutional-Interstitial Diffusion (SID) and, in particular, Kick-out mechanism. To explain the observed concentration profiles and related diffusion mechanisms, a Generalized Substitutional-Interstitial Diffusion model is proposed. A simultaneous diffusion by Dissociative and Kick-out mechanisms is suggested. Good agreements between experimental depth profiles and simulated curves have been obtained.

Be diffusion mechanisms in InGaAs during post-growth annealing

Be diffusion during post-growth annealing has been studied in InGaAs epitaxial layers, grown between two undoped InGaAs layers. To explain the observed concentration profiles and related diffusion mechanisms, a general substitutional–interstitial model is proposed. On the one hand, a simultaneous diffusion by dissociative and kick-out models is suggested and, on the other hand, the Fermi-level effect is used to explain the functional dependence change of the effective diffusion coefficient of beryllium species with its concentration. The concentration dependent diffusivity has also been covered to perform an improved data fitting of Be diffusion profiles.

Interface recombination and photoluminescence efficiency of thick (> 3 μm) InGaAs prepared by LPE with InAs-enriched composition

In 0.97 Ga 0.03As with a room temperature energy gap of 0.375 eV, corresponding to the fundamental hydrocarbon absorption band at 3.32 μm, was prepared by liquid phase epitaxy (LPE). The epitaxial layer thickness was consistent with a diffusion-limited growth mechanism. Since gallium was the minor constituent in the indium-based melt, growth of the InGaAs epitaxial layer was found to be limited by the rate of arrival of gallium at the melt-solid interface. A value of D Ga = 3 × 10 −4cm 2 s −1 (597.5 °C) was found for the gallium diffusion coefficient in the melt. The LPE grown material was characterized as a function of layer thickness using photoluminescence spectroscopy. The peak photoluminescence emission intensity was found to increase with epitaxial layer thickness. This was correlated with a reduction in interface recombination at the InGaAs/InAs heterojunction. The improved surface morphology observed in thicker layers was thought to be related to a decreased threading dislocation density in the thicker samples. In the extrinsic p-type InGaAs material a value of 5.2 × 10 −8 s was calculated for the 80 K radiative lifetime.

Misfit dislocation generation mechanisms in InGaAs/GaAs heterostructures

An experimental investigation of misfit dislocation generation mechanisms at an InGaAs/GaAs heterointerface is reported. InGaAs epitaxial layers were grown by low-pressure organometallic vapor-phase epitaxy on patterned and unpatterned GaAs substrates having etch-pit densities (EPD) of 200, 1400, and 10 000 cm−2. After epitaxial growth, the samples were annealed at temperatures between 650 and 750 °C, and analyzed by optical and transmission electron microscopy. For the range of substrate EPD studied, it was found that the substrate EPD controls the onset of misfit dislocation generation for low-temperature epitaxy (<600 °C) on unpatterned substrates. When epilayers were annealed at 750 °C, the density of misfit dislocations was independent of the substrate EPD. These studies also show that the dominant misfit dislocation generation mechanism for films grown on patterned substrates is nucleation at the growth-mesa edge. The density of preexisting threading dislocations has little influence on misfit dislocation generation for films selectively deposited within 100×100 μm2 growth windows. For selective heteroepitaxy, misfit dislocation generation strongly depends on the crystallographic orientation of the growth-mesa edge.

Stability of phase-separated microstructure in LPE-grown InGaAs epitaxial layers

Phase-separated In 0.53Ga 0.47As epitaxial layers were grown by liquid phase epitaxy on (001) InP substrates. To investigate the stability of the microstructure, two types of Zn addition experiments were carried out: Zn in-diffusion anneals, and growth of Zn-doped layers followed by annealing (pre-doping experiment). The in-diffusion anneal resulted in a homogeneous microstructure, indicating that the phase-separated microstructure is not the ground state in as-grown InGaAs layers. However, annealing of the pre-doped layer has little effect on the microstructure. Arguments have been developed to rationalize these observations.

Effects of V/III ratio on Si-doping in InGaAs lattice-matched to InP grown by organometallic vapor phase epitaxy

Effects of the V/III ratio on Si-doping characteristics in InGaAs lattice-matched to InP were studied using organometallic vapor phase epitaxy (OMVPE). As the V/III ratio decreases from 81 to 2.5, an increase in the carrier density from 2.7×10 16 to 8.2×10 16 cm -3 is observed. A secondary ion mass spectrometry (SIMS) profile shows that the increase in the carrier density is due to an increase in Si concentration in the InGaAs epitaxial layer. Incorporation kinetics of Si plays an important role in Si doping of InGaAs and V/III ratio makes an effect on the incorporation kinetics.

Interfacial defects and morphology of InGaAs epitaxial layers grown on tilted GaAs substrates

Systematic transmission electron microscopy studies of In0.2Ga0.8As layers grown by molecular-beam epitaxy on [001] GaAs substrates and on substrates tilted up to 10° toward [110], [120], [100], [1̄10], and [1̄20] are described. Three different layer thicknesses were investigated: 6, 20, and 40 nm. In the 40 nm layers misfit dislocations were formed which partially relaxed the elastic strain. Three-dimensional growth with an indication of microscale In segregation (dendritic growth) occurred for substrates tilted toward [1̄10] and [1̄20] directions. Smooth growth surfaces were observed when the substrates were tilted toward [110] and [120] showing that the chemistry of the predominant surface growth steps can play an important role in the growth mode and quality of the epitaxial layer. Indium segregation to misfit dislocation cores, the presence of In-rich platelets near the interface, and diffuseness of the interface are considered as evidence for short range migration of In during and after growth.

Phase Separated Microstructure and its Stability in InGaAs Epitaxial Layers Grown by LPE

In 0.53 Ga 0.47 As layers were grown by LPE on (001) InP substrates in the temperature range of 480~780°C. The fine speckle contrast, which is attributed to two-dimensional phase separation, was observed in layers grown at temperatures as high as 780°C. Since the critical temperature for bulk miscibility gap is predicted to be around 450°C, this suggests that in the presence of the surface the critical temperature for the two-dimensional surface decomposition is higher than that for the bulk. The wavelength of the fine modulations does not change with the growth temperature. This could be due to the fast diffusion kinetics in the solid-liquid interface and the balance between the driving force and the gradient energy. To examine the stability of the above microstructures, zinc diffusion experiments were carried out in the temperature range of 390 to 540°C using the ampoule technique. The diffused layers exhibit homogenous microstructures. This demonstrates that critical temperature for phase separation in bulk is below 390°C and is comparable to that predicted by theory.

Electron mobility measured in undoped InGaAs epitaxial layer grown on n-InP substrate

Electron mobility in an epitaxial layer grown on a highly conductive substrate can be determined by angle dependent magnetoresistance measurements. In this letter electron mobility has been measured first in the literature in undoped InGaAs epitaxial layers (n∼1×1015 cm−3) on n-InP substrate (n∼1018 cm−3). Because of the low resistance of the InGaAs layer compared to the series resistance of the structure, mainly that of the contacts, the acceptable range of contact resistances is indicated.

Properties of InGaAs epitaxial layers lattice matched to InGaAs single crystal substrates

The structural and optical properties of In x Ga 1 − x As (0 < x < 0.09) epitaxial layers grown by low pressure metalorganic vapour phase epitaxy (MOVPE) on lattice matched single crystal ternary alloy substrates with the same composition have been studied. They are compared with layers grown simultaneously on GaAs (lattice mismatch up to 0.06%) and on ternary substrates. The lattice matched layers are characterized by a narrow photoluminescence and X-ray linewidth which reflects their excellent structural quality. As the indium concentration increases, the improvement of these properties over those of mismatched layers and of the substrates themselves emphasizes the importance of lattice matched epitaxial buffer layers for the control of strain and defect density in strained structure devices even at relatively small indium concentration.

The thermal stability of lattice mismatched InGaAs grown on patterned GaAs

Patterning and etching substrates into mesas separated by trenches before the growth of mismatched (by about 1% or less) epitaxial layers considerably reduces the interface misfit dislocation density when the layer thickness exceeds the critical thickness. Such films are in a metastable state, since misfit dislocations allow the epitaxial layers to relax to an in-plane lattice parameter closer to its strain-free value. Thermal annealing (from 600 to 850° C) has been used to study the stability of these structures to explore the properties of the misfit dislocations and their formation. The misfit dislocation density was determined by counting the dark line defects at the InGaAs/GaAs interface, imaged by scanning cathodoluminescence. InGaAs epitaxial layers grown on patterned GaAs substrates by organometallic chemical vapor deposition possess a very small as-grown misfit dislocation density, and even after severe annealing for up to 300 sec at 800° C the defect density is less than 1500 cm−1 for a In0.04Ga0.96As, 300 nm thick layer (about 25% of the dislocation density found in unpatterned material that has not been annealed). The misfit dislocation nucleation properties of the material are found to depend on the trench depth; samples made with deeper (greater than 0.5 μm) trenches are more stable. Molecular beam epitaxially grown layers are much less stable than the above material; misfit dislocations nucleate in much greater numbers than in comparable organo-metallic chemical vapor deposited material at all of the temperatures studied.

Heavily-doped n-type InP and InGaAs grown by metalorganic chemical vapor deposition using tetraethyltin

Heavily doped n-type InP and InGaAs epitaxial layers have been grown by metalorganic chemical vapor deposition at atmospheric pressure using tetraethyltin (TESn) as a dopant source. Sn-doped InP and InGaAs layers have been grown with doping concentrations as high as n300K∼3.3×1019 cm−3 and n300 K∼6.1×1019 cm−3, respectively. Hall measurements of Nd-Na at 300 and 77 K indicate that the Sn is uncompensated up to these concentrations. Analysis of the Sn concentration in InP:Sn and InGaAs:Sn layers using secondary ion mass spectrometry, shows that all of the Sn is ionized in InP and InGaAs until a limit is reached that corresponds to the electrical limits. SIMS profiles also show that the use of TESn for the growth n+ InP and InGaAs layers results in no severe memory effects and that abrupt Sn doping profiles can be achieved.