Search

We have used positron annihilation spectroscopy to study the high-pressure annealing induced thermal recovery of vacancy defects in free-standing GaN grown by hydride vapor phase epitaxy (HVPE). The results show that the in-grown Ga vacancy complexes recover after annealing at 1500-1700 K. Comparison of the experimental positron data with ab-initio calculations indicates that the Doppler broadening measurement of the electron momentum distribution is sensitive enough to distinguish between the N and O atoms surrounding the Ga vacancy. We show that the difference between the isolated V Ga in electron irradiated GaN and the V Ga -O N complexes in highly O-doped GaN is clear, and the Ga vacancy related defect complexes that start dissociating at 1500 K can be identified as V Ga -O N pairs.

We have performed a systematic study of magnesium and oxygen doping as well as the effect of annealing in metal organic chemical vapor deposition (MOCVD) and hydride vapor phase epitaxy (HVPE) grown Gallium nitride (GaN) by positron annihilation spectroscopy. The formation of vacancy-type defects identified here depends on the presence of impurity atoms as well as their concentrations. O doping generates vacancies in the Ga sublattice with concentrations proportional to the amount of O, whereas Mg atoms introduce vacancies in the N sublattice. In addition at very high Mg doping levels ( > 10 19 cm - 3 ) vacancy clusters with inhomogeneous depth profile are formed. Annealing dissociates vacancy-impurity pairs in MOCVD GaN:Mg and the defect profile is made homogeneous by vacancy migration. Vacancies in the Ga sublattice of HVPE GaN:O are shown to be complexed with O atoms by comparing the GaN:O data with the one obtained from electron-irradiated GaN, which contains isolated Ga vacancies. This comprehensive study shows that vacancy-type defects decorated by impurities are formed in both sublattices of GaN depending on the doping of the material.

Using positron annihilation measurements we observed the formation of thermal vacancies in highly As-doped Si. The vacancies form already at 700 K and upon cooling down the vacancies form stable vacancy-impurity complexes such as V – As 3 . Only a fraction of thermal vacancies is observed during equilibrium temperature measurements, but the quenching to room temperature reveals that the concentration of thermally generated V – As 3 defects is an order of magnitude larger. This mismatch is explained by positron detrapping from V – As 3 at high temperatures.

Positron lifetime and coincidence Doppler broadening measurements have been conducted to identify the defects formed during the annealing of electron irradiated highly As- and P-doped Si. We show that the vacancy–donor pairs (V–D1) formed during the electron irradiation migrate to form V–D2 at 450 K which in turn transform to V–D3 at 700 K. We furthermore show that the V–As3 defects anneal at 1100 K and that the formation and annealing of the V–D3 defects can explain the electrical compensation observed in highly doped Si.

We show that the detailed atomic structure of vacancy complexes in Si can be experimentally determined by combining positron lifetime and electron momentum distribution measurements. The divacancies, vacancy–oxygen complexes (A-centres) and vacancies paired with a donor impurity, V–P and V–As (E-centres), are identified in electron-irradiated Si. The formation of native vacancy defects is observed in highly As-doped Si at the doping level of 1020 cm−3. The defects are identified as monovacancies surrounded by three As atoms. We verify the formation mechanism of these defects by annealing experiments in irradiated Si. We show that the migration of V–As pairs at 450 K leads to the formation of V–As2 complexes, which convert to V–As3 defects at 700 K. These results explain the electrical deactivation and clustering of As in epitaxial or ion-implanted Si during postgrowth heat treatment at 700 K.

Gallium vacancy complexes have been identified in n-type bulk GaN by applying positron annihilation spectroscopy. Their formation is suppressed when the material becomes resistive by Mg doping, as expected from the behavior of the VGa formation energy as a function of the Fermi level. In Be-doped GaN vacancies are observed even in resistive material. The positron lifetimes show that their open volume is larger than expected for the N vacancy. A possible identification is a VN–BeGa complex, where the atoms neighboring the N vacancy are strongly relaxed outwards, thus increasing the open volume.

Positrons, which scatter back from the target and annihilate in chamber walls near the detectors, may cause a significant error in annihilation parameters. We have constructed a new UHV target chamber for slow positron beam studies. In our design special care has been taken to reduce the effect of backscattered positrons. Detector wells are designed for two-detector coincidence measurements and they are situated on both sides of the target. The distance of the wells from the target can be adjusted by simple manipulators. This enables optimization regarding the count rate and the rate of backscattered positrons hitting the detector wells. The magnetic field in front of the target is increased by permanent magnets situated behind the target. The increased magnetic field guides the backscattered positrons effectively away from the detectors. The increased magnetic field also focuses the beam spot strongly.

Positron annihilation measurements show that negative Ga vacancies are the dominant acceptors in n-type gallium nitride grown by hydride vapor phase epitaxy. The concentration of Ga vacancies decreases, from more than 1019 to below 1016 cm−3, as the distance from the interface region increases from 1 to 300 μm. These concentrations are the same as the total acceptor densities determined in Hall experiments. The depth profile of O is similar to that of VGa, suggesting that the Ga vacancies are complexed with the oxygen impurities.

We have used positron annihilation spectroscopy to study the introduction and recovery of point defects introduced by 0.45 and $2\phantom{\rule{0.3em}{0ex}}\mathrm{MeV}$ electron irradiation at room temperature in $n$-type GaN. Isochronal annealings were performed up to $1220\phantom{\rule{0.3em}{0ex}}\mathrm{K}$. We observe vacancy defects with specific lifetime of ${\ensuremath{\tau}}_{V}=190\ifmmode\pm\else\textpm\fi{}15\phantom{\rule{0.3em}{0ex}}\mathrm{ps}$ that we tentatively identify as N vacancies or related complexes in the neutral charge state in the samples irradiated with $0.45\phantom{\rule{0.3em}{0ex}}\mathrm{MeV}$ electrons. The N vacancies are produced at a rate ${\ensuremath{\Sigma}}_{\mathrm{N}}^{0.45}\ensuremath{\simeq}0.25\phantom{\rule{0.3em}{0ex}}{\mathrm{cm}}^{\ensuremath{-}1}$. The irradiation with $2\phantom{\rule{0.3em}{0ex}}\mathrm{MeV}$ electrons produces negatively charged Ga vacancies and negative nonopen volume defects (negative ions) originating from the Ga sublattice, at a rate ${\ensuremath{\Sigma}}_{\mathrm{Ga}}^{2.0}\ensuremath{\simeq}5\phantom{\rule{0.3em}{0ex}}{\mathrm{cm}}^{\ensuremath{-}1}$. The irradiation-induced N vacancies anneal out of the samples at around $600\phantom{\rule{0.3em}{0ex}}\mathrm{K}$, possibly due to the motion of the irradiation-induced N interstitials. Half of the irradiation-induced Ga vacancies anneal out of the samples also around $600\phantom{\rule{0.3em}{0ex}}\mathrm{K}$, and this is interpreted as the isolated Ga vacancies becoming mobile with a migration barrier of ${E}_{M}^{V,\mathrm{Ga}}=1.8\ifmmode\pm\else\textpm\fi{}0.1\phantom{\rule{0.3em}{0ex}}\mathrm{eV}$. Interestingly, we observe a change of charge state of the irradiation-induced negative ions from $2\ensuremath{-}$ to $1\ensuremath{-}$ likely due to a reconstruction of the defects in two stages at annealing temperatures of about 600 and $700\phantom{\rule{0.3em}{0ex}}\mathrm{K}$. The negative ions anneal out of the samples together with the other half of the Ga vacancies (stabilized by, e.g., N vacancies and/or hydrogen) in thermal annealings at $800--1100\phantom{\rule{0.3em}{0ex}}\mathrm{K}$.

Positron annihilation spectroscopy, supported by ab initio theory, has been applied to verify the decoration of Ga vacancies in $\mathrm{GaN}$ by oxygen and hydrogen. Our results indicate that the Doppler broadening measurement of electron momentum distribution is sensitive enough to distinguish even between N and O atoms neighboring the Ga vacancy. We identify isolated ${V}_{\mathrm{Ga}}$ in electron irradiated $\mathrm{GaN}$ and ${V}_{\mathrm{Ga}}\text{\ensuremath{-}}{\mathrm{O}}_{\mathrm{N}}$ complexes in highly O-doped high-purity $\mathrm{GaN}$. Evidence of H decoration of Ga vacancies is obtained in epitaxial $\mathrm{GaN}$ grown by metalorganic chemical-vapor deposition.

Using positron annihilation measurements we observed the formation of thermal vacancies in highly As and P doped Si. The vacancies start to form at temperatures as low as 650 K and are mainly undecorated at high temperatures. Upon cooling the vacancies form stable vacancy-impurity complexes such as V-As3. We determine the vacancy formation energy of E(f)=1.1(2) eV and the migration energy of E(m)=1.2(1) eV in highly doped Si. By associating these values with the vacancy-impurity pair, we get an estimate of 2.8(3) eV for the formation energy of an isolated neutral monovacancy in intrinsic Si.

We have used positron annihilation spectroscopy to determine the nature and the concentrations of the open volume defects in as-grown and electron irradiated (${E}_{\mathrm{e}\mathrm{l}}=2\text{ }\mathrm{M}\mathrm{e}\mathrm{V}$, fluence $6\ifmmode\times\else\texttimes\fi{}{10}^{17}\text{ }{\mathrm{c}\mathrm{m}}^{\ensuremath{-}2}$) ZnO samples. The Zn vacancies are identified at concentrations of $[{V}_{\mathrm{Z}\mathrm{n}}]\ensuremath{\simeq}2\ifmmode\times\else\texttimes\fi{}{10}^{15}\text{ }{\mathrm{c}\mathrm{m}}^{\ensuremath{-}3}$ in the as-grown material and $[{V}_{\mathrm{Z}\mathrm{n}}]\ensuremath{\simeq}2\ifmmode\times\else\texttimes\fi{}{10}^{16}\text{ }{\mathrm{c}\mathrm{m}}^{\ensuremath{-}3}$ in the irradiated ZnO. These concentrations are in very good agreement with the total acceptor density determined by temperature dependent Hall experiments. Thus, the Zn vacancies are dominant acceptors in both as-grown and irradiated ZnO.

We have applied positron annihilation and Hall-effect/resistivity experiments to identify and quantify vacancy-impurity complexes formed in molecular-beam epitaxially grown Si at very high As doping levels of $g{10}^{20}{\mathrm{cm}}^{\ensuremath{-}3}.$ The results indicate that $\mathrm{V}\ensuremath{-}{\mathrm{As}}_{3}$ is the dominant vacancy defect which exists at concentrations relevant to the electrical deactivation of doping. Larger complexes, tentatively identified as ${\mathrm{V}}_{2}\ensuremath{-}{\mathrm{As}}_{5},$ are also present at high concentrations. The $\mathrm{V}\ensuremath{-}{\mathrm{As}}_{3}$ and ${\mathrm{V}}_{2}\ensuremath{-}{\mathrm{As}}_{5}$ defects are removed by annealings at $800\ifmmode^\circ\else\textdegree\fi{}\mathrm{C}$ and $900\ifmmode^\circ\else\textdegree\fi{}\mathrm{C},$ respectively. The rapid thermal annealing leads to lowest concentrations of these deactivating defects, indicating that their formation is limited by migration processes during the cooling down.

Positron annihilation experiments have been applied to verify the formation mechanism of electrically inactive vacancy-impurity clusters in highly n-type Si. We show that the migration of V-As pairs at 450 K leads to the formation of V-As 2 complexes, which in turn convert to stable V-As 3 defects at 700 K. These processes manifest the formation of V-As 3 as the dominant vacancy-impurity cluster in highly n-type Si. They further explain the electrical deactivation and clustering of As in epitaxial or ion-implanted Si during postgrowth heat treatment at 700 K.

We have applied a low-energy positron beam and secondary ion mass spectrometry to study defects in homoepitaxial and heteroepitaxial GaN layers. Positron experiments reveal high concentrations of Ga vacancies in nominally undoped n-type GaN, where the conductivity is due to unintentional oxygen incorporation. Ga vacancies are observed in both homoepitaxial and heteroepitaxial layers, indicating that their formation is independent of the dislocation density. No Ga vacancies are detected in p-type or semi-insulating samples doped with Mg, as predicted by the theoretical formation energies. In samples where n-type conductivity is due to Si doping and the incorporation of oxygen impurities is suppressed, the concentration of Ga vacancies is much lower than in n-type samples containing oxygen. This indicates that the presence of oxygen donor in GaN promotes the formation of Ga vacancy. We suggest that this effect is due to the creation of ${\mathrm{V}}_{\mathrm{Ga}}{\ensuremath{-}\mathrm{O}}_{\mathrm{N}}$ complexes during the epitaxial growth.

Gallium vacancies and negative ions are observed in GaN bulk crystals by applying positron lifetime spectroscopy. The concentration of Ga vacancies decreases with increasing Mg doping, as expected from the behavior of the VGa formation energy as a function of the Fermi level. The concentration of negative ions correlates with that of Mg impurities determined by secondary ion mass spectrometry. We thus attribute the negative ions to Mg−Ga. The negative charge of Mg suggests that Mg doping converts n-type GaN to semi-insulating mainly due to the electrical compensation of ON+ donors by MgGa− acceptors.