Epitaxy Of InP Research Articles

Selectively Grown III-V Lasers for Integrated Si-Photonics

Epitaxially integrating III-V lasers with Si-photonics is the key for compact, efficient, and scalable photonic integrated circuits (PICs). Here we present an investigation of a path forward in integrating III-V functionalities on industry-standard silicon-on-insulator (SOI) platforms using selective area hetero-epitaxy. Based on our recently developed methods of selectively growing device quality InP on (001)-oriented SOI wafers, we demonstrated InP stripes and segments, with dimensions varying from a few hundred nanometers to a few micrometers. The flexible epitaxy of InP on SOI together with the unique “bufferless” trait will enable efficient light interfacing with Si-based photonic devices using either evanescent or butt coupling schemes. We simulated the possibility of employing the micrometer-scale InP on insulator to realize electrically driven lasers and found that the metal induced optical loss is negligible when the InP dimension exceeds 4.0 μm. The potential of utilizing this selective area growth method to realize fully integrated Si-photonics is illustrated.

Selective lateral epitaxy of dislocation-free InP on silicon-on-insulator

Efficient on-chip laser sources of Si photonics can be built from direct epitaxy of dislocation-free III–V alloys on industrial-standard (001) Si wafers. Here, we report on selective lateral epitaxy of InP on patterned (001) silicon-on-insulators (SOIs) by metal organic chemical vapor deposition. Based on the conventional “aspect ratio trapping” approach, we created undercut patterns to alter the growth front to the lateral direction. Growth of InP inside the nano-scale SOI trenches results in dislocation-free InP crystals right atop the buried oxide layer. The intimate placement of the InP crystals with the Si device layer points to the development of dislocation-free nano-ridges for integration of efficient III–V light emitters with Si-based photonic components on SOI.

Selective Epitaxy of InP on Si and Rectification in Graphene/InP/Si Hybrid Structure

The epitaxial integration of highly heterogeneous material systems with silicon (Si) is a central topic in (opto-)electronics owing to device applications. InP could open new avenues for the realization of novel devices such as high-mobility transistors in next-generation CMOS or efficient lasers in Si photonics circuitry. However, the InP/Si heteroepitaxy is highly challenging due to the lattice (∼8%), thermal expansion mismatch (∼84%), and the different lattice symmetries. Here, we demonstrate the growth of InP nanocrystals showing high structural quality and excellent optoelectronic properties on Si. Our CMOS-compatible innovative approach exploits the selective epitaxy of InP nanocrystals on Si nanometric seeds obtained by the opening of lattice-arranged Si nanotips embedded in a SiO2 matrix. A graphene/InP/Si-tip heterostructure was realized on obtained materials, revealing rectifying behavior and promising photodetection. This work presents a significant advance toward the monolithic integration of graphene/III-V based hybrid devices onto the mainstream Si technology platform.

Selective epitaxy of InP on Si(1 0 0) substrates prepared by liquid-phase epitaxy

Selective liquid-phase (LPE) epitaxial growth of InP was performed on patterned Si(1 0 0) substrates. The growth temperature was rapidly cooled, then the crystal growth proceeded at constant temperature. As a result, area-selective epitaxy (ASE) and slight epitaxial lateral overgrowth (ELO) layers were achieved in narrow openings and small nuclei were observed in the wide openings. X-ray diffraction results show that the lattice strain in the InP nuclei on the Si substrate was fully relaxed. Huber etching revealed the dislocation-related etch pits in ASE layers, and it is shown that the etch pit density (EPD) is dependent on the length of open seed area.

Area selective epitaxy of indium phosphide on silicon (100) substrates without buffer layers by liquid phase epitaxy

Area selective epitaxy (ASE) of InP was performed by liquid phase epitaxy on Si (100) substrates without the use of buffer layers. At the start of epitaxy, the growth temperature was rapidly lowered to obtain high supersaturation. Epitaxial growth was then carried out at constant temperature. While small InP nuclei were observed in the broader open areas, ASE layers and layers showing slight epitaxial lateral overgrowth were formed in the narrow openings. Etch pit density of the InP ASE layers on Si is dependent on the length of open seed region and X-ray diffraction results indicate that the lattice strain in the InP nuclei on the Si substrate was fully relaxed.

Anisotropy of lateral growth rate in liquid phase epitaxy of InP and its association with kink-step structures on the surface

The angular dependence of the lateral growth rate as a function of crystal growth temperature (Tg) was determined for the liquid phase epitaxy (LPE) of (001) and (111)A,B InP. Low temperature LPE growth of InP was performed at constant growth temperatures between 330°C and 490°C. Post-growth variations in the shapes of pre-formed mesa structures show that lateral growth rates exhibit strong anisotropy in the [110] direction at low growth temperatures (Tg < 400°C) on InP (001) surfaces. This implies that the kink-step density is highest in the [110] direction on InP (001). On InP (111)B surfaces, the maximum lateral growth rate occurred at Tg = 450°C in the 〈112〉 direction, indicating that the kink-step density was highest in this direction. However, at Tg = 450°C the lateral growth rate shows isotropic tendencies on (111)A surfaces, and the anisotropy decreases as the growth temperature increases.

Fabrication of a P-stabilized InP(0 0 1) surface at low pressure and temperature using t-butylphosphine (TBP)

Fabrication of a P-stabilized InP(0 0 1)-(2 × 1) surface at low pressure and temperature using t-butylphosphine (TBP) has been studied. The (2 × 1) surface was fabricated at 260 °C using TBP (5 × 10 −6 Torr) cracked with a hot W-filament although the (2 × 4) structure was not changed by exposing to TBP (5 × 10 −6–1 × 10 −5 Torr) at 260–280 °C without the filament. Mass spectra of the cracked TBP indicate that the relative spectral intensities of P and P 2 are increased by 1.3–1.5 times higher than TBP without the filament, leading to the result that the surface can be fabricated at low pressure and temperature. The formation rate of the (2 × 1) surface by our method is higher by about one order of magnitude even at lower pressure and temperature than the reported ALE method. The (2 × 1) surface is stable at 260–280 °C in an ultrahigh vacuum. This suggests that atomic layer epitaxy (ALE) of InP can be performed at low pressure and temperature using the hot W-filament. The Auger electron spectroscopy (AES) result indicates that carbon contamination does not occur on the surface by adsorption of TBP cracked by the filament.

Anisotropy of lateral growth rate in liquid phase epitaxy of {0 0 1} InP

The angular dependences of the lateral growth rate are determined as the functions of growth temperature and growth time in liquid phase epitaxy of (0 0 1) InP at low growth temperature of 330–450 °C. From the deformation of artificially made tables after epitaxy, it is shown that the lateral growth rate has a strong anisotropy in [1 1 0] direction especially at low growth temperature ( T g<400 °C). This indicated that kink density was high in [1 1 0] direction on InP(0 0 1) surface. It is also shown that the anisotropy of lateral growth rate decreases as the growth temperature becomes high.

Area-selective epitaxy and epitaxial lateral overgrowth of InP by liquid phase epitaxy on three integral axes

The area-selective epitaxy (ASE) and epitaxial lateral overgrowth (ELO) were performed on {0 0 1}-, {1 1 1}A,B- and {1 1 0}-oriented InP by liquid phase epitaxy at constant growth temperature (450–650°C). The as-grown surface and cross section were investigated by Nomarski interference optical microscope, SEM and confocal laser scanning microscope. From the cross sectional structure, area-selective epitaxy and ELO layers have shown {1 0 0} and {1 1 1}A,B facets on these surface. When compared with three integral axes, the vertical growth rate on {1 1 0} plane is the fastest among the three integral axes. According to the observations of cross sectional shape, the orientation dependence of vertical growth rate was determined to be {1 1 0}>{1 1 1}A,B>{1 0 0} under the present experimental conditions.

Determination of angular dependence of lateral growth rate in liquid phase epitaxy of (0 0 1) InP

The angular dependences of the lateral growth rate are determined as the functions of growth temperature and growth time in liquid phase epitaxy of (0 0 1) InP at low growth temperature of 330–450°C. From the observations of the deformation of artificially made tables after epitaxy, the lateral growth rate has a strong anisotropy in [1 1 0] direction especially at low growth temperature ( T g<400°C). This indicated that kink density was high in [1 1 0] direction on InP (0 0 1) surface. It is also shown that the anisotropy of lateral growth rate decreases as the growth temperature becomes high.

Observation of anisotropic initial growth nucleation in liquid phase epitaxy of InP

Liquid-phase epitaxial growth of InP was performed at constant growth temperature by the temperature difference method under controlled vapor pressure on InP substrates. Growth temperature was kept constant from 550°C to 280°C. The as-grown surface morphology was investigated by Nomarski interference optical microscope and electron microscope. The anisotropic initial growth nuclei are investigated on {0 0 1},{1 1 1}A,B and {1 1 0} oriented InP substrates. This anisotropic behavior is discussed in view of the anisotropy of lateral growth rate due to the preferential surface migration and sticking at specific kinks and steps on the surface.

Growth mechanisms and modeling for metalorganic chemical vapor deposition selective-area epitaxy on InP substrates

Selective area epitaxy of InP on masked (100) InP substrates is studied. InP layers are deposited between pairs of SiO2 stripes using low-pressure metalorganic chemical vapor deposition. Layer thickness is investigated by surface profiling and scanning electron microscopy. For growth between oxide stripes, the growth velocity is enhanced by lateral diffusion of growth species from the masked region to the exposed region. Two transport mechanisms are known to exist: vapor phase diffusion and surface migration. However, most existing quantitative models focus only on the former. A new computational model, based on the diffusion equation with time dependent boundary conditions, is presented which describes the growth enhancement component due to surface migration. The role played by surface migration is shown to depend on nominal film thickness. The model correctly predicts a super growth enhanced region adjacent to the oxide. Previous quantitative models have not successfully described this aspect of growth near the oxide film.

Coalescence in microchannel epitaxy of InP

We have studied the lateral coalescence in microchannel epitaxy (MCE) of InP by LPE. In the present study, we suggest that there are two modes of the coalescence, “one-zipper” and “two-zipper”. New patterns of microchannels were designed to realize these two coalescence modes. The coalesced MCE islands were chemically etched and it was found that dislocations were usually generated in the coalesced region when the growth was carried out in the “two-zipper” mode; however, when the growth occurred in the “one-zipper” mode, dislocations were not found in the coalesced region. It is concluded that by conducting the growth in the “one-zipper” mode, the formation of dislocations in the coalesced region can be avoided.

Cyclohexylphosphine: synthesis and application to metalorganic vapor-phase epitaxy of InP

We have investigated the synthesis and the thermal decomposition of cyclohexylphosphine and its application to the metalorganic vapor-phase epitaxy of InP. Cyclohexylphosphine is a liquid compound with a 20°C vapor pressure of 7.2 hPa. The thermal decomposition reaction is detected at temperatures as low as 400°C. By conducting coupled experiments of thermolysis and mass spectrometry the formation of PH n species can be proven. Details of the decomposition pathways are given. InP layers with good surface morphologies could be grown at atmospheric pressure and at substrate temperatures between 500°C and 560°C. The surface morphology reflects the role of PH n and P 2/P 4 species during the growth process.

Chemical Beam Epitaxy of InP with Ar + Laser Irradiation

A study on laser‐assisted chemical beam epitaxy of InP with trimethylindium and tertiarybutylphophine is reported here. In situ monitoring of the laser‐assisted growth surface is carried out using reflection high‐energy electron diffraction. The growth rates over laser‐irradiated and nonirradiated areas are interpreted in terms of physical chemistry of the physisorbed and chemisorbed layers. Two postulations on the photolytic mechanism responsible for growth enhancement with laser are made, which describe the roles of physisorbed group V and group III species in the mechanism. © 1999 The Electrochemical Society. All rights reserved.

Liquid-phase epitaxy (LPE) microchannel epitaxy of InP with high reproducibility achieved by predeposition of In thin layer

The wettability of a solution with the surface of line seeds is a key factor to improve the yield and uniformity of microchannel epitaxy (MCE) by liquid phase epitaxy (LPE). To improve the wetting, a special technique is proposed. In this technique, a thin In layer is deposited before the growth on the substrate with narrow line seeds in SiO 2 mask. The predeposited In film in the seed area greatly helps the solution to contact with the surface inside the line seeds even when the width and the length are very small. As a result, the yield of MCE has been drastically increased up to almost 100%.

Interface supersaturation in microchannel epitaxy of InP

By employing microchannel epitaxy (MCE), it became possible to reduce the dislocation density in liquid phase epitaxy (LPE) of InP so that steps supplied from only one screw dislocation can cover the whole surface of MCE island. This enabled us to measure interstep distance by AFM and to calculate interface supersaturation even in metallic solution with the help of Cabrera and Levine formula by assuming appropriate value of interface free energy. The interface supersaturation was found being ranged from 0.02 to 0.05 upon various experimental conditions. A minimum interface supersaturation was realized when the growth temperature and cooling rate were chosen as 500°C and 0.05°C/min, respectively. The ratio of width to thickness of the grown MCE layer, which is defined as W/ T ratio, was found to increase rapidly with the decrease of the interface supersaturation. A W/ T ratio as high as 20 was accomplished under the growth condition where minimum interface supersaturation can be realized.

Interface supersaturation dependence of step velocity in liquid-phase epitaxy of InP

By employing microchannel epitaxy (MCE), it became possible to reduce the dislocation density in liquid-phase epitaxy (LPE) of InP, so that steps supplied from only one screw dislocation can cover whole surface of MCE island. This enabled us to measure interstep distance by AFM and to calculate interface supersaturation even in metallic solution with the help of the formula given by Cabrera and Levine. From the interstep distance, the step velocity was calculated by employing the macroscopic vertical growth velocity. It was found that the step velocity increases linearly with the increase of the interface supersaturation, and the slope of the curve, which gives a step kinetic coefficient of the growth, increases with the increase of the growth temperature. From the Arrhenius plot of the kinetic coefficient, the activation energy of step advance was found as E=32.8 kJ/mol. A critical value of the interface supersaturation σ c, below which steps do not move, has also been found to exist. It was also found that σ c decreases with the increase of the growth temperature. The presence of σ c probably means that step advance is stopped by the impurity pinning effect.

Formation of a transition layer during heteroepitaxy of III–V compound semiconductors in gas-phase epitaxy systems

It is known that during gas-phase epitaxy of InP and GaAs the region of the film close to the heteroboundary is formed with altered phase, structural, and electrophysical properties. The main cause for this is formation of a transition layer due to a change in the growth mechanism (transition from nucleational to layer-step growth) and the structure of the growing surface (height and density of growth steps).

Si-doping of MOVPE grown InP and GaAs by using the liquid Si source ditertiarybutyl silane

The liquid Si-compound ditertiarybutyl silane (DTBSi) has been investigated as a doping source for the metal organic vapour-phase epitaxy (MOVPE) of InP and GaAs using TBP or DTBP and TBAs as less hazardous group-V-sources. For both material systems the measured carrier concentration is directly proportional to the DTBSi/group-III-partial pressure ratio with a slope of unity. Uncompensated n-type InP as well as GaAs layers are achieved up to the 10 18 cm −3 doping range. Additional calibrated SIMS-studies of InP layers show a coincidence of the Si-concentration in the InP layers with the measured net n-type carrier concentration within the experimental errors. With a reduction of the growth temperature from 610°C to 570°C a significant reduction of the Si-doping efficiency is observed. However, for temperature in excess of 610°C up to 700°C an almost constant doping efficiency for GaAs is obtained. For GaAs a slight reduction in Si-incorporation efficiency with increasing V/III-ratio is detected. The successful doping characteristics for GaAs have been applied to realize n-GaAs : Si Hall sensor device structures. The lateral homogeneity of the sheet resistance is evaluated as a function of the V/III-ratio and the substrate material. Under optimum conditions the normalized variance of the sheet resistance can be as low as 1%.