V. Krishnamoorthy

Anisotropic strain relaxation of GaInP epitaxial layers in compression and tension

We have investigated the strain relaxation of intentionally lattice mismatched (±0.5%) GaInP layers grown on GaAs substrates by organometallic vapor phase epitaxy. Double axis x‐ray diffraction was used to measure the relaxation in these epitaxial layers in perpendicular 〈110〉 directions as a function of thickness. For samples in tension, the difference in relaxation between [110] and [110] increases from 10% to 48% as the layer thickness increases from 7 to 28 times the critical thickness, hc. For samples in compression this difference is 28% at 24hc while no relaxation is measured for a sample at 6hc. These results indicate that strain relaxes anisotropically and that the anisotropy is more pronounced for samples in tension than in compression. Furthermore, the major relaxation axis was found to be [110] regardless of the sign of the strain. Reciprocal space maps, generated using triple axis x‐ray diffraction, showed that the amount of microtilt of the epitaxial layers was also anisotropic. This aniso...

Strain relief study concerning the InxGa1−xAs/GaAs (0.07<x<0.5) material system

The evolution of dislocations in the InxGa1−xAs/GaAs material system has been studied as a function of the ternary alloy comparison in the range 0.07<x<0.5. Cross‐sectional transmission electron microscope observations indicate that for x<0.18, threading dislocations are absent in the epilayer and dislocations propagate from the heterointerface into the GaAs material, while, for 0.18<x<0.28, dislocations appear to propagate into both the epilayer and the GaAs. Furthermore, for x ≳0.28, all the dislocations are observed in the epilayer while the GaAs appears to be dislocation‐free. We propose a model involving the balance of forces acting on misfit dislocations generated at the heterointerface which includes a surface image force term to explain our observations of dislocation evolution as a function of the ternary alloy composition.

Application of ‘‘critical compositional difference’’ concept to the growth of low dislocation density (<104/cm2) InxGa1−xAs (x≤0.5) on GaAs

Multilayer epitaxial structures consisting of InxGa1−xAs layers of various compositions were grown on GaAs substrates by the molecular beam epitaxy technique. Dislocation evolution and residual strain in these heterostructures were studied using cross‐sectional transmission electron microscopy (XTEM) and high‐resolution x‐ray diffraction analyses, respectively. The multilayer heterostructures were designed such that the compositional difference between two adjacent InxGa1−xAs layers in the stack was less than a critical compositional difference of Δx=0.18, taking partial lattice‐relaxation into account. XTEM studies of the stacked structures indicated dislocation evolution to be confined to the GaAs substrate and the InxGa1−xAs layers underlying the top InxGa1−xAs layer in the stack, the top InxGa1−xAs layer being essentially dislocation‐free. This phenomenon is attributed to a monotonic increase in the yield strength of InxGa1−xAs at the appropriate growth temperatures with increasing values of x. Such b...

Diffusion of ion implanted boron in preamorphized silicon

Transient enhanced diffusion of boron in preamorphized and subsequently regrown Si was studied by secondary ion mass spectrometry (SIMS) and transmission electron microscopy (TEM). A comparison of 4 keV, 1× 1014/cm2 boron implants into crystalline and Ge+ preamorphized silicon was undertaken. Upon annealing the B+ implant into crystalline material exhibited the well‐known transient enhanced diffusion (TED). In this case the peak of the boron distribution was relatively immobile and only B in the tail showed TED. In the second set of samples, the surface was first preamorphized by a 180 keV, 1×1015/cm2 Ge+ implant which produced an amorphous layer 2300 A deep, which then was implanted with boron. After implantation the tail of the B distribution extended to only 700 A. Upon annealing, TED of the boron in the regrown Si was also observed, but the diffusion profile was very different. In this case the peak showed no clustering, so the entire profile diffused. The time for the TED to decay was around 15 min a...

Residual strain analysis of InxGa1- xAs/GaAs heteroepitaxial layers

InxGa1−xAs/GaAs heteroepitaxial layers, having various compositions and thicknesses, have been analyzed using the high resolution x‐ray diffraction technique which has revealed that the residual strain in the epilayers is strongly dependent on both the epilayer composition as well as thickness. However, published theoretical models concerning residual strain in InxGa1−xAs/GaAs epilayers suggest that the extent of relaxation is independent of epilayer composition. In this letter, we present an empirical model based on our findings which can be used to accurately predict the extent of lattice relaxation in InxGa1−xAs/GaAs epilayers which includes the influence of epilayer composition.

Studies of the interactions between (311) defects and type I and II dislocation loops in Si+ implanted silicon

Silicon wafers were implanted with Si+ at doses of 2 × 1014 and 1 × 1015 /cm2. Annealing treatments were done at temperatures between 700°C and 1000°C for times between 15 min and 16 h, both with and without an SiO2 cap. Plan-view TEM micrographs were taken and the density of interstitials trapped in both the (311) defects and the type I and II perfect loops were measured. The results showed that for the 2 × 1014 /cm2 Si+ dose, which is below the amorphization threshold, the dominant defect at 700°C is the (311) defect with a much smaller concentration of type I loops. The total trapped interstitial concentration in both kinds of defects was around 7 × 1013 /cm2 for 700°C 1 h anneals. The (311) defects begin dissolving after several hours at 700°C but their dissolution rate is slower than previously reported by Stolk et al. [MRS Symp. Proc. 354 (1995)] for lower dose (5 × 1013 /cm2) implants. It is not believed that this slower dissolution rate is due to the increased dose. The reduced dissolution rate does not change with capping and may be due to a difference in furnace calibration methods. The type I loops show some growth during the (311) dissolution but quantitatively less than half of the released interstitials appear to be trapped by the type I loops. For the 1 × 1015 /cm2 sample amorphization occurs and both type II (end of range) loops and (311) defects are observed for 700°C anneals. The total number of trapped interstitials for 700°C 1 h anneals is also around 7 × 1013 /cm2. However, the ratio of (311) to loops has switched such that the dominant defect is the type II loop. Upon annealing, the (311) defects again show a reduced dissolution rate and the type II loops are in the growth regine. Increasing the anneal temperature to 800°C results in further growth of the type II loops and all of the (311) defects have either dissolved or unfaulted. The growth of the type II loops appears to be greater than can be quantitatively accounted for by the (311) defects. In addition there is a high level of strain in the lattice that cannot be accounted for by the (311) defects. Both of these results imply there may be an additional source of interstitials besides the (311) defects for amorphizing implants.

Device quality In0.4Ga0.6As grown on GaAs by molecular beam epitaxy

A novel approach to growing device quality In0.4Ga0.6As epilayers on GaAs is reported which involves the controlled propagation of dislocations via a carefully designed multistage strain‐relief buffer system. Cross‐sectional transmission electron microscopy analysis revealed the In0.4Ga0.6As epilayers to be threading dislocation free in contrast to the heavily dislocated material obtained by growing In0.4Ga0.6As directly on GaAs. Hall effect measurements performed on unintentionally doped buffered In0.4Ga0.6As epilayers indicated the room‐temperature electron concentrations in the epilayers to be around 1×1015 cm−3 while electron mobilities were around 4700 cm2 V−1 s−1. In addition, strong band‐edge photoluminescence was recorded from the buffered epilayers, the luminescence peak occurring at 1304 nm having a linewidth around 7 meV at 13 K.

The effect of boron implant energy on transient enhanced diffusion in silicon

Transient enhanced diffusion (TED) of boron in silica after low energy boron implantation and annealing was investigated using boron-doping superlattices (DSLs) grown by low temperature molecular beam epitaxy. Boron ions were implanted at 5, 10, 20, and 40 keV at a constant dose of 2×1014/cm2. Subsequent annealing was performed at 750 °C for times of 3 min, 15 min, and 2 h in a nitrogen ambient. The broadening of the boron spikes was measured by secondary ion mass spectroscopy and simulated. Boron diffusivity enhancement was quantified as a function of implant energy. Transmission electron microscopy results show that 〈311〉 defects are only seen for implant energies ⩾10 keV at this dose and that the density increases with energy. DSL studies indicate the point defect concentration in the background decays much slower when 〈311〉 defects are present. These results imply there are at least two sources of TED for boron implants (B-I): short time component that decays rapidly consistent with nonvisible B-I pai...

Reactivation of acceptors and trapping of hydrogen in GaN/InGaN double heterostructures

The apparent thermal stability of hydrogen passivated Mg acceptors in GaN is a function of the annealing ambient employed, with H2 leading to a reactivation temperature approximately 150 °C higher than N2. The dissociation of Mg–H complexes and the loss of hydrogen from GaN are sequential processes, with reactivation occurring at ⩽700 °C for annealing under N2, while significant concentrations of hydrogen remain in the crystal even at 900 °C in implanted samples. The hydrogen is gettered to regions of highest defect density such as the InGaN layer in GaN/InGaN double heterostructure.

Growth of AlN by metalorganic molecular beam epitaxy

Thin film AlN has been grown on Al2O3 and GaAs substrates by metalorganic molecular beam epitaxy using amine bonded alane precursors and either tertiarybutylamine or nitrogen from a compact electron cyclotron resonance (ECR) plasma source operating at 2.45 GHz. Typical growth pressures were in the 0.5–1×10−4 Torr range. The growth rates, impurity backgrounds, and surface morphologies were examined for both nitrogen sources and both the solid and liquid alanes. In general, growth efficiencies were good for both alane precursors, allowing for deposition of the low temperature, ∼400 °C, AlN buffers needed for subsequent growth of GaN and InGaAlN alloys. Low growth temperatures could not be obtained using tertiarybutylamine, presumably due to poor decomposition efficiency of the source at low temperatures. The structural quality of material grown at high temperatures from the ECR plasma was measured by atomic force microscopy, high resolution x‐ray diffraction, and transmission electron microscopy, indicating...