Hydrothermal synthesis of tin doped ceria-zirconia solid solutions with enhanced thermal stability and oxygen storage capacity

Abstract

A facile hydrothermal synthesis of tin doped ceria-zirconia (Ce_0.5Zr_0.43Sn_0.07O₂) solid solutions was carried out using Ce(NH₄)₂(NO₃)₆, Zr(NO₃)₃·2H₂O and SnCl₄·5H₂O as the starting materials. The synthesized Ce_0.5Zr_0.43Sn_0.07O₂ particles were characterized for their oxygen storage capacity (OSC) for automotive catalysis applications. For the characterization, X-ray diffraction, transmission electron microscopy and the Brunauer-Emmet-Teller (BET) technique were employed. The OSC values of all samples were measured using thermogravimetric-differential thermal analysis. Ce_0.5Zr_0.43Sn_0.07O₂ solid solutions with a BET surface area of 246 m² g⁻¹ exhibited a considerably high OSC of 1425 μmol-O g⁻¹. The incorporation of tin ions in the lattice of the ceria based catalyst greatly enhanced the thermal stability and OSC. The influence of cation radius on the thermal stability and OSC was also discussed.

---

1. Introduction

Ceria (CeO₂)-based materials have been recognized as an important component of three-way catalysts (TWCs) because of their excellent oxygen storage capacity (OSC).^1,2 The major role of ceria-based materials is the OSC and maintaining an air/fuel ratio around 14.7 during engine operation, since TWCs have high pollutant conversion efficiency only under a narrow operation window of the air/fuel ratio.³ Since the 1990s, CeO₂–ZrO₂ solid solutions have gradually replaced pure CeO₂ as OSC materials in TWCs to reduce the emission of toxic pollutants (CO, NO_x, hydrocarbons, etc.) from automobile exhausts, because of their enhanced OSC performance and improved thermal stability at elevated temperatures.^4–8

The redox property of CeO₂ can be greatly enhanced by incorporating zirconium ions (Zr⁴⁺) into the lattice to form a solid solution.^9–11 Nagai et al. have suggested that enhancing the homogeneity of Ce and Zr atoms in the CeO₂–ZrO₂ solid solutions can improve the OSC performance.¹² The detailed structure and property of CeO₂–ZrO₂ solid solutions were reported in review article by Monte and Kašpar.⁷ This review included the results of reducing performance for a series of samples with gradually elevated Ce contents, and a possible mechanism of structural changes in the reducing process was proposed. Fornasiero et al. have reported that an optimum composition, like Ce_0.5Zr_0.5O₂ (molar ratio of Ce : Zr = 1 : 1) can exist as a cubic phase, which can have a considerably high redox property.¹³ Using density functional theory, Wang et al. found that in a series of Ce_1−xZr_xO₂ solutions with a content of 50% ZrO₂ possess the lowest formation energy of the O vacancy, therefore, Ce_0.5Zr_0.5O₂ exhibits the best OSC performance.¹⁴ Recently, many researchers have paid a lot of attention to preparing Ce_0.5Zr_0.5O₂ solutions with a homogeneous composition, good dispersion of particles, narrow particle size distribution, better crystallinity and high surface area in order to improve OSC and redox properties for catalytic applications.^15–20

Although Ce_0.5Zr_0.5O₂ solid solutions have been studied extensively, there are few reports on the preparation of Ce_0.5Zr_0.5−xM_xO₂ in the literature. SnO₂ has been widely used as an oxidation catalyst as it can reversibly undergo the Sn⁴⁺ ⇔ Sn²⁺ reaction at relatively lower temperatures.^21,22 Sasikala et al. have reported the preparation of Ce_1−xSn_xO₂ by a co-precipitation method and they observed low surface area for the as-prepared Ce_1−xSn_xO₂ with low oxidation/reduction temperature.²² Through a single step solution combustion method using tin oxalate precursor, Ce_1−xSn_xO₂ solid solutions have been prepared, which show high oxygen storage capacity compared to Ce_1−xZr_xO₂, however, the Ce_0.8Sn_0.2O₂ solid solution is only stable up to 700 °C in air.²³ Considering the smaller cation radius (eight-coordination) of Sn⁴⁺ (0.081 nm) than those of Zr⁴⁺ (0.084 nm) and Ce⁴⁺ (0.097 nm),²⁴ the incorporation of Sn⁴⁺ into Ce–Zr solid solutions may enhance the oxygen release reaction to form larger Ce³⁺.²⁵ In the present work, we describe the preparation and characterization of Ce_0.5Zr_0.43Sn_0.07O₂ solid solutions with high surface area via a facile hydrothermal route. Further experimental results show that introducing tin ions enhances the thermal stability and OSC even after calcination at 1000 °C for 20 h. The OSC of CeO₂ and Ce_0.5Zr_0.5O₂ prepared via the same method is compared.

2. Experimental

All chemicals used were of analytical grade and were purchased from Kanto Chemical Co. Inc., Japan (purity 99.999%). The chemicals were used without further purification.

2.1 Catalysts preparation

The stoichiometric amounts of (NH₄)₂Ce(NO₃)₆ (6 mmol), ZrO(NO₃)₂ (4.8 mmol) and SnCl₄·5H₂O (1.2 mmol) were dissolved in 60 ml distilled water. NH₄OH solution was slowly dropped into the above mixed solution, and the pH value was maintained at 9. The yellow mixed solution was introduced in a 100 ml Teflon®-lined autoclave, which was maintained at 200 °C for 24 h, then cooled to room temperature naturally. The obtained products were washed with distilled water three times, and dried in air at 100 °C for 12 h to form the as-prepared fresh samples. Finally, the fresh samples were calcined at 1000 °C for 20 h in air atmosphere to evaluate the thermal stability. The same synthesis route was employed for the preparation of the CeO₂ and Ce_0.5Zr_0.5O₂.

2.2 OSC analysis

The OSC of fresh samples and calcined samples at 1000 °C for 20 h was determined using a thermogravimetric-differential thermal analysis (TG-DTA, Rigaku TAS-200). Before the measurement, the samples were held in flowing air at 600 °C for 30 min to remove residual water and other volatile gases. The mixed gas of H₂–N₂ (H₂, 17.1%), (100 cm³ min⁻¹) and air (100 cm³ min⁻¹) was flowed alternatively at 600 °C. Finally OSC was analyzed after getting the TGA profile.

2.3 Characterization

The phase composition of the sample was determined by X-ray diffraction analysis (XRD, Shimadzu XD-D1) using graphite-monochromized Cu-Kα radiation. The morphology and size of the samples were determined by a transmission electron microscopy (TEM, JEOL JEM-2010). The specific surface area was measured using a BET (NOVA 4200e) surface area and pore size analyzer. The element analysis of the sample was carried out by an inductively coupled plasma atomic emission spectrometry measurement (ICP-AES, SPECTRO ARCOS EOP) after dissolving the sample by an acid digestion method. The sample shows the composition of Ce/Zr/Sn = 0.5 : 0.43 : 0.07.

3. Results and discussion

All products of (a) CeO₂, (b) Ce_0.5Zr_0.5O₂ and (c) Ce_0.5Zr_0.43Sn_0.07O₂, consisted of single phase of fluorite structure (Fig. 1(a)–(c)). All the diffraction patterns exhibited broad peaks suggesting that fresh samples were nanocrystalline materials. No phase separation was noticed even at such high calcination temperatures of 1000 °C for 20 h, except the increase of particle size (Fig. 1(a′)–(c′)). The crystal sizes of fresh CeO₂, Ce_0.5Zr_0.5O₂ and Ce_0.5Zr_0.43Sn_0.07O₂ calculated by Scherer's formula were 9, 5 and 3 nm, while those of calcined CeO₂, Ce_0.5Zr_0.5O₂ and Ce_0.5Zr_0.43Sn_0.07O₂ were 35, 10 and 8 nm, respectively. The calcined samples had a slight shift in diffraction peaks when compared to the pure CeO₂ XRD pattern. This shows the formation of the corresponding solid solutions. The calculated lattice parameters of the calcined samples of Ce_0.5Zr_0.5O₂ (a = 0.5384 nm) and Ce_0.5Zr_0.43Sn_0.07O₂ (a = 0.5371 nm) are smaller than that of CeO₂ (a = 0.5413 nm). The shrinkage of lattice cells may be due to the substitution of smaller cation radius of Zr⁴⁺ (0.084 nm) and Sn⁴⁺ (0.081 nm) with Ce⁴⁺ (0.097 nm).

Fig. 1 XRD patterns of fresh samples (a) CeO₂, (b) Ce_0.5Zr_0.5O₂, (c) Ce_0.5Zr_0.43Sn_0.07O₂, and calcined samples (a′) CeO₂, (b′) Ce_0.5Zr_0.5O₂ and (c′) Ce_0.5Zr_0.43Sn_0.07O₂.

The morphology and size of the fresh and calcined samples (1000 °C for 20 h) were observed by TEM as shown in Fig. 2. For fresh samples, the particles seem to be partly dispersed and formed small agglomerates (Fig. 2(a)–(c)), and single particles exhibited a spherical-like morphology with diameters of 9–12 nm, 5–8 nm and 3–5 nm for CeO₂, Ce_0.5Zr_0.5O₂ and Ce_0.5Zr_0.43Sn_0.07O₂, respectively, which are in agreement with the crystallite size calculated from Scherer's formula. The particle size increased after calcination at 1000 °C for 20 h because of aggregation, and the particle sizes were found to increase to 90–100 nm, 50–55 nm and 30–35 nm for the CeO₂, Ce_0.5Zr_0.5O₂ and Ce_0.5Zr_0.43Sn_0.07O₂ samples as shown in Fig. 2(a′)–(c′), respectively.

Fig. 2 TEM images of fresh samples (a) CeO₂, (b) Ce_0.5Zr_0.5O₂, (c) Ce_0.5Zr_0.43Sn_0.07O₂, and calcined samples (a′) CeO₂, (b′) Ce_0.5Zr_0.5O₂ and (c′) Ce_0.5Zr_0.43Sn_0.07O₂.

BET nitrogen adsorption-desorption analysis was undertaken to measure the specific surface area of all samples. As a result, the fresh sample of Ce_0.5Zr_0.43Sn_0.07O₂ showed much higher surface area (246 m² g⁻¹) than that of CeO₂ (119 m² g⁻¹) and Ce_0.5Zr_0.5O₂ (168 m² g⁻¹, Fig. 3(a)–(c)). After calcination at 1000 °C for 20 h in air the specific surface areas of CeO₂ (3 m² g⁻¹) and Ce_0.5Zr_0.5O₂ (8 m² g⁻¹) decreased to less than 10 m² g⁻¹, but the sample of Ce_0.5Zr_0.43Sn_0.07O₂ exhibited a relatively high BET specific surface area of 24 m² g⁻¹ (Fig. 3(a′)–(c′)).

Fig. 3 BET specific surface areas of fresh samples (a) CeO₂, (b) Ce_0.5Zr_0.5O₂, (c) Ce_0.5Zr_0.43Sn_0.07O₂, and calcined samples (a′) CeO₂, (b′) Ce_0.5Zr_0.5O₂ and (c′) Ce_0.5Zr_0.43Sn_0.07O₂.

The OSC values of fresh samples and the calcined samples were determined at 600 °C with a continuous flow of H₂–N₂ gas and air alternately.^26,27Fig. 4 shows the typical TG profiles of the CeO₂, Ce_0.5Zr_0.5O₂ and Ce_0.5Zr_0.43Sn_0.07O₂ samples. The TG profile shows the oxygen release/storage performance of the CeO₂, Ce_0.5Zr_0.5O₂ and Ce_0.5Zr_0.43Sn_0.07O₂ samples at 600 °C with time. As a result, Ce_0.5Zr_0.43Sn_0.07O₂ exhibited the higher OSC of 1425 μmol-O g⁻¹, when compared to that of CeO₂ (84 μmol-O g⁻¹) and Ce_0.5Zr_0.5O₂ (721 μmol-O g⁻¹) sample (Table 1). It is accepted that the OSC is dependent on the specific surface area, it is obvious that Ce_0.5Zr_0.43Sn_0.07O₂ exhibited the highest specific surface area and highest OSC values. Considering platinum pan was used in TG-DTA, the pan could not contribute to OSC even at high temperatures. Therefore, the highest OSC value of 1425 μmol-O g⁻¹ is close to the theoretical maximum value of OSC (1710 μmol-O g⁻¹).

Fig. 4 TG profiles of fresh and calcined samples (1000 °C, 20 h) at 600 °C, which show oxygen release/storage properties. Fresh samples: (a) CeO₂, (b) Ce_0.5Zr_0.5O₂ and (c) Ce_0.5Zr_0.43Sn_0.07O₂. Calcined samples: (a′) CeO₂, (b′) Ce_0.5Zr_0.5O₂ and (c′) Ce_0.5Zr_0.43Sn_0.07O₂.

Table 1 OSC of CeO₂, Ce_0.5Zr_0.5O₂, Ce_0.5Zr_0.43Sn_0.07O₂, Ce_0.5Zr_0.4Ti_0.1O₂ and Ce_0.5Zr_0.4 Fe_0.05Nb_0.05O₂

Chemical composition	OSC of fresh samples (μmol-O g^−1)	OSC of calcined samples (μmol-O g^−1)
CeO2	84	13
Ce0.5Zr0.5O2	721	449
Ce0.5Zr0.43Sn0.07O2	1425	1067
Ce0.5Zr0.4Ti0.1O2	906	339
Ce0.5Zr0.4Fe0.05Nb0.05O2	1112	654

---

The OSC of the CeO₂, Ce_0.5Zr_0.5O₂ and Ce_0.5Zr_0.43Sn_0.07O₂ samples decreased after calcination at 1000 °C for 20 h to 13 μmol-O g⁻¹, 449 μmol-O g⁻¹ and 1067 μmol-O g⁻¹ (Table 1), respectively. The decrease in OSC is due to the increase in particle size and decrease in BET surface area because of the sustained high temperature calcination. In order to examine the OSC performance and stability, oxygen release/storage cycles were measured, and Ce_0.5Zr0_.43Sn_0.07O₂ retained the same OSC value and the specific surface area (23 m² g⁻¹) even after 118 cycles (Fig. S1†). These results indicate that Ce_0.5Zr_0.43Sn_0.07O₂ has good OSC performance stability. Moreover, compared with Fig. 2(c) and (c′), the size increased to 9–12 nm for the fresh samples of Ce_0.5Zr_0.43Sn_0.07O₂ (Fig. S2(a)†), however, calcined samples (1000 °C for 20 h) kept almost the same morphology and size after treating at 600 °C for 20 h (Fig. S2(b)†).

Recently, incorporating tin into CeO₂ has been reported to improve the redox property and oxygen storage capacity at low temperatures.^28–30 As the cation radius of Sn⁴⁺ (0.077 nm) is smaller than both of Zr⁴⁺ (0.084 nm) and Ce⁴⁺ (0.097 nm), the incorporation of Sn⁴⁺ into Ce–Zr solid solutions may enhance the oxygen release reaction to form larger Ce³⁺. In the present work, higher surface area is exhibited by Ce_0.5Zr_0.43Sn_0.07O₂ solid solutions both before calcination and after calcination compared with Ce_0.5Zr_0.5O₂. The experimental results show that the incorporation of tin ions into Ce–Zr solid solutions also enhances the thermal stability. The Ce_0.5Zr_0.43Sn_0.07O₂ solid solutions show considerably higher oxygen storage capacity than Ce_0.5Zr_0.5O₂. The enhanced OSC is not only due to the high surface area but also due to the involvement of Ce⁴⁺/Ce³⁺ and Sn⁴⁺/Sn²⁺ redox couples utilized for H₂ oxidation.²³ It is proposed that SnO₂, being easily reducible, gives out its lattice oxygen for the oxidation reaction, which possibly gets rejuvenated by subtracting oxygen from the adjacent CeO₂ molecules.²²

Other cations, such as Fe³⁺ (0.078 nm), Nb⁵⁺ (0.074 nm) and Ti⁴⁺ (0.074 nm), whose radii are also smaller than both of Zr⁴⁺ and Ce⁴⁺,²⁴ could also enhance the OSC by incorporating into Ce–Zr solid solutions as shown in Fig. 5 and Table 1. Nb has a similar cation radius to that of Fe. It is better to balance the valency by adding Nb compared to the case of Fe³⁺ only. It is accepted that Fe³⁺ might be obtained under hydrothermal treatment conditions, because usually hydrothermal reactions provide an oxidative environment.^31,32 The relationship between OSC and cation radius is summarized in Fig. 6. It is indicated that incorporating cation with the radius from 0.074 to 0.084 nm into Ce–Zr solid solutions could enhance the OSC. The OSC (Fig. 5) can be greatly improved by incorporating Sn⁴⁺ compared to other cations such as Ti⁴⁺, Nb⁵⁺ and Fe³⁺. It was found that the samples incorporating Ti⁴⁺, Nb⁵⁺ and Fe³⁺ caused phase separation after calcination at 1000 °C for 20 h as shown in Fig. 7, which resulted in low thermal stability. It was indicated that these cations are too small to stabilize the fluorite structure of Ce–Zr solid solutions. The fresh samples exhibited high specific surface areas, however, they decreased to less than 10 m² g⁻¹ after calcination as shown in Fig. 8. Therefore, the radius of the incorporated cation might be an important factor which affects thermal stability.

Fig. 5 TG profiles after measuring the OSC at 600 °C for fresh and calcined samples (1000 °C, 20 h), which show oxygen release/storage properties. Fresh samples: (a) Ce_0.5Zr_0.4Ti_0.1O₂ and (b) Ce_0.5Zr_0.4Fe_0.05Nb_0.05O₂. Calcined samples: (a′) Ce_0.5Zr_0.4Ti_0.1O₂ and (b′) Ce_0.5Zr_0.4Fe_0.05Nb_0.05O₂.

Fig. 6 The relationship between OSC and incorporating cation radius. For fresh samples: (a) CeO₂, (b) Ce_0.5Zr_0.5O₂, (c) Ce_0.5Zr_0.43Sn_0.07O₂, (d) Ce_0.5Zr_0.4Fe_0.05Nb_0.05O₂ and (e) Ce_0.5Zr_0.4Ti_0.1O₂. For calcined samples: (a′) CeO₂, (b′) Ce_0.5Zr_0.5O₂, (c′) Ce_0.5Zr_0.43Sn_0.07O₂, (d′) Ce_0.5Zr_0.4Fe_0.05Nb_0.05O₂ and (e′) Ce_0.5Zr_0.4Ti_0.1O₂.

Fig. 7 XRD patterns of fresh samples (a) Ce_0.5Zr_0.4Ti_0.1O₂ and (b) Ce_0.5Zr_0.4 Fe_0.05Nb_0.05O₂, and calcined samples (a′) Ce_0.5Zr_0.4Ti_0.1O₂ and (b′) Ce_0.5Zr_0.4 Fe_0.05Nb_0.05O₂.

Fig. 8 BET specific surface areas of fresh samples (a) Ce_0.5Zr_0.4Ti_0.1O₂ and (b) Ce_0.5Zr_0.4Fe_0.05Nb_0.05O₂, and calcined samples (a′) Ce_0.5Zr_0.4Ti_0.1O₂ and (b′) Ce_0.5Zr_0.4Fe_0.05Nb_0.05O₂.

4. Conclusions

Ce_0.5Zr_0.43Sn_0.07O₂ solid solutions with high surface area were successfully synthesized via a facile hydrothermal method. The structures of the fresh samples and calcined samples were characterized by X-ray diffraction. The lattice parameters of Ce_0.5Zr_0.43Sn_0.07O₂ solid solution are smaller than those of CeO₂ and Ce_0.5Zr_0.5O₂, suggesting the incorporation of Sn⁴⁺ into the Ce–Zr solid solutions. The fresh particles showed a spherical-like morphology with diameters of 3–5 nm determined by TEM. Ce_0.5Zr_0.43Sn_0.07O₂ solid solutions exhibited a remarkably higher oxygen storage capacity than CeO₂ and Ce_0.5Zr_0.5O₂ samples prepared via the same method, even after calcination at 1000 °C for 20 h, indicating the improvement of the OSC and thermal stability due to the incorporation of tin. An appropriate radius range of the incorporating cation is also suggested.

Acknowledgements

This work was partly supported by Rare Metal Substitute Materials Development Project of New Energy and Industrial Technology Development Organization (NEDO), Japan and the Management Expenses Grants for National Universities Corporations from the Ministry of Education, Culture, Sports and Science for Technology of Japan (MEXT).

References

1. H. S. Gandhi, A. G. Piken, M. Shelef and R. G. Delosh, SAE Paper, 1976, 55, 760201.
2. X. Wu, X. Wu, Q. Liang, J. Fan, D. Z. Weng Xie and S. Wei, Solid State Sci., 2007, 9, 636.
3. J. Kašpar, P. Fornasiero and N. Hickey, Catal. Today, 2003, 77, 419.
4. J. Kašpar, P. Fornasiero and M. Graziani, Catal. Today, 1999, 50, 285.
5. J. Kašpar and P. Fornasiero, J. Solid State Chem., 2003, 171, 19.
6. R. Di Monte and J. Kašpar, Catal. Today, 2005, 100, 27.
7. R. Di Monte and J. Kašpar, J. Mater. Chem., 2005, 15, 633.
8. P. Fornasiero, G. Balducci, R. Di Monte, J. Kašpar, V. Sergo, G. Gubitosa, A. Ferrero and M. Graziani, J. Catal., 1996, 164, 173.
9. M. H. Yao, R. J. Baird, F. W. Kunz and T. E. Hoost, J. Catal., 1997, 166, 67.
10. K. Kenevey, F. Valdivieso, M. Soustelle and M. Pijolat, Appl. Catal., B, 2001, 29, 93.
11. F. Zhang, C. H. Chen, J. C. Hanson, R. D. Robinson, I. P. Herman and S. W. Chan, J. Am. Ceram. Soc., 2006, 89, 1028.
12. T. Nagai, T. Nonaka, A. Suda and M. Sugiura, R&D Rev. Toyota CRDL, 2002, 37, 20.
13. P. Fornasiero, R. Di Monte, G. R. Rao, J. Kašpar, S. Meriani, A. Trovarelli and M. Graziani, J. Catal., 1995, 151, 168.
14. H. F. Wang, X. Q. Gong, Y. L. Guo, Y. Guo, G. Z. Lu and P. Hu, J. Phys. Chem. C, 2009, 113, 10229.
15. T. Taniguchi, T. Watanabe, N. Matsushita and M. Yoshimura, Eur. J. Inorg. Chem., 2009, 2054.
16. M. K. Devaraju, X. W. Liu, K. Yusuke, S. Yin and T. Sato, Nanotechnology, 2009, 20, 405606.
17. M. Sanchez-Domingueza, L. F. Liotta, G. D. Carlod, G. Pantaleob, A. M. Veneziab, C. Solansa and M. Boutonnet, Catal. Today, 2010, 158, 35.
18. R. O. Fuentes and R. T. Baker, J. Phys. Chem. C, 2009, 113, 914.
19. J. O. Yang and H. M. Yang, J. Phys. Chem. C, 2009, 113, 6921.
20. M. L. Teng, L. T. Luo and X. M. Yang, Microporous Mesoporous Mater., 2009, 119, 158.
21. A. Boulahouache, G. Kons, H. G. Lintz and P. Schulz, Appl. Catal., A, 1992, 91, 115.
22. R. Sasikala, N. M. Gupta and S. K. Kulshreshtha, Catal. Lett., 2001, 71, 69.
23. T. Baidya, A. Gupta, P. A. Deshpandey, G. Madras and M. S. Hegde, J. Phys. Chem. C, 2009, 113, 4059.
24. R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751.
25. Q. Dong, S. Yin, C. S. Guo and T. Sato, Catal. Sci. Technol., 2012 10.1039/C2CY20425H.
26. D. Y. Wang, Y. J. Kang, V. Doan-Nguyen, J. Chen, R. KÜngas, N. L. Wieder, K. Bakhmutsky, R. J. Gorte and C. B. Murray, Angew. Chem., Int. Ed., 2011, 50, 4378.
27. J. Zhang, H. Kumagai, K. Yamamura, S.. Ohara, S. Takami, A. Morikawa, H. Shinjoh, K. Kaneko, T. Adschiri and A. Suda, Nano Lett., 2011, 11, 361.
28. R. Lin, Y. J. Zhong, M. F. Luo and W. P. Liu, Indian J. Chem., 2001, 40A, 36.
29. T. B. Nguyen, J. P. Deloume and V. Perrichon, Appl. Catal., A, 2003, 249, 273.
30. Y. Z. Chen, B. J. Liaw and C.W. Huang, Appl. Catal., A, 2006, 302, 168.
31. M. T. Liang, S. H. Wang, Y. L. Chang, H. I. Hsiang, H. J. Huang, M. H. Tsai, W. C. Juan and S. F Lu, Ceram. Int., 2010, 36, 1131.
32. J. Zhang, X. L. Wang, X. H. Xia, C. D. Gu and J. P. Tu, Sol. Energy Mater. Sol. Cells, 2011, 95, 2107.

---

Footnote

† Electronic supplementary information (ESI) available: See DOI: 10.1039/c2ra21766j

---