Selective area growth of GaAs using a Ga beam with a step-function lateral intensity profile

Selective area growth of GaAs has been carried out in order to investigate the surface diffusion of Ga atoms using molecular beam epitaxy (MBE) with the aid of a Ga beam with a lateral step-function intensity profile. This step-function profile was obtained using a closely fitted GaAs shadow mask. When the mask edge was parallel to [01 ], a (311)A facet was typically observed near the edge of the Ga beam, while in the case of the mask edge parallel to [011], a (111)B facet was formed. MBE growth simulation based on the diffusion model was carried out in order to understand the mechanism of this selective area growth. The calculated results were in good agreement with the experimental results, and the diffusion lengths of Ga atoms were determined to be 0.10 μm along [011] direction on the (100) GaAs surface, 0.37 μm along [233] direction on the (311)A GaAs surface and 0.17 μm along [21 ] direction on the (111)B GaAs surface during MBE growth. These diffusion lengths seem to be smaller than those previously observed, which is probably due to a large V/III ratio in the region of the substrate close to the mask edge.     

GaAs initial growth on InAs(001) vicinal surfaces observed by scanning tunneling microscopy

GaAs initial growth on InAs surfaces misoriented by 2° toward the [110] and [1 0] directions was investigated by scanning tunneling microscopy (STM). In the STM images of both InAs vicinal surfaces after GaAs deposition, white lines running in the [1 0] direction, corresponding to the grown GaAs surface, were observed. Almost all of the lines were attached only to steps running in the [110] direction (B-type steps) on both InAs surfaces; that is, the lines were seldom attached to steps running in the [1 0] direction (A-type steps). These results indicate that the B-type steps are more favorable for the sticking of deposited Ga atoms than the A-type steps during GaAs initial growth on InAs vicinal surfaces.     

Lateral OMVPE growth of GaAs on patterned substrates

GaAs was grown on patterned 1 0 0 on- and off-axis GaAs substrates by organometallic vapor-phase epitaxy (OMVPE). Patterned mesas were observed to change shape because lateral growth rates varied by more than an order of magnitude in different crystallographic directions. For this study, misoriented GaAs (1 0 0) wafers were polished 3° toward the nearest [1 1 0] or [1 1 1] family of directions, and 320 nm high cross-shaped mesas were fabricated. OMVPE growth was performed between 550°C and 650°C for 1 h at a vertical growth rate of approximately 1.3 μm/h. Atomic force microscopy showed that three effects have a powerful influence on lateral growth initiated at mesa sidewalls. First, the symmetry of the dominant surface reconstruction has a major effect on the diffusion of Ga adatoms. Rapid Ga diffusion occurs along the 0 1 1–0−1−1 axis in OMVPE, or the perpendicular 0−1 1–0 1−1 axis in molecular beam epitaxy, and appears to be a result of the different surface reconstructions which exist in the two growth ambients. Second, misorientation of the wafer causes a growth asymmetry as Ga adatoms move preferentially from high-to-low terraces. When terrace steps descend toward a mesa wall, rapid lateral growth away from the wall is always observed. When terrace steps descend away from a mesa wall, little lateral growth occurs and even reduced vertical growth may be observed. When the misorientation and reconstruction symmetries align, the surface acts like an atomic diode and the rapid lateral growth can exceed the vertical growth rate by more than an order of magnitude. Third, on misoriented substrates, step bunching increases with increasing temperature, and this can lead to significant changes in the original shape of a mesa. A growth model is presented which relates the lateral growth rate in different crystallographic directions to the substrate misorientation, the growth temperature, and the partial pressure of As during growth. It is also shown that different surface reconstruction patterns are related to chemical species with continuously varying concentrations rather than thermodynamically distinct phases.     

InxGa1−xAs/GaAs quantum wire structures grown on GaAs (100) patterned substrates with [001] ridges

In_xGa_1−xAs/GaAs (x = 0.12-0.23) quantum well (QW) structures were grown by molecular beam epitaxy (MBE) on [001] ridges with various widths (1.1-12 μm) of patterned GaAs (100) substrate. The smallest lateral width of the InGaAs/GaAs quantum wire (QWR) structures was estimated to be about 0.1 μm by high-resolution scanning electron microscope (SEM). The In contents of the grown InGaAs/GaAs QWs on the ridges were studied as a function of ridge top width (ridge width of the MBE grown layer) by cathodoluminescence (CL) measurements at 78 K. Compared to the InGaAs QW grown on a flat substrate, the In content of the InGaAs/GaAs QW on the ridge increases from 0.22 to 0.23 when the ridge top width decreases to about 2.9 μm, but it decreases steeply from 0.23 down to 0.12 with a further decrease of the ridge width from 2.9 to 0.05 μm. A simulation of MBE growth of InGaAs on the [001] ridges shows that this reduced In content for narrow ridges is due to a large migration of Ga atoms to the (100) ridge top region from {110} side facets.     

The growth of GaAs and InAs dots on etched mesas: The effect of substrate temperature on mesa profile and surface morphology on dot distribution

The molecular beam epitaxy (MBE) growth of GaAs and InAs quantum dots on etched mesas has been studied using scanning electron microscopy (SEM) and atomic force microscopy (AFM). The [0 1 1]-oriented mesas are etched into (1 0 0) GaAs substrates, exposing (5 3 3)B sidewall facets. At a substrate temperature of 610 °C a top (1 0 0) plane is seen to evolve on a ridge mesa structure. Alternatively, if the overgrowth is carried out at 630 °C no such facet is seen, and the top ridge remains unchanged during GaAs growth. By controlling the mesa shape, either ordered lines of dots can be grown or the dot density can be varied from <5×10⁸ cm⁻² to >1×10¹¹ cm⁻² on the same substrate in pre-defined regions. The dot distribution observed on the mesa sidewalls and top is discussed in terms of net migration of adatoms from different facets, underlying step density, step height and surface curvature of the mesa top.     

Anisotropic lateral growth in GaAs MOCVD layers on (001) substrates

For metalorganic chemical vapor deposition, a fast lateral growth rate is observed for the first time on (001) GaAs having round mesas. The lateral growth rate is greater than the vertical growth rate by a factor of 3–5. The lateral growth rates have anisotropy with respect to the crystallographic directions on the (001) surfaces. The fastest growth direction is the [110] and the slowest one is the [ 10]. The [110] and [ 10] growth rates were found to be strongly dependent on growth conditions, though the vertical one is independent. The [110] growth rate decreases with decreasing As pressure, while the [ 10] remains constant. As growth temperature increases, both the [110] and the [ 10] growth rates decrease. A simple model for the lateral growth mechanism is proposed from the consideration of atomic arrangements and the number of dangling bonds at [110] and [ 10] step sites. According to the model, the lateral growth rate is proportional to the number of bonds available for binding Ga atoms at step sites. The model can explain well the anisotropy in the lateral growth rate and its dependence on the growth conditions.     

Selective growth and other applications of hydrogen-assisted molecular beam epitaxy

The usefulness of atomic hydrogen in molecular beam epitaxy has been demonstrated, centering around selective growth. Atomic hydrogen is effective for low-temperature cleaning of substrates, surfactant effects such as restrain of island growth and suppression of the surface migration of the adatoms and selective growth on masked or V-grooved substrates. These effects are dependent on substrate temperatures. The selective growth of GaAs has been successfully demonstrated at the conventional growth temperature and growth rate with the aid of atomic hydrogen. The main mechanism of the selective growth is the re-evaporation of Ga and As from mask materials such as SiN_x or SiO₂. Selective growth has also been observed on low-index crystal facets. On (111)A and (110) facets, no GaAs was deposited in the presence of atomic hydrogen, the flux of which is approximately the same as that of Ga. GaAs quantum wire structures have been fabricated on the substrates with V-shaped grooves. The efficient capture and confinement of carriers into wire regions have been observed by photolumenescence.     

Investigation of X‐ray Satellite Reflections Attributed to Antiphase Boundaries in Epitaxial (Ga,In)P Layers grown on (001) GaAs Substrates

We report on the analysis of additional X‐ray reflections that probably arise from antiphase domain boundaries within (Ga,In)P/(001) GaAs heteroepitaxial layers. Due to the preferred cation ordering along the crystallographic directions [1‐1 1] and [‐1 1 1] which belong to the [110] zone the original sphalerite‐type structure of (Ga,In)P changes into a CuPt‐like of the cation sublattice. This ordering phenomenon causes a loss of symmetry, i.e. the cubic structure is converted into a rhombohedral one. The antiphase boundaries between ordered domains are assumed to behave similar to lattice planes at X‐ray diffraction. Therefore, additional reflections may occur spatially neighboured to the [001] direction. The presented results of X‐ray experiments are discussed in relation to TEM experiments published in the literature in order to explain the origin of the satellite reflections. In the case of the investigated samples (grown on GaAs substrates misoriented 2° towards the azimuthal [010] direction) the APBs run preferentially in directions tilted up to angles of 20° with respect to growth direction. A preferential occurrence of satellite reflections in <13 2 1> directions was observed coinciding with {13 2 1} "lattice planes" whose normals enclose the same angle to the [001] growth direction as the normals of the average planes characterized by APBs. The appearence of the phenomenon in other directions that are also spatially neighboured to the <13 2 1> directions was determined on the basis of the shift of the reflection positions due to tilting the sample around an axis geometrically included in the scattering plane.     

Hillock formation observed in MOMBE of InGaAs grown on a patterned GaAs substrate

We observed hillock formation during metalorganic molecular beam epitaxy (MOMBE) of InGaAs on a mesa-grooved (100) GaAs substrate. Hillocks were formed under specific growth conditions and comprised mostly InAs. The distribution of hillocks formed in InGaAs MOMBE using trimethylindium (TMIn) and metal Ga depended strongly on the widths of mesa-grooves; the density decreased with decreasing width and hillocks were hardly observed on the ridges. The hillock density also varied, depending on the off-angle of the substrate from the (100) plane. This indicates that the observed anomalous distribution of InGaAs hillocks was caused by both the formation of facets and a vicinal tilted surface near the edge of mesa-grooves, due to the growth of a GaAs buffer layer on a patterned substrate.     

Ti-adsorption on Se-prereacted GaAs (110) surfaces studied by SXPS

This study is embedded in the broader context of reactive metal overlayers and passivation on GaAs. High resolution synchrotron-radiation photoemission experiments for Ti coverages on Se-reacted GaAs (110) surfaces show that, contrary to clean Ti-reacted GaAs (110) interface, there is no initial disruption (submonolayer reaction) of the surface involving both Ga and As atoms during early stage of interface formation. However, a delayed Ti involved reaction appears at a trigger coverage of about 1 ML involving only As atoms from the prereacted As Se interface configuration continued by Stranski-Krastanov growth mode. A second reacted phase starts to form two Ga Ti involved configurations replacing Ga Se bonds near a Ti coverage of 4 ML. The preferential chemical trapping of Ti by Se atoms is associated with a smaller interface thickness very likely <10 ML instead of 50 ML in the case of clean Ti/GaAs (110) interface caused by mobile As atoms.     

In-situ fabrication of three-dimensionally confined GaAs and InAs volumes via growth on non-planar patterned GaAs(001) substrates

Three-dimensionally confined GaAs/AlGaAs and InAs/GaAs structures on 100 oriented square mesas patterned onto GaAs(001) substrates are realized, in-situ, via size-reducing molecular beam epitaxy. Two stages of mesa top pinch-off involving {103} and subsequently {101} side facets are revealed. GaAs and InAs quantum boxes with lateral linear dimensions down to 40 nm and confined by AlGaAs and GaAs, respectively, are reported. For InAs, the strain relief in mesas is found to enhance the well known 2 ML thickness for three-dimensional island formation on unpatterned substrates to, remarkably, >5 ML for mesa size 75 nm. Cathodoluminescence emission from the InAs on the mesa top attests to its good optical quality.     

Twins in GaAs Crystals Grown by the Vertical Gradient Freeze Technique

The growth of facets and the generation of twins on <100> VGF (vertical gradient freeze technique) grown GaAs were investigated using DSL (diluted Sirtl-like etchant with light) photoetching and transmission X-ray topography. Due to the polarity of the (111) plane in GaAs, As facets are larger and more irregular than Ga facets and twins always occur on As facets. Twins are initiated at the change of boundary condition which is affected by temperature gradient and crucible shape. The mechanism of twin generation is explained by considering the edge concavity at the solid-liquid interface and the supercooling required for initial nucleation of a facet. Twins are more often produced in Si-doped crystals than undoped ones due to the constitutional supercooling.     

Epitaxial growth of CuInSe2 single crystal by halogen transport method

CuInSe₂ single crystals were epitaxially grown on (001) GaP, (001) GaAs, and (110) GaP by the halogen transport method. The orientation relationships in the growth on the (001) and (110) faces were (a) [001]_CuInSe2 [001]_sub and [100]_CuInSe2 [100]_sub (c-axis orientation growth), and (b) [110]_CuInSe2 [110]_sub and [001]_CuInSe2 [001]_sub, respectively. On (001) InP, the orientation relationships between the layer and substrate consist of two sets: (c) [100]_CuInSe2 [001]_sub and [001]_CuInSe2 [010]_sub, and (d) [100]_CuInSe2 [001]_sub and [001]_CuInSe2 [ 00]_sub (a-axis orientation growth). The above results, i.e., c-axis growth on (001) GaP and (001) GaAs and a-axis growth on (001) InP, could be explained by a criterion of the minimum lattice mismatch between grown layers and substrates. A series of growth experiments on (001) GaAs indicated that appropriate gas etching of the substrate surface and growth temperature were required for obtaining twin-free single-crystal epitaxial CuInSe₂.     

Surface reconstructions and growth mode transitions of AlAs(100)

The surface reconstructions of AlAs(100) layers grown by molecular beam epitaxy (MBE) on GaAs(100) were mapped as a function of substrate temperature and arsenic flux. Three main reconstructions were observed - a c(4×4) at lower temperatures and higher arsenic fluxes, a (2×4) at middle temperatures, and a (3×2) at higher temperatures and lower arsenic fluxes. Growth of AlAs on AlAs(100) is layer-by-layer for the high temperature and low temperature reconstructions. In the mid-temperature region, AlAs grows rough on (2×4) reconstructed AlAs(100) as indicated by rapidly damped reflection high-energy electron diffraction (RHEED) intensity oscillations and the appearance of three-dimensional (3D) features. The addition of fractional layers of Ga enhances the smooth growth of AlAs. A metastable (5×2) reconstruction was observed when a fraction of a layer of Ga was present on the surface. The results indicate that Ga segregates during the growth of AlAs on GaAs(100) at temperatures at least as low as 500°C, and that annealing at temperatures above 700°C removes most of the Ga from the surface.     

Surface diffusion lenght during MOMBE and CBE growth measured by μ-RHEED

The growth rates of layers grown on a mesa-etched (001) GaAs surface were measured by in-situ scanning microprobe reflection high-energy electron diffraction (μ-RHEED) from the period of the RHEED intensity oscillation in real time. The diffusion lenght of the surface adatoms of column III elements was determined from the gradient of the variation of the growth rates in the cases of MBE, MOMBE using trimethylgallium (TMGa) and CBE using TMGa or triethylgallium (TEGa) and arsine (AsH₃). The obtained values of the diffusion lengths were of the order of a micrometer in every case of the source-material combination. In the case of metalorganic materials as Ga source, it was found that the diffusion length was larger than that of Ga atom from metal Ga source. Since the substrate temperature of the present experiment is high enough to decompose TMGa and TEGa on the surface, Ga adatoms are considered to be responsible to the surface diffusion. Therefore, it is considered that the derivatives of the metalorganic molecules such as methyl radicals affect the diffusion of Ga adatoms.     

Silicon compensation and scattering mechanisms in two-dimensional electron gases on (110)GaAs

A study of the MBE growth of (001) and (110) (Al,Ga)As is reported, and the efficiency of Si as an n-type dopant in (110)GaAs is accessed. A 40 nm spacer two-dimensional electron gas (2DEG) structure grown on (110)GaAs gives a mobility of 540,000 cm² V⁻¹ s⁻¹ at 4 K after illumination. The dominant scattering mechanisms in 2DEGs on (110) and (001)GaAs grown under the separate optimum growth conditions for the two orientations are compared.     

Diffusion constant of Ga adatom on GaAs (001) surface: Molecular dynamics calculations

A valence force field to reproduce both the phonon dispersion curves of crystalline GaAs and first principle derived interaction energies of Ga adatom on GaAs (001) surface has been optimized. Calculations of diffusion constant of isolated Ga atoms on the GaAs surface have been performed by molecular dynamics classical trajectory simulations.     

Growth of embedded ErAs nanorods on (4 1 1)A and (4 1 1)B GaAs by molecular beam epitaxy

The low solubility of Er in GaAs results in the formation of ErAs nanostructures when GaAs is grown with 5–6 at% Er/Ga ratio by molecular beam epitaxy on GaAs surfaces. For growth on the (4 1 1)A GaAs surface, cross-sectional scanning transmission electron microscopy images show the presence of ErAs nanorods embedded in a GaAs matrix extending along the [2 1 1] direction with a spacing of roughly 7 nm and a diameter of roughly 2 nm. Growth on the GaAs (4 1 1)B surface resulted in only nanoparticle formation. Variation of the polarized optical absorption with in-plane polarization angle is consistent with coupling to surface plasmon resonances of the semimetallic nanostructures.     

MOVPE growth of spontaneously ordered (GaIn) and (AlIn)P layers lattice matched to GaAs substrates

A systematic study of the metal-organic vapour-phase epitaxial growth of (GaIn)P and (AlIn)P layers deposited on GaAs substrates with (001) and (110) orientation is presented. Special attention has been paid to the growth on (001)-oriented wafers with different misorientations to the growth direction. The influence of the growth conditions on the properties of the epitaxial layers such as lattice mismatch, alloy composition, photoluminescence (PL) wavelength, FWHMs of PL peaks and atomic ordering is discussed. Layers with mirrorlike surfaces and various degrees of order could be deposited at growth temperatures T_g ranging from 595 °C to 750 °C for (GaIn)P and 720 °C to 800 °C for (AlIn)P. In addition to the influence of T_g on the Ga incorporation during the (GaIn)P growth we found the Ga distribution coefficient k_Ga to be affected by the misorientation of the substrates. k_Ga correlates presumably with the number of kinks and steps on the substrate surface. Transmission electron diffraction (TED) and PL investigations show that the degree of order — often described by the ordering paramter η — depends strongly on T_g the ordering is more pronounced when the layers are deposited on substrates misoriented towards the (1 11) lattice plane. Strong ordering has been observed for (GaIn)P samples grown at 680 °C on substrates 2° misoriented towards the [1 10] direction and at 650 °C on substrates 6° misoriented towards the same direction. For the (AlIn)P samples striking ordering has been found when they were grown at 720 °C.     

A scanning tunnelling microscopy study of the deposition of Si on GaAs(001); Implications for Si δ-doping

Atomic resolution scanning tunnelling microscopy (STM) has been used to study the adsorption of Si on GaAs(001) surfaces, grown in situ by molecular beam epitaxy (MBE), with a view to understanding the incorporation of Si in δ-doped GaAs structures. Under the low-temperature deposition conditions chosen, the clean GaAs surface is characterized by a well-defined c(4×4) reflection high-energy electron diffraction (RHEED) pattern, a structure involving termination with two layers of As. Filled states STM images of this surface indicate that the basic structural unit, when complete, consists of rectangular blocks of six As atoms with the As-As bond in the surface layer aligned along the [110] direction. Deposition of <0.05 ML of Si at 400°C onto this surface shows significant disruption of the underlying structure. A series of dimer rows are formed on the surface which, with increasing coverage, form anisotropic "needle-like" islands which show no tendency to coalesce even at relatively high coverages (0.5 ML). The formation of these islands accompanies the splitting of the 1/2 order rods in the RHEED pattern along [110]. As the Si is known to occupy only Ga sites, the Si atoms displace the top layer As atoms of the c(4×4) structure, with the displaced As atoms forming dimers in a new top layer. The results are consistent with a recently proposed site exchange model and subsequent island formation for surfactant mediated epitaxial growth.