One-step anion-assisted electrodeposition of ZnO nanofibrous networks as photoanodes for dye-sensitized solar cells

Abstract

A highly efficient ZnO photoanode consisting of three-dimensional (3D) nanofibrous-like networks (NFs) for dye-sensitized solar cells (DSSCs) was synthesized using a one-step and seed-layer-free electrodeposition on an indium-tin-oxide (ITO) substrate at 75 °C in a solution containing Zn(NO₃)₂, KCl, NaCH₃COO, and Na₃C₆H₅O₇. In this solution, KCl as the supporting electrolyte promotes the reduction of NO₃⁻ and the diffusion of Zn²⁺, whereas CH₃COO⁻ and C₆H₅O₇³⁻ anions act as the capping agents to selectively inhibit ZnO growth along the c-axis. The photoelectrochemical results reveal that the DSSC based on ZnO NFs has the highest power conversion efficiency (3.78%) in comparison with those of DSSCs based on nanosheets (1.36%), nanorods (2.18%), and microplates (2.55%). This can be attributed to the large dye adsorption amount, efficient light scattering and direct electron transfer networks, which lead to a significant improvement in solar cell performance. Therefore, the ZnO NFs structure can be considered as a promising and efficient photoanode for DSSCs.

---

Introduction

Recently, ZnO-based nanomaterials, important II–VI semiconductors with a wide band gap of 3.37 eV and a large exciton binding energy of 60 meV,¹ have attracted considerable interest due to their range of potential applications such as dye-sensitized solar cells (DSSCs),² light-emitting diodes,³ sensors,⁴ and nanogenerators.⁵ In the early reports on ZnO-based DSSCs, ZnO nanoparticles were often used as the photoanode.⁶ Photoexcited electrons diffuse through the nanoparticles to the conductive substrate via a series of interparticle hopping steps.⁷ However, excess of electron hopping through the interparticle barriers results in a long dwell time within the individual particles and thus increases the probability of charge recombination between the injected electrons and oxidized dye or redox electrolyte.⁸ Recently, 1D ZnO nanomaterials, such as nanorods and nanotubes, with direct electric pathways have been particularly attractive as photoanodes in DSSCs due to their high electron mobility and increased electron diffusion length, which reduces the number of interparticle hops; moreover, these nanostructures can be easily tailored via a mild wet-chemical method as compared to conventional TiO₂.^9–11 However, the insufficient internal surface area of these 1D nanomaterials inhibits any significant improvement in their photovoltaic efficiency due to their weak capability for dye loading and light harvesting.⁹ To overcome these problems, multi-scale hierarchical structures have been used as photoanodes, including 3D nanofiber networks,¹² nanorods/nanosheets,¹³ nanoforests,¹⁴ nanowires/nanoparticles,¹⁵ and microspheres/nanowires.¹⁶ Among these, 3D nanofiber networks are particularly attractive as photoanodes due to their highly conductive network for efficient electron transport, high roughness factor for dye loading, and large pore size for enhanced light harvesting through light scattering.¹⁷

Several methods have been previously used for the preparation of ZnO nanofibers as photoanodes for use in DSSCs, including electrospinning,¹² thermal evaporation,¹⁸ and magnetron sputtering.¹⁹ However, to best of our knowledge, there is no study reporting the electrodeposition of ZnO nanofibers or fibrous-like nanostructures as photoanodes in DSSCs. Electrodeposition is a simple and effective method to synthesize nanomaterials because of its economic nature, ease of scale-up, fine selectivity, ability to work at low temperatures, and environmental friendliness.^20–24

In this study, we report a one-step anion-assisted electrodeposition of 3D ZnO nanofibrous-like networks (NFs) in the presence of KCl, NaCH₃COO, and Na₃C₆H₅O₇. The deposition potential was selected using cyclic voltammetry experiments. The structural and morphological characteristics of the products are presented and discussed. The photovoltaic properties of the DSSC based on the NFs photoanode are discussed on the basis of incident photon-to-current efficiency (IPCE), photocurrent–voltage (J–V), and electrochemical impedance spectroscopy (EIS), and compared to other DSSCs based on multi-scale ZnO structures such as nanorods, nanosheets and microplates. The enhanced conversion efficiency of the DSSC based on NFs was 3.78%, which is a large improvement compared with those found for other DSSCs.

Experimental

Materials synthesis

The electrodeposition of the ZnO samples and cyclic voltammetry studies were performed in an electrochemical cell immersed in a water bath held at 75 °C using a PARSTAT 2273 model potentiostat/galvanostat (Princeton Applied Research). The electrochemical cell consisted of an indium-tin-oxide (ITO, 8–12 Ω cm⁻²) coated glass working electrode, a zinc plate counter electrode,²⁵ and an Ag/AgCl reference electrode (Bioanalytical Systems). All the solutions used in this study were prepared using deionized water (resistivity ≥ 18 MΩ cm). Prior to each experiment, the solutions were purged with purified N₂ gas. The solutions were moderately stirred using a magnetic stirrer during both the electrodeposition and cyclic voltammetry studies. The deposition of the ZnO samples was performed in a solution containing 15 mM Zn(NO₃)₂·4H₂O (Merck) and various concentrations of KCl (Merck), NaCH₃COO (Merck), and Na₃C₆H₅O₇·2H₂O (Merck) at pH = ∼6.0. All the samples were synthesized at a constant potential of −0.950 V. To obtain the same thickness for all the samples, the samples were deposited using different deposition times from solutions 1–4 (Table 1). After deposition, the sample-coated ITO substrates were thoroughly rinsed with deionized water and dried under a stream of N₂ gas. Finally, the samples were annealed at 350 °C for 30 min through a rapid annealing process prior to their assembly. The active areas of the photoanodes were set as 0.16 cm² by scraping-off the excess area.

Table 1 Experimental conditions and thickness of the ZnO samples

Solution	Composition of solutions (mM)				Deposition time (h)	Thickness (μm)
	Zn(NO3)2	KCl	CH3COO^−	C6H5O7^3−
1	15	100	—	—	3.0	3.2
2	15	—	1	—	3.5	3.0
3	15	—	—	1	3.5	2.9
4	15	100	1	1	3.0	3.1

---

Assembly of DSSCs

The resulting ZnO samples were sensitized by immersing them in 0.3 mM N719 dye (Ruthenizer 535-bisTBA, Solaronix) in an acetonitrile/tert-butyl alcohol (v/v = 1 : 1) solution for 30 min at room temperature, then rinsed with absolute ethanol to remove the excess dye and dried in air. The counter electrode was prepared by spreading a H₂PtCl₆ solution (Platisol T, Solaronix) on an FTO (fluorine-doped tin oxide, 13 Ω cm⁻²) coated glass substrate with heating at 450 °C for 15 min. The counter electrode was then placed directly on the top of the dye-sensitized ZnO electrode sealed with a polymer foil (Meltonix 1170-25, Solaronix) as a spacer frame, and then thermally treated at 100 °C for 5 min to bind the two electrodes. I⁻/I₃⁻ electrolyte (Iodolyte AN-50, Solaronix) was injected from one of the two pre-drilled holes on the counter electrode into the space between the sandwiched cells. Then, the holes were sealed using parafilm at an elevated temperature.

Characterization

The prepared samples were subjected to SEM analysis using a JEOL JSM-6060LV to investigate the surface morphology. The XRD patterns for the samples were recorded on a Rigaku Advance Powder X-ray diffractometer (λ = 1.54050 Å) in the span of angle between 20° and 70°. The average thickness of the samples was measured using a profiler (KLA Tencor P6). The amount of dye loaded onto the samples was determined by desorbing the dye from the sample surface (1 × 2 cm²) into 0.1 M NaOH aqueous solution and measuring its absorption spectrum by a UV-vis spectrophotometer (UV-2401, Shimadzu). The diffused reflectance spectra of the ZnO samples were obtained using a UV-vis spectrophotometer (UV-2600, Shimadzu). The J–V curves of the DSSCs were recorded using the above-mentioned potentiostat/galvanostat under illumination of simulated sunlight (100 mW cm⁻²) provided by a solar simulator (Model 96000, Newport) equipped with an AM 1.5G filter. A 150 W xenon lamp (Oriel) was used as the light source. The intensity of the incident light was calibrated using a standard Si detector (Model 918D-SL-OD3, Newport) and an optical power meter (Model 1917-R, Newport). The fill factor (FF) and the power conversion efficiency (η) of the DSSCs were calculated according to the following equations.²⁶J_sc is the short-circuit current density (mA cm⁻²), V_oc is the open-circuit voltage (V), P_in is the incident light intensity (mW cm⁻²) and J_max (mA cm⁻²) and V_max (V) are the current density and voltage at the point of maximum power output in the J–V curves, respectively. The electrochemical impedance spectroscopy (EIS) measurements of the DSSCs were performed using the potentiostat/galvanostat by applying bias of open-circuit voltage (V_oc) under a constant light illumination of 100 mW cm⁻² and the curves were recorded over a frequency range of 10⁻¹ to 10⁵ Hz with an AC amplitude of 10 mV. The incident photon-to-current efficiency (IPCE) curves for the DSSCs were obtained using the potentiostat/galvanostat and solar simulator equipped with a monochromator (Model 74004, Oriel). The IPCE was calculated using the following equation.²⁷J_sc is the short-circuit current density (mA cm⁻²), λ is the monochromatic incident light wavelength (nm) and P_in is the incident light intensity at that wavelength (mW cm⁻²).

Results and discussion

Cyclic voltammetry

To determine the electrodeposition potential, cyclic voltammetry experiments were performed on ITO substrates. The cyclic voltammogram (CV) for the solution containing only 15 mM Zn(NO₃)₂ is shown in Fig. 1a. A relatively broad cathodic peak is observed between −0.500 and −1.250 V, where the ZnO electrodeposition is expected to occur by the following reactions.²⁸

Zn(NO₃)₂ ↔ Zn²⁺ + 2NO₃⁻ (4)

NO₃⁻ + H₂O + 2e⁻ → NO₂⁻ + 2OH⁻ (5)

Zn²⁺ + 2OH⁻ → Zn(OH)₂ (6)

Zn(OH)₂ → ZnO + H₂O (7)

Fig. 1 Cyclic voltammograms of 15 mM Zn(NO₃)₂ (a), 15 mM Zn(NO₃)₂ + 0.1 M KCl (b), 15 mM Zn(NO₃)₂ + 1 mM Na₃C₆H₅O₇ (c), and 15 mM Zn(NO₃)₂ + 1 mM NaCH₃COO (d).

The steep peak observed from −1.250 to −1.500 V is attributed to the metallic deposition of Zn and H₂ evolution reaction. In the reverse scan, the anodic current is negligible, indicating the high stability of the deposited ZnO.²⁹ The CV for the solution containing a mixture of 15 mM Zn(NO₃)₂ and 0.1 M KCl is shown in Fig. 1b. It can be seen that the reduction of NO₃⁻ and the onset of metallic Zn deposition are slightly shifted to more positive potentials as compared with that of the solution of Zn(NO₃)₂. This could be due to the fact that KCl increases the electrolyte conductivity and accelerates the diffusion of Zn²⁺ and reduction of NO₃⁻.³⁰ Furthermore, in contrast to the behavior observed for the solution of Zn(NO₃)₂, an anodic peak appeared at around −0.700 V in the reverse scan. Similar anodic peaks were observed for the solutions containing 15 mM Zn(NO₃)₂ + 1 mM Na₃C₆H₅O₇ and 15 mM Zn(NO₃)₂ + 1 mM NaCH₃COO, as shown in Fig. 1c and d, respectively. These anodic peaks appear to result from the oxidation of Cl⁻, CH₃COO⁻ and C₆H₅O₇³⁻ anions adsorbed on the ZnO surface. The oxidation processes may be represented as follows:^31–33

Cl⁻ → Cl + e⁻ (8)

CH₃COO⁻ → CH₃COO + e⁻ (9)

C₆H₈O₇ ↔ C₆H₅O₇³⁻ + 3H⁺ (10)

In the last process, the subsequent step involves the oxidation of the adsorbed C₆H₅O₇³⁻ ion.³³ Based on the above mentioned results, if the deposition is carried out in a solution containing 15 mM Zn(NO₃)₂ + 0.1 M KCl + 1 mM NaCH₃COO + 1 mM Na₃C₆H₅O₇ over a potential range from −0.400 to −1.05 V, ZnO will be formed on the ITO surface. For Cl⁻ adsorption, the CVs were also obtained at concentrations lower than 0.1 M KCl but a noticeable anodic peak was not observed (not shown here). When the concentrations of the anions were considered, Cl⁻ anion showed the lowest adsorption on ZnO. Pradhan and Leung carried out a study on the electrodeposition of ZnO films in the presence of KCl solutions.³⁴ They inferred that the slow hydroxylation reaction (eqn (6)) allows sufficient time for ion exchange, hence hinders any significant surface capping by Cl⁻ anions in the solution containing a mixture of 0.1 M KCl and 10 mM Zn(NO₃)₂. However, at a sufficiently high concentration of KCl (>0.1 M), the capping effect of Cl⁻ will occur.

Morphological and structural characterization

The influence of KCl on the morphology of the deposited ZnO samples was investigated. Fig. 2a and b show the SEM images of the ZnO samples deposited at −0.950 V for 3 h in the absence and presence of 0.1 M KCl (solution 1), respectively. In the Zn(NO₃)₂ solution with KCl, the nanorod arrays (NRs) along the c-axis direction and hexagonal planes were obtained, whereas spike-like nanorods are observed in the solution containing only Zn(NO₃)₂. This obviously confirms the role of Cl⁻ adsorption on the positive polar (0001) crystal plane, causing the formation of a hexagon on the top of the ZnO nanorods. However, an apparent change in the morphology of the samples is not found in Fig. 2a and b, revealing that Cl⁻ has no effect on the growth along the c-axis when the concentration of Cl⁻ is 0.1 M. The weak adsorption of Cl⁻ on ZnO is in agreement with the literature.³⁴ However, it was found that the thickness of the deposited ZnO sample (2.6 μm) in the absence of KCl was thinner than that deposited in the presence of KCl (Table 1). This may be due to the fact that KCl increases solution conductivity and accelerates the diffusion of Zn²⁺ and reduction of NO₃⁻. Thus, it was decided that the solution with KCl (solution 1) would be more appropriate for the deposition of nanorods.

Fig. 2 SEM images of ZnO nanorods (a, b), microplates (c), nanosheets (d), and nanofibrous-like networks (e, f).

To define the effect of CH₃COO⁻ and C₆H₅O₇³⁻ on the shape and size of the ZnO samples, depositions were performed from solutions 2 and 3, respectively (Fig. 1c and d). In the solution with CH₃COO⁻, vertically aligned porous microplate-like structures comprising of nanoparticles (MPs) were obtained, whereas nanosheet-like networks (NSs) were observed in the solution containing CH₃COO⁻ and C₆H₅O₇³⁻. Similar structures have been observed for the electrodeposition of ZnO in the presence of CH₃COO⁻ or C₆H₅O₇³⁻.^35,36 This clearly confirms the strong adsorption of CH₃COO⁻ and C₆H₅O₇³⁻ on the (0001) plane, causing the formation of plate or sheet structures. The strong adsorption behavior is also verified by the CV results. On the other hand, a clear difference in the morphology of the samples is found in Fig. 2c and d. The difference may be related to the adsorption mechanism of C₆H₅O₇³⁻, which can form the [Zn(C₆H₅O₇)₄]¹⁰⁻ complex with Zn²⁺.³⁶ Thus, this complex is preferred over the (0001) plane, which decreases the growth rate of the (0001) plane or the corresponding (002) peak; this leads to the formation of ZnO NSs with the contribution of CH₃COO⁻.

The SEM images of the ZnO sample obtained from the solution containing KCl, CH₃COO⁻ and C₆H₅O₇³⁻ (solution 4) at different magnifications are shown in Fig. 2e and f and S1.† It can be clearly seen that the morphology evolves from nanosheet networks to 3D dense porous nanofibrous-like networks (NFs). These results indicate that less conducting electrolytes give rise to less dense nanosheets, while more conducting electrolytes produce more dense nanofibrous-like networks. Briefly, highly conducting media enhances the diffusion of Zn²⁺ and reduction of NO₃⁻, resulting in the growth of dense nanofibrous-like networks. The effect of the electrolyte on ZnO is in agreement with previously reported studies.^30,37

Fig. 3 shows the corresponding XRD patterns of ZnO NRs, MPs, NSs, and NFs deposited on ITO substrates from solutions 1–4, respectively. All the presented diffraction peaks, except for those of ITO, can be indexed to the hexagonal wurtzite ZnO structure (JCPDS 01-089-0511). It can be seen that the peak intensities vary with the morphology of the ZnO samples. For ZnO NRs, the relative intensity of the (002) peak is increased significantly, with the (100) and (101) peak intensities reduced, which is as expected from their c-axis growth. However, in Fig. 3b–d, the relative intensity of the (002) peak is found to be drastically decreased, suggesting the transformation from NRs to NFs, as shown in Fig. 2. This revealed that the CH₃COO⁻ and C₆H₅O₇³⁻ ions adsorb on positive polar (0001) plane and suppress ZnO growth along the c-axis or the intensity of corresponding (002) peak. Therefore, the resulting NFs have the preferred growth direction of the (100) peak or non-polar (101̄0) plane. Similar results have been reported for ZnO nanostructures.^34,38

Fig. 3 XRD patterns of NRs, MPs, NSs, and NFs. The diffraction peaks marked with solid stars correspond to the ITO substrates.

Performance of the DSSCs

To determine the effect of the morphologies of the ZnO samples on the performance of the DSSCs, we fabricated cells with NRs, MPs, NSs and NFs as photoanodes. Fig. 4 compares the typical J–V curves of the DSSCs based on the photoanodes, and the corresponding photoelectrochemical parameters are listed in Table 2. For all the samples, the V_oc was found to be ca. 0.55 V, indicating that there was no significant effect of morphology on the V_oc of the DSSCs used in this study. This was attributed to the fact that V_oc was proportional to the energy difference between the Fermi level of the semiconductor and the redox potential of the electrolyte.³⁹ However, the value of the J_sc increased from 8.12 to 14.42 mA cm⁻² with the morphology of photoanode changing from the NSs to the NFs. Consequently, the power conversion efficiency remarkably improved from 1.36% to 3.78%.

Fig. 4 J–V curves of DSSCs based on NFs, NSs, MPs, and NRs.

Table 2 The photoelectrochemical parameters of DSSCs with different ZnO samples

Samples	Adsorbed dye (nmol cm^−2)	Jsc (mA cm^−2)	Voc (V)	FF	η (%)	R2 (Ω)	τe (ms)
NRs	48.9	10.7	0.55	0.37	2.18	35.4	4.7
MPs	61.0	11.9	0.55	0.39	2.55	27.4	6.6
NSs	37.5	8.1	0.54	0.31	1.36	41.6	3.3
NFs	80.2	14.4	0.57	0.46	3.78	16.3	8.2

---

The IPCE curves give more detailed information on the photoelectrochemical performance of the ZnO samples, as shown in Fig. 5. The maximum IPCE values of the four samples with N719 dye occur at approximately 520 nm. All the integrated IPCE values are in good agreement with the measured J_sc shown in Fig. 4. Compared with the other samples, the ZnO NFs have a higher IPCE in the wavelength range from 400 to 750 nm. The enhanced IPCE for the NFs at shorter wavelengths (400–600 nm) can be attributed to high dye adsorption, while the improved IPCE at the longer wavelengths (600–750 nm) can be ascribed to efficient light scattering. As a result, the highest IPCE value of the DSSC based on NFs indicates the maximized use of solar light, which further enhances the light harvesting efficiency and improves the J_sc.⁴⁰

Fig. 5 IPCE curves of DSSCs with NFs, NSs, MPs, and NRs.

To verify the high dye adsorption, the amount of adsorbed dye on the ZnO samples was calculated from the UV-vis spectra using the Lambert–Beer's law as follows (Fig. 6 and Table 2):

A = εlc (11)

where A is the absorbance of the UV-vis spectra at 515 nm, ε = 14 100 M⁻¹ cm⁻¹ is the molar extinction co-efficient of the dye at 515 nm, l is the path length of the light beam, and c is the dye concentration.⁴¹ As expected, the NFs sample shows the highest dye adsorption due to its relatively high specific area, which is in good agreement with the IPCE results.

Fig. 6 Optical absorption spectra of the solutions containing N719 desorbed from the sensitized ZnO photoanodes composed of different structures.

To further understand light scattering ability, the diffused reflectance spectra of the ZnO samples with different morphologies were obtained (Fig. 7). The reflectance spectrum of the NFs was about 68% in the range of 300–800 nm, which was higher than that found for the other samples, indicating that the incident light was efficiently scattered within the NFs, which leads to a higher photocurrent in the DSSC based on NFs. The higher light scattering ability might be caused by the repeated light reflection in the 3D porous structure.

Fig. 7 Diffused reflectance spectra of the different ZnO structures.

To better understand the effects of morphology on the electron transfer properties of the DSSCs, EIS was used, and the Nyquist and Bode plots are shown in Fig. 8a and b, respectively. The internal impedances were determined by fitting the experimental data with an equivalent circuit (inset of Fig. 8a). As shown, there are two semicircles on the Nyquist plots, of which the one in the low frequency region shows a bigger radius than the other one in the higher frequency region. The large semicircles are assigned to the charge transfer process occurring at the ZnO/dye/electrolyte interface (R₂) and the small semicircles correspond to the resistance of the Pt/redox (I⁻/I₃⁻) interface charge transfer (R₁). R_s is the series resistance, including the sheet resistance of the ITO glass and the contact resistance of the cell, whose semicircle cannot be seen.⁴² It can be observed that the second semicircle of the DSSC based on NFs is smaller than that found for the DSSCs based on other samples. Therefore, the resistance of electron transfer (R₂) is lowest in the DSSC based on NFs (Table 2). The smaller electron transfer impedance implies faster electron transfer, which can hinder the electron recombination and result in higher photocurrent and efficiency. This mainly originates from the direct electron transfer pathway and the tight connection of fibrous-like structures in the ZnO NFs.

Fig. 8 Nyquist (a) and Bode plots (b) of the DSSCs based on NSs, NRs, MPs, and NFs. The inset in panel a shows the equivalent circuit.

This result is also supported by the corresponding Bode plots (Fig. 8b), which display the characteristic frequency peaks for the charge transfer process of DSSCs. Based on the maximum frequency (f_max) of the peak at an intermediate frequency, the electron lifetime (τ_e) for these four DSSCs can be calculated using the following formula.⁴³

As can be seen from Table 2, the DSSC based on NFs shows the longest electron lifetime among the DSSCs tested. Therefore, the enhanced electron lifetime suggests that the photogenerated electrons can diffuse further without any interruption.⁴⁴ Consequently, the lower resistance and the longer electron lifetime favor a higher charge collection rate of produced electrons, which improves the efficiency of the DSSC.

Conclusions

In summary, we have developed a one-step anion-assisted electrodeposition method for fabricating 3D ZnO nanofibrous-like networks as photoanodes for use in DSSCs. Detailed morphological and structural investigations show that the addition of KCl, NaCH₃COO and Na₃C₆H₅O₇ plays a key role in driving the growth of the nanofibrous-like networks. Compared with microplates, nanorods and nanosheets, the 3D nanofibrous-like networks show the highest efficiency of 3.78%, with a high J_sc of 14.4 mA cm⁻², a FF of 0.46, a τ_e of 8.2 ms and a low R₂ of 16.3 Ω, which can be mainly ascribed to the higher charge transfer rate, sufficient dye adsorption and better light scattering capacity. This simple method for the fabrication of the ZnO nanofibrous-like networks is promising for the development of low-cost and eco-friendly devices with high power conversion efficiency.

Acknowledgements

The authors are grateful for financial support provided by the Commission of Science Research of Sakarya University (No. 2014-50-01-011).

References

1. M. H. Huang, S. Mao, H. Feick, H. Q. Yan, Y. Y. Wu, H. Kind, E. Weber, R. Russo and P. D. Yang, Science, 2001, 292, 1897–1899.
2. S. B. Zhu, L. M. Shan, X. N. Chen, L. He, J. J. Chen, M. Jiang, X. L. Xie and Z. W. Zhou, RSC Adv., 2013, 3, 2910–2916.
3. J. M. Bao, M. A. Zimmler, F. Capasso, X. W. Wang and Z. F. Ren, Nano Lett., 2006, 6, 1719–1722.
4. C. W. Na, H. S. Woo and J. H. Lee, RSC Adv., 2012, 2, 414–417.
5. Z. L. Wang and J. Song, Science, 2006, 312, 242–246.
6. K. Keis, E. Magnusson, H. Lindstrom, S. E. Lindquist and A. Hagfeldt, Sol. Energy Mater. Sol. Cells, 2002, 73, 51–58.
7. J. B. Baxter, A. M. Walker, K. van Ommering and E. S. Aydil, Nanotechnology, 2006, 17, S304–S312.
8. F. Xu, M. Dai, Y. N. Lu and L. T. Sun, J. Phys. Chem. C, 2010, 114, 2776–2782.
9. M. Law, L. E. Greene, J. C. Johnson, R. Saykally and P. D. Yang, Nat. Mater., 2005, 4, 455–459.
10. Q. F. Zhang, C. S. Dandeneau, X. Y. Zhou and G. Z. Cao, Adv. Mater., 2009, 21, 4087–4108.
11. A. I. Hochbaum and P. D. Yang, Chem. Rev., 2010, 110, 527–546.
12. I. D. Kim, J. M. Hong, B. H. Lee, D. Y. Kim, E. K. Jeon, D. K. Choi and D. J. Yang, Appl. Phys. Lett., 2007, 91, 163109.
13. J. H. Qiu, M. Guo and X. D. Wang, ACS Appl. Mater. Interfaces, 2011, 5, 2358–2367.
14. S. H. Ko, D. Lee, H. W. Kang, K. H. Nam, J. Y. Yeo, S. J. Hong, C. P. Grigoropoulos and H. J. Sung, Nano Lett., 2011, 11, 666–671.
15. L. Y. Chen and Y. T. Yin, RSC Adv., 2013, 3, 8480–8488.
16. X. H. Kang, C. Y. Jia, Z. Q. Wan, J. Zhuang and J. Feng, RSC Adv., 2015, 5, 16678–16683.
17. N. Tetreault, E. Horvath, T. Moehl, J. Brillet, R. Smajda, S. Bungener, N. Cai, P. Wang, S. M. Zakeeruddin, L. Forro, A. Magrez and M. Gratzel, ACS Nano, 2010, 4, 7644–7650.
18. S. Li, X. Z. Zhang, X. J. Jiao and H. Lin, Mater. Lett., 2011, 65, 2975–2978.
19. O. Lupan, V. M. Guerin, L. Ghimpu, I. M. Tiginyanu and T. Pauporte, Chem. Phys. Lett., 2012, 550, 125–129.
20. G. H. A. Therese and P. V. Kamath, Chem. Mater., 2000, 12, 1195–1204.
21. M. Biçer, H. Köse and İ. Şişman, J. Phys. Chem. C, 2010, 114, 8256–8263.
22. İ. Şişman and H. Öz, Electrochim. Acta, 2011, 56, 4889–4894.
23. M. Biçer and İ. Şişman, Langmuir, 2012, 28, 15736–15742.
24. M. Arpacık, M. Biçer and İ. Şişman, J. Electrochem. Soc., 2013, 160, A2250–A2257.
25. M. Zi, M. Zhu, L. Chen, H. M. Wei, X. P. Yang and B. Q. Cao, Ceram. Int., 2014, 40, 7965–7970.
26. G. T. Yue, F. R. Tan, F. M. Li, C. Chen, W. F. Zhang, J. H. Wu and Q. H. Li, Electrochim. Acta, 2014, 149, 117–125.
27. Y. G. Lin, Y. K. Hsu, Y. C. Chen, S. B. Wang, J. T. Miller, L. C. Chen and K. H. Chen, Energy Environ. Sci., 2012, 5, 8917–8922.
28. J. S. Wellings, N. B. Chaure, S. N. Heavens and I. M. Dhannadasa, Thin Solid Films, 2008, 516, 3893–3898.
29. T. Mahalingam, V. S. John, M. Raja, Y. K. Su and P. J. Sebastian, Sol. Energy Mater. Sol. Cells, 2005, 88, 227–235.
30. S. J. Jiao, K. J. Zhang, S. S. Bai, H. L. Li, S. Y. Gao, H. T. Li, J. Z. Wang, Q. J. Yu, F. Y. Guo and L. C. Zhao, Electrochim. Acta, 2013, 111, 64–70.
31. J. Elias, R. Tena-Zaera and C. Levy-Clement, J. Phys. Chem. C, 2008, 112, 5736–5741.
32. T. Fukuda and A. Aramata, J. Electroanal. Chem., 1999, 467, 112–120.
33. O. I. Gonzalez-Pena, T. W. Chapman, Y. M. Vong and R. Antano-Lopez, Electrochim. Acta, 2008, 53, 5549–5554.
34. D. Pradhan and K. T. Leung, Langmuir, 2008, 24, 9707–9716.
35. C. W. Kung, H. W. Chen, C. Y. Lin, Y. H. Lai, R. Vittal and K. C. Ho, Prog. Photovoltaics, 2014, 22, 440–451.
36. L. Ding, R. X. Zhang and L. Z. Fan, Nanoscale Res. Lett., 2013, 8, 78.
37. M. Abd-Ellah, N. Moghimi, L. Zhang, N. F. Heinig, L. Y. Zhao, J. P. Thomas and K. T. Leung, J. Phys. Chem. C, 2013, 117, 6794–6799.
38. X. Bai, L. Yi, D. L. Liu, E. Y. Nie, C. L. Sun, H. H. Feng, J. J. Xu, Y. Jin, Z. F. Jiao and X. S. Sun, Appl. Surf. Sci., 2011, 257, 10317–10321.
39. X. J. Feng, K. Shankar, M. Paulose and C. A. Grimes, Angew. Chem., Int. Ed., 2009, 48, 8095–8098.
40. J. Y. Liao, J. W. He, H. Y. Xu, D. B. Kuang and C. Y. Su, J. Mater. Chem., 2012, 22, 7910–7918.
41. J. Zhi, A. Chen, H. L. Cui, Y. Xie and F. Q. Huang, Phys. Chem. Chem. Phys., 2015, 17, 5103–5108.
42. J. Liu, A. X. Wei, Y. Zhao, K. B. Lin and F. Z. Luo, J. Mater. Sci.: Mater. Electron., 2014, 25, 1122–1126.
43. K. S. Kim, H. Song, S. H. Nam, S. M. Kim, H. Jeong, W. B. Kim and G. Y. Jung, Adv. Mater., 2012, 24, 792–798.
44. K. Mahmood, H. W. Kang, S. B. Park and H. J. Sung, ACS Appl. Mater. Interfaces, 2013, 5, 3075–3084.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra13623g

---