Structure and electrical properties of lead-free Sr_1−x(K,Ce)_x/2(Na_0.5Bi_0.5)Bi₄Ti₅O₁₈ piezoelectric ceramics

Abstract

Bismuth layer-structured ferroelectrics with a formula of Sr_1−x(K,Ce)_x/2(Na_0.5Bi_0.5)Bi₄Ti₅O₁₈ (KCSNBT-x, x = 0.0, 0.1, 0.2 and 0.3), were prepared by a conventional solid-state reaction method. The X-ray diffraction analysis suggested that the substitution led to the formation of a layered perovskite structure. Plate-like morphologies of the grains were clearly observed in all samples. The activation energy value of the ionic conductivity suggested that defects were related to oxygen vacancies. Excellent electrical properties (e.g., d₃₃ ∼ 21 pC N⁻¹, 2P_r ∼ 16.4 μC cm⁻² and T_c ∼ 567 °C) were simultaneously obtained in the ceramic where x = 0.3. Additionally, thermal annealing studies indicated that the piezoelectric constant (d₃₃) of the KCSNBT-0.1 ceramic remains almost unchanged (22 pC N⁻¹, only decreased by 4%) at temperatures below 400 °C, demonstrating that this ceramic is a promising candidate for high-temperature applications.

---

1. Introduction

Bismuth layer-structured ferroelectrics (BLSFs), also known as the Aurivillius family of oxides, have been considered as promising lead-free piezoelectric materials in high-temperature applications.^1,2 Compared with the traditional perovskite structure, the structure of these ferroelectrics, represented as (Bi₂O₂)²⁺(A_m−1B_mO_3m+1)²⁻, consists of [Bi₂O₂]²⁺ layers interleaved with perovskite blocks of [A_m−1B_mO_3m+1]²⁻ units along the c-axes. The A-site in the perovskite blocks can be occupied by 12-coordinated cations such as Bi³⁺, La³⁺, Ba²⁺, Sr²⁺, Ca²⁺, etc., while the B-site can be occupied by 6-coordinated cations such as Ti⁴⁺, Nb⁵⁺, Ta⁵⁺, W⁶⁺, Mo⁶⁺, Co³⁺, etc. In addition, m, in the range of 1–5, represents the number of perovskite blocks, and the ferroelectric and dielectric properties of BLSFs are strongly dependent on its value.^3–7 BLSF ceramics possess some outstanding electrical properties, such as a high Curie temperature, low dielectric loss, excellent fatigue endurance and low aging rate, making BLSFs a promising candidate for ferroelectric non-volatile random access memory (FRAM) storage devices.⁸

SrNa_0.5Bi_4.5Ti₅O₁₈ (abbreviated as SBT), is the m = 5 member of the Aurivillius family, where Bi₂O₂ layers alternate with (SrNa_0.5Bi_2.5Ti₅O₁₆) perovskite blocks built by five TiO₆ octahedral layers and hosting Sr and Bi at the A site. This material has drawn special attention because of its high Curie temperature.⁹ However, similar to other BLSF compounds, this material exhibits relatively low piezoelectric coefficients (d₃₃) and low remnant polarization (2P_r) because spontaneous polarization is restricted in the a–b plane.^9,10 To overcome this shortcoming, many efforts have been made to improve its dielectric and piezoelectric properties by adding some additives in the A-site and/or B-site.^11,12 In these studies, A-site substitution is more effective than that of B-site substitution because the cations in the B-site have a similar size and do not make a major contribution to the polarization process in BLSFs. For instance, Ji et al.¹³ reported that the SrNa_0.5Bi_4.5Ti₅O₁₈ + 1 wt% CeO₂ sample shows an excellent Curie temperature (586 °C) and very large piezoelectric coefficient (24 pC N⁻¹). Chen et al.¹⁰ investigated the properties of Sr_2−x(Na, K)_xBi₄Ti₅O₁₈ composite ceramics and found that the piezoelectric constant d₃₃ and remnant polarization 2P_r were improved, with values of 20 pC N⁻¹ and 20 μC cm⁻² respectively, whereas the Curie temperature was only 324 °C. Additionally, it was found that (MCe) (M = Li, Na, K) dopants can efficiently enhance the piezoelectric coefficients for even-layer structured Aurivillius type compounds compared to the pure compositions.^14,15

Nevertheless, very little is known about the detailed electrical properties of (KCe) modified Sr(Na_0.5Bi_0.5)Bi₄Ti₅O₁₈ (SNBT) ferroelectric ceramics. Therefore, in the present work, A-site (KCe)-doped Sr(Na_0.5Bi_0.5)Bi₄Ti₅O₁₈ ceramics were prepared using a conventional solid-state sintering method, and the effect of (KCe) modification in the A-site on their electrical properties was investigated. Additionally, the underlying physical mechanisms for enhanced piezoelectricity and electrical conductivity have been addressed.

2. Experimental details

Sr_1−x(K,Ce)_x/2(Na_0.5Bi_0.5)Bi₄Ti₅O₁₈ (abbreviated as KCSNBT-x, x = 0, 0.1, 0.2 and 0.3) ceramics were prepared by a conventional solid-state reaction method using reagent-grade metal oxides or carbonate powders of K₂CO₃ (99%, Sinopharm Chemical Reagent Co., Ltd., China), CeO₂ (99.95%, Sinopharm Chemical Reagent Co., Ltd., China), Bi₂O₃ (99.99%, Sinopharm Chemical Reagent Co., Ltd., China), SrCO₃ (99%, Sinopharm Chemical Reagent Co., Ltd., China), TiO₂ (99.5%, Sinopharm Chemical Reagent Co., Ltd., China) and Na₂CO₃ (99.49%, Sinopharm Chemical Reagent Co., Ltd., China) as the starting materials. All raw materials were weighed in stoichiometric proportions and then mixed using planetary ball milling in polyethylene with stabilized zirconia balls for 15 h, using ethanol as the solvent. After drying, the mixed powders were calcined at 800 °C for 2 h. After calcination, the mixture was milled again for 12 h, and then dried. The powders were mixed with an appropriate amount of polyvinyl butyral (PVB) binder, and pressed into pellets with a diameter of 12 mm and a thickness of 0.7 mm under a pressure of about 200 MPa. After burning off the PVB at 850 °C, the ceramics were sintered in an alumina crucible at 1180 °C for 3 h. For the electric measurements, disk samples with about 0.3 mm thickness were used.

The density of the sintered ceramics was measured by means of the Archimedes method. The crystal structure of the ceramics was determined by X-ray diffraction (XRD) using Cu Kα radiation (λ = 1.54178 Å) (D8 Advance, Bruker Inc., Germany). The surface morphology of the ceramics was observed using a scanning electron microscope (SEM) (JSM-6380, Japan). The ferroelectric hysteresis loops were measured through a standardized ferroelectric test system (TF2000, Germany). The temperature dependence of dielectric properties and impedance spectroscopy for the samples was performed using a Broadband Dielectric Spectrometer (Novocontrol Germany). The samples were polarized in silicon oil in the range of 150–180 °C for 20 min, and piezoelectric measurements were carried out with a quasi-static d₃₃-meterYE2730 (SINOCERA, China).

3. Results and discussion

The X-ray diffraction spectra of the KCSNBT-x ceramics (x = 0.0, 0.1, 0.2, 0.3) in the 2θ range of (a) 20–50° and (b) 28–32° are plotted in Fig. 1. As shown in Fig. 1(a), the Aurivillius structure can be identified by indexing all the diffraction peaks on the basis of an orthorhombic cell (PDF#no. 05-0626),¹³ indicating that the K⁺ and Ce³⁺ ions diffused into the A-site lattice and formed solid solutions as expected. Meanwhile, it is obvious that the highest diffraction peak in KCSNBT-x ceramics corresponds to the (101̲1̲) orientation, which is consistent with the fact that it is the strongest diffraction peak in BLSFs.¹⁶ In addition to the main Aurivillius phase, secondary phase Bi₂Ti₂O₇ (PDF#no. 32-0118, labeled by *) in the KCSNBT-x samples can also be detected regardless of the existence of (KCe), as shown in Fig. 1(b). It seems that this peak could be related to the volatilization of bismuth species at elevated temperatures.¹⁷ Similar deficiency behavior of bismuth species was also found during the sintering of another bismuth-layered compound, SrBi₄Ti₄O₁₅.^18,19 In order to further confirm the crystallographic evolution of KCSNBT-x ceramics, the lattice parameters of the samples were calculated and are shown in the inset part of Fig. 1(a). It was found that the lattice parameters a, b, and c varied with increasing x values, demonstrating that the (KCe) substitution reduced the lattice distortion of SNBT-based ceramics. Such a lattice distortion can be attributed to the replacement of the A-site Sr²⁺ by K⁺ and Ce³⁺, and it is supposed to lead to enhanced electrical properties.

Fig. 1 XRD patterns of the KCSNBT-x ceramics sintered at 1180 °C; inset: detailed information on the response of the variation of lattice parameters as a function of x.

The SEM images of the surface morphologies of the KCSNBT-x ceramics (x = 0.0, 0.1, 0.2, 0.3) are displayed in Fig. 2. It was found that all grains in all of the samples were well-packed, and pore-free microstructures were observed. In addition, (KCe)-substituted SNBT ceramics have uniform grain sizes, indicating that the increased doping with (KCe) does not change the microstructure dramatically. Due to the highly anisotropic grain growth rate in the direction of the a–b plane and the lower surface energies of the (001) planes induced predominantly during sintering, all the specimens are composed of plate-like grains, which is a typical characteristic of Aurivillius ceramics.¹³ Moreover, all ceramics have a high relative density ρ_rd (>96%), suggesting that all samples have been well sintered.

Fig. 2 SEM micrographs of the KCSNBT-x ceramics: (a) x = 0, (b) x = 0.1, (c) x = 0.2, (d) x = 0.3.

The temperature dependence of the dielectric constant (ε) and loss (tan δ) of the KCSNBT-x ceramics (x = 0.0, 0.1, 0.2, 0.3) measured at 10 kHz is depicted in Fig. 3. Two phase transitions were observed at ∼400 °C and ∼550 °C, corresponding to the ferroelectric–ferroelectric and ferroelectric–paraelectric phase transitions.²⁰ Such a phenomenon was also reported in Ca-doped SrNa_0.5Bi_4.5Ti₅O₁₈ ceramics.¹³ In addition, Fig. 3 clearly shows that the value of the Curie temperature (T_c) corresponding to the ferroelectric–paraelectric phase transition of ferroelectrics was slightly increased from 562 °C to 570 °C, which is higher than the reported results for other Aurivillius material systems (Table 1). Moreover, as shown Fig. 3, the dielectric loss values were found to be very low (less than 7%) and very stable when the measurement temperature is below 400 °C for all compositions. These results show that the KCSNBT-x ceramics possess high stability in their dielectric behavior, which is of great importance for high-temperature device applications. When the temperature was above 500 °C, the tan δ value significantly increased, arising from the space charge carriers induced by the increase of electrical conductivity.²¹

Fig. 3 Temperature dependence of the dielectric constant (ε) and loss (tan δ) of the KCSNBT-x ceramics measured at 10 kHz. Inset: ln(1/ε − 1/ε_m) as a function of ln(T − T_m) at a frequency of 10 kHz for the KCSNBT-x ceramics.

Table 1 Comparison of the polarization (2P_r), Curie temperature (T_c) and piezoelectric constant (d₃₃) of the KCSNBT-x ceramics with other Aurivillius material systems

Materials	2Pr (μC cm^−2)	Tc (°C)	d33 (pC N^−1)	Reference
KCSNBT-x	16.4	567	21	Current work
SBTi-(Na,K)x	20	324	20	10
CxSBN	15	552	18	16
SBBT	1.4	535	12	22
SNBT	—	535	17	23
SBN	—	440	20	24

---

It is well known that the nature of the phase transition is determined by calculating the degree of diffusion (γ) in the measured dielectric constant by modifying Curie–Weiss law:⁵

where, ε is the dielectric constant, ε_m is the dielectric constant maximum, T is the temperature (>T_c), T_m is the temperature at the dielectric peak, γ is the relaxation strength and C is the Curie constant. The value of γ ranges from 1 (normal ferroelectrics) to 2 (complete diffuse phase transition). The inset of Fig. 3 shows the plots of ln(1/ε − 1/ε_m) versus ln(T − T_m) at 10 kHz for SNBT, KCSNBT-0.1, KCSNBT-0.2, and KCSNBT-0.3 ceramics. It is observed that the values of the degree of diffusion (γ) for all ceramics can be described by two different temperature regions with inflexion at ln(T − T_m) ∼ 3.0, under which the degree of diffusion (γ) is obtained γ = 1.82–1.87, while above which the values of the degree of diffusion are much lower (γ = 1.23–1.41). This phenomenon could be related to the structural distortion originating from the increasing oxygen vacancies in the SNBT ceramic at high temperature.²⁵ Additionally, the values of γ for all samples are found to increase with increasing (KCe) concentration, and all these values are close to the degree of diffusion (γ) reported for BLSF ceramics, indicating that the addition of (KCe) in the SNBT ceramic intensifies the diffusive-type phase transition.²² This can be attributed to the grain size dependence of the relaxor behaviour in micro-compositional fluctuations.⁵ Such a transition from relaxor-like behavior to relaxor-ferroelectric behavior has been observed in other BLSFs.²⁶

Fig. 4 shows the DC resistivity of the KCSNBT-x ceramics as a function of reciprocal temperature. The resistivity of all compositions is higher than 10⁵ Ω cm at 525 °C. This is important for high temperature piezoelectric sensor applications. Furthermore, the behavior of the temperature dependent resistivity follows the Arrhenius relationship:

ρ = A exp(−E_a/k_BT) (2)

where A is a pre-exponential factor constant, E_a is the activation energy of the mobile charge carriers, k_B is the Boltzmann constant and T is the absolute temperature.¹⁶ The activation energy is shown in the inset of Fig. 4 and Table 2. It is found that the values of the activation energy E_a for the KCSNBT-x ceramics, calculated by the linear fitting of the data points, were calculated to be 0.8 eV, 0.78 eV, 0.66 eV and 0.54 eV for SNBT, KCSNBT-0.1, KCSNBT-0.2 and KCSNBT-0.3 ceramics, respectively. The results were close to the activation energy values of the ionic conductivity by oxygen vacancies in perovskite type ferroelectric oxides.^27–29

Fig. 4 Temperature dependence of the resistivity of the KCSNBT-x ceramics.

Table 2 Room temperature electrical properties of the KCSNBT-x ceramics

Samples (x)	ε	tan δ (%)	Tc (°C)	Ea (eV)	ρrd (%)	Ec (kV cm^−1)	2Pr (μC cm^−2)	d33 (pC N^−1)
0.0	267	3.4	562	0.80	96.7	34	8	20
0.1	291	1.2	563	0.78	96.9	36	11	23
0.2	293	1.4	564	0.66	97.3	50	12.8	21
0.3	291	3.0	567	0.54	97.4	59	16.4	21

---

Fig. 5(a) shows P–E hysteresis of the KCSNBT-x ceramics measured at a frequency of 10 Hz under a maximum electric field of 140 kV cm⁻¹ and at a temperature of 180 °C. The measured remnant polarization (2P_r) and coercive field (E_c) values with varying (KCe) content (x) in SNBT ceramics are listed in Table 2. It is evident that the remnant polarization (2P_r) and coercive field (E_c) both increase gradually with an increase of the (KCe) doping content. When the doping content is 0.3, 2P_r and E_c for KCSNBT-x simultaneously exhibit maximum values of 16.4 μC cm⁻² and 59 kV cm⁻¹, respectively. 2P_r is much higher than that of the pure SNBT ceramic, indicating that the ferroelectric properties of the SNBT ceramic have been enhanced by (KCe)-doping. The remarkable enhancement in ferroelectric polarization could be mainly attributed to the crystal lattice distortion originating from the replacement of the A-site Sr²⁺ by K⁺ and Ce³⁺, which is also substantiated by the XRD results, in which the lattice parameters a, b and c varied with increasing x values. Moreover, the ferroelectricity in BLSFs is related to the tilting of oxygen in octahedral BO₆ from the c axis and its rotation in the a–b plane.^30,31 In view of the difference of the mean ionic radius caused by the introduction of (KCe) into the SNBT ceramics, the tilting and rotation of oxygen in octahedral TiO₆ have been increased, which should be responsible for the enhanced ferroelectric properties.³² Meanwhile, the increased E_c is consistent with the assumption (Fig. 4) that oxygen vacancies are formed. Under high electric fields, mobile oxygen vacancies can assemble at the low energy domain walls, and thereby hinder domain switching due to domain pinning.³³ In order to further realize the polarization state of the KCSNBT-x ceramics, the polarization current curves of the ceramics were collected, as shown in Fig. 5(b). One sharp polarization current peak could be observed when the applied electric field reached E_c, indicating that the ferroelectric domain could be easier to switch when the driving electric field increases to E_c.³⁴

Fig. 5 (a) P–E hysteresis loops of the KCSNBT-x ceramics with different values of x at 180 °C; (b) I–E loops of the KCSNBT-x ceramics at 180 °C.

Fig. 6 gives the thermal annealing behavior of the piezoelectric constant (d₃₃) of the KCSNBT-x ceramics depolarized at different temperatures held for 10 min. An excellent piezoelectric coefficient of 23 pC N⁻¹ is found when x is 0.1 as shown in Table 2, which is higher than the reported results in other Aurivillius material systems (Table 1).^22–24 In addition, the values of d₃₃ were not zero when the depolarization temperature was higher than the first dielectric anomalies (∼400 °C). However, when the depolarization temperature was higher than the second dielectric anomalies, d₃₃ was zero. Therefore, in the x range of 0–0.3, the KCSNBT-x piezoelectric materials underwent a ferroelectric–ferroelectric transition at the temperature of the first dielectric anomalies and a ferroelectric–paraelectric transition at the temperature of the second dielectric anomalies, as shown in Fig. 3. Moreover, the inset of Fig. 6 clearly shows that the piezoelectric constant (d₃₃) of the KCSNBT-0.1 ceramic remains almost unchanged (22 pC N⁻¹, a decrease of only 4%) at temperatures below 400 °C, indicating that the ceramic has excellent temperature stability, so that it is very tolerant to thermal annealing and might be an appropriate candidate for high temperature applications.

Fig. 6 Annealing temperature dependence of the piezoelectric coefficient (d₃₃) for the KCSNBT-x specimens. The inset shows the relative d₃₃ (%) values of the KCSNBT-0.1 sample.

4. Conclusion

KCSNBT-x ceramics were prepared by a conventional solid-state sintering method, and their structure and electrical properties were studied. The KCSNBT-x ceramics presented a typical layered perovskite structure. The morphologies of these Aurivillius ceramics show the grains in all the samples are all well-defined and the plate-like morphologies of the samples can be observed. The T_c increases slightly from 562 °C to 567 °C. The activation energy (E_a) values of the ionic conductivity suggested that defects were related to oxygen vacancies. The 2P_r increased up to a value of 16.4 μC cm⁻² with (KCe)-modifications. Meanwhile, the ceramics showed excellent thermal stability when the annealing temperature was below 400 °C. As a result, a high T_c and d₃₃, and good thermal stability have been attained in the KCSNBT-x ceramics. Therefore, such a material system is a potential candidate for high-temperature piezoelectric applications.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 51372110, 51402144, 51302124, 51302025), National High Technology Research and Development Program of China (No. 2013AA030801), Science and Technology Planning Project of Guangdong Province, China (No. 2013B091000001), Independent innovation and achievement transformation in Shandong Province special, China (No. 2014CGZH0904), The Project of Shandong Province Higher Educational Science and Technology Program (No. J14LA11, No. J14LA10) and Research Foundation of Liaocheng University (No. 318011301, No.318011306).

References

1. B. H. Park, B. S. Kang, S. D. Bu, T. W. Noh and J. Lee, Nature, 1999, 401, 682–684.
2. P. Xiao, Y. Guo, M. Tian, Q. Zheng, N. Jiang, X. Wu, Z. Xia and D. Lin, Dalton Trans., 2015, 44, 17366–17380.
3. J. S. Kim, M. S. Jang, I. W. Kim and K. S. Lee, J. Electroceram., 2006, 17, 129–133.
4. J. Xiao, H. Zhang, Y. Xue, Z. Lu, X. Chen, P. Su, F. Yang and X. Zeng, Ceram. Int., 2015, 41, 1087–1092.
5. S. Ye, J. Fuh, L. Lu, Y.-l. Chang and J.-R. Yang, RSC Adv., 2013, 3, 20693–20698.
6. J. Wu, Y. Wang and H. Wang, RSC Adv., 2014, 4, 64835–64842.
7. A. Kaushal, S. M. Olhero, B. Singh, R. Zamiri, V. Saravanan and J. M. F. Ferreira, RSC Adv., 2014, 4, 26993–27002.
8. Y. Zhong-Ran, C. Rui-Qing, X. Zhi-Jun, H. Ji-Gong and L. Guo-Rong, J. Inorg. Mater., 2015, 30, 989–994.
9. L. Fei, Z. Zhou, S. Hui and X. Dong, Ceram. Int., 2015, 41, 9729–9733.
10. Q. Chen, Z. Xu, R. Chu, J. Hao, Y. Zhang, G. Li and Q. Yin, Phys. B, 2010, 405, 2781–2784.
11. J. D. Bobić, M. M. Vijatović Petrović, J. Banys and B. D. Stojanović, Ceram. Int., 2013, 39, 8049–8057.
12. D. Peng, H. Zou, C. Xu, X. Wang and X. Yao, J. Alloys Compd., 2013, 552, 463–468.
13. W. Ji, R. Chu, Z. Xu, J. Hao, R. Cheng, X. Chen and Y. Xu, J. Mater. Sci.: Mater. Electron., 2015, 26, 5686–5689.
14. L. Sun, Q. Chen, D. Wu, J. Wu, Z. Tan, D. Xiao and J. Zhu, J. Alloys Compd., 2015, 625, 113–117.
15. L. Sun, Q. Chen, J. Wu, Z. Peng, Z. Tan, D. Xiao and J. Zhu, Ceram. Int., 2014, 40, 14159–14163.
16. Z. Yao, R. Chu, Z. Xu, J. Hao, D. Wei and G. Li, J. Mater. Sci.: Mater. Electron., 2015, 26, 8740–8746.
17. L. Nibou, A. Aftati, M. El Farissi and J.-P. Mercurio, J. Eur. Ceram. Soc., 1999, 19, 1383–1386.
18. C.-H. Lu and C.-H. Wu, J. Eur. Ceram. Soc., 2002, 22, 707–714.
19. E. Khomyakova, J. Pavlic, M. Makarovic, H. Ursic, J. Walker, T. Rojac, B. Malic and A. Bencan, J. Eur. Ceram. Soc., 2015, 35, 4163–4171.
20. Z. Xu, R. Chu, J. Hao, Y. Zhang, Q. Chen, L. Zhao, G. Li and Q. Yin, J. Alloys Compd., 2009, 487, 585–590.
21. Z. Peng, Q. Chen, D. Liu, Y. Wang, D. Xiao and J. Zhu, Curr. Appl. Phys., 2013, 13, 1183–1187.
22. P. Nayak, S. R. Mohapatra, P. Kumar and S. Panigrahi, Ceram. Int., 2015, 41, 9361–9372.
23. X. Meng, W. Ma, T. Chen, M. Wang and Y. Guo, Electron. Mater. Lett., 2015, 11, 902–905.
24. R. Singh, V. Luthra, R. S. Rawat and R. P. Tandon, Ceram. Int., 2015, 41, 4468–4478.
25. C. Diao, H. Li, Z. Chen and H. Zheng, Ceram. Int., 2016, 42, 621–626.
26. V. Senthil, T. Badapanda, A. Chandrabose and S. Panigrahi, Mater. Lett., 2015, 159, 138–141.
27. A. Peláiz-Barranco and Y. González-Abreu, Solid State Commun., 2009, 149, 2082–2084.
28. A. Palanduz and D. Smyth, J. Electroceram., 2003, 11, 191–206.
29. S. Bharadwaja and S. Krupanidhi, J. Appl. Phys., 1999, 86, 5862–5869.
30. J. Rödel, W. Jo, K. T. Seifert, E. M. Anton, T. Granzow and D. Damjanovic, J. Am. Ceram. Soc., 2009, 92, 1153–1177.
31. Z. Yao, H. Liu, Y. Liu, Z. Wu, M. Cao and H. Hao, Appl. Phys. Lett., 2008, 92, 2905.
32. T.-L. Zhao, C.-M. Wang, C.-L. Wang, Y.-M. Wang and S. Dong, Mater. Sci. Eng., B, 2015, 201, 51–56.
33. A. Khokhar, P. K. Goyal, O. P. Thakur, A. K. Shukla and K. Sreenivas, Mater. Chem. Phys., 2015, 152, 13–25.
34. R. Cheng, C. Wang, Z. Xu, R. Chu, J. Hao, H. Li, W. Li, J. Du and G. Li, RSC Adv., 2015, 5, 90508–90514.

---