Luminescence properties of Ca₂Si₅N₈:Eu²⁺ prepared by gas-pressed sintering using BaF₂ as flux and cation substitution

Abstract

Eu²⁺ doped Ca₂Si₅N₈ phosphors were successfully prepared by gas-pressed sintering. The red-shift of the emission band from 608 nm to the longer wavelength 622 nm of the Ca₂Si₅N₈:Eu²⁺ phosphor under blue excitation has been achieved, and a large enhancement in the emission intensity has been obtained by using BaF₂. XRD data revealed that the lattice of Ca₂Si₅N₈:Eu²⁺ was expanded with Ba²⁺ ion doping. XPS results suggested that there were more Eu²⁺ ions incorporated into the lattice of Ba²⁺ doped samples than those of the undoped samples. The doping effect of Ba²⁺ ions has been discussed in detail.

---

1. Introduction

Light emitting diode (LED)-based solid-state lighting has recently received worldwide attention, owing to the characteristics of high efficiency, simple structure, and long life.¹ The need for new phosphors arose in the past decade with the invention of the near-UV to blue emitting GaN-based LEDs by Nakamura in 1991 and the development of high-power LEDs (HP-LEDs) in the same spectral region.² It is well known that the spectral properties of rare-earth ions with 5d–4f transitions (e.g., Eu²⁺, Ce³⁺) strongly depend on the surrounding environment (e.g., symmetry, covalence, coordination, bond length, site size, crystal-field strength, etc.), due to the fact that the 5d excited state is not shielded from the crystal field by the 5s² and 5p⁶ electrons.^3–5 These phosphors can efficiently absorb in the near UV to blue spectral range and emit light in the visible range. And the small Stokes shift leads to high conversion efficiency. For a LED phosphor to be applied in commercial products, several criteria have to be met such as high efficiency in light conversion, high thermal quenching temperature, and the possibility to adjust the color point by means of varying the chemical composition. Host lattices with a high degree of covalence and/or a large crystal field splitting at the site for which Eu²⁺ substitute can lead to efficient visible emission while absorbing light in the near UV to blue range of the electromagnetic spectrum. Presently Y₃Al₅O₁₂:Ce³⁺ (YAG:Ce³⁺) is applied in phosphor converted LEDs (pc-LEDs) as luminescent converter.^6,7 The most common commercial white LEDs (WLEDs) are combination of the blue-emitting InGaN chip and the yellow YAG:Ce³⁺ phosphor. However, they have a low Colour Rendering Index (CRI) because of their lack of emission in the red region, and high color temperature due to the deficiency of emission in the visible spectrum. An alternative way to overcome this weakness is the incorporated with red phosphors. In this respect, a promising new class of LED phosphor materials is the nitridosilicates. The N³⁻ in this lattice is a soft Lewis base, which results in a high covalence. This shifts the energy of the 4f–5d absorption and emission for Eu²⁺ ions to sufficiently low energies.^8,9 In addition, nitridosilicates are known to be highly stable with oxidation and hydrolysis.¹⁰ At present the commercial red nitrides phosphors are CaAlSiN₃:Eu²⁺ (ref. 11 and 12) and Sr₂Si₅N₈:Eu²⁺.^13–15 However, the sintering process of CaAlSiN₃:Eu²⁺ phosphor needs critical preparation conditions (higher temperature, higher N₂ pressure, and air sensitive starting powders). For the latter, Piao et al.¹⁶ reported on the carbothermal reduction and nitridation method to synthesize Sr₂Si₅N₈:Eu²⁺ phosphor. In this method, residual carbon is inevitably incorporated into the phosphor, which reduces its intensities of absorption and emission. Thus these nitrides phosphors cannot meet the requirements of orange-red phosphors at present. So there is a still need to discover novel orange-red phosphors. Currently, there is a increasing interest in nitridosilicates Ca₂Si₅N₈:Eu²⁺ due to the potential application in warm white LEDs.^17,18 And the preparation conditions are easier than the other commercial nitrides phosphors, such as the lower sintering temperature and pressure as well as the cheaper raw materials. Recently the researches on Ca₂Si₅N₈:Eu²⁺ mainly concentrate on the different doping concentrations of Eu²⁺ to tune the emission color and increase the emission intensity. But there is no report on the cations substitution in Ca₂Si₅N₈ host to adjust the luminescence property. BaF₂ is a potential flux.^19,20 So we focus on the Ca₂Si₅N₈:Eu²⁺ by using a BaF₂ as flux and cations substitution.

In this paper we report the possibility to tune the emission color of Ca₂Si₅N₈:Eu²⁺ by incorporating BaF₂, to determine the most promising way to design new compositions that can serve as efficient phosphors in pc-LEDs with a warmer color. The luminescence and thermal quenching properties have been estimated. And the mechanism for the emission increasing and wavelength shift after BaF₂ doping is discussed.

2. Experimental section

2.1 Materials and synthesis

A series of nitridosilicate phosphors, Ca₂Si₅N₈:Eu²⁺ with different concentration of BaF₂ were prepared by gas-pressed sintering. Stoichiometric amounts of powder BaF₂ (AR), Ca₃N₂ (Aldrich, >95.0%), Si₃N₄ (Aldrich, 99.5%), and Eu₂O₃ (Aldrich, 99.99%) were ground in an agate mortar for 30 min in a glove box to form a homogeneous mixture. The concentrations of both moisture and oxygen in the glovebox were <1 ppm. Thereafter, the powder mixture was transferred into a BN crucible and heated at 1500 °C for 4 h under high-purity nitrogen (99.9995%) atmosphere at a pressure of 0.2 MPa. The sintered products were ground again, yielding crystalline powder.

2.2 Characterization

All measurements were made on finely ground powder. The phase purity of samples were analyzed by X-ray diffraction (XRD) using a Rigaku D/Max-2400 X-ray diffractometer with Ni-filtered CuKα radiation. Photoluminescence (PL) and PL excitation (PLE) spectra were measured at room temperature using an FLS-920T fluorescence spectrophotometer equipped with a 450 W Xe light source and double excitation monochromators. High temperature luminescence intensity measurements were carried out by using an aluminum plaque with cartridge heaters; the temperature was measured by thermocouples inside the plaque and controlled by a standard TAP-02 high temperature fluorescence controller. X-ray photoelectron spectroscopy (XPS) measurements were performed on an ESCALAB250xi high-performance electron spectrometer using a monochromatized AlKα excitation source (hν = 1486.6 eV).

3. Results and discussion

Fig. 1 shows the XRD patterns of the Ca₂Si₅N₈:Eu²⁺ phosphors as a function of Eu²⁺ concentration. The detailed analysis of the XRD patterns of samples (Ca_1−xEu_x)₂Si₅N₈ (x = 0.05, 0.1, 0.15, 0.2 and 0.25) shows that their phase compositions depend on the concentration of Eu²⁺. When 0 ≤ x ≤ 0.15, the samples are almost the single phase of (Ca_1−xEu_x)₂Si₅N₈ with a small amount of α-Si₃N₄ as impurity. A further increase of Eu²⁺ concentration leads to the formation of α-Si₃N₄ and EuSiO₃, both as impurities. These observations indicate that the solubility of Eu²⁺ in Ca₂Si₅N₈ is very limited, probably, to a range of 0–0.15, i.e. 0 ≤ x ≤ 0.15, which is in consistent well with the result reported by Li et al.²¹ Such a limitation may be due to two reasons. One is the difference in the crystal structure that Ca₂Si₅N₈ is monoclinic while Eu₂Si₅N₈ is orthorhombic. The other is the difference between the ionic radii: the radius of Eu²⁺ (1.17 Å) (1 Å = 0.1 nm) is obviously larger than that of Ca²⁺ (1.00 Å).²²

Fig. 1 XRD patterns of the Ca₂Si₅N₈:Eu²⁺ phosphors as a function of Eu²⁺ concentration.

Fig. 2(a) shows the excitation spectrum (λ_em = 617 nm) of Ca_1.85Eu_0.15Si₅N₈. The Ca_1.85Eu_0.15Si₅N₈ phosphor exhibited a typical broad excitation band resulting from the crystal field splitting of the 5d orbital due to the 4f⁷-ground state to the 4f⁶5d-excited state of the Eu²⁺ ion electronic transitions.²³ Fig. 2(b) shows the emission spectra of the Ca_2−xEu_xSi₅N₈ phosphors synthesized at 1500 °C for 4 h excited at 460 nm. The relative emission peak originated from the transitions of the 5d to the 4f states. As the Eu²⁺ doping concentration increases, the relative emission intensity increases continuously. The highest emission intensity is observed for the 0.15 mol of Eu²⁺ sample. However, when the Eu²⁺ concentration exceeds 0.15 mol, there was a sudden decrease in the emission intensity due to concentration quenching.²⁴ As the Eu²⁺ contents increase, the distance between the Eu²⁺ ions becomes smaller, which leads to the probability of energy transfer among Eu²⁺ ions.²⁵ When the Eu²⁺ concentration increases, the emission band shifts to the red side. This may be ascribed to the lattice distortion caused by Eu²⁺ ions introducing the mismatch between the small Ca²⁺ and large Eu²⁺ ionic radius in the lattice.²⁶

Fig. 2 PLE spectrum (a) (monitored at 617 nm) and PL spectra (b) (excited at 460 nm) of the Ca₂Si₅N₈:Eu²⁺ phosphor with different Eu²⁺ contents.

Fig. 3 shows the XRD patterns of Ca_1.95Eu_0.05Si₅N₈ with different weight of BaF₂. When BaF₂ is doped, the sample is the single phase of (Ca_1−xEu_x)₂Si₅N₈. The positions of the peaks move to lower angles and the volumes of lattice parameters show smooth evolution as the BaF₂ content increases, which means that the higher the Ba²⁺ content is, the larger the lattice parameters are (seen in Fig. S1 ESI†) and the Ba²⁺ should occupy the Ca²⁺ position. The crystallinity has been improved with the addition of BaF₂. When 8 wt% BaF₂ is added, the crystallinity of the sample reaches the highest. And the crystallinity begins to decline when the content of BaF₂ exceeds 8 wt%.

Fig. 3 XRD patterns of the Ca₂Si₅N₈:Eu²⁺ phosphors with different concentrations of BaF₂.

Fig. 4 shows the experimental, calculated, and difference results of Rietveld refinement XRD patterns of Ca_1.95Eu_0.05Si₅N₈ with 8 wt% BaF₂ at room temperature. The crystal structure of Ca_1.95Eu_0.05Si₅N₈ with 8 wt% BaF₂ was analyzed by the Materials Studio program on the basis of the XRD data. The pattern factor R_p, and the weighted pattern factor R_wp, are 10.33% and 14.22%, respectively. The XRD patterns of Ca_1.95Eu_0.05Si₅N₈ with 8 wt% BaF₂ obtained herein indicate that single phase is formed. The Ca_1.95Eu_0.05Si₅N₈ with 8 wt% BaF₂ synthesized crystallized as a monoclinic structure with the space group of Cc. The refined structure parameters of BaF₂ doped Ca₂Si₅N₈:Eu²⁺ are given in Table 1.

Fig. 4 Rietveld refinement results of the XRD patterns of the Ca₂Si₅N₈:Eu²⁺ with 8 wt% BaF₂, including the experimental and calculated intensities as well as differences in intensity between experimental and calculated data.

Table 1 The refined structure parameters of BaF₂ doped Ca₂Si₅N₈:Eu²⁺

Atom	Site	x/a	y/b	z/c
Ca1	0.95	−0.00658	0.75426	0.05207
Ba	0.05	−0.00658	0.75426	0.05207
Ca2	1	0.61050	0.73056	0.26069
Si1	1	0.05412	0.79009	0.41769
Si2	1	0.75221	0.19935	0.34537
Si3	1	0.76388	0.51462	0.11491
Si4	1	0.35762	0.21081	0.42527
Si5	1	0.85405	0.02307	0.17204
N1	1	0.94559	0.55880	0.44382
N2	1	0.12258	0.12972	1.08940
N3	1	0.81030	0.25662	0.23980
N4	1	0.79107	0.85548	0.15102
N5	1	0.98494	0.98304	0.27838
N6	1	0.86102	0.17452	1.06034
N7	1	0.62438	0.04133	0.36236
N8	1	0.79600	0.49423	0.41610

---

When compared the two images of (a) Ca_1.95Eu_0.05Si₅N₈ and (b) Ca_1.95Eu_0.05Si₅N₈ with 8 wt% BaF₂ in Fig. 5, we can find that the addition of the BaF₂ in the host is helpful for enhancing the crystallization degree and decreasing surface defects. This indicates that the Ca_1.95Eu_0.05Si₅N₈ phosphor has a good dispersion, a regular shape, and the particle size of the synthesized powder was about 6–12 μm. The corresponding EDX spectra analysis (Fig. 5(c) and (d)) and elemental mappings (Fig. 6(a) and (b)) indicates that the products have a chemical composition of Ca, Si, O and N and Ca, Ba, Si, O and N. And the Ba²⁺ is incorporated into the Ca_1.95Eu_0.05Si₅N₈.

Fig. 5 SEM and EDX spectra of the (a and c) Ca_1.95Eu_0.05Si₅N₈ and (b and d) Ca_1.95Eu_0.05Si₅N₈ with 8 wt% BaF₂.

Fig. 6 Elemental mappings of Ca_1.95Eu_0.05Si₅N₈ and Ca_1.95Eu_0.05Si₅N₈ with 8 wt% BaF₂.

Fig. 7 shows the excitation and emission spectra of Ca_1.95Eu_0.05Si₅N₈. It is obvious that all the spectral features of the as-synthesized samples are similar. The excitation spectra consist of three broad bands peaking at about 295, 397 and 467 nm, which mainly arise from the 4f⁶5d¹ multiplets of Eu²⁺ excitation states. And the remarkable enhancement of the emission intensity is observed with increasing the content of BaF₂. After incorporating 8 wt% BaF₂, the emission intensity reaches twice than that of the sample without BaF₂. In addition, it is noticeable that the emission peak of the Eu²⁺ shifts to longer wavelength (608 nm to 622 nm) with increasing the concentration of BaF₂. This would be beneficial to the color point tuning. The excitation, emission, centre of gravity and Stokes shift crystal filed splitting of Ca_1.95Eu_0.05Si₅N₈ with different weight of BaF₂ are listed in Table 2. The enhancement can be explained by the fluxing agent (BaF₂). The redshift could be explained by the fact that the Ca₂Si₅N₈:Eu²⁺ structure is preserved while a part of the Ca²⁺ ions are replaced by the larger Ba²⁺ ions. To accommodate these larger cations, the distance between Ca²⁺ (or Eu²⁺) and the anions could not increase or even become slightly smaller, thus leading to the increase of the crystal field splitting, and then causing a red shift of the emission.²⁷

Fig. 7 PL/PLE spectra of Ca_1.95Eu_0.05Si₅N₈ with different weight of BaF₂.

Table 2 Excitation, emission, centre of gravity and Stokes shift crystal filed splitting of Ca_1.95Eu_0.05Si₅N₈ with different weight of BaF₂

Samples	λex (nm)	λem (nm)	Center of gravity (cm^−1)	Stocks shift (cm^−1)
Without BaF2	295, 397, 467	608	26 830	4966
2% BaF2	295, 397, 467	613	26 830	5100
4% BaF2	295, 397, 467	615	26 830	5153
6% BaF2	295, 397, 467	618	26 830	5232
8% BaF2	295, 397, 467	622	26 830	5336

---

Table 3 XPS quantitative results of Ca_1.95Eu_0.05Si₅N₈ with and without BaF₂

Peak	Position BE (eV)	Atomic conc.%	Mass conc.%
N	402	15.44	10.23
O	536	21.34	16.16
N	402	18.81	12.68
O	536	20.56	15.84

---

Fig. 8 shows the XPS of Ca_1.95Eu_0.05Si₅N₈ with 8 wt% BaF₂ and without BaF₂. The lattice parameters are greatly affected by the occupation of Eu²⁺ and Ba²⁺ ions in the critical structure, which depends on the difference in electronegativity and ionic radii compared with the replaced ions. The two ions have a similar possibility to replace Ca²⁺ ions and be incorporated into the structure. However, it is well known that the vacancy formation caused by charge imbalance and lattice strain can self-limit the inclusion of guest ions into a host lattice.²⁸ Thus there is a propensity for the ions to migrate to less strained surface sites, rather than incorporate in the crystal lattice, which can be confirmed by XPS data. In Fig. 6, the peak at about 135.6 eV attributed to Eu_4d is assigned to Eu₂O₃ which was formed on the surface of the sample. That means more Eu²⁺ ions are incorporated into the lattice, lead to emission intensity increasing and red shift of the emission. Another reason is that the N/O ratio is higher in the Ca_1.95Eu_0.05Si₅N₈ with 8 wt% BaF₂ sample. The nitrogen ion (N³⁻) has a higher effective charge compared with the oxygen ion (O²⁻), and the electronegativity of nitrogen (3.04) is smaller than that of oxygen (3.50). Therefore, coordinating with nitrogen would cause a stronger nephelauxetic effect (covalence), the centre of gravity of the 5d states of the activator ions shift to longer wavelength, and the crystal-field splitting larger than that in a similar oxygen environment,^29,30 which leads to the red shift of the emission (Table 3).

Fig. 8 XPS survey spectrum of Ca_1.95Eu_0.05Si₅N₈ with and without BaF₂.

Fig. 9(a) shows the crystal structure of Ca₂Si₅N₈ viewed along [010] and Fig. 9(b) depicts the proposed model of substitution of Eu²⁺ and Ba²⁺ for Ca²⁺. Furthermore, a random ion displacement model can be used to clarify the modification of the lattice. This allows the use of an analysis similar to Vegard's law, which is an empirical law that relates the statistical substitution of a guest ion into the host lattice with the experimentally observed degree of lattice change with increasing defect ion concentration. Statistical substitution into a lattice site is predicted to lead to a lattice contraction for smaller ions and a lattice expansion for larger ions. When there are only Eu²⁺ ions doped in the structure, the cell lattice will be expanded, since the radius of Eu²⁺ ions is larger than that of Ca²⁺ ions. A strain may arise in the lattice around the Eu²⁺ ions, and may limit the stability of the Eu²⁺ ions that had been incorporated into the lattice. Then, when Ba²⁺ was doped into the structure, the Ba²⁺ ions with larger radius than that of Eu²⁺ could make the lattice expand. So more Eu²⁺ would incorporate into the lattice because of the larger lattice expended by Ba²⁺. This means that the structure doped with Ba²⁺ ions can make more Eu²⁺ ions incorporate into the lattice.

Fig. 9 (a) Crystal structure of Ca₂Si₅N₈ viewed along [010]. (b) The proposed model of substitution of Eu²⁺ and Ba²⁺ for Ca²⁺.

Fig. 10 shows the temperature dependence of the integrated emission intensity for Ca_1.95Eu_0.05Si₅N₈ with and without 8 wt% BaF₂, which shows an identical thermal stability. It is believed that thermal ionization is responsible for quenching of the luminescence of Eu²⁺ at high temperatures in Ca₂Si₅N₈ host,¹⁷ because the excited 5d electrons are easily ionized by the absorption of thermal energy and entrance into the bottom of the conduction band of the host through the top of the Eu²⁺ excitation levels. At 200 °C, the integral emission intensity of the both phosphors is about 30% of that measured at room temperature.

Fig. 10 Temperature dependent integrated emission intensities of Ca_1.95Eu_0.05Si₅N₈ with and without 8 wt% BaF₂. (The inset show the emission spectra with increasing temperature).

Fig. 11 represents the Commission International de I'Eclairage (CIE) chromaticity coordinates for Ca_1.95Eu_0.05Si₅N₈ with different amounts of BaF₂ (0–8 wt%). With increasing the content of Ba²⁺, the chromaticity coordinates (x, y) vary systematically from (0.556, 0.437) to (0.591, 0.407), corresponding to color points of the samples change gradually from orange-yellow to orange-red. Therefore, it is expected that the white light with good rendering could be obtained when the tunable emission phosphors Ca_1.95Eu_0.05Si₅N₈ with different amounts of BaF₂ (0–8 wt%) for white LEDs.

Fig. 11 CIE chromaticity coordinates of Ca_1.95Eu_0.05Si₅N₈ with different BaF₂ (0–8 wt%).

4. Conclusions

A remarkable orange-red nitridosilicate phosphor Ca_1.95Eu_0.05Si₅N₈ with 8 wt% BaF₂ was synthesized by gas-pressed sintering at 1500 °C using the raw materials BaF₂, Ca₃N₂, Si₃N₄, and Eu₂O₃. This method promises the use of inexpensive, commercially available and powder handling at ambient pressure, thus offering a simple, efficient, and high-yield way to obtain orange-red emitting phosphors. The color of the emission can be tuned with adding BaF₂ into the Ca₂Si₅N₈:Eu²⁺ host lattice. We have demonstrated that adding BaF₂ can make better single phase, emission peak shift to longer wavelength (608 nm to 622 nm) and significantly enhance the emission intensities of Ca₂Si₅N₈:Eu²⁺ phosphors. This is mainly due to the fact that the lattice structure of Ca₂Si₅N₈:Eu²⁺ phosphors can be modified and the solubility of the Eu²⁺ ions can be increased by adding BaF₂. More Eu²⁺ ions would incorporate the lattice because of the larger lattice expended by Ba²⁺. This novel Ca₂Si₅N₈:Eu²⁺ with 8 wt% BaF₂ phosphor is expected to be useful for phosphor converted white LEDs.

Acknowledgements

This work is supported by Specialized Research Fund for the Doctoral Program of Higher Education (no. 20120211130003) and the National Natural Science Funds of China (Grant no. 51372105).

Notes and references

1. E. F. Schubert and J. K. Kim, Science., 2005, 308, 1274–1278.
2. S. Nakamura, Jpn. J. Appl. Phys., 1991, 30(10A), L1705.
3. G. Blasse and B. C. Grabmaier, Luminescent Materials, Springer, 1994, vol. 44.
4. J. W. H. van Krevel, H. T. Hintzen, R. Metselaar and A. Meijerink, J. Alloys Compd., 1998, 268, 272–277.
5. R. J. Xie, N. Hirosaki and M. Mitomo, J. Electroceram., 2008, 21, 370–373.
6. P. Schlotter, J. Baur, C. Hielscher, M. Kunzer, H. Obloh, R. Schmidt and J. Schneider, J. Mater. Sci. Eng. B, 1999, 59, 390–396.
7. Y. Hu, W. Zhuang, H. Ye, S. Zhang, Y. Fang and X. Huang, J. Lumin., 2005, 111, 139–145.
8. J. Ballato, J. S. Lewis III and P. Holloway, Mater. Res. Bull., 1999, 24(9), 51.
9. Z. J. Zhang, O. M. ten Kate, A. Delsing, E. van der Kolk, P. H. L. Notten and H. T. Hintzen, J. Mater. Chem., 2012, 22, 9813–9820.
10. C. Hecht and W. Schnick, Chem. Mater., 2009, 21(8), 1595–1601.
11. X. Piao, K. Machida, T. Horikawa, H. Hanzawa, Y. Shimomura and N. Kijima, Chem. Mater., 2007, 19, 4592–4599.
12. Y. Li, N. Hirosaki, R. Xie, T. Takeda and M. Mitomo, Chem. Mater., 2008, 20, 6704–6714.
13. X. Piao and K.-i. Machida, J. Lumin., 2010, 13, 08–12.
14. H. Nersisyan, H. Il Won and C. W. Won, Chem. Commun., 2011, 47, 11897–11899.
15. Y. Q. Li and H. T. Hintzen, J. Solid State Chem., 2008, 181, 515–524.
16. X. Piao, K.-i. Machida, T. Horikawa and H. Hanzawa, J. Electrochem. Soc., 2008, 155(1), 17–22.
17. X.-D. Wei and Q.-L. Liu, Chin. Phys. B, 2009, 18, 3555–3563.
18. B. Lei, Y. Shimomura and H. Yamamoto, J. Electrochem. Soc., 2010, 157, J196–J201.
19. S. Ho Lee, H. Young Koo, D. Soo Jung, J. Man Han and Y. Chan Kang, Opt. Mater., 2009, 31, 870–875.
20. S. Ho Lee, D. Soo Jung, J. Man Han, H. Young Koo and Y. Chan Kang, J. Alloys Compd., 2009, 477, 776–779.
21. Y. Q. Li, G. de With and H. T. Hintzen, J. Alloys Compd., 2006, 417, 273.
22. R. D. Shannon, Acta Crystallogr., 1976, A32, 751–767.
23. Y. Gu, Q. Zhang, Y. Li, H. Wang and R. J. Xie, Mater. Lett., 2009, 63, 1448–1450.
24. D. L. Dexter and J. H. Schulman, J. Chem. Phys., 1954, 22, 1063.
25. L. G. Van Uitert, J. Electrochem. Soc., 1967, 114, 1048–1053.
26. H. Zhang, T. Horikawa, H. Hanzawa, A. Hamaguchi and K.-I. Machida, J. Electrochem. Soc., 2007, 154(2), J59–J61.
27. V. Bachmann and A. Meijerink, Chem. Mater., 2009, 21, 316–325.
28. Y. Gu, Q. Zhang, Y. Li, H. Wang and R. J. Xie, Mater. Lett., 2009, 63, 1448–1450.
29. Z. Y. Zhao, Z. G. Yang, Y. R. Shi, C. Wang, B. Liu, G. Zhu and Y. H. Wang, J. Mater. Chem. C, 2013, 1, 1407–1412.
30. C. W. Yeh, W. T. Chen, R. S. Liu, S. F. Hu, H. S. Sheu, J. M. Chen and H. T. Hintzen, J. Am. Chem. Soc., 2012, 134, 14108–14117.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra04683h

---