F. Genova

Investigations on the interface abruptness in CBE-grown InGaAs/InP QW structures

In this work we present a detailed analysis of chemical beam epitaxy-grown (CBE) InGaAs/InP multi quantum well (MQW) interfaces to explain experimental data from high quality single and multi-QWs. Our results compare well with the best published data we have obtained some outstanding results. For example, the very intense absorption peak and the high number of satellite peaks in the diffraction rocking curve, were obtained even on samples grown in non-optimized conditions. A careful use of growth interruption at the interfaces allows us to obtain monolayer (ml) interfaces. Nevertheless, the switching of the group V element at each interface leads to strain formation. This effect could become dramatic in superlattice structures with periods smaller than about 5 nm and barriers of less than 3–4 nm. More generally, the conditions for the growth of high quality single and multiple QWs is discussed in this work and these will be correlated with fourier transform photoluminesence (FTPL), high resolution x-ray diffraction (HRXRD), absorption, photo-absorption and photo-current (in PIN structures) measurements.

Influence of growth parameters on the interface abruptness in CBE-grown InGaAs/InP QWs and SLs

Abstract A simple thermodynamic model for As and P incorporation at the CBE-grown InGaAs/InP and InP/InGaAs interfaces has been developed. This model agrees with the X-ray diffraction and the photoluminescence features experimentally obtained from high-quality single quantum wells (SQWs) and multi-quantum wells (MQWs). Our experimental results compare well with the best published data and clearly show that monolayer interfaces can be obtained in this material system only by chosing the proper growth interruption (GI) conditions and accepting a strong mismatch at each interface. This effect could become dramatic in superlattice structures in which the QW period is smaller than 5 nm and the resulting strain could lead to poor crystal quality and optical properties.

The effects of roughness and composition variation at the InP/InGaAs and InGaAs/InP interfaces on CBE grown quantum wells

Abstract Quantum wells of the InGaAs/InP system grown with CBE and MOVPE techniques show compositional changes at the heterointerfaces. The interface layers (strained on the InP substrate) modify the energy profile of the well and the strain can be the cause of deviations from the simple layer-by-layer growth mechanism. Using FTPL, HREM and HRXRD characterization techniques, we will discuss the results of a model for the prediction of the InP/InGaAs/InP interface composition of CBE structures. The role of the growth interruption in order to obtain highly uniform QWs will be clearly emphasized.

Misfit dislocations in InGaAs/InP mbe single heterostructures

Abstract Misfit dislocations in In 1− x Ga x As/InP single heterostructures grown by molecular beam epitaxy were studied by X-ray topography, cathodoluminescence and chemical etching. It was found that they are parallel to the 〈110〉 directions lying on the (001) growth plane, that they are 60° type with Burgers vector at 45° to the (001) plane and that they are driven by the misfit stress into the InP substrate up to a depth of a few μm. The growth conditions, i.e. the layer thickness and lattice mismatch, under which misfit dislocation-free heterostructures can be obtained were determined. Finally, the correlation between cross hatch patterns on the epilayer surface and misfit dislocations was studied.

Effect of InP substrate thermal degradation on MBE InGaAs layers

Abstract A common cleaning technique of InP substrates is the thermal in situ stabilization under As fluxes at the oxide desorption temperature. Since this desorption temperature is well above the upper limit of congruent sublimation, preferential desorption of phosphorus occurs with the formation of indium droplets. In the presence of arsenic flux these droplets precipitate as InAs crystallite with pyramidal habit which can be at the origin of the oval and whisker defects commonly reported in previous works.

Cracks in InP-based heterostructures

Abstract InGaAs/InP and InAlAs/InP single heterostructures grown by molecular beam epitaxy and InGaAsP/InP single heterostructures grown by liquid phase epitaxy under large negative lattice mismatch conditions have been investigated in order to clarify the physical mechanisms of crack formation and propagation. It has been found that cracks generated at the epilayer-substrate interface propagate into both the epilayer and the substrate. It has been observed that cracks parallel to the [1 1 0] direction have a higher mobility than cracks parallel to the [110] direction. The previously reported cracking model, involving glide, combination and lock up of misfit dislocations, has been confirmed.

On the location of the misfit dislocations in InGaAs/InP mbe single heterostructures

Abstract The depth location of the misfit dislocations in InGaAs/InP single heterostructures grown by molecular beam epitaxy has been studied; selective chemical etching and defect observation by cathodoluminescence have been used. It has been found that the misfit dislocations are located inside the InP substrate over a region of a few μm. This result is the opposite of that predicted by the commonly accepted theoretical models for the calculation of the misfit stress in heterostructures.

Growth of In. 53Ga. 47as layers on InP substrates for I.R. detectors by MBE

Abstract The increasing interest in the 1.3–1.55 μm region for optical fibre communications assignes to In . 53 Ga . 47 As a prominent role for detectors operating at these wavelengths. In this contribute we mainly discuss the origin of several kinds of morphological defects on this material and the technical solutions used to reduce the defect density to less than 1×10 4 / cm 2 . Using chemical etching techniques combined with transmission cathodoluminescence and optical microscope observations, we show that there is no relation between single dislocations in the substrate and morphological defects in the layer. Most of defects originate at the interface between the substrate and the layer, and are strongly dependent on the substrate cleaning procedures. Moreover we discuss the influence of growth conditions (growth rate, substrate temperatures and As pressure) on the evolution of these defects. Hall measurements on InGaAs single layers give a residual doping concentration as low as n = 1.10 14 cm −3 and a mobility (at n = 2 10 15 cm −3 ) as high as 10000 cm 2 ·V −1 ·sec −1 at room temperature. InGaAs PIN detectors were fabbricated, and dark currents as low as 3 nA at 10V were obtained over 65 μm mesa diameter.

CBE growth of InGaAs for optoelectronic applications

In this work we studied the growth conditions Three different group III flux controls were of InGaAs single layer lattice matched to InP, tested: (1) pinhole (0.3—0.8 mm in diameter), (2) using TEGa, TEIn, TMIn and AsH3, in order to manual high precision leak valves without any investigate: (a) thickness and composition unifeedback; (3) automatic high precision leak valves formity; (b) role of different MO flux control with a feedback control to keep a constant pressystems on crystal quality; (c) effect of growth sure in the line connecting this valve to the effutemperature on growth rate and alloy composision cell. No carrier gas was used for metalorganic tion. transport. The effusion cell was common to all the The growth apparatus consisted of a standard four group III lines and specially designed in diffusion pumped VG 80H MBE, connected to a order to allow proper metalorganic premixing only pressure controlled gas system. While the group V at the end of the cell. gas line and furnace are standard, the group III To compare the different group III flux control, introduction system was entirely home made. The over 100 samples have been grown and analyzed growth was performed under an As2 flux of 5 X by DCXRD. The spectra of layers grown using 1014_3 x i0~mol/cm 2 s, at growth rates of 0.8—2 the pinhole were very reproducible, peaks of 50 sm/h and at substrate speed rotations of 20—45 arc sec FWHM were obtained, but the presence of rpm. Uniformity on 2 inch wafers has been meawider peaks and/or tails in most of the samples sured on an InGaAs 1.5 ~tm thick epilayer grown implies the onset of unwanted parasitic effects on both In-soldered and In-free mounted subalong the line. Some of the layers grown using the strates. Results are summarized in table 1. As the manual high precision leak valve showed narrow, maximum deviation results were always lower than even if structured, DC X-ray diffraction peaks 1% in thickness and 1.5% in composition, CBE (FWHM of 33 arc sec was obtained), but part of looks very promising from this point of view.

Experimental study of misfit dislocations in InP-based heterostructures

Abstract X-ray topography has been used in order to study the location of the misfit dislocations in InGaAs/InP, InAlAs/InP and InGaAlAs/InP single heterostructures grown by molecular beam epitaxy. It has been found that the dislocations originate at the layer-substrate interface, but propagate mainly into the layer in InGaAlAs/InP structures, into the substrate in InGaAs/InP structures and into both the layer and the substrate in InAlAs/InP structures.