Inorganic salt templated porous TiO₂ photoelectrode for solid-state dye-sensitized solar cells

Abstract

TiO₂ photoelectrodes with channels (or pores) are prepared by using an inorganic salt (NaHCO₃) as a template. NaHCO₃ doped TiO₂ slurries deposited on a compacted TiO₂ film are heated at 150 °C, which releases CO₂ gas and water vapour, thus forming channels (or pores) inside a TiO₂ light-scattering photoelectrode. These channels (pores) improve the penetration degree of solid-state electrolytes into the TiO₂ photoelectrode and enhance the photovoltaic efficiency of resulting DSSCs (dye sensitized solar cells). With a channel (pore) size of ∼500 nm diameter, an ionic liquid plastic crystal, 5-azoniaspiro[4.4]nonane bis(trifluoromethanesulfonyl)imide (N₄₄TFSI), shows power conversion efficiencies of 6.1% and 5.4% under 1.5 solar spectrum illuminations at 50 and 100 mW cm⁻², respectively.

---

Introduction

Dye-sensitized solar cells (DSSCs) have attracted both academic and industrial attention in the past decade due to their high-efficiency, easy fabrication, and relatively low production cost.¹ A typical DSSC is generally composed of a dye-coated mesoporous TiO₂ film supported on transparent conducting glass, and a counter electrode separated by an electrolyte.^2,3 Although a new record in power conversion efficiency of 13% has recently been achieved based on liquid electrolyte,⁴ the leakage and volatilization of organic liquid electrolyte still needs to be overcome for the practical application of DSSCs.^5–7 In order to solve the problems caused by the use of organic liquid electrolytes, alternative electrolytes, including solid-state electrolytes,^8–10 organic and inorganic hole conductors,¹¹ polymer gel electrolytes,^12,13 and p-type semiconductors,¹⁴ have been extensively studied.^15,16

Recently, a power conversion efficiency (PCE) exceeded 10% for solid-state DSSCs has been reported.¹⁷ However, efforts to further improve the photovoltaic performance are still necessary. It has been recognized that the complete filling of porous TiO₂ films with solid-state electrolyte is a key feature in efforts to enhance the efficiency of such devices.^18,19 It has been realized that a light-scattering TiO₂ film may reduce the light transmittance and enhance the reflectance of TiO₂ photoanodes.²⁰ Therefore, modification of TiO₂ photoelectrodes, such as single crystal structures,²¹ composite photoelectrodes, etc.,^22–25 have been considered as feasible methods to improve the photoelectrode/electrolyte interfacial properties, and thus increase the photovoltaic performances of DSSCs.

Organic plastic crystals are materials composed of molecules with long-range molecular order and with short-range rotational disorder,⁹ which endues these materials with plasticity and high diffusivity properties. Therefore, organic plastic crystals are usually soft and can be easily deformed under an applied force without fracture. These unique properties endow organic plastic crystals with good electrode/electrolyte interfacial contact when they are applied for solid-state electrolytes electrochemical devices.^26,27

In order to improve filling of the light-scattering TiO₂ film with solid-state electrolyte, herein we report a facile and effective method for the preparation of light-scattering TiO₂ film with porous channels (or pores) using a salt template method. An inorganic salt, NaHCO₃, was chosen as template salt in this work. Thermal decomposition of NaHCO₃ inside the TiO₂ light-scattering film generates porous channels (or pores), which boost the penetration of electrolyte (especially solid-state electrolyte) into the porous TiO₂ photoelectrode and, therefore, improve the photovoltaic efficiency of the solar cells due to enhanced interfacial contact at the TiO₂ photoelectrode/electrolyte interface. An ionic liquid plastic crystal, 5-azoniaspiro[4.4]nonane bis(trifluoromethanesulfonyl)imide (N₄₄TFSI) based solid electrolyte interdigitated in a porous TiO₂ photoelectrode was used to fabricate solid-state DSSCs, which exhibited power conversion efficiencies of 6.1% and 5.4% under 1.5 solar spectrum illuminations at 50 and 100 mW cm⁻², respectively.

Experimental

Materials

Iodine (I₂), N-butylbenzimidazole (NBB), tetrahydropyrrole, sodium bicarbonate (NaHCO₃) and 1,4-dibromobutane were purchased from Alfa Aesar. TiCl₄ and H₂PtCl₆ were bought from Sinopharm Chemical Reagent Co., Ltd. Lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) was kindly provided by Rhodia. cis-Bis(isothiocyanato)-bis(2,20-bipyridyl-4,40-dicarboxylato)ruthenium(II) ([RuL₂(NCS)₂]) was purchased from Solaronix SA (Switzerland) and used as received. Slurries containing 20 nm-sized mesoporous and 200 nm-diameter light-scattering TiO₂ colloid and fluorine-doped tin oxide overlayer (FTO) glass (7 Ω sq⁻¹) were obtained from Dalian Hepat Chroma Solar Tech. Co., Ltd.

Synthesis of ionic plastic crystal

5-Azoniaspiro[4.4]nonane bromine (N₄₄Br) was synthesized as follows: a mixture containing 4.40 g (6.2 mmol) tetrahydro pyrrole and 13.11 g (12.4 mmol) potassium carbonate was dissolved in 40 mL acetonitrile, in which 13.36 g (6.2 mmol) 1,4-dibromobutane was added dropwise and stirred at 70 °C for 24 h. The precipitate was removed by filtration (see Scheme 1). The solvent was evaporated under reduced pressure and the solid compound was washed with diethyl ether and dry ethyl acetate. The obtained compound was dried at 50 °C under vacuum to give a white solid: N₄₄Br (8.7 g, 68% yield). ¹H NMR (D₂O): 2.2 (s, 4H), 3.6 (m, 4H).

Scheme 1 General synthetic route for ionic liquid plastic crystal, N₄₄TFSI.

5-Azoniaspiro[4.4]nonane bis(trifluoromethanesulfonyl)imide (N₄₄TFSI) was synthesized by stirring a mixture of 1.5 g (5.2 mmol) LiTFSI and 1.07 g (5.2 mmol) N₄₄Br in 50 mL water at room temperature for 4 h. The precipitate was washed by water and then dried at 70 °C under vacuum to give a white solid: N₄₄TFSI (1.91 g, 90% yield). ¹H NMR (d₆-DMSO): 1.9 (s, 4H), 3.4 (m, 4H).

Preparation of electrolytes

Plastic crystal solid-state electrolyte was prepared by stirring a mixture containing 20 wt% 1-propyl-3-methylimidazolium iodide (PMII), 0.5 M N-butylbenzimidazole (NBB), 5 wt% lithium bis(trifluoromethanesulfonyl)imide (LiTFSI), and 0.05 M iodine (I₂) in N₄₄TFSI at 70 °C for 24 h.

Preparation of TiO₂ colloid

To a slurry (0.4 g) containing 200 nm-diameter TiO₂ colloid, 0.012 g, 0.02 g, 0.04 g NaHCO₃, respectively, were added. Then 4 mL isopropanol was added to the respective mixtures, and they were stirred at room temperature for 24 h. The isopropanol was removed under vacuum at 35 °C.

Characterization

A DSC-Q200 was used to characterize the thermal properties of the prepared electrolytes at a scan rate of 10 °C min⁻¹ under a N₂ atmosphere. The electrochemical impedance spectra (EIS) of fabricated devices were conducted on a CHI660c electrochemical workstation in the dark. A bias voltage of −0.80 V was used for the impedance measurement, and the frequency was varied from 0.01 to 10⁵ Hz. Scanning electron microscopy (SEM) images and energy-dispersive X-ray (EDX) spectroscopy measurements were performed with the spectrometer attached to the Hitachi Model S-4700 field-emission SEM system.

Photocurrent density–voltage (J–V) of the fabricated DSSCs (shielded with an aluminum foil mask with an area of 0.1 cm²) was characterized by a digital source meter (Keithley, model 2612) under simulated air mass (AM) illumination of 1.5 solar spectrum at 100, 50, 15 and mW cm⁻², respectively. A source meter (Keithley 1612) was applied to measure the voltage dynamics. The electrolyte conductivities were characterized in a Teflon tube with two identical stainless steel (diameter of 1 cm) as the electrodes. The conductivity was recorded on a CHI660c electrochemical work-station, and calculated by the following equation:²⁸

where R is the ohmic resistance of the electrolyte, σ is the conductivity in S cm⁻¹, S is the area of the electrodes and l is the distance between the two electrodes.

Device fabrication

FTO glasses were cleaned with acetone, isopropanol, and deionized water sequentially, and then dried under vacuum at 90 °C. A 10 μm thick film of 20 nm-sized TiO₂ particles was firstly coated onto the cleaned FTO glass by the doctor-blade technique. After the coated TiO₂ film was annealed at 125 °C, a second 200 nm light-scattering TiO₂ particles (10 μm thick) were coated onto the previous TiO₂ layer in the same way. The fabricated TiO₂ films were annealed at 500 °C. The fabricated photoanodes were immersed in a 0.5 mM Z-907 dye solution in ethanol for 16 h at room temperature. The electrode was then rinsed with anhydrous ethanol and dried. Pt counter electrode was prepared by annealing of 5 mM H₂PtCl₆ ethanol solution at 400 °C. The working and counter electrodes were assembled by a 25 μm thick hot melt ring (Surlyn, Dupont) and sealed by heating. A methanol solution containing 20 wt% PMII and 80 wt% N₄₄TFSI was injected into the cell using a vacuum backfilling system. Then the methanol was removed under vacuum at 80 °C.

Results and discussion

Scanning electron microscopy

Fig. 1 shows the schematic of representation for light-scattering TiO₂ film with porous channels (or pores). Inorganic salt, NaHCO₃, was stirred and dispersed in a TiO₂ (200 nm in diameter) isopropanol solution. A NaHCO₃/TiO₂ mixture was deposited on the compacted TiO₂ film by the doctor-blade technique. The isopropanol was removed under vacuum at 35 °C and then heated at 150 °C for 2 h. Upon heating, NaHCO₃ decomposes with the release of carbon dioxide gas and water vapour. Therefore, the channels were opened inside the TiO₂ light-scattering film by escaping carbon dioxide and water bubbles. It is anticipated that the formed channels could improve the penetration degree of the solid-state electrolyte, which therefore enhance the photovoltaic performances of the DSSCs.

Fig. 1 Schematic of representation for TiO₂ photoelectrode with channels (or pores). A mixture of NaHCO₃ and TiO₂ colloid was deposited on the FTO glass by a doctor-blade technique. These deposited TiO₂ films were then heated at 150 °C for 2 h. The channels inside the TiO₂ light-scattering film were formed by escaping carbon dioxide and water bubbles.

The formation of the light-scattering TiO₂ film with channels (pores) was investigated by scanning electron microscopy (SEM), as shown in Fig. 3. Fig. 3a shows SEM images of the TiO₂ film prepared without the addition of NaHCO₃. Pores with a diameter of ∼60 nm could be observed from the top surface of the TiO₂ film (Fig. 3a). While the channels (pores) with larger size (∼500 nm in diameter) and higher density were observed upon the addition of 5 wt% of NaHCO₃ (Fig. 3b), due to the release of carbon dioxide gas and water vapour generated from the decomposition of NaHCO₃. Further addition of NaHCO₃ (10 wt%) would increase the channels (pores) sizes formed in the light-scattering TiO₂ film (∼5 um in diameter, Fig. 3c). The formed pores and channels may increase the filling of the solid-state electrolyte. The elemental mapping of the cross section of the solar cell were further characterized by the energy-dispersive X-ray (EDX) spectroscopy (Fig. 2). The appearance of N, S elements confirmed the successful penetration of solid-state electrolyte into TiO₂ film.

Fig. 2 The N, S, Ti elemental mapping of the cross section of TiO₂ films filled with ionic liquid plastic crystal, N₄₄TFSI. The TiO₂ films were prepared through the addition of 5 wt% of NaHCO₃.

Fig. 3 Scanning electron microscopy (SEM) images of light-scattering TiO₂ films: (A) light-scattering TiO₂ film prepared without addition of NaHCO₃; (B) and (C) TiO₂ films prepared through the addition of 5 wt% and 10 wt% of NaHCO₃, respectively. The TiO₂ photoanodes were washed with water to remove salt residue and then dried under vacuum.

Thermal properties of electrolytes

Organic plastic crystals are compounds composed of rotationally disordered organic molecules. They are considered to be one of the best solid-state electrolytes for electrochemical devices, because they could eliminate these leakage problems, provide a good electrolyte/electrode interfacial contact, and provide high ionic conductivity.⁹ The ionic liquid plastic crystal, N₄₄TFSI, with two exothermic solid–solid phases at 1.2 °C and 70.2 °C, and a melting point around 115.3 °C (Fig. 4), was synthesized and used for the preparation of solid-state electrolyte. Under the outdoor applications, the temperature might reach 60 °C under full sunlight; such relatively high melting point makes it possible for N₄₄TFSI to be used as a solid-state electrolyte. Previous studies have revealed that proper dopant can drastically increase the conductivity of ionic plastic crystals.²⁹ Therefore, in this work, PMII was used to dope N₄₄TFSI to form high conductive solid-state electrolytes. The typical thermal analysis traces (Fig. 4a) showed that a second solid–solid phase transition temperature and melting point of N₄₄TFSI were slightly lowered from 71.5 °C to 65.8 °C, and from 115.3 °C to 81.8 °C, respectively, upon addition of 20 wt% PMII. This result suggests that a solid solution of PMII and N₄₄TFSI has been successfully prepared. However, the first solid–solid phase transition temperature of N₄₄TFSI doped with 30 wt% PMII disappeared, indicating that the N₄₄TFSI plastic crystal bulk matrix was completely disrupted, resulting in the absence of a bulk N₄₄TFSI phase. Fig. 4b shows the ionic conductivity over 30–80 °C of N₄₄TFSI doped with 20 wt% PMII electrolytes. It can be seen that the ionic conductivity of the electrolyte increases with the temperature rise. For instance, the conductivity of 20 wt% PMII doped N₄₄TFSI electrolyte at 70 °C (∼3.1 × 10⁻³ S cm⁻¹) is higher than that at 30 °C (∼3.2 × 10⁻⁴ S cm⁻¹). Such a change in conductivity may be due to the free volume changes in the plastic crystal.³⁰ The doping of PMII has increased the vacancy concentration or volume fraction of the plastic crystal electrolyte. Herein, the TiO₂ photoelectrodes with channels (pores) were used for ionic liquid plastic crystal electrolyte DSSCs. The prepared solid-state electrolyte composed of 20 wt% PMII, 0.05 M I₂, 0.5 M NBB, 5 wt% LiTFSI in N₄₄TFSI was characterized and used for devices.

Fig. 4 (a) Differential scanning calorimetry (DSC) thermograms for PMII doped N₄₄TFSI electrolytes; (b) DSC trace and conductivity–temperature plot for 20 wt% PMII doped N₄₄TFSI electrolyte; J–V curves of DSSCs containing N₄₄TFSI based solid-state electrolyte (20 wt% 1-propyl-3-methylimidazolium iodide (PMII), 0.05 M I₂, 0.5 M N-butylbenzimidazole (NBB), 5 wt% LiTFSI in N₄₄TFSI). (c) Under simulated AM 1.5 solar spectrum irradiation at 100 mW cm⁻², and (d) under dark condition. All the devices were tested with a mask (test area: 0.1 cm²). The light-scattering TiO₂ films were prepared with 0 wt%, 3 wt%, 5 wt%, and 10 wt% NaHCO₃ for cells A, B, C and D, respectively.

Photovoltaic performance of DSSCs

The J–V curves of DSSCs based on the TiO₂ photoanodes with and without channels (pores) were shown in Fig. 4. The detailed photovoltaic parameters, including short-circuit current density (J_sc), fill factor (FF), open circuit voltage (V_oc), as well as power conversion efficiency (PCE) for the fabricated devices are summarized and listed in Table 1. As can be seen that the average efficiency of cell A (control device, without the channels (pores)) is about 4.23%. The efficiency of the (cells B and C). For instance, the overall power conversion efficiency was increased from 4.23% to 4.47% and 5.43% for cells B and C, respectively. The enhancement of the photovoltaic efficiency is perhaps due to the increased penetration degree of the electrolytes into the porous TiO₂ photoelectrode. However, cell D with a channel (pore) size of ∼5 μm could not improve the photovoltaic performances of the cells, because the larger size channels (pores) may weaken the light scattering intensity (see Fig. 3C).

Table 1 Solid-state DSSC performance parameters under simulated AM 1.5 solar spectrum irradiation at 100 mW cm⁻² (average of three cells). The solid-state electrolyte contained 20 wt% PMII, 0.05 M I₂, 0.5 M NBB, 5 wt% LiTFSI in N₄₄TFSI. The light-scattering TiO₂ films were prepared with 0 wt%, 3 wt%, 5 wt%, and 10 wt% of NaHCO₃ for cells A, B, C and D, respectively

Cell	Jsc [mA cm^−2]	Voc [mV]	FF (%)	PCE [%]
A	9.99 ± 0.2	582 ± 2	72.7 ± 0.1	4.23 ± 0.2
B	10.86 ± 0.1	588 ± 1	72.4 ± 0.2	4.47 ± 0.1
C	12.58 ± 0.2	591 ± 3	73.0 ± 0.2	5.43 ± 0.1
D	7.45 ± 0.3	649 ± 1	76.5 ± 0.1	3.70 ± 0.1

---

Fig. 5 shows the IPCE values of the cells based on solid-state electrolyte. It can be seen that the maximum IPCE values are 49.7%, 54.2%, 60.1%, and 37.5% for cells A–D at 530 nm, respectively. The results of the IPCE values are consistent with the J_sc values. The PCE values of the device under various light intensities are listed in Table 2. The best performance devices were achieved at low levels of simulated sunlight, might due to the inefficient electron transport within the TiO₂ layer. A maximum PCE value of 6.08% was obtained under AM 1.5 irradiance at 50 mW cm⁻².

Fig. 5 IPCE vs. wavelength profiles for DSSC containing N₄₄TFSI based solid-state electrolyte (20 wt% PMII, 0.05 M I₂, 0.5 M N-butylbenzimidazole, 5 wt% LiTFSI in N₄₄TFSI). The TiO₂ light-scattering films were prepared with 0 wt%, 3 wt%, 5 wt%, and 10 wt% of NaHCO₃ for cells A, B, C and D, respectively.

Table 2 Photoconversion efficiency (PCE) values of DSSCs based on different electrolytes under simulated AM 1.5 solar irradiation (average of three cells). The light-scattering TiO₂ films were prepared via doping with 0 wt%, 3 wt%, 5 wt%, and 10 wt% of NaHCO₃ for cells A, B, C and D, respectively

Cell	η under different incident light intensities^a
	0.15 sun	0.5 sun	1.0 sun
a The spectral distribution of the lamp mimicked AM 1.5 solar spectrum. The intensity “1.0 sun” corresponds to an intensity of 100 mW cm^−2.
A	4.40 (±0.02)	4.82 (±0.01)	4.23 (±0.02)
B	4.64 (±0.03)	5.09 (±0.01)	4.47 (±0.01)
C	5.63 (±0.02)	6.08 (±0.02)	5.43 (±0.03)
D	4.12 (±0.02)	4.25 (±0.01)	3.70 (±0.02)

---

Electrochemical impedance spectra (EIS) of DSSCs

However, it is noteworthy that the light-scattering TiO₂ film with channels (pores) could slightly increase the V_oc values of the devices if compared with the control device (cell A). It has already been demonstrated that the enhanced V_oc is due to the suppression of the dark current at electrolyte/TiO₂ photoelectrode interface.³¹ Fig. 4d shows that the onset of dark current shifted about 14 and 23 mV from cell A to cells C and D, respectively. The lowered onset of dark current value indicates a lower I³⁻ reduction rate, which leads to an increase of V_oc. Here, our understanding is that the channels (pores) could inhibit the dark current which originated from the I³⁻ reduction.

EIS measurements were further performed to understand the influence of the channels (pores) on the dye-coated TiO₂ photoelectrode/electrolyte interfacial properties (see Fig. 6).³² The EIS spectra comprises three semicircles, which represent the resistances of electrolyte/counter electrode interface (R₁), electrolyte/TiO₂ layer interface (R₂), and the Nernst diffusion in electrolyte (R₃), respectively. The smaller the semicircle in the 12.72, and 9.67 Ω (see Table 3), respectively, indicating that the generated channels (pores) of the light-scattering film can increase the resistance of the charge-transfer elements at the recombination.³³ This result agrees with the dark current and V_oc measurements. To calculate effective lifetime of electrons (τ_e) before the recombination in a TiO₂ photoelectrode. Fig. 6b shows the Bode phase plots of EIS spectra of all the fabricated devices. The effective lifetime of electrons (τ_e) before the recombination in TiO₂ photoelectrodes can be determined follow the equation:³⁴

where f_max is the maximum frequency of the low-frequency peak. Here, the τ_e values extracted from the f_max are determined to be 7.02, 22.02, 30.92 and 17.91 ms for cells A, B, C and D, respectively (see Table 3), indicating that the electron lifetimes are enhanced after increasing the penetration degree of the solid-state electrolytes into the TiO₂ photoelectrode. Here, the cell C shows the longest τ_e and yields the highest efficiency of 5.43%.

Fig. 6 Electrochemical impedance spectroscopy (EIS) Nyquist plots (a), and Bode phase plots (b), obtained in the dark for the cells A–D with the light-scattering TiO₂ films were prepared with 0 wt%, 3 wt%, 5 wt%, 10 wt% NaHCO₃ for cells A, B, C and D, respectively.

Table 3 EIS parameters obtained for cells A, B, C and D. The light-scattering TiO₂ films were prepared with 0 wt%, 3 wt%, 5 wt%, and 10 wt% of NaHCO₃ for cells A, B, C and D, respectively

Cell	R1 (Ω)	R2 (Ω)	R3 (Ω)	fmax (Hz)	τe (ms)
A	2.83	9.32	8.72	22.5	7.02
B	2.42	10.14	9.67	7.23	22.02
C	2.63	12.72	9.22	5.15	30.92
D	2.50	9.67	10.13	8.89	17.91

---

The long-term stability of the fabricated solid-state cells was studied by an accelerating test. The PCE values of all the cells were measured every seven days. The efficiencies are normalized to those determined on the first day. The fabricated devices were stored in air (humidity below 10%) at room temperature and 60 °C in the dark (Fig. 7). During the first seven days, the photovoltaic efficiency of the cells was increased due to the enhanced dye regeneration rate, which resulted in an increased J_sc value. The cell A (control device, light-scattering TiO₂ film without porous channels) maintained about 91% of the initial PCE value after 42 days. A similar result was observed for cell C (TiO₂ light-scattering film with ∼500 nm porous channels), indicating that reasonably sized channels (pores) in a light-scattering TiO₂ film can keep the stability of the devices. The enhancement of the efficiency at the first few days for cell B and cell C is perhaps due to the increased penetration degree of the electrolytes into the porous TiO₂ photoelectrode. The stability of the cell is in the order of A > B > C. However, the efficiency of cell D decreases gradually after 12 days probably due to the larger channels (pores) formed which destroy the light-scattering TiO₂ film and decreases long-term stability.

Fig. 7 Time-course variation of PCE of N₄₄TFSI based solid-state DSSCs employing the TiO₂ photoanodes prepared with 0 wt%, 3 wt%, 5 wt%, and 10 wt% of NaHCO₃ for cells A, B, C and D, respectively. The cells were stored at room temperature in the dark and in air with low humidity (below 10%) and tested under the simulated AM 1.5 solar spectrum irradiation at 100 mW cm⁻², (a) tested at room temperature; and (b) tested at 60 °C.

Conclusions

In conclusion, we reported an effective method for the performance enhancement of solid-state electrolyte DSSCs. TiO₂ photoelectrodes with channels (or pores) were prepared via a salt templating method. The channels (or pores) enhance the penetration degree of solid-state and quasi-solid-state electrolytes into the TiO₂ photoelectrodes and improve the photovoltaic performances of DSSCs. This method could also be applied for plastic crystal type solid-state electrolytes with low melting points, and provides a feasible and practical method for fabricating solid-state DSSCs.

Acknowledgements

This work was supported by the National Science Foundation for Distinguished Young Scholars (No. 21425417), the Natural Science Foundation of China (No. 21274101), the National Basic Research Program of China (973 Program) (No. 2012CB825800), and the Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions.

References

1. B. O'regan and M. Grätzel, Nature, 1991, 353, 737–740.
2. S. K. Balasingam, M. Lee, M. G. Kang and Y. Jun, Chem. Commun., 2013, 49, 1471–1487.
3. M. Nasrollahzadeh, RSC Adv., 2014, 4, 29089–29093.
4. S. Mathew, A. Yella, P. Gao, R. Humphry-Baker, B. F. Curchod, N. Ashari-Astani, I. Tavernelli, U. Rothlisberger, M. K. Nazeeruddin and M. Grätzel, Nat. Chem., 2014, 6, 242–247.
5. Q. Yu, Y. Wang, Z. Yi, N. Zu, J. Zhang, M. Zhang and P. Wang, ACS Nano, 2010, 4, 6032–6038.
6. K. Kakiage, Y. Aoyama, T. Yano, T. Otsuka, T. Kyomen, M. Unno and M. Hanaya, Chem. Commun., 2014, 50, 6379–6381.
7. A. Hilmi, T. A. Shoker and T. H. Ghaddar, ACS Appl. Mater. Interfaces, 2014, 6, 8744–8753.
8. T. Taguchi, X.-t. Zhang, I. Sutanto, K.-i. Tokuhiro, T. N. Rao, H. Watanabe, T. Nakamori, M. Uragami and A. Fujishima, Chem. Commun., 2003, 2480–2481,  10.1039/b306118c.
9. V. Armel, M. Forsyth, D. R. MacFarlane and J. M. Pringle, Energy Environ. Sci., 2011, 4, 2234–2239.
10. J. Y. Kim, T. H. Kim, D. Y. Kim, N.-G. Park and K.-D. Ahn, J. Power Sources, 2008, 175, 692–697.
11. M. Yu, T. I. Draskovic and Y. Wu, Phys. Chem. Chem. Phys., 2014, 16, 5026–5033.
12. J. E. Benedetti, M. A. de Paoli and A. F. Nogueira, Chem. Commun., 2008, 1121–1123,  10.1039/b717278h.
13. S. Senthilkumar, R. K. Selvan, N. Ponpandian and J. Melo, RSC Adv., 2012, 2, 8937–8940.
14. Z. Bian, T. Tachikawa, S.-C. Cui, M. Fujitsuka and T. Majima, Chem. Sci., 2012, 3, 370–379.
15. H. J. Snaith and L. Schmidt-Mende, Adv. Mater., 2007, 19, 3187–3200.
16. R. D. Costa, F. Lodermeyer, R. Casillas and D. M. Guldi, Energy Environ. Sci., 2014, 7, 1281–1296.
17. H. Wang, J. Li, F. Gong, G. Zhou and Z.-S. Wang, J. Am. Chem. Soc., 2013, 135, 12627–12633.
18. C.-Y. Cho and J. H. Moon, Adv. Mater., 2011, 23, 2971–2975.
19. P.-Y. Chen, C.-P. Lee, R. Vittal and K.-C. Ho, J. Power Sources, 2010, 195, 3933–3938.
20. D. Xu, C. Shi, L. Wang, L. Qiu and F. Yan, J. Mater. Chem. A, 2014, 2, 9803–9811.
21. X. Feng, K. Shankar, O. K. Varghese, M. Paulose, T. J. Latempa and C. A. Grimes, Nano Lett., 2008, 8, 3781–3786.
22. N. G. Park, M. G. Kang, K. M. Kim, K. S. Ryu, S. H. Chang, D. K. Kim, J. van de Lagemaat, K. D. Benkstein and A. J. Frank, Langmuir, 2004, 20, 4246–4253.
23. S. Ito, Y. Makari, T. Kitamura, Y. Wada and S. Yanagida, J. Mater. Chem., 2004, 14, 385–390.
24. Y. Kim, B. J. Yoo, R. Vittal, Y. Lee, N.-G. Park and K.-J. Kim, J. Power Sources, 2008, 175, 914–919.
25. S. Rai, A. Ikram, S. Sahai, S. Dass, R. Shrivastav and V. R. Satsangi, RSC Adv., 2014, 4, 17671–17679.
26. D. R. MacFarlane, J. Huang and M. Forsyth, Nature, 1999, 402, 792–794.
27. S. Li, L. Qiu, C. Shi, X. Chen and F. Yan, Adv. Mater., 2014, 26, 1266–1271.
28. X. Chen, Q. Li, J. Zhao, L. Qiu, Y. Zhang, B. Sun and F. Yan, J. Power Sources, 2012, 207, 216–221.
29. J. M. Pringle, P. C. Howlett, D. R. MacFarlane and M. Forsyth, J. Mater. Chem., 2010, 20, 2056–2062.
30. A. J. Hill, J. Huang, J. Efthimiadis, P. Meakin, M. Forsyth and D. R. MacFarlane, Solid State Ionics, 2002, 154–155, 119–124.
31. D. Kuang, P. Wang, S. Ito, S. M. Zakeeruddin and M. Grätzel, J. Am. Chem. Soc., 2006, 128, 7732–7733.
32. W. Zhang, C. Yuan, J. Guo, L. Qiu and F. Yan, ACS Appl. Mater. Interfaces, 2014, 6, 8723–8728.
33. J. Xia, N. Masaki, M. Lira-Cantu, Y. Kim, K. Jiang and S. Yanagida, J. Am. Chem. Soc., 2008, 130, 1258–1263.
34. C. Shi, L. Qiu, X. Chen, H. Zhang, L. Wang and F. Yan, ACS Appl. Mater. Interfaces, 2013, 5, 1453–1459.

---

Footnote

† Authors with equal contributions.

---