H/He interaction with vacancy-type defects in α-Al₂O₃ single crystals studied by positron annihilation

Abstract

To probe the interaction of H and He, produced by tritium decay, with vacancy-type defects of α-Al₂O₃ as a tritium permeation barrier (TPB) in fusion reactors, α-Al₂O₃ single crystals were treated in pure Ar gas, D₂ gas and T₂ gas with subsequent tritium aging, respectively, and then their positron annihilation lifetimes and the type of defects that may contribute to the observed positron lifetime components were studied, in combination with DFT results. More monovacancies and vacancy clusters were formed in the thermally hydrogenated samples when compared to the fresh and Ar-annealed samples, indicating the stabilizing effect of hydrogen; this was consistent with the Fermi level position of α-Al₂O₃ moving towards the conduction band minimum (CBM) in the presence of hydrogen impurities, resulting in V_Al³⁻ and [V_Al³⁻–H⁺]²⁻ becoming more stable, as observed by DFT calculations. The monovacancies were slightly eliminated when the samples were thermally annealed and then aged in T₂ gas at room temperature, indicating that He filled the vacancies. This was consistent with it being favourable for He atoms to occupy Al vacancies, with He_Al³⁻ forming most readily, whilst more vacancy clusters were continuously induced, suggesting that Al–O bonds weakened and thus nano-hardness decreased with an external load. This study provides the first evidence that Al vacancies can be stabilized by H and filled with He, which will provide further novel TPB design opportunities.

---

1. Introduction

Various oxide materials, such as Al₂O₃, Y₂O₃, ZrO₂ and so on, are used in fusion reactors as plasma diagnosis windows, electric insulators, oxide dispersion-strengthened (ODS) ferritic steels and tritium permeation barriers (TPBs).¹ The use of TPBs of α-Al₂O₃ as the structural material, due to its chemical stability and low solubility for hydrogen, is an efficacious way to suppress hydrogen isotopic permeation through the steel walls of ducts for hydrogen storage and distribution, protection against hydrogen embrittlement and control of the tritium inventory in future fusion reactors like the ITER.^2,3

The effectiveness of TPBs depends critically on the thermodynamics and kinetics of hydrogen transport within the barrier materials.^2–4 It has been proposed that the properties of metal oxide materials are directly or indirectly connected to the presence of defects, such as vacancies, impurities, dislocations and grain boundaries.^3,5 However, to the authors’ knowledge, interactions between hydrogen and defects in α-Al₂O₃, and their effects on the thermodynamics and kinetics of the H mass transport within α-Al₂O₃, have not been studied well; this is essential to explicitly identify and control the defect roles in H transport within α-Al₂O₃. Moreover, a deep knowledge of the physical mechanisms of deuterium, tritium and helium presenting in the selected materials due to tritium decay and the (n, a) reaction, and of their synergy determining the long-term stability of the materials, is important to assess long-term predictions of behaviour of operating components in fusion reactors.³

It is well known that vacancies are one of the simplest defect types in crystal structures, and so represent a benchmark for experimental and theoretical understanding, which is the foundation of exploring the fascinating properties of materials. What’s more, theoretical calculations have predicted that H and He-related vacancy defects in α-Al₂O₃ should play an enhancement role in low H transport in TPBs.^6,7 H in α-Al₂O₃ is present mainly in the H_i⁺ state, and it is favourable for it to exist in the [V_Al³⁻–H⁺]²⁻ and H_O⁺ forms. The positive binding energies of [V_Al³⁻–H⁺] and H_O⁺ should increase the activation energy of H migration, decreasing H mobility, which is favored for low H transport in α-Al₂O₃ TPBs.⁶ In the presence of He, forms of H–He complex defects in α-Al₂O₃ are [He_i–H⁺]⁺, [He_Al³⁻–H⁺]²⁻ and [HO⁺–He_i]⁺. [He_Al³⁻–H⁺]²⁻ and [H_O⁺–He_i]⁺, with their positive binding energies, will increase the activation energy of H migration in α-Al₂O₃, which is also favorable for low H transport in TPBs.⁷ It can be seen from the theoretical findings outlined that the preferred positions and dominant migration modes for H(He) in α-Al₂O₃ depend on the presence of intrinsic defects acting as traps, especially vacancies. Thus, it is urgent to experimentally investigate H/He interactions with vacancy-type defects in α-Al₂O₃, providing further novel TPB design opportunities.

The experimental configuration applied for the study of H/He behaviour in the near-surface region of solids generally follows a three step approach:³ (1) H/He incorporation in the material, (2) thermal annealing under fully controlled conditions and (3) H/He measurement. Due to the very low H/He dissolution content in α-Al₂O₃ and sensitivity, the experimental H/He measurement signal of α-Al₂O₃ with H/He incorporated is always weak with conventional methods. Consequently, the identification and interactions of H/He-related defects are still debatable issues. Infrared (IR) spectroscopy has proven to be a powerful tool to experimentally probe H structures. Based on known IR absorption peaks for molecules containing hydroxyl groups, O–H (O–D) absorption bands are in the range of 3183.7–3309.3 cm⁻¹ (2375.8–2469.6 cm⁻¹), with the angle of OH⁻ (OD⁻) ions protruding from the basal plane (most are <15°), depending on the sample conditions.^8,9 One can compare IR spectra for samples immediately following hydrogenation and after certain periods at different temperature ‘anneals’, investigating the stability of OH complexes. However, He-related defects cannot be visibly detected by IR spectroscopy. From desorption peaks of thermal desorption spectrometry (TDS), one can extract the values of the different dissociation energy H(He)-related defects.³ However, which H/He-related defects can be attributed to IR modes or TDS dissociation energies must be identified with the aid of other tools, for example DFT calculations.

Positron annihilation spectroscopy (PAS), with its sub-ppm sensitivity, is a powerful technique to identify neutral or negatively charged vacancy-type defects in materials of interest, and has been extensively used for characterization of such defects.¹⁰ Due to the lack of positive ion cores, positrons are trapped by the attractive potential, which leads to characteristic changes in the measured annihilation parameters. Therefore, positron lifetime measurements can be traced to the H and He decorating vacancies that induce the positron traps, which can give some indication about the nature of positron traps that are established during TPB operation.

Thus, in the present work, using PAS, positron annihilation experiments of α-Al₂O₃ single crystals were carried out at different annealing conditions, including annealing samples in Ar gas, D₂ gas and T₂ gas with subsequent tritium aging; the type of defects that may contribute to the observed positron lifetime components are discussed, with DFT results. Importantly, these results complement the systematic investigation of H/He interactions with defects of α-Al₂O₃, playing an important role in understanding the roles of hydrogen/helium in α-Al₂O₃ and in the design of efficient T-permeation-resistance materials for oxide TPBs. Moreover, the method can be extended to the oxide materials used in fusion reactors.

2. Experimental details

The α-Al₂O₃ piece samples, purchased from MTI Corporation (China), were single crystals of 17 mm external diameter and 0.5 mm thickness (purity: >99.99%), with a density of 3.98 g cm⁻³. They were polished to optical grade.

To introduce additional hydrogen or deuterium and determine the H effects on defect behavior, samples were sealed in quartz ampoules filled with 1.5 atm of deuterium gas (99.999%). The annealing was performed at 950 °C for 100 h and samples were quenched to room temperature by immersing the ampoules in water. IR absorption spectra were measured with a Fourier-transform IR spectrometer (Nicolet IS50R). The instrumental resolution was 2 cm⁻¹. The IR absorption spectra of the D₂-annealed samples are shown in Fig. 1. Peaks at around 2440–2480 cm⁻¹, with a main frequency of 2458.70 cm⁻¹, were observed. This observation was consistent with frequencies of OD peaks obtained earlier, shown in Table 1. Thus, D penetrated into the bulk α-Al₂O₃ under these annealing conditions.

Fig. 1 IR spectra of the deuterium-annealed α-Al₂O₃ samples just after annealing for 100 h at 950 °C. A spectrum from a sample without a D₂ anneal was used as a reference to calculate the absorbance spectra. All spectra were obtained at a sample temperature of 25 °C.

Table 1 OD frequencies of deuterium-annealed α-Al₂O₃ samples

	T, °C	OD^−, cm^−1	H charging method	Ref.
α-Al2O3	25	2458.7	Thermal annealing	This work
α-Al2O3	25	2436.9	Thermal annealing	8
α-Al2O3	25	2374.3, 2406.4, 2437.4, 2458.6	Electric-field assisted diffusion	9

---

Similarly, α-Al₂O₃ pieces were subsequently annealed in Ar gas (99.999%) at 950 °C for 100 h in the tube furnace, to identify the temperature effects on defect behavior and remove the defects induced by cutting and polishing. Two pieces were hydrogenated in tritium gas (99.9%) at 950 °C for 100 h in a stainless steel container, and stored at room temperature for 1.5 years under tritium gas, to identify the effects of helium-3 from tritium decay on defect behavior.

PAS was measured using a conventional fast coincident system at room temperature. The coincidence spectrometer had a prompt time resolution of 230 ps (FWHM) for the γ-rays from a ⁶⁰Co source selected under the experimental conditions. During the measurement, the positron source was sandwiched between two identical sample pieces. The source (20 μCi) of ²²Na was sandwiched between two identical sample disks. The positron lifetime spectrum containing 10⁶ counts was analyzed by the POSITRONFIT program to decompose several lifetime components.

The spectra were analyzed using a program based on a sum of several exponential decay terms:¹¹

Three lifetime components, τ₁, τ₂ and τ₃, and their corresponding percentage intensities, I₁, I₂ and I₃, were resolved in the case of the single crystal samples. Owing to the low intensity (<1%) of τ₃, which was attributed to positron annihilation from the surface state and/or macroscopic defects,¹¹ the effect of the longest lifetime τ₃ has not been discussed in this work.

τ_av gives the overall defect information, and it is calculated as follows:

3. Results

The shorter lifetime component (τ₁) and the corresponding intensity (I₁) in different α-Al₂O₃ crystals are shown in Fig. 2. In the fresh α-Al₂O₃ sample, the τ₁ was 111 ps. For the α-Al₂O₃ annealed in D₂ gas, the τ₁ was longer at 115 ps, 4% and 7% greater than in the fresh and Ar-annealed samples, respectively, with the relative intensity (I₁) increasing from 52% in the Ar-annealed sample to 54% in the D₂-annealed sample. For the α-Al₂O₃ annealed and then aged in T₂ gas, with tritium in α-Al₂O₃ decaying to helium-3, at room temperature for 1.5 years, the τ₁ was shorter at 113 ps, 2% smaller than in the D₂-annealed sample. Correspondingly, the relative intensity (I₁) of the sample decreased from 54% in the D₂-annealed sample to 52% in the T₂-aged sample.

Fig. 2 Positron lifetime component τ₁ (a) and the corresponding intensity I₁ (b) in different α-Al₂O₃ crystals: (1) fresh, (2) Ar-annealed, (3) D₂-annealed and (4) T₂-annealed and subsequently tritium-aged. The inset figures in (a) are local structures of V_Al³⁻ without and with H/He decoration.

The longer lifetime component (τ₂) and the corresponding intensity (I₂) in different α-Al₂O₃ crystals are shown in Fig. 3. In the fresh α-Al₂O₃, the τ₂ was 320 ps, with a intensity (I₂) of 46% (Fig. 3a). For the Ar-annealed sample, the τ₂ decreased to 313 ps, whereas the intensity (I₂) increased to 47%. In the D₂-annealed α-Al₂O₃, the τ₂ increased to 321 ps, with a intensity (I₂) of 45%. For the T₂-annealed and subsequently aged sample, the τ₂ slightly increased to 323 ps, with the intensity (I₂) increasing to 47%.

Fig. 3 Positron lifetime component τ₂ (a) and corresponding intensity I₂ (b) in different α-Al₂O₃ crystals: (1) fresh, (2) Ar-annealed, (3) D₂-annealed and (4) T₂-annealed and subsequently tritium-aged.

4. Discussion

4.1 The shorter lifetime component of α-Al₂O₃

The τ₁ is generally due to the free annihilation of positrons in a defect-free crystal, denoted as the first lifetime of positrons.^11,12 In a disordered system, smaller vacancies (such as monovacancies, etc.) or shallow positron traps (such as oxygen vacancies in oxides) may reduce the surrounding electron density,¹³ which increases the τ₁ lifetime.

As shown in Fig. 2, the τ₁ for the fresh sample was 111 ps, slightly shorter than the experimental values of the shorter lifetime, 117 ps in a single crystal,¹⁴ which could be ascribed to the inherent free positron lifetime in pristine α-Al₂O₃. However, the τ₁ for the Ar-annealed sample was smaller at 108 ps, 3% smaller than in the fresh sample. Correspondingly, the relative intensity (I₁) of the Ar-annealed sample decreased from 53% in the fresh sample to 52% (Fig. 2b). Therefore, the reduced τ₁ of the fresh sample reflected that a few monovacancies in the α-Al₂O₃ lattice were incorporated or annihilated during the annealing of the sample in the Ar gas. It has been reported that the relative stability of intrinsic point defects in α-Al₂O₃ is V_Al³⁻ > V_O⁰ > O_i²⁻ > Al_i³⁺ at the Fermi level under Al-rich conditions.^6,15 Thus, τ₁ of the Ar-annealed sample could be a mixed contribution from the annihilation of trapped positrons in V_Al³⁻ or V_O⁰. Correspondingly, the reduction of τ₁ could be caused by thermally-activated mutual annihilation of intrinsic point defects.¹⁶ This is a typical phenomenon of materials annealed at high temperature.

The prolonged τ₁ of the D₂-annealed sample (Fig. 2a) demonstrated that monovacancies were introduced into the α-Al₂O₃ lattice by the thermal hydrogenation. Correspondingly, the relative intensity (I₁) of the D₂-annealed sample also increased by 2% (Fig. 2b). Our DFT results suggested that H_i⁺, V_Al³⁻, V_O⁰ and O_i²⁻ are likely to exist in the form of [V_Al³⁻–H⁺]²⁻, H_O⁺ and HO_i⁻, but also in isolated point defects under Al-rich conditions.^6,17 In the thermal hydrogenation, H_O⁺, due to H-trapping of V_O⁰, cannot trap positrons efficiently, and the increase in positron lifetimes essentially points towards the generation or H-decoration of Al vacancies. Thus, the τ₁ of the D₂-annealed sample could be a mixed contribution from the annihilation of trapped positrons in V_Al³⁻, [V_Al^3––H⁺]²⁻ or V_O⁰. Correspondingly, the increase of τ₁, meaning that more monovacancies formed, demonstrated the stabilizing effect of hydrogen, in agreement with the reduction of the formation energies of V_Al³⁻ and [V_Al³⁻–H⁺]²⁻; this is due to the Fermi level position of α-Al₂O₃ moving towards the conduction band minimum (CBM) in the presence of the H impurity (Fig. 3 in ref. 6), and thus V_Al³⁻ and [V_Al³⁻–H⁺]²⁻ become more stable.

The reduced τ₁ of the T₂-aged sample (Fig. 2a) demonstrated that a few of the Al monovacancies were filled by helium atoms. This was consistent with DFT results which indicated that He atoms preferentially occupy Al vacancies, centers of octahedral interstitial site (OIS) or form a dumbbell around Al vacancies, forming He_i, He_Al³⁻, He_i–He_Al³⁻, [V_O⁰–He_i]⁰ and [O_i²⁻–He]²⁻ complexes.⁷ During the tritium aging at room temperature, tritium being present as T⁺, T_O⁺ and [V_Al³⁻–T⁺]²⁻ in α-Al₂O₃, [V_O⁰–He_i]⁰, as a result of the trapped T from T_O⁺ decaying, can trap positrons efficiently. The large increase in positron lifetimes essentially points towards the annihilation of Al vacancies, forming He_Al³⁻ and [He_Al³–He]³⁻, due to the trapped T from [V_Al³⁻–T⁺]^2– decaying.

It is worth mentioning here that the shorter positron lifetime in α-Al₂O₃ is still a debatable issue, as is evident from the large spread in the values reported by several authors. G. Moya et al. have reported the lowest values of 117 and 122 ps for sintered and single crystal α-Al₂O₃ respectively,¹⁴ whereas values of 150 and 130 ps for single crystal α-Al₂O₃ were reported by G. Molnár¹⁸ and K. P. Muthe,¹⁹ respectively. A value of 110 ps was obtained in the present work.

In addition, the local configurations of V_Al³⁻ without and with H/He decoration are shown in Fig. 2a. A V_Al³⁻ has 6 O neighbours, all of which are equivalent by the cubic symmetry. A H⁺ can bind to one of these O atoms, pointing toward the vacancy centre and forming a [V_Al³⁻–H⁺]²⁻ complex with a fully relaxed O–H bond length of 0.98 Å. Thus, [V_Al³⁻–H⁺]²⁻ can trap positrons efficiently and, hence, the large increase in positron lifetimes was found. A He atom substitutes on the Al site, locating at the point equidistant from three oxygen atoms of the close-packed O layer above V_Al³⁻ and forming a He_Al³⁻ complex. The distances between the He and the oxygen atoms are 2.084 Å. Upon adding a second He atom into He_Al³⁻, the 2nd He atom occupies the centre of the octahedral site along the c axis, forming a dumbbell with a He–He distance of 1.990 Å.⁷ Thus, the He_Al³⁻ can decrease the positron lifetimes.

4.2 The longer lifetime component of α-Al₂O₃

The τ₂ is attributed to positrons captured by defects of larger size.¹² The average electron density in larger-sized defects, such as vacancy clusters, vacancy complexes and dislocations, is lower than that in small size defects, which decreases the annihilation rate and increases the positron lifetime correspondingly. Therefore, the values of τ₂ are much larger than the values of τ₁.

In the fresh α-Al₂O₃, the τ₂ was 320 ps (Fig. 3a). This lifetime is comparable to those obtained in small vacancy clusters (300–500 ps), corresponding to irradiation-induced defects in single crystals after annealing at high temperatures.²⁰ The lifetime of 359 ps corresponded to the annihilation of positrons at oxygen vacancies,²¹ whereas the lifetime of 310–340 ps was associated with the annihilation of positrons trapped in Al-related vacancies.^14,18 Our DFT results show that the V_Al³⁻ and Schottky defects of {3V_O¹⁺ : V_Al³⁻} are dominant in pure α-Al₂O₃ under the Al-rich conditions.²² Thus, the lifetime (312–323 ps) measured for the α-Al₂O₃ was estimated to be due to mainly clusters and complexes of Al vacancies. Obviously, the larger-sized defects account for a smaller proportion than that of monovacancies in the fresh α-Al₂O₃, as I₂ < I₁.

For the Ar-annealed sample, the τ₂ decreased to 313 ps due to annealing the sample in Ar gas (Fig. 3a), which demonstrated that clusters and complexes of Al vacancies were annihilated. This resulted from flock interaction between V_Al³⁻ and V_O⁰ or their associates. There was a clear decrease in τ_av as shown in Fig. 4. τ_av decreased to 205 ps for the α-Al₂O₃ annealed in the Ar gas. Compared to the fresh α-Al₂O₃, the intensity of τ₁ and τ₂ for the Ar-annealed α-Al₂O₃ decreased and increased respectively, suggesting that Al monovacancies were predominately annihilated during the annealing of α-Al₂O₃ in the Ar gas.

Fig. 4 The variation of average lifetime (τ_av) of positrons annihilated in the α-Al₂O₃ annealed in different conditions: (1) fresh, (2) Ar-annealed, (3) D₂-annealed and (4) T₂-annealed and subsequently tritium-aged.

In the D₂-annealed α-Al₂O₃, the τ₂ increased to 321 ps, with the intensity (I₂) being 45% (Fig. 3); this demonstrated that hydrogenation induces the formation of some clusters or complexes of Al vacancies in the D₂-annealed α-Al₂O₃, which resulted from the stabilizing effect of hydrogen. There was a clear increase in τ_av, as shown in Fig. 4. τ_av increased to 209 ps, revealing that the vacancy-type defects were generated. In contrast to the Ar-annealed α-Al₂O₃, the increased τ₁ intensity and decreased τ₂ intensity of the D₂-annealed α-Al₂O₃ suggests that generation of Al monovacancies was preferred over the formation of clusters or complexes of Al vacancies.

For the T₂-annealed and subsequently aged sample, the τ₂ slightly increased to 323 ps because of tritium decay, with the intensity (I₂) increasing to 47% (Fig. 3). This demonstrated that more larger-sized defects were introduced inside the T₂-aged α-Al₂O₃, which may have resulted from disorder of the larger-sized defects, due to more and more helium atoms occupying them. There was a clear increase in τ_av, as shown in Fig. 4. τ_av continuously increased to 213 ps. Compared to the D₂-annealed α-Al₂O₃, the decreased τ₁ intensity and increased τ₂ intensity suggested that Al vacancy clusters were predominately introduced as the α-Al₂O₃ aged in the tritium gas. As a result of such larger-sized defects forming, the Al–O bands of α-Al₂O₃ would be weakened and the lattice would even be disordered, hence one can assume that the mechanical performance of the α-Al₂O₃ would degenerate.

Fig. 5 shows the nano-hardness of α-Al₂O₃ annealed at 950 °C for 100 h in deuterium and in tritium with subsequent tritium aging. The nano-hardness of the fresh α-Al₂O₃ was 21.43 GPa. When the α-Al₂O₃ was annealed in the D₂ gas, the nano-hardness decreased by 30% from 21.43 GPa to 15.3 GPa. The nano-hardness continuously decreased by 70% as the α-Al₂O₃ was annealed in tritium with subsequent aging for 1.5 years. Taking into consideration the clear increase of τ_av for the overall defects shown in Fig. 4, after tritium and helium atoms were introduced into α-Al₂O₃, one can assume that the hardness decrease could be caused by the weakening of the Al–O bands or even lattice disorder of the α-Al₂O₃. An uneven distribution of H/He-related defects would cause a stress concentration to be produced by an external load, which could also be the reason behind such a hardness decrease.

Fig. 5 Nano-hardness of α-Al₂O₃ annealed at 950 °C for 100 h in different conditions: (1) fresh, (2) D₂-annealed and (3) T₂-annealed and subsequently tritium-aged.

5. Conclusions

α-Al₂O₃ single crystals were subjected to different anneals in pure Ar gas, D₂ gas and T₂ gas with subsequent tritium aging, and then they were investigated by positron annihilation spectroscopy to provide evidence for interactions of H and He with vacancy defects of α-Al₂O₃.

More monovacancies and vacancy clusters were formed in the thermally hydrogenated samples, due to the stabilizing effect of hydrogen when compared to the fresh sample and the sample annealed in pure Ar gas. This was consistent with H impurities leading to the Fermi level position of α-Al₂O₃ moving towards the conduction band minimum (CBM), observed by DFT calculations, and thus with V_Al³⁻ and [V_Al³⁻–H⁺]²⁻ becoming more stable. The monovacancies were slightly eliminated when a sample was annealed and then aged in T₂ gas, which was consistent with He atoms preferentially filling Al vacancies, with He_Al³⁻ formation, and more vacancy clusters being continuously induced. The results are an important complement to our understanding of the roles of hydrogen/helium in α-Al₂O₃ and in designing TPBs of oxide materials.

Acknowledgements

This work is supported by the National Magnetic Confinement Fusion Science Program (No. 2013GB110006B) and National Natural Science Foundation (No. 21471137, 11275175) of China and the fund from Center for Fusion Energy Science Technology, CAEP (No. 2014-0304-03). We appreciate Dr Hua Luhui and Mr Jing Wenyong for PAS and tritium-annealed experiments.

Notes and references

1. K. Hashizume, K. Ogata, M. Nishikawa, T. Tanabe, S. Abe, S. Akamaru and Y. Hatano, J. Nucl. Mater., 2013, 442, S880.
2. G. W. Hollenberg, E. P. Simonen, G. Kalinin and A. Terlain, Fusion Eng. Des., 1995, 28, 190.
3. A. Causey Rion, A. Karnesky Richard and S. Marchi Chris, Compr. Nucl. Mater., 2012, 4, 511.
4. X. Xiang, X. Wang, G. Zhang, T. Tang and X. Lai, Int. J. Hydrogen Energy, 2015, 40, 3696.
5. K. Arshak and O. Korostynska, Mater. Sci. Eng., B, 2006, 133, 1.
6. G. Zhang, Y. Lu and X. L. Wang, Phys. Chem. Chem. Phys., 2014, 16, 17523.
7. G. Zhang, X. Xiang, F. Yang, X. Peng, T. Tang, Y. Shi and X. Wang, Phys. Chem. Chem. Phys., 2016, 18, 1649.
8. R. Ramírez, R. González, I. Colera and Y. Chen, Phys. Rev. B: Condens. Matter Mater. Phys., 1997, 55, 237.
9. R. Ramírez, I. Colera, R. González and Y. Chen, Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 69, 14302.
10. K. Rehberg and H. S. Leipner, Positron annihilation in semiconductors: defect studies, Springer, Berlin, 1999.
11. J. Kansy, Nucl. Instrum. Methods Phys. Res., Sect. A, 1996, 374, 235.
12. D. Sanyal and D. Banerjee, Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 58, 15226.
13. E. Lavrov, F. Herklotz and J. Weber, Phys. Rev. Lett., 2009, 102, 185502.
14. G. Moya, J. Kansy, A. Si Ahmed, J. Liebault, F. Moya and D. Goeuriot, Phys. Status Solidi A, 2003, 198, 215.
15. M. Katsuyuki, T. Tomohito, Y. Takahisa and I. Yuichi, Phys. Rev. B: Condens. Matter Mater. Phys., 2003, 68, 085110.
16. J. Čížek, I. Procházka, J. Kuriplach, W. Anwand, G. Brauer, T. E. Cowan, D. Grambole, H. Schmidt and W. Skorupa, Defect Diffus. Forum, 2012, 331, 113.
17. A. M. Holder, K. D. Osborn, C. J. Lobb and B. Charles, Phys. Rev. Lett., 2013, 111, 065901.
18. G. Molnár, J. Borossay, M. Benabdesselam, P. Icconi, D. Lappaz, K. Süvegh and A. Vértes, Phys. Status Solidi A, 2000, 179, 245.
19. K. P. Muthe, K. Sudarshan, P. K. Pujari, M. S. Kulkarni, N. S. Rawat, B. C. Bhatt and S. K. Gupta, J. Phys. D: Appl. Phys., 2009, 42, 105405.
20. Y. Nagashima, K. Kawashima, T. Hyodo, M. Hasegawa, B. T. Lee, K. Hiraga, S. Yamaguchi, M. Forster and H. E. Schaefer, Mater. Sci. Forum, 1995, 461, 175.
21. M. L. Chithambo, E. J. Sendezera and A. T. Davidson, Radiat. Prot. Dosim., 2002, 100, 269.
22. X. Xiang, G. Zhang, X. Wang, T. Tang and Y. Shi, Phys. Chem. Chem. Phys., 2015, 17, 29134.

---