Dominant Native Defects Research Articles

An interatomic potential study of the properties of gallium nitride

A set of interatomic pair potentials were derived for gallium nitride within the shell-model approach. It was shown that the potential set successfully reproduces the properties of the fourfold-coordinated wurtzite and zinc-blende structures as well as the sixfold-coordinated rock-salt structure. The high-pressure phase transition from wurtzite to rock-salt structure is correctly described yielding the phase transition pressure of 50 GPa. The calculated formation energies of intrinsic point defects reveal that vacancies are the dominant native defects in GaN. Lastly, the calculated structure relaxation of the zinc-blende (110) surface predicts a layer rotation angle of in agreement with the results of recent first-principles calculations.

Ab initio study of oxygen point defects in GaAs, GaN, and AlN.

We have studied oxygen point defects with the plane-wave pseudopotential method in GaAs, GaN, and AlN. The calculations demonstrate a qualitatively different behavior of oxygen impurities in these materials. ${\mathrm{O}}_{\mathrm{As}}$ in GaAs acts as a deep center with an off-center displacement and negative-U behavior, in agreement with the experimental data. ${\mathrm{O}}_{\mathrm{N}}$ in GaN is found to be a shallow donor with a low formation energy, and is suggested to act as a partial source for the unintentional n-type conductivity commonly observed in GaN. O in AlN is also found to easily substitute for N, which is consistent with the experimentally observed large oxygen concentrations in AlN. However, ${\mathrm{O}}_{\mathrm{N}}$ in AlN is shown to be a deep center due to the wide band gap, in contrast with ${\mathrm{O}}_{\mathrm{N}}$ in GaN. Our calculations thus predict that isolated oxygen acts as a DX-type center in ${\mathrm{Al}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{\ensuremath{-}}\mathit{x}}$N alloys. Results for other oxygen point defect configurations and for the dominant native defects are also presented. \textcopyright{} 1996 The American Physical Society.

Compositional dependence of cation impurity gettering in Hg1−xCdxTe

Cation impurity gettering in Hg1−xCdxTe is described in the context of process models which include the interactions of the impurities and the dominant native point defects. Experimental results are presented using secondary ion mass spectroscopy (SIMS) profiles of Au redistribution in Hg1−xCdxTe (x = 0.2,0.3,0.4) following Hg anneals and ion mills, which are processes known to inject excess Hg interstitials. In either process, the IB impurity distributes preferentially to high vacancy regions. The junction depth of the low to high impurity transition is determined by SIMS. For Hg-rich anneals of Au-doped high vacancy concentration material, the impurity junction behavior with respect to anneal time and temperature is compared to that expected for type converted electrical junctions in vacancy-only material. For milled Au-doped Hg0.7Cd0.3Te with a high vacancy concentration, the impurity junction depths are approximately proportional to the amount of material removed, as was the case with x = 0.2 material. Hg anneal type-conversion rates are found to have a strong compositional dependence which compares favorably with the strong self-diffusion coefficient dependence on x-value. In contrast, the mill conversion rate has a weak x-value dependence. Effects of trace vs dominant Au levels compared to the background vacancy concentration are quantified. True decoration of intrinsic defect processes requires Au <<[Cation Vacancies].

Optical Characterization of GaSb-Based Ternary and Quaternary Alloys Grown by Liquid-Phase Epitaxy at Low Temperatures

GaSb and GaSb-based alloys are grown by liquid-phase epitaxy (LPE) and characterized by photoluminescence (PL) and Raman spectroscopies. Although the concentration of the dominant native defect decreases with decreasing growth temperature, the crystalline quality of GaSb grown at 270°C is very poor: near-band-edge emission is not observed in the PL measurement and a peak due to forbidden scattering strongly appears in the Raman spectrum. On the other hand, AlxGa1-xSb and Ga1-xInxSb ternary alloys and AlxGa1-x-yInySb quaternary alloys grown at the same temperature exhibit much better crystal quality than the binary: the near-band-edge emission is clearly observed in the PL spectra and only the allowed mode is detected in the Raman measurement. Thus we can improve the quality of the low-temperature-LPE-grown layer through alloying.

Pressure as a Probe of Deep Levels and Defects in Semiconductors: GaAs, GaP and Their Alloys

Measurements of the effects of pressure on the thermal electron emission rate and capture cross section for a variety of deep electronic levels in GaAs, GaP and their alloys have yielded the pressure dependences of the energies of these levels in the bandgaps, allowed evaluation of the breathing mode lattice relaxations accompanying carrier emission or capture by these levels and revealed trends which lead to new insights into the nature of the responsible defects. Emphasis is on deep levels believed to be associated with simple defects. Specifically, results will be summarized for the donor levels of the dominant native defect known as EL2 in GaAs, which is believed to be associated with the arsenic antisite, and on the radiation-induced El and E2 levels in GaAs, GaP and their alloys, which are believed to be due to arsenic (or phosphorous) vacancies. The results are discussed in terms of models for the defects responsible for these deep levels.

Model of EL2 formation in GaAs

It is demonstrated that existing thermodynamic data on the native deep donor, EL2, in melt-grown and epitaxially grown GaAs are consistent with that defect having the atomic structure AsGaVGa. In melt-grown GaAs at high temperatures (∼1200 °C) arsenic antisite defects appear as the complex AsGaVAsVGa. As the temperature drops toward 1000 °C and the equilibrium concentration of divacancies decreases this complex dissociates, the divacancies outdiffusing and the antisites capturing gallium vacancies to form EL2. In GaAs grown by organometallic vapor-phase epitaxy it is suggested that the arsenic interstitial is the dominant native defect produced in equilibrium with the vapor and that it dictates the deviation from stoichiometry of the epilayer. Below the growth interface these interstitials rapidly react with indiffusing divacancies to form primarily arsenic antisites. Other divacancies then react with the antisites to briefly form the complexes AsGaVAsVGa which, in the nonuniform temperature regime of the epilayer, dissociate into EL2 and arsenic vacancies. The model predicts [EL2]∝(As/Ga)1/2 in agreement with selected data and predicts that the EL2 concentration will increase under a nonuniform thermal anneal. It also accounts for the formation of EL2 in GaAs grown by molecular-beam epitaxy when subsequently annealed at ∼800 °C in a nonuniform temperature environment.

Mechanisms of incorporation of donor and acceptor dopants in (Hg,Cd)Te alloys

The usefulness of the quasichemical approach in predicting the defect structure of undoped and doped (Hg,Cd)Te crystals is emphasized. To begin with, the defect structure of undoped (Hg,Cd)Te crystals is established along with mass action constants for the intrinsic excitation process and the formation of the dominant native acceptor defects. Subsequently, the defect structures of donor doped and acceptor doped (Hg,Cd)Te crystals are established. With the values of the mass action constants established for the undoped crystals along with those for the doped crystals, the concentrations of defects present at the high temperature are calculated along with the carrier concentrations expected in the cooled crystals. The results of the carrier concentration calculations are compared with the experimental values to verify the validity of the incorporation mechanisms established for the doped crystals.

Kinetics of formation and dissociation of a dominant native defect (EL2) in GaAs

It is shown that a simple kinetic model can account for existing data both on the formation of the native defect EL2 in the temperature range 644–800 °C in GaAs samples from which EL2 was eliminated by a 1200 °C anneal/quench and on the disappearance of EL2 during anneals in the temperature range 1000–1200 °C. Our analysis suggests that EL2 consists of VGa bound to an unidentified ‘‘kernel’’ which, if not actually stable at temperatures up to 1200 °C, forms relatively rapidly at the lower temperatures and dictates the final concentration of EL2 in the sample. The change in enthalpy involved in the capture or release of VGa by the kernel is estimated to be 5.6 eV.

Concentration of native point defects in Si single crystals at high temperatures.

The concentration of dominant native defects in a float-zoned Si single crystal at high temperatures was directly determined from the difference between the macroscopic linear thermal expansion and the lattice-parameter thermal expansion. The concentration fraction of native defects in thermal equilibrium at 1300 K is (3.6\ifmmode\pm\else\textpm\fi{}2.7)\ifmmode\times\else\texttimes\fi{}${10}^{\mathrm{\ensuremath{-}}7}$, or (1.8\ifmmode\pm\else\textpm\fi{}1.3)\ifmmode\times\else\texttimes\fi{}${10}^{16}$ atoms ${\mathrm{cm}}^{\mathrm{\ensuremath{-}}3}$. It is found that vacancies are predominant over interstitial atoms in concentration.

Hg1-xCdxTe: Defect Structure Overview

ABSTRACTNative point defects play an important role in HgCdTe. Here we discuss some of the relevant mass action equations, and use recently calculated defect formation energies to discuss relative defect concentrations. In agreement with experiment, the Hg vacancy is found to be the dominant native defect to accommodate excess tellurium. Preliminary estimates find the Hg antisite and the Hg interstitial to be of comparable densities. Our calculated defect formation energies are also consistent with measured diffusion activation energies, assuming the interstitial and vacancy migration energies are small.

Dopant-dependent formation and annealing of the dominant native deep-level defect in liquid-phase epitaxial AlGaAs

In liquid-phase epitaxial GaAs the hole traps with levels at EV+0.40 eV and EV+0.70 eV (labeled A and B) are commonly observed. Here, AlGaAs LPE layers doped with Mg, Zn, Si, Ge, Sn, or Te are investigated in order to assess the solubility and the annealing characteristic of both hole traps. It is shown that the concentration of the traps is strongly dependent on the incorporated impurity. The formation of the underlying native defect is determined by the type of conductivity, the amount of the dopant, as well as the lattice site occupied by the impurity atom. The mechanism of dopant-induced solubility is clearly demonstrated for a particular deep-level defect in a semiconductor. Experimental evidence is given that the formation of the native defect in n-type material is enhanced by Ga-site dopants. The solubility characteristic unambiguously reveals that the native defect is mobile at temperatures above 550 °C and that the deep levels are of acceptorlike character. By incorporating high concentrations of As-site dopants the formation of the native defect is significantly suppressed in n-type as well as in p-type material. Therefore, it is suggested that the native defect studied here is linked with the As lattice site. The GaAs antisite defect model is corroborated with regard to the charge character, the defect site, and the two coupled charge states. Usually, the native defect is stable against thermal treatment. Annealing, most likely due to a defect reaction, is observed only in the presence of Ge or Si.

Native Point Defects and Nonstoichiometry in GaAs (I). Homogeneity Region of Gallium Arsenide

AbstractHomogeneity region and the concentration of dominant native point defects along its boundary are calculated by the method of quasichemical reactions with the use of experimental data on the structure type of solid solutions formed by the components in GaAs. The entropy and enthalpy of the formation of dominant native point defects in GaAs are obtained. Temperature ranges of the formation of microdefects‐finely dispersed Ga and As precipitates are considered.

Defects in GaAs

A comprehensive review of defects in GaAs with focus on native point defects and dislocations is given. The effects of these defects on devices are also considered. It is pointed out that a unified point defect model cannot at present be drawn from the available information. The importance of including anti-site defects in the point defect models which have hitherto only considered vacancies and interstitials is stressed. Attention is drawn to the need for understanding the dominant equilibrium native defects in GaAs both from fundamental and technological considerations. In this respect new experimental techniques are suggested to understand and control the defect structure. The current understanding of dislocations in GaAs is very much in its infancy compared to that in elemental semiconductors. Both theoretical work and careful experiments are wanting. This is essential since dislocations have been directly implicated in the degradation of GaAs devices.

Conductivity type conversion in (Hg,Cd)Te

The behavior of native defects in Hg0.70Cd0.30Te and Hg0.77Cd0.23Te prepared by solid state recrystallization has been investigated. Galvanomagnetic data on n‐type samples reannealed under various conditions in Hg vapor are presented. Reconversion from n‐type to p‐type conductivity is found to depend on sample composition and temperature, ambient Hg pressure, and on the residual extrinsic donor impurity concentration. It is concluded that conductivity‐type conversion in (Hg,Cd)Te may be mainly attributed to residual extrinsic donor impurities and dominant native acceptor defects in agreement with previously published results.

Defect Structure of α‐Al2O3 Doph with Magnesium

The conductivity of single crystals of Al2O3+ Mg and the ionic and electronic transference numbers were measured at high temperatures as a function of orientation, oxygen pressure, and temperature. Optical absorption in the visible range was measured on cooled annealed crystals. The results are interpreted on the basis of a model with either Ali3‐ or VO as the dominant native defects and lead to expressions for the ionic mobility as a function of T and orientation, and for the position of the MgAl’level in the forbidden gap.

Defect structure and electronic donor levels in stannic oxide crystals

By measuring the conductivity of stannic oxide crystals as a function of oxygen partial pressure at elevated temperatures, it is shown that the dominant native defect in SnO2 is a doubly ionizable oxygen vacancy. Both donor levels of this defect, the first 30 meV deep and the second 150 meV deep, are identified and a model is presented that explains previous results. The behavior in hydrogen is contrasted to that in oxygen, and preliminary results are presented indicating that hydrogen introduces a donor 50 meV deep.

Diffusion of Cd in CdS

The self-diffusion of Cd in CdS has been measured under a variety of doping and firing conditions. Under saturated Cd pressure the diffusion coefficient is given by $D=3\mathrm{exp}(\ensuremath{-}2.0 \mathrm{eV}/kT)$. Under sulfur pressure at 800\ifmmode^\circ\else\textdegree\fi{}C the diffusion coefficient is found to be linearly dependent on the donor-impurity concentration. A new mechanism for diffusion is proposed to explain the results. No evidence for interstitial Cd was observed and it is suggested that Cd and sulfur vacancies (Schottky defects) are the dominant native defects in CdS.