Structural transformation and multiferroic properties in Gd-doped Bi₇Fe₃Ti₃O₂₁ ceramics

Abstract

Bismuth layer-structured Bi_7−xGd_xFe₃Ti₃O₂₁ (0.00 ≤ x ≤ 1.50) ceramics were synthesized by Pechini's method, in which gadolinium doping was used with a goal to enhance the magnetic response of the material. With increase in the Gd content of x, an obvious structural transformation, changing gradually from the originally designed six-layer structure to five-layers, was initially shown by X-ray diffraction patterns, and then by Raman scattering spectra and high-resolution transmission electron microscopy images. Substituting Bi sites with Gd³⁺ ions was found to be able to effectively suppress the leakage current, and its resistivity was found to be about two orders of magnitude higher than that of the un-doped sample. Improved magnetic properties and a clear magnetic anomaly were observed in the sample with a composition of x = 1.00, indicating the behaviour of transforming from anti-ferromagnetism with weak ferromagnetism at the room temperature into a complex magnetism at low temperature.

---

Introduction

Single-phase multiferroic materials, simultaneously possessing electric dipole and magnetic ordering in a uniform single-phase, have recently attracted remarkable interest. Due to the coupling between electric order and magnetic order in such materials they contain rich fundamental physics and have great potential applications such as multi-state storage, sensing, and quantum controlling devices.^1–3 Major efforts undertaken in the field of multiferroics have been focused on a typical perovskite ABO₃ structure such as bismuth ferrite⁴ and bismuth manganite.⁵ However, realizing a strong ferromagnetism or a strong magneto-electric effect at or above room temperature (RT) is still very challenging.⁶ Recently, layered bismuth-based Aurivillius phase compounds⁷ have been gaining attention as one of the potential candidates able to become single-phase multiferroic materials. The use of their layered structures can avoid difficulties arising from the fact that in a single-phase material normally ferroelectric (FE, usually from unoccupied d orbitals) and ferromagnetic (FM, from partially occupied d orbitals) are mutually exclusive. Especially, the compounds from the Bi₄Ti₃O₁₂–BiFeO₃ system with a general formula of Bi_n+1Fe_n−3Ti₃O_3n+3 (here n denotes an integer corresponding to the number of perovskite layers)⁸ are of special interest. These compounds, which are described by the regular stacking of fluorite-like (Bi₂O₂)²⁺ layers and perovskite-like (Bi_n−1Fe_n−3Ti₃O_3n+1)²⁻ slabs, possess unusual dielectric properties, elevated Curie temperatures and enhanced magnetism (even though it is normally at low temperatures).^9,10

Recently, we have initiated a new method via magnetic ion B-site substitution in a four-layer Bi₅Fe_0.5Co_0.5Ti₃O₁₅ that can achieve an improvement in both the magnetic properties and FE response at RT.¹¹ Similar behaviour was found in higher layer number Aurivillius oxides, Bi₇Fe_3−xNi_xTi₃O₂₁.¹² However, A-site modification has not been reported in the six-layer Bi₇Fe₃Ti₃O₂₁ according to the past literature. Interestingly, questions, such as could the A-site modification be useful in further enhancing the RT magnetic response and in which manner, should be answered, and this requires additional focus on the synthesis of new multiferroic materials. Past research on Bi₇Fe₃Ti₃O₂₁ is rare, despite the fact that the Bi_7−xSr_xFe_3−xTi_3+xO₂₁ ceramics have been investigated, in which magnetic study has not been carried out.¹³ Very encouragingly, previously studies have shown that substitution of Bi³⁺ with rare-earth elements in BiFeO₃ resulted in a remarkable improvement in both FE and FM.¹⁴ Unfortunately, MFeO₃ (M = Sm-Dy) that crystallizes in the perovskite structure possess dominant antiferromagnetic behaviour.¹⁵ Therefore, it will be reasonable to assume that using the magnetically active Gd ion as the A-site substituting element could give rise to an even larger magnetization, because the ionic radius of Gd³⁺ is smaller than that of Bi³⁺, and the coupling between magnetic Gd³⁺ and Fe³⁺ ions might enhance the magnetization.

In light of above discussions, we have synthesized the Bi_7−xGd_xFe₃Ti₃O₂₁ (0.00 ≤ x ≤ 1.50) ceramics by Pechini's method, systematically analyzed their structural changes and investigated their basic multiferroic properties. Interestingly, an obvious structural transformation, changing gradually from the originally designed six-layer perovskite structure to five-layers, was found. Improved magnetic properties were observed at x = 1.00, possibly arising from the structural distortion as a result of the substitution, the coupling between Gd³⁺ and Fe³⁺ ions, or the coupling contribution from the co-existed phases around the boundary of the structural transformation. A magnetic anomaly was also observed at x = 1.00. The compound with x = 1.00 shows antiferromagnetism behaviour with a weak ferromagnetism at RT and a complex magnetism at the low temperatures.

Experimental

Polycrystalline Bi_7−xGd_xFe₃Ti₃O₂₁ (BGFT-x, 0.00 ≤ x ≤ 1.50) powders were prepared as follows: firstly, appropriate amounts of Ti(OC₄H₉)₄, Bi(NO₃)₃·5H₂O, Fe(NO₃)₃·9H₂O and Gd(NO₃)₃·6H₂O were dissolved in HNO₃ as precursors. C₆H₈O₇·H₂O and C₁₀H₁₆N₂O₈ were then added to the solution as the reducing agent and complexing agent, respectively. Secondly, NH₄OH was dropped into the above solution to reach a pH = 6–7, and then the precursor was heated until combusted inside the crucible. Thirdly, the formed gels were pre-heated at 1023 K for 2 h in order to remove the organic residues. Finally, the obtained powders were cold-pressed into pellets, and then hot-pressed at 1173 K for 3 h in oxygen and argon mixture atmosphere (O₂/Ar = 1/4) under a pressure of 12.56 MPa.

Structural characterization was performed by powder X-Ray diffraction (XRD) (Cu-K_α radiation, D/Max-gA, Japan), scanning electron microscopy (SEM) images (JEOL JSM-6400), high-resolution transmission electron microscopy (HRTEM) images (JEOL JEM-2010 field emission electron microscope) and Raman spectroscopy (SPEX-1403, Ar⁺ laser, 514.5 nm). Ferroelectric measurements were conducted using a Precision LC ferroelectric analyzer (Radiant Technology product, USA). Magnetic properties were characterized by vibrating sample magnetometer (VSM) option of the Quantum Design physical property measurement system (PPMS) (Quantum Design, USA).

Results and discussion

Fig. 1a shows the RT XRD patterns of the BGFT-x sintered at 1173 K. All the detectable diffraction peaks for the samples match well to a typical Aurivillius phase compared with the standard powder diffraction data from Bi₇Fe₃Ti₃O₂₁ (F2mm no. 42). However, detailed analyses were performed in a small section of the XRD patterns in the 2θ range of 29.3°–34.2°, 45°–55° and 10°–30.5° (Fig. 1c–e, respectively), and we observed that the (200) (or (020)) and (220) diffraction peaks shift toward a larger angle with increasing Gd content (compared with x = 0.00, those peaks of x = 1.50 shift about 0.5° higher), which is beyond the scope arising only from the substitution of Bi ions. The (0012) peak shifts toward the lower angle and the (0014) peak remains unchanged (Fig. 1e). Compared with the XRD patterns for the five-layer Bi₆Fe₂Ti₃O₁₈ (Fig. 1b), it is more reasonable to suggest that a structural transformation occurs with an increase in Gd substitution. Additional evidence is shown in Fig. 1f, the samples with x = 0.50 and x = 0.75 show a broad peak composed of the reflections from (119) and (0016) for Bi₆Fe₂Ti₃O₁₈ (n = 5), and (1111) for Bi₇Fe₃Ti₃O₂₁ (n = 6). However, the samples with x = 1.00 and x = 1.50 show a broad peak composed of only (119) and (0016) for n = 5. The above results indicate that the samples between x = 0.50 and 1.00 appear to have a coexistence of six- and five-layer perovskite structures and that the sample with x = 1.50 has mainly a five-layer perovskite structure, indicating that the structural transformation changes gradually from a six-layer perovskite structure to five-layers. According to Fig. 1b, it may change to a four-layer if x > 1.50.

Fig. 1 (a) Room temperature XRD patterns of the sintered BGFT-x (0.00 ≤ x ≤ 1.50) ceramics. (b) Room temperature XRD patterns of the sintered Bi₅FeTi₃O₁₅ (n = 4), Bi₆Fe₂Ti₃O₁₈ (n = 5) and Bi₇Fe₃Ti₃O₂₁ (n = 6) ceramics. Selected sections of the XRD patterns of the sintered BGFT-x (0.00 ≤ x ≤ 1.50) ceramics in the 2θ range of (c) 29.3°–34.2°, (d) 45°–55° and (e) 10°–30.5°. (f) Selected sections of the XRD patterns of the sintered BGFT-x (0.00 ≤ x ≤ 1.50) ceramics in the 2θ range of 27.6°–29.6° (black lines are the measured patterns and the red lines are fitted patterns) and the decomposed peaks (the reflections (119) and (0016) for Bi₆Fe₂Ti₃O₁₈ (n = 5), and (1111) for Bi₇Fe₃Ti₃O₂₁ (n = 6)).

Since the Raman scattering spectrum is sensitive to the atomic displacements, evolvement of Raman normal modes with increasing Gd content may provide valuable information about lattice properties and structural transitions. Fig. 2a shows the measured Raman spectra of the BGFT-x (x = 0.00–1.50) ceramics at RT. For the Raman spectrum of bismuth layer structured crystals, low frequency modes below 200 cm⁻¹ are ascribed to the vibration of heavy mass Bi³⁺ ions, and the high frequency modes, which are above 200 cm⁻¹, are known to result from the torsion bending and stretching modes of octahedral BO₆.^16,17 The modes in 78 cm⁻¹ and 117 cm⁻¹ originate from the vibrations of the Bi³⁺ at the A-site of the perovskite slabs. The continuous decrease in the intensity of the 117 cm⁻¹ mode and an opposite shifting of the 78 cm⁻¹ mode with increasing Gd³⁺ doping content are shown in Fig. 2c, indicating that the Gd entered the Bi sites of the perovskite slabs. This phenomenon is similar to that of doping Nd³⁺ into a Bi₇Ti₄NbO₂₁ ceramic,¹⁸ in which a continuous decrease in the intensity of the 115 cm⁻¹ mode was identified when the Nd³⁺ doping content increased. This demonstrates a stronger vibration along the z-axis and intensification along the z-axis, accompanying by a weakened effect along the x- and y-axis reflected from the 82 cm⁻¹ mode with an opposite shift. In addition, compared with the obvious difference in the Raman spectra of Aurivillius Bi₅FeTi₃O₁₅ (n = 4), Bi₆Fe₂Ti₃O₁₈ (n = 5) and Bi₇Fe₃Ti₃O₂₁ (n = 6) compounds (the mode in 206 cm⁻¹ for n = 6, 220 cm⁻¹ for n = 5 and 233 cm⁻¹ for n = 4, Fig. 2b), the 206 cm⁻¹ mode shifts gradually to a higher frequency with increase in the Gd concentration (Fig. 2c) may indicate the discussed perovskite-layer structural transformation.

Fig. 2 (a) Raman spectra of the BGFT-x (x = 0.00–1.50) ceramics at the RT. (b) Raman spectra of the Bi₅FeTi₃O₁₅ (n = 4), Bi₆Fe₂Ti₃O₁₈ (n = 5) and Bi₇Fe₃Ti₃O₂₁ (n = 6) compounds. (c) A selected section of the Raman spectra of BGFT-x (x = 0.00–1.50) ceramics in the frequency range of 50–400 cm⁻¹.

Fig. 3a shows HRTEM images of the BGFT-x (x = 0.00, 0.50, 0.75, 1.00) lattice structures, in which the bright spots stand for the location of the bismuth atoms. Fig. 3b and c are the 2D atomic structures of Bi₇Fe₃Ti₃O₂₁ and Bi₆Fe₂Ti₃O₁₈ compounds, respectively. Intriguingly, the coexistence of the five-layer and six-layer Aurivillius structure (x = 0.50 and 0.75), single well-periodic six-layer Aurivillius structure (x = 0.00) and the five-layer Aurivillius structure (x = 1.00) were further confirmed by the well-defined electron diffraction patterns. This further confirms that the Gd-doping in Bi₇Fe₃Ti₃O₂₁ leads to a structural transformation changing gradually from the six-layer perovskite structure to five-layers.

Fig. 3 (a) HRTEM images of the BGFT-x (x = 0.00, 0.50, 0.75, 1.00) samples. (b) The 2D atomic structure of the Bi₇Fe₃Ti₃O₂₁ (n = 6) compound. (c) The 2D atomic structure of the Bi₆Fe₂Ti₃O₁₈ (n = 5) compound.

P–E hysteresis loops at RT of the BGFT-x under an electric field of 110 kV cm⁻¹ and a measuring frequency of 100 Hz are presented in Fig. 4a. With increasing the amount of Gd (x ≤ 1.00), the remnant polarization value of the samples, especially for x = 0.75, are actually enhanced slightly when compared to that in the un-doped sample. The reason for this small enhancement is due to the fact that Gd-doping causes a small BO₆ octahedral structural distortion, consistent with the XRD result in which the (1113) main peak position almost remains unchanged with an increasing Gd content (Fig. 2c). However, Gd-doping in the samples can effectively suppress the leakage problem, according to P–E loop of the sample with x = 1.00, ascribed to the high relative density of the doped sample. This can be verified by the current–density versus electric-field curves at RT (Fig. 4b) and the SEM micrographs of the samples (Fig. 4c). The resistivity estimated from the linear characteristics is about 7 × 10⁷ and 2 × 10⁹ Ω cm for the samples with x = 0.00 and 1.00, respectively. The SEM images show that a dense microstructure could be realized in the Gd-doped sample. The relative densities of x = 0.00 and x = 1.00 are about 89% and 97%, respectively, with the help of the hot-press sintering process.

Fig. 4 (a) P–E hysteresis loops at RT of the BGFT-x under an electric field of 110 kV cm⁻¹ and a measuring frequency of 100 Hz. (b) The current density of the BGFT-x (x = 0.00 and 1.00) vs. electric-field curves at RT. (c) The SEM micrographs of the BGFT-x (x = 0.00 and 1.00).

The magnetic hysteresis loops at RT of the BGFT-x ceramics are shown in Fig. 5a. The nearly linear field dependence of magnetization (M) for the BGFT-0.00 indicates its antiferromagnetic (AFM) nature. A small gap exists with the remnant magnetization (2M_r) ∼2.7 memu g⁻¹, as shown in the inset of Fig. 5a, similar to the previous result.¹⁰ The magnetic moment versus H plot for the BGFT (x > 0) samples clearly indicate the enhanced magnetism at higher magnetic fields, demonstrating that Gd-doping in the samples can enhance the magnetic properties. Especially, the sample with x = 1.00 shows weak ferromagnetism with no saturation magnetization at RT, and 2M_r is about 6.1 memu g⁻¹. This value is close to that of Bi₅Fe_0.5Co_0.5Ti₃O₁₅ (2M_r ∼ 7.8 memu g⁻¹)¹¹ and SmBi₅Fe₂Ti₃O₁₈ (2M_r ∼ 7.7 memu g⁻¹), and about one order of magnitude higher than that of LaBi₅Fe₂Ti₃O₁₈ (2M_r ∼ 0.4 memu g⁻¹).¹⁹ The reason for the enhanced magnetic properties may include: (1) structural distortion by the substitution of Bi³⁺ by Gd³⁺ ions will change the bond angle of Fe³⁺–O²⁻–Fe³⁺, which leads to the suppression of the spatially modulated spiral spin structure and hence the appearance of weak ferromagnetism, similar to that in Bi_1−xGd_xFeO₃;²⁰ (2) the contribution was attributed to the anti-parallel spin clusters or canted nature of spin in the AFM material, which are gradually turned towards the field direction giving rise to the FM nature by Gd substitution;¹⁹ and (3) the coupling contribution from the co-existed phases around the boundary of the structural transformation, which is similar to morphotropic phase boundary (MPB) effect in piezoelectric solid solution ceramics.^21,22

Fig. 5 (a) M–H loops at the RT of the BGFT-x ceramics. Inset: enlarged curves showing the changes. (b) Temperature dependences of magnetization for the BGFT-x (x = 0.00, 0.50, 0.75, 1.50) in zero field cooling (ZFC) and field cooling (FC) modes under a magnetic field of 100 Oe. Inset: enlarged curve of the BGFT-0.00. (c) ZFC–FC curves of the BGFT-1.00 under a magnetic field of 100 Oe. Inset: enlarged curve of blue box.

Fig. 5b and c show the temperature dependence of magnetization for the BGFT-x in zero field cooling (ZFC) and field cooling (FC) modes under a magnetic field of 100 Oe. The ZFC–FC curve of BGFT-0.00 (inset of Fig. 5b) is consistent with that of previous research.^23,24 The divergence between ZFC and FC curves starts at about 200 K, and then becomes noticeable at 25–100 K. Because such materials exhibit AFM ordering of the magnetic ions in the lattice, the reason for this divergence may result from the onset of some magnetic ordering or domain orientation.²³ Compared with the un-doped sample, the Gd-doped samples have a similar trend except for the BGFT-1.00 sample, which has one anomaly in the ZFC–FC curve (Fig. 5c). A sharp increase in the magnetization was observed with a decrease in the temperature is in accordance with that in Bi_1−xGd_xFeO₃, which was explained as the f–d exchange interaction.²⁵

The anomaly observed for x = 1.00 is a broad valley around 172 K, as shown in the inset of Fig. 5c. In order to study the further details about the magnetic anomaly, the FC curve from 5 K to 300 K and the reversing curve (from 300 K to 5 K) under a magnetic field of 100 Oe are shown in Fig. 6a. The broad valley, around 172 K upon increasing the temperature and around 115 K when decreasing the temperature, does not trace back on reversing the temperature. We also observed a shift in the broad valley with different magnetic fields. The temperature dependence of χ⁻¹ for BGFT-1.00 from 300 K to 5 K and from 5 K to 300 K with different magnetic fields is shown in Fig. 6b and c, respectively. With increase in the magnetic field, the abnormal peak shifts to a high temperature if the measuring temperature is from 300 K to 5 K, and shifts to a low temperature if the measuring temperature is from 5 K to 300 K. Moreover, the shape of the peak will gradually disappear, if we continue increasing the magnetic field. Previous studies have found that those materials have complicated magnetic behaviour such as AFM, weak FM, paramagnetic and AFM spin glass-like.²⁴ Suryanarayana et al. have reported that there appears a magnetic anomaly for Bi₇Fe₃Ti₃O₂₁ (n = 6) at 190 K and for Bi₆Fe₂Ti₃O₁₈ (n = 5) at 160 K, possibly arising from the magnetic fluctuations in the sub-lattice, and the temperature at which the slope changes in the χ⁻¹ versus T varies with the number of perovskite layers.²⁶ It may be natural to infer that the origin of the magnetic anomaly in the BGFT-1.00 sample should be similar to that in those materials previously reported;²⁶ however, this anomaly was not found in our others samples. The reason for this is still not clear, but it might be related to the structural transformation; thus, more effort is needed to understand the mechanism. In short, the occurrence of the weak hysteresis loops at RT and the significant curvature of the χ⁻¹ line indicate the coexistence of AFM and weak FM behaviours in the sample with x = 1.00, which is in accordance with previous results.^23,24,27

Fig. 6 (a) The FC curves of the BGFT-1.00 from 5 K to 300 K and from 300 K to 5 K under a magnetic field of 100 Oe. The temperature dependence of reciprocal susceptibility (χ⁻¹) for the BGFT-1.00, (b) from 300 K to 5 K and (c) from 5 K to 300 K under a magnetic field of 100 Oe, 250 Oe, 500 Oe, 750 Oe and 1000 Oe.

Conclusions

In summary, we have studied with details in the synthetic process using Pechini's method, the structural transformation and multiferroic properties of Aurivillius phase BGFT-x. The obviously structural transformation, changing gradually from the originally designed six-layer perovskite structure to five-layers, can be successfully exhibited by XRD, Raman and HRTEM studies. By Gd-substitution, the ferroelectric property was slightly enhanced and the leakage current can be effectively suppressed. Magnetic measurements indicate that substituting partial Bi sites with Gd ions is effective in enhancing the magnetic properties. Especially, the sample with x = 1.00 having a clear magnetic anomaly, shows the coexistence of AFM and weak FM behaviours.

Acknowledgements

The authors are grateful for financial support from the National Basic Research Program of China (2012CB922001), the Natural Science Foundation of China (51072193), and the Fundamental Research Funds for the Central Universities.

Notes and references

1. N. A. Spaldin and M. Fiebig, Science, 2005, 309, 391–392.
2. N. A. Spaldin, S. W. Cheong and R. Ramesh, Phys. Today, 2010, 63, 38–43.
3. W. Eerenstein, N. D. Mathur and J. F. Scott, Nature, 2006, 442, 759–765.
4. J. Wang, J. B. Neaton, H. Zheng, V. Nagarajan, S. B. Ogale, B. Liu, D. Viehland, V. Vaithyanathan, D. G. Schlom, U. V. Waghmare, N. A. Spaldin, K. M. Rabe, M. Wuttig and R. Ramesh, Science, 2003, 299, 1719–1722.
5. T. Kimura, S. Kawamoto, I. Yamada, M. Azuma, M. Takano and Y. Tokura, Phys. Rev. B: Condens. Matter Mater. Phys., 2003, 67, 180401.
6. M. Fiebig, J. Phys. D: Appl. Phys., 2005, 38, R123–R152.
7. B. Aurivillius, Arki. Kemi, 1948, 26, 1–6.
8. M. Krzhizhanovskaya, S. Filatov, V. Gusarov, P. Paufler, R. Bubnova, M. Morozov and D. C. Meyer, Z. Anorg. Allg. Chem., 2005, 631, 1603–1608.
9. R. S. Singh, T. Bhimasankaram, G. S. Kumar and S. V. Suryanarayana, Solid State Commun., 1994, 91, 567–569.
10. A. Srinivas, M. M. Kumar, S. V. Suryanarayana and T. Bhimasankaram, Mater. Res. Bull., 1999, 34, 989–996.
11. X. Mao, W. Wang, X. Chen and Y. Lu, Appl. Phys. Lett., 2009, 95, 082901.
12. S. Sun, Y. Ling, R. Peng, M. Liu, X. Mao, X. Chen, R. J. Knize and Y. Lu, RSC Adv., 2013, 3, 18567–18572.
13. A. Kan, H. Ogawa, Y. Inami and T. Moriyama, Phys. B: Condens. Matter, 2011, 406, 3170–3174.
14. A. Lahmar, S. Habouti, M. Dietze, C. H. Solterbeck and M. Es-Souni, Appl. Phys. Lett., 2009, 94, 012903.
15. Z. V. Gabbasova, M. D. Kuzmin, A. K. Zvezdin, I. S. Dubenko, V. A. Murashov, D. N. Rakov and I. B. Krynetsky, Phys. Lett. A, 1991, 158, 491–498.
16. S. Kojima, R. Imaizumi, S. Hamazaki and M. Takashige, Jpn. J. Appl. Phys., 1994, 33, 5559–5564.
17. P. R. Graves, G. Hua, S. Myhra and J. G. Thompson, J. Solid State Chem., 1995, 114, 112–122.
18. C. Shao, Y. Lu, D. Wang and Y. Li, J. Eur. Ceram. Soc., 2012, 32, 3781–3789.
19. N. V. Prasad and G. S. Kumar, J. Magn. Magn. Mater., 2000, 213, 349–356.
20. V. A. Khomchenko, D. A. Kiselev, I. K. Bdikin, V. V. Shvartsman, P. Borisov, W. Kleemann, J. M. Vieira and A. L. Kholkin, Appl. Phys. Lett., 2008, 93, 262905.
21. Y. L. Lu, J. J. Zheng, M. C. Golomb, F. L. Wang, H. A. Jiang and J. Zhao, Appl. Phys. Lett., 1994, 74, 3764–3766.
22. O. Dieguez and J. Iniguez, Phys. Rev. Lett., 2011, 107, 057601.
23. S. J. Patwe, S. N. Achary, J. Manjanna, A. K. Tyagi, S. K. Deshpande, S. K. Mishra, P. S. R. Krishna and A. B. Shinde, Appl. Phys. Lett., 2013, 103, 122901.
24. E. Jartych, T. Pikula, M. Mazurek, A. Lisinska-Czekaj, D. Czekaj, K. Gaska, J. Przewoznik, C. Kapusta and Z. Surowiec, J. Magn. Magn. Mater., 2013, 342, 27–34.
25. V. A. Khomchenko, V. V. Shvartsman, P. Borisov, W. Kleemann, D. A. Kiselev, I. K. Bdikin, J. M. Vieira and A. L. Kholkin, Acta Mater., 2009, 57, 5137–5145.
26. S. V. Suryanarayana, A. Srinivas and R. S. Singh, Proc. SPIE, 1999, 3903, 232–265.
27. A. R. James and T. Bhimasankaram, Mod. Phys. Lett. B, 1998, 12, 785–795.

---