James Griffin

Annealing of ion implantation damage in SiC using a graphite mask

For p-type ion implanted SiC, temperatures in excess of 1,600 C are required to activate the dopant atoms and to reduce the crystal damage inherent in the implantation process. At these high temperatures, however, macrosteps (periodic welts) develop on the SiC surface. In this work, the authors investigate the use of a graphite mask as an anneal cap to eliminate the formation of macrosteps. N-type 4H- and 6H-SiC epilayers, both ion implanted with low energy (keV) Boron (B) schedules at 600 C, and 6H-SiC substrates, ion implanted with Aluminum (Al), were annealed using a Graphite mask as a cap. The anneals were done at 1,660 C for 20 and 40 minutes. Atomic force microscopy (AFM), capacitance-voltage (C-V) and secondary ion mass spectrometry (SIMS) measurements were then taken to investigate the effects of the anneal on the surface morphology and the substitutional activation of the samples. It is shown that, by using the Graphite cap for the 1,660 C anneals, neither polytype developed macrosteps for any of the dopant elements or anneal times. The substitutional activation of Boron in 6H-SiC was about 15%.

Studies of deep centers in dilute GaAsN and InGaAsN films grown by molecular beam epitaxy

Abstract Deep levels spectra DLTS, 77 K photoluminescence (PL) spectra and photosensitivity were measured for GaAsN and InGaAsN films with low N and In concentration grown by molecular beam epitaxy and in GaAs films grown on GaAsN buffer. It is shown that the bandedge luminescence intensity is greatly decreased in GaAsN, GaAs/GaAsN and particularly in InGaAsN structures compared to the homoepitaxial GaAs. Comparison of the DLTS and PL spectra strongly suggests that the main recombination center in such films is the EL3-like electron trap whose concentration greatly increases upon In and N incorporation into the solid solution. Based on published results the trap is associated with substitutional oxygen on As site and the results are discussed in view of such possible assignment.

Nitrogen activated bowing parameter of GaAs1-xNx (x ≤ 1%) obtained from photoreflectance spectra

Abstract Photoreflectance studies were made for a series of GaAs 1− x N x samples with x ⩽0.01. The photoreflectance studies reveal that the composition dependent bowing parameter of the band gap obtained from the photoreflectance measurements are consistent with those from other optical measurements and also from the first principle supercell calculation.

Nitrogen-activated bowing of dilute InyGa1-yAs1-xNx based on photoreflectance studies

The dependence of the fundamental band gap and higher-lying critical-point energies of dilute-nitrogen Ga1−yInyAs1−xNx epilayers on nitrogen mole fraction (x), for x⩽0.0125, and temperature, from 20 to 300 K, was investigated by photoreflectance spectroscopy. The band gap, EG, was found to decrease with increasing x in a highly nonlinear manner. The bowing parameter (the second-order parameter b in a quadratic expression for the dependence of EG on x) was found to become less negative with increasing x; the value of b changed from −50 eV, at very low nitrogen fraction, to −20 eV, at x>0.01. These results strongly suggest that nitrogen-related impurity levels arise within the band gap of dilute-nitrogen Ga1−yInyAs1−xNx alloys.The dependence of the fundamental band gap and higher-lying critical-point energies of dilute-nitrogen Ga1−yInyAs1−xNx epilayers on nitrogen mole fraction (x), for x⩽0.0125, and temperature, from 20 to 300 K, was investigated by photoreflectance spectroscopy. The band gap, EG, was found to decrease with increasing x in a highly nonlinear manner. The bowing parameter (the second-order parameter b in a quadratic expression for the dependence of EG on x) was found to become less negative with increasing x; the value of b changed from −50 eV, at very low nitrogen fraction, to −20 eV, at x>0.01. These results strongly suggest that nitrogen-related impurity levels arise within the band gap of dilute-nitrogen Ga1−yInyAs1−xNx alloys.

Interface properties and deep levels in InGaAsN/GaAs and GaAsN/GaAs heterojunctions

Abstract Interface properties of dilute slightly lattice mismatched GaAsN/GaAs (0.35 at.% N) and closely lattice matched InGaAsN (1 at.% In, 0.35 at.% N) heterojunctions (HJs) were studied by means of capacitance–voltage profiling, deep levels transient spectroscopy (DLTS) and current–voltage measurements. It is found that the lattice matched HJs show no electrical breakdown when the space charge region crosses the interface. The carrier concentration profiles in such HJ show, as expected, the accumulation region on the low-bandgap side and the depletion region on the high-bandgap side of the HJ. This is not the case for the GaAsN/GaAs (GaAsN layer on top) and the GaAs/GaAsN (GaAs layer on top) HJ. The density of deep traps in GaAsN, InGaAsN films and in GaAs films grown on GaAsN underlayers was very much higher than in epitaxial GaAs films. The dominant deep centers were the EL6 and the EL3 electron traps. The interface regions of the GaAs/GaAsN and the InGaAsN/GaAs HJs were shown to be enriched by EL3 traps, while for the GaAsN/GaAs HJ those regions were enriched by EL6 traps which was associated with the former films being Ga-rich and thus facilitating incorporation of oxygen on As sites.

Optical properties of GaAs1−xNxalloys grown by molecular beam epitaxy

Optical properties of GaAs1−x N x alloys grown by molecular beam epitaxy using GaAs (001) as the substrate have been studied. These include photoluminescence (PL), cathodoluminescence (CL), photocurrent and photomemory effects. The low-temperature (77 K) PL characteristics were measured on samples with 0–0.105% N content. The wide emission band indicates the defective nature of the materials. The widening of the band for materials with increasing nitrogen concentration also suggested that the concentration of defect states in these materials dramatically increased with increasing nitrogen content. The PL and CL spectra for GaAs1−x N x layer 1854 did not show identical characteristics. Some layers showed a very sharp fall in photocurrent at low temperatures, indicating a very sharp photoquenching and an interaction between antisite, interstitial and vacancy defects. The photomemory effect, which causes photoquenching and the transition from the EL2 to the EL2* metastable state, was strongly influenced by the optical exposure and thermal history of the sample.

Study of SiC polytype heterojunctions

Abstract The results of a study of the chemical vapor deposition (CVD) growth and characterization of SiC polytype heterojunctions using ON-AXIS 4H and 6H SiC substrates are presented. Characterization results along with a pre-growth hydrogen/propane etch study were used as a baseline for developing a growth process for the SiC polytype heterojunctions. Grown heterojunctions show almost no polytype inclusions. An oxidation study to map polytype inclusions in the layers grown under different conditions show a high (>95%) polytype homogeneity. Transmission electron microscopy (TEM) shows no visible defects and/or columnar growth features. TEM diffraction patterns obtained are very strong, further indicating crystalline 3C–SiC epitaxial layers. The FWHM of the X-ray scans are very narrow, between 0.01 and 0.02°, an excellent figure further indicating single crystalline 3C–SiC epitaxial layers. Iso-intensity contours from the reciprocal space maps show very little broadening in ω which indicates less mosaicity in the samples. The ω–2θ direction shows almost no broadening implying that there is almost no strain in the material.

Defects and localized states in MBE-grown GaAs1−xNxsolid solutions prepared by molecular-beam epitaxy

Defects and localized states have been studied for molecular-beam-epitaxy (MBE)-grown high-resistivity and undoped GaAs1−x N x films with a N concentration not exceeding approximately 1.0 at.%. The crystalline quality of the films and hence the defects and localized states were determined by high-resolution X-ray diffraction, photoluminescence spectra, capacitance versus voltage measurements and photoinduced current transient spectra of GaAs and GaAs1−x N x layers. It was concluded that incorporation of low concentrations of N into MBE-grown GaAs1−x N x films promotes the formation of high densities of deep centres similar to EL2 donors, leading to heavy compensation of the films by some unidentified acceptors. GaAs antisite acceptors were believed to be responsible for the said compensation. A prominent defect band near 1.33-1.38 eV also appeared to be associated with these defects. The most prominent centres in dilute GaAs1−x N x films with N content less than 0.35 at.% seem to be the EL2 donors and the hole traps located near E v +0.3 eV.