Novel fabrication of TiO₂/ZnO nanotube array heterojunction for dye-sensitized solar cells

Abstract

TiO₂ nanotube arrays (TNTs) were prepared by using an electrochemical anodization method, and then the detached TNTs were successfully transferred onto a fluorine-doped tin oxide glass substrate with the help of a tetrabutyl titanate-containing sol. In order to improve the photoelectric conversion efficiency (PCE) of the as-fabricated dye-sensitized solar cells (DSSCs), an easily-operated immersing method was employed to fabricate a TiO₂/ZnO nanotube array (TNT/ZnO) heterojunction, which has advantages of a high aspect ratio, low recombination rate and high absorption of visible light. The results indicate that under AM 1.5 illumination, the DSSCs based on the TNT/ZnO heterojunction exhibit a better short circuit current density (J_sc) of 8.67 mA cm⁻² and a higher PCE of 3.98%. Electrochemical impedance spectroscopy analysis shows that the TNT/ZnO heterojunction-based DSSCs have optimized properties, such as a longer electron lifetime, lower impedance of electron transport, higher impedance of electron recombination, and electron collecting rate as high as 95.2%.

---

Introduction

Since Grätzel fabricated a low-cost dye-sensitized solar cell based on colloidal titanium dioxide in 1991,¹ this new research field has attracted considerable attention and made great progress in recent years. For example, Grimes et al. prepared TiO₂ nanotube arrays (TNTs) by an electrochemical anodization method.² Considering that the TNTs have a good light absorption and electron transport ability, they received much attention and many research scientists carried out the preparation and characterization of the TNTs and their application for dye-sensitized solar cells (DSSCs).³ For improving the photoelectric conversion efficiency (PCE) of DSSCs, many ideas were proposed, such as O₂ plasma treatment,⁴ increasing the length of the TNTs to as long as 33.0 μm (ref. 5) and transferring the TNTs onto a fluorine-doped tin oxide (FTO) glass substrate.⁶ However, the PCE of the DSSCs based on the TNTs is still relatively low and difficult to be further improved, which is due to a high recombination rate of the photo-generated electron and hole pairs.⁷ An attractive way to overcome this issue is the fabrication of a heterojunction, which has become a research focus for DSSCs in recent years. For example, by the electrochemical deposition of ZnO, Xie et al.⁸ and Liu et al.⁹ successfully applied coaxial TiO₂/ZnO nanotube arrays as photoanodes for DSSCs. However, the DSSCs were based on non-transferred TNTs, allowing back light irradiation only, which severely limited the PCE of the DSSCs. In this paper, DSSCs were fabricated based on transferred TNTs and a step forward by replacing the electrochemical deposition by immersing TNTs in Zn(NO₃)₂-based solution, thus simplifying the formation of the TiO₂/ZnO nanotube arrays and also improvingthe PCE of the DSSCs. Fig. 1 shows a flowchart of the DSSC fabrication.

Fig. 1 Flowchart of the DSSC fabrication.

As is well known, DSSCs can be divided into two different modes, including back light irradiation¹⁰ and front light irradiation.¹¹ Thus, the purpose of transferring TNTs onto a fluorine-doped tin oxide (FTO) glass substrate is to realize the mode of front light irradiation, so as to make it possible for the light to reach the dye molecules that are absorbed to the TNTs or the TNT/ZnO heterojunction films directly, and thus avoid the unnecessary loss of light due to the absorption or reflection from the Pt-coated FTO glass substrate and the I⁻₃/I⁻-containing electrolyte.

ZnO has a similar energy band structure to TiO₂, that is, it is possible that ZnO and TiO₂ can form a steady heterojunction.⁹ Another important reason for choosing ZnO to fabricate the TiO₂/ZnO heterojunction is to make use of its high electron mobility, which is 2–3 orders of magnitude higher than that of TiO₂.¹² Hence, the TNT/ZnO heterojunction presents advantages for DSSCs. Firstly, the surface of the TNTs is very smooth, but ZnO can exist in the form of nanoparticles on the surface of TNTs,¹³ as depicted in Fig. 1, thus increasing the aspect ratio to a large extent, which is a great help to absorb more dye molecules; secondly, the existence of ZnO can prevent the recombination rate of the electrons with dye molecules¹⁴ and the I⁻₃/I⁻-containing electrolyte, due to a higher conduction band of ZnO,¹⁵ as shown in Fig. 2; thirdly, the ZnO layer plays the role of a scattering layer, thus increasing the absorption of visible light.¹⁶

Fig. 2 Comparison of (a) TNT-based DSSC and (b) TNT/ZnO-based DSSCs.

Experimental

Materials

The following chemicals were employed for the fabrication of the DSSCs based on the TNT/ZnO heterojunction in the present work; titanium foils (50 mm × 15 mm × 0.3 mm, 99.9%), deionized water (18.0 MΩ cm in resistivity), zinc nitrate hexahydrate [Zn(NO₃)₂·6H₂O], ethylene glycol (CH₂OH)₂, ammonium fluoride (NH₄F), ethanol, ethanolamine (C₂H₇NO, MEA), polyethylene glycol 600, tetrabutyl titanate, acetylacetone, tert-butyl alcohol (C₄H₁₀O), acetonitrile (C₂H₃N), 4-tert-butylpyridine, guanidine thiocyanate (C₂H₆N₄S), 1-methyl-3-propylimidazolium iodide (C₇H₁₃IN₂), valeronitrile (C₅H₉N), and iodine (I₂).

Transferring of TNTs onto FTO glass substrate

Highly ordered TNTs film were prepared by using a two-step anodization method. Here, the Ti foils were firstly cleaned by ultra-sonication in acetone, ethanol and deionized water for 10 min, respectively. After drying them at 80 °C, two pieces of the cleaned Ti foils were used as the anode and cathode, respectively. A mixed solution, which contained 0.3 wt% NH₄F, 2 vol% deionized water and 98 vol% ethylene glycol,¹⁸ was employed as the electrolyte. The anodization process was performed at a constant voltage of 60 V at 15 °C for 5 hours. Then, the anode Ti foil was taken out and ultrasonicated for several minutes until the TNTs were detached from the Ti foil, which resulted in the Ti foil having a shiny surface.¹⁹ This process was recognized as the first-step of anodization. Then, the Ti foil with a shiny surface was used for the second-step anodization, similarly, the as-obtained Ti foil were still ultrasonically cleaned by acetone, ethanol and deionized water in turn, and the anodization temperature and anodization voltage were still controlled at 15 °C and 60 V, but the anodization time was shortened to 3 h. At the end of the second-step anodization, the anodization voltage was uniformly and slowly decreased to 10 V in 20 seconds, followed by a sudden increase to 100 V and maintained for 20 seconds. It should be mentioned here that this final anodization processes allows the detachment of the entire TNTs film from the Ti foil. After shutting off the power, the Ti foil was taken out of the electrolyte and immersed in deionized water immediately, to enable the water to get into the TNTs, which is also quite helpful to detach the TNTs film. In addition, the tetrabutyl titanate-contained sol was employed to transfer the TNTs film onto a FTO glass substrate. Then, the FTO glass substrate covered by the TNTs film was dried in an oven, and sintered at 500 °C for 1 h, at a heating rate of 3 °C min⁻¹.

Preparation of TNT/ZnO heterojunction

A facile process was used to achieve a TNT/ZnO heterojunction. Typically, a Zn(NO₃)₂·6H₂O-based solution, which contained a certain amount of ethanolamine (MEA), was used for immersing the FTO glass substrate with the TNTs film (called the transferred TNTs) to fabricate the TNT/ZnO heterojunction. Here, the MEA was used to prevent an agglomeration of the ZnO nanoparticles²⁰ and also made the solution alkaline, which ensures that Zn element can exist in the form of Zn(OH)₂. In order to achieve an ideal concentration of Zn(NO₃)₂, four different solutions with molar concentrations (MC) of Zn(NO₃)₂ controlled accurately at 0.25 M, 0.50 M, 0.75 M, and 1.00 M were prepared. Another important factor to affect the preparation of the TNT/ZnO heterojunction is the immersing time of the transferred TNTs into the Zn(NO₃)₂·6H₂O-based solution. In the present work, 13 samples were prepared and the immersion time was set at 1 h, 2 h and 3 h, respectively. Finally, all the samples were put into a furnace to carry out a heat treatment at 500 °C for 2 h, at a heating rate of 3 °C min⁻¹, and then they were carefully rinsed by ethanol. It should be noted that one of the 13 samples was temporarily stored, for comparison.

Preparation of counter electrode

The FTO glass substrate was employed as the counter electrode and was prepared as follows. Firstly, the FTO glass substrate was ultrasonically cleaned in acetone, ethanol and deionized water for 10 min, respectively. Secondly, two tiny holes, which were used for injection of the electrolyte and kept at about 5 mm apart from each other, were drilled and perforated, with the help of an ultrasonic punching machine. Finally, the FTO counter electrodes were obtained by sputtering platinum for 8 min, under a DC sputtering instrument.

Fabrication of DSSCs

The as-prepared TNT/ZnO heterojunction films were cut into squares, 5 mm × 5 mm, using a nicking tool and they were then immersed in a solution of 5 × 10⁻⁴ M N719 in acetonitrile and tert-butyl alcohol, with the volume ratio 1 : 1,²¹ at room temperature for dye adsorption. Five hours later,²² that is, when the dye molecules were fully attached to the TNT/ZnO heterojunction, the dye-sensitized TNT/ZnO heterojunction films were sandwiched together by heating and pressing a 25 μm thick hot melt film for 1 min to obtain the DSSCs. The electrolyte was made up of 0.03 M iodine, 0.6 M 1-methyl-3-propylimidazolium iodide, 0.1 M guanidine thiocyanate, 0.5 M 4-tert-butylpyridine, acetonitrile and valeronitrile, with the volume ratio 17 : 3.²¹

Characterization

A field emission scanning electron microscopy (JSM-6700F, JEOL Inc., Japan) was used for the observation of the morphological properties of the TNTs and the TNT/ZnO heterojunction. X-ray diffraction (XRD) from a D/max 2400 X Series X-ray diffractometer, whose radiation source was Cu Kα, obtained at 40 kV and 100 mA, was used to study the phase structural properties of the as-prepared samples. A solar simulator (AM 1.5, 100 mW cm⁻², Newport) was used for the illumination of the as-fabricated DSSCs to record their current–voltage (I–V) curves. A Corrtest CS electrochemical workstation was used for the measurements of the electrochemical impedance spectroscopy (EIS).

Results and discussion

Fig. 3(a) shows the bottom view of the as-fabricated TNTs. It can be clearly seen that the diameter of the nanotubes is around 100 nm. The cross-sectional view of the TNTs is shown in Fig. 3(b), which shows the length of the TNTs is about 16 μm, and the inset of Fig. 3(b) shows the well-aligned nanotubes. Fig. 3(c) shows the front morphology of the TNT/ZnO heterojunction with an uneven surface, which was acquired by immersing the TNTs into a 0.25 M Zn(NO₃)₂ solution, while the inset of Fig. 3(c) in the white circle shows the pure TNTs, with a very smooth surface and the one in the black circle shows the TNT/ZnO heterojunction that is relatively clearer, where ZnO nanoparticles with a diameter of below 50 nm can be seen. These results indicate that the ZnO does exist in the formation of nanoparticles on the surface of the TNTs.

Fig. 3 SEM images of (a) bottom view of TNTs, (b) cross-sectional view of TNTs, and (c) front view of TNT/ZnO heterojunction by immersing into 0.25 M Zn(NO₃)₂ solution.

The XRD patterns of the pure TNTs and the TNT/ZnO heterojunction sintered at 500 °C are shown in Fig. 4. It can be clearly seen that both Fig. 4(a) and (b) show the peaks of anatase TiO₂ of (101), (004), (200), (211), (204), while the TNT/ZnO heterojunction shown in Fig. 4(b) has three peaks that belong to the wurtzite ZnO crystal, which includes (100), (002) and (101) that shown no obvious preferred orientation, but indicating that the TNT/ZnO heterojunction is formed.¹⁷ Moreover, the XRD studies also reveal that with the presence of the ZnO nanoparticles, the intensity of the diffraction peaks for the anatase TiO₂ are weakened relative to the pure TNTs'. The greatly weakened diffraction peak at (101) makes the one at (004) in intensity become very strong and the full width at half maximum (FWHM) of this main peak quite small, indicating that the TNTs are well crystallized.

Fig. 4 XRD patterns of (a) TNTs, and (b) TNT/ZnO heterojunction after annealing.

Fig. 5 shows the current–voltage (I–V) curves of the TNT-based DSSC and the TNT/ZnO heterojunction-based DSSCs fabricated from the solutions with different Zn(NO₃)₂ molar concentrations, which include 0.25 M, 0.50 M, 0.75 M, and 1.00 M, respectively. Obviously, all the TNT/ZnO heterojunction-based DSSCs show a better value of PCE. It can be seen that there is a significant decline of the short circuit current densities of the TNT/ZnO heterojunction-based DSSCs by a comparison between Fig. 5(a) and (b) and Fig. 5(c) and (d), indicating that a lower molar concentration of Zn(NO₃)₂ will lead to a better performance. In order to better understand these results, Fig. 6 is drawn based on the data in Table 1, that listed the photovoltaic parameters of the DSSCs under AM-1.5 illumination. We can see from Fig. 6 that most of the curves show an overall tendency of first increasing and then decreasing; only two curves of the open voltage are exception. For the DSSCs fabricated from the low molar concentration of Zn(NO₃)₂, as shown in Fig. 6(a), all four photovoltaic parameters achieve their best values when the immersion time is set at 2 h. However, when the molar concentration of Zn(NO₃)₂ gets higher, as shown in Fig. 6(b) and (c), it takes only 1 h for the immersion time to achieve the best values for most of the photovoltaic parameters of the DSSCs besides the open voltage, then these parameter values decrease with increase the immersion time. However, as seen in Fig. 6(d), a slight increase can be observed for the value of PCE when the immersion time is further increased from 1 h to 2 h. It is probably related to an increase of the molar concentration of Zn(NO₃)₂;the growth of the TNT/ZnO heterojunction becomes faster and non-uniform, thus resulting in an unfavorable influence. Even the TiO₂ nanotubes are blocked due to the ZnO nanoparticles, so as to prevent the adsorption of the dye molecules. It can be concluded, based on the above results and analysis that a lower molar concentration of Zn(NO₃)₂ is expected to be better for the fabrication of the TNT/ZnO heterojunction. Fig. 7 shows a comparison of the photovoltaic parameters among the TNT-based DSSC and four TNT/ZnO heterojunction-based DSSCs, which are the best PCE of the DSSCs fabricated from each molar concentration of Zn(NO₃)₂. However, the TNT/ZnO heterojunction-based DSSC fabricated from 0.25 M Zn(NO₃)₂ solution does not present the best performance parameter values in all the four factors, but its PCE is higher than those of the other four DSSCs, and this generally accords with the above-mentioned conclusions.

Fig. 5 I–V characteristics of TNT based DSSC and TNT/ZnO heterojunction-based DSSCs in (a) 0.25 M, (b) 0.50 M, (c) 0.75 M, and (d) 1.00 M Zn(NO₃)₂ solution.

Fig. 6 Changing trends of photovoltaic parameters between TNTs-based and TNT/ZnO heterojunction-based DSSCs in (a) 0.25 M, (b) 0.50 M, (c) 0.75 M, and (d) 1.00 M Zn(NO₃)₂ solution with the increase of immersing time.

Table 1 Photovoltaic parameters of DSSCs under AM-1.5 illumination

MC of Zn(NO3)2 (mol L^−1)	Immersing time (h)	Dye adsorption (h)	Voc (mV)	Jsc (mA cm^−2)	FF	PCE
—	—	5	718.94	5.65	48.33%	1.96%
0.25	1	5	739.01	7.71	51.83%	2.96%
0.25	2	5	759.63	8.20	63.86%	3.98%
0.25	3	5	756.29	7.04	60.08%	3.20%
0.50	1	5	777.05	8.67	57.44%	3.87%
0.50	2	5	754.70	8.34	56.81%	3.60%
0.50	3	5	816.20	7.67	47.77%	3.00%
0.75	1	5	784.54	5.89	67.68%	3.13%
0.75	2	5	768.21	5.86	63.98%	2.88%
0.75	3	5	798.73	4.61	66.59%	2.45%
1.00	1	5	760.60	5.86	64.33%	2.85%
1.00	2	5	790.63	6.32	62.20%	3.11%
1.00	3	5	783.15	6.16	59.02%	2.88%

---

Fig. 7 Comparison of photovoltaic parameters of four best DSSCs selected from 0.25 M, 0.50 M, 0.75 M and 1.00 M Zn(NO₃)₂ solution and the TNT-based DSSC

The analysis of the electrochemical impedance spectroscopy is an easy way to assess the electron properties of DSSCs.²³ Generally, we can observe three regions from a Nyquist plot, namely, a high frequency region, middle frequency region and low frequency region. According to the reports by Adachi,²⁴ Grätzel and Bisquert,²⁵ the arc in the high frequency range is caused by the impedance of the electron transfer at the Pt-coated counter electrode, the arc in the low frequency range results from the finite Warburg impedance²⁴ of the I⁻₃/I⁻-containing electrolyte, while the remaining largest semicircle in the middle region includes both the impedance of the electron transport in the TNTs and the impedance of the electron recombination with the dye molecules and the I⁻₃/I⁻-containing electrolyte, as depicted in Fig. 8. Thus, the value of the first semicircle to intersect with the x-axis at the left side, represents the impedance of the FTO glass.

Fig. 8 Influence analysis of arcs of a Nyquist plot.

Based on the above analysis, we can confirm that the largest semicircle, which corresponds to the middle frequency range, should be the most important part, since it contains some information directly related to the transferred TNTs and the TNT/ZnO heterojunction,²⁶ while the FTO, Pt-coated FTO and I⁻₃/I⁻-containing electrolyte are considered to be fixed factors. The equivalent circuit as shown in the inset of Fig. 9 and the following equations are used:²⁷

where τ (s) is the electron lifetime, R_t (Ω) is the impedance of the electron transport in the TNTs or the TNT/ZnO heterojunction, R_c (Ω) is the impedance of the electron recombination with the electrolyte, D_eff (cm² s⁻¹) is effective diffusion coefficient, L_n (μm) is the diffusion length and η_cc (%) is the electron collecting rate, k_eff (s⁻¹) is the effective electron recombination constant and L (μm) is the thickness of the film.

Fig. 9 EIS Nyquist plots of the DSSCs, with and without immersion into 0.25 M Zn(NO₃)₂ solution.

It should be mentioned that above parameters of only the TNT-based DSSC and the TNT/ZnO heterojunction-based DSSCs fabricated by immersion into a 0.25 M Zn(NO₃)₂ solution were calculated. Fig. 9 shows the EIS Nyquist plots of these DSSCs and Table 2 presents the corresponding calculated properties of the transport and recombination electrons. Through the fabrication of the TNT/ZnO heterojunction, we can observe an obvious promotion of the electron lifetime from 0.124 s to 0.476 s. Meanwhile, both the impedance of the electron transport (R_t) and the impedance of the electron recombination with the electrolyte (R_c) are optimized from 1.51 Ω and 6.01 Ω to 0.76 Ω and 15.90 Ω, respectively. The effective diffusion coefficient (D_eff) and diffusion length (L_n) of the electrons also improve and their values are as high as 2.56 × 10⁻⁴ cm² s⁻¹ and 74.3 μm, respectively. The variation of above these parameters leads to a higher electron collecting rate (η_cc), and the best one reaches 95.2%. The above results and analyses indicate that the TNT/ZnO heterojunction is helpful in preventing the recombination of the electrons with the electrolyte and the dye molecules.

Table 2 Properties of electrons calculated by the measurements of electrochemical impedance spectroscopy

MC of Zn(NO3)2 (mol L^−1)	Immersing time (h)	Keff (s^−1)	τ (s)	Rc (Ω)	Rt (Ω)	Deff (cm^2 s^−1)	Ln (μm)	ηcc (%)
0.25	0	7.03	0.142	6.01	1.51	7.43 × 10^−5	32.5	74.9
0.25	1	3.36	0.298	11.3	1.31	7.44 × 10^−5	47.9	88.4
0.25	2	2.10	0.476	15.9	0.764	1.12 × 10^−4	74.3	95.2
0.25	3	6.82	0.147	8.28	0.565	2.56 × 10^−4	62.3	93.2

---

Conclusions

Highly ordered TNT-based TNT/ZnO heterojunctions have been successfully prepared by using a HF-free and easily-operated method. The results indicate that the resultant DSSCs based on the TNT/ZnO heterojunction can achieve a PCE of 3.98%, which is much higher than that of the original pure TNT-based DSSC (1.96%). By analyzing the EIS plots and calculating the impedance model of Adachi et al., it is also clearly indicated that the DSSCs based on the TNT/ZnO heterojunction obtained from the Zn(NO₃)₂ molar concentration of 0.25 M have better properties, which include an electron lifetime (τ) of 0.476 s, impedance of electron transport (R_t) as low as 0.764 Ω, impedance of electron recombination with the electrolyte (R_c) as high as 15.9 Ω, diffusion length (L_n) as long as 74.3 μm, and the electron collecting rate (η_cc) improved from 74.9% to 95.2%.

Acknowledgements

This work was supported by the Research Fund for the Doctoral Program of Higher Education of China under grant no. 20120201130004, the National Natural Science Foundation of China under Grant no. 61078058, and the Science, Technology Developing Project of Shaanxi Province (2012KW-11) and the Fundamental Research Funds for the Central Universities. The SEM work was done at the International Center for Dielectric Research (ICDR), Xi'an Jiaotong University, Xi'an, China; the authors also thank Ms. Dai for her help in using the SEM.

Notes and references

1. B. O'Reagan and M. Grätzel, Nature, 1991, 353, 737.
2. D. W. Gong, A. Grimes and K. Varghese, J. Mater. Res., 2001, 16, 3331.
3. E. V. A. Premalal, N. Dematage and G. R. A. Kumara, Electrochim. Acta, 2012, 63, 375; M. Paulose, H. E. Prakasam and O. K. Varghese, J. Phys. Chem. C, 2007, 111, 14992; Y. K. Li, H. G. Yu and C. K. Zhang, Electrochim. Acta, 2013, 107, 313; K. Lee and P. Schmuki, Electrochim. Acta, 2013, 100, 229.
4. M. D. Ye, X. K. Xin, C. J. Lin and Z. Q. Lin, Nano Lett., 2011, 8, 3214.
5. O. K. Varghese, M. Paulose and A. Grimes, Nat. Nanotechnol., 2009, 4, 592.
6. H. Park, W. R. Kim and H. T. Jeong, Sol. Energy Mater. Sol. Cells, 2011, 95, 184.
7. G. D. Yang, Z. F. Yan and T. C. Xiao, Appl. Surf. Sci., 2012, 258, 8704; Y. Bessekhouad, R. Brahimi and F. Hamdini, J. Photochem. Photobiol., A, 2012, 248, 15.
8. Y. L. Xie, Z. X. Li and Z. G. Xu, Electrochem. Commun., 2011, 13, 788.
9. R. Liu, W. D. Yang and L. S. Qiang, J. Power Sources, 2012, 220, 153.
10. X. Y. Wang, L. D. Sun and S. Zhang, RSC Adv., 2013, 3, 11001; J. Wang and Z. Q. Lin, Chem. Mater., 2010, 22, 579.
11. G. K. Mor, K. Shankar and M. Paulose, Nano Lett., 2006, 6, 215; X. L. Liu, J. Lin and X. F. Chen, RSC Adv., 2013, 3, 4885; K. Kant and D. Losic, Phys. Status Solidi RRL, 2009, 3, 139; Y. Y. Kuo and C. H. Chien, Electrochim. Acta, 2013, 91, 337.
12. H. Pan, N. Misra and S. H. Ko, Appl. Phys. A, 2009, 94, 111; S. H. Ko, D. Lee and H. W. Kang, Nano Lett., 2011, 11, 666.
13. F. X. Xiao, ACS Appl. Mater. Interfaces, 2012, 4, 7055; R. Liu, W. D. Yang and L. S. Qiang, J. Power Sources, 2012, 220, 153.
14. C. S. Chou, F. C. Chou and Y. G. Ding, Sol. Energy, 2012, 86, 1435; S. S. Kim, S. I. Na and Y. C. Nah, Electrochim. Acta, 2011, 58, 503.
15. C. M. Firdaus, S. Rizam and M. Rusop, Procedia Eng., 2012, 41, 1367; S. S. Kanmani and K. Ramachandran, Renewable Energy, 2012, 43, 149.
16. Z. L. He, W. X. Que and Y. C. He, Ceram. Int., 2013, 39, 5545; Z. L. Zhu, Z. L. He, J. Q. Li, J. Q. Zhou, N. Wei and D. G. Liu, J. Mater. Sci., 2011, 46, 931; Z. L. He and W. X. Que, J. Nanoeng. Nanomanuf., 2012, 2, 17; Z. L. He and W. X. Que, Phys. Chem. Chem. Phys., 2013, 15, 16768.
17. Z. He, W. Que, J. Chen, X. Yin, Y. He and J. Ren, ACS Appl. Mater. Interfaces, 2012, 4, 6816; Z. He, Z. Zhu, J. Li, N. Wei and J. Zhou, J. Hazard. Mater., 2011, 190, 133; Z. He, W. Que and Y. He, Mater. Lett., 2013, 94, 136; Z. L. He, W. X. Que, H. X. Xie, J. Chen, Y. Yuan and P. Sun, J. Am. Ceram. Soc., 2012, 95, 3941; Z. L. He, W. X. Que, J. Chen, Y. C. He and G. F. Wang, J. Phys. Chem. Solids, 2013, 74, 924.
18. H. E. Prakasam, K. Shankar and M. Paulose, J. Phys. Chem. C, 2007, 111, 7235.
19. X. Y. Wang, S. Zhang and L. D. Sun, Thin Solid Films, 2011, 519, 4694.
20. C. H. Chao, J. S. Huang and C. F. Lin, J. Phys. Chem. C, 2009, 113, 512.
21. J. H. Park, T. W. Lee and M. G. Kang, Chem. Commun., 2008, 2867.
22. H. M. Cheng, W. H. Chiu and C. H. Lee, J. Phys. Chem. C, 2008, 112, 16359.
23. Y. P. Zhang, H. B. Li and L. K. Pan, J. Electroanal. Chem., 2009, 634, 68.
24. M. Adachi, M. Sakamoto and J. Jiu, J. Phys. Chem. B, 2006, 110, 13872.
25. Q. Wang, S. Ito and M. Grätzel, J. Phys. Chem. B, 2006, 110, 25210; J. Bisquert, G. B. Germà and F. S. Francisco, J. Electroanal. Chem., 1999, 475, 152.
26. J. Bisquert, J. Phys. Chem. B, 2002, 106, 325.
27. P. Zhong, W. X. Que, J. Zhang, Q. Y. Jia, W. J. Wang, Y. L. Liao and X. Hu, J. Alloys Compd., 2011, 509, 7808.

---