Graphene oxide/poly(3,4-ethylenedioxythiophene):polystyrenesulfonate layers on silver nanowire working electrodes enhance the power conversion efficiencies of dye-sensitized solar cells in a low temperature process

Abstract

In this study, silver nanowire (AgNW) networks were applied, for the first time, as transparent conductive layers for the working electrodes of dye-sensitized solar cells (DSSCs). To avoid corrosion of the AgNWs by iodine in the electrolyte, graphene oxide (GO) was used as a protective layer that isolated the AgNWs from the electrolyte; the effective prevention of corrosion probably arose from electrostatic repulsion between the electrolyte and the charged surface of the protective layer. After incorporation of 25% poly(3,4-ethylenedioxythiophene):polystyrenesulfonate (PEDOT:PSS) into the GO, the power conversion efficiency (PCE) of the resulting GO/PEDOT:PSS–AgNW DSSC improved by 324%. The PEDOT:PSS doping led to the work function of the GO from 5.63 to 5.24 eV. The result may be because of decreasing the strength of the C–O dipole moment of the GO. The enhanced PCE can be attributed to the lower Schottky barrier between the AgNWs and the GO, due to the decreased work function of the GO.

---

1. Introduction

The consumption of fossil fuels is a significant issue because of their limited supply on Earth and the resulting impact on the environment. Photovoltaic technologies have promise as a route toward converting solar energy, an alternative energy source, into electrical power. Among the photovoltaic technologies, dye-sensitized solar cells (DSSCs) are particularly interesting because of their inexpensive production and relatively high power conversion efficiencies (PCEs).^1,2 In the last two decades, the component materials and the mechanism of electronic transport in DSSCs have been investigated widely, resulting in the commercialization of DSSCs becoming more feasible. Recently, there has been increasing attention drawn toward flexible DSSCs, replacing the rigid glass substrates with light-weight plastic substrates that can be produced inexpensively on a large scale using roll-to-roll coating techniques.^3–7

Regardless of the nature of rigid or flexible DSSCs, indium tin oxide (ITO) remains the material most commonly used to prepare the transparent conductive electrodes of solar cells, due to its low electrical resistance and high transparency. Nevertheless, ITO is brittle and over-bending will result in significant and irreversible increase in electrical resistance. For plastic substrates, the limitation of low-temperature processing impedes any attempts to enhance the electrical conductivity of ITO. To meet the requirements for the electrodes of solar cells, expensive processes can be used to lower the electrical resistance of plastic substrates.^8,9 Accordingly, several new materials, including carbon nanotubes,^10,11 graphene nanosheets,¹² and silver nanowires (AgNWs),^13,14 have been investigated as potential replacements for ITO. Among them, AgNWs, which feature the highest ratio of direct current conductivity to optical conductivity, appear to be the most suitable replacements for ITO in products requiring low resistance.¹⁵ Taking advantage of their rapid electron transfer along the axial direction and localized surface plasmon resonance effects, AgNWs have been applied successfully in the porous TiO₂ layers of DSSCs to minimize recombination between the excited electrons and the holes/tri-iodide ions or to enhance light excitation, thereby improving PCEs.^16,17 The electron transfer and localized surface plasmon resonance characteristics can also enhance the photocurrent of organic photovoltaics when AgNWs are applied in the hole transport¹⁸ and active¹⁹ layers. Moreover, AgNWs have successfully replaced ITO as the transparent conductive electrodes for the counter electrodes of DSSCs^20–23 and as the cathodes of organic photovoltaics.^24–26 To prevent the iodine electrolyte from corroding the AgNWs, a protective material {e.g., poly(3,4-ethylenedioxythiophene):polystyrenesulfonate (PEDOT:PSS),²⁰ TiO₂,^16,17 SiO₂ (ref. 27)} can be coated as on their surfaces. However, iodine ions could penetrate the thin coating layers within a very short time period.²⁸ Very recently, solid state DSSCs (ss-DSSCs) featuring 2,2′,7,7′-tetrakis(N,N-di-p-methoxyphenylamino)-9,9′ spirobifluorene (spiro-MeOTAD) as a hole transport material have been prepared, avoiding the issue of leakage of the electrolyte solution and increasing their applicability in flexible devices.²⁹ With perovskite absorbers, such ss-DSSCs can display very high PCEs.^30,31 Although spiro-MeOTAD does not corrode AgNWs, it is very expensive and unstable under long periods of irradiation with light.³² Cheap and stable CuI is a promising alternative to spiro-MeOTAD,³³ but, unfortunately, it can damage AgNWs. Regardless of whether an iodine electrolyte or a new hole transport material is used, overcoming the issue of corrosion of AgNWs will presumably benefit their application within DSSCs.

To the best of our knowledge, no attempts have been made previously to use transparent conductive films of AgNWs as working electrodes in DSSCs. Again, the main limitation may be the issue of corrosion²⁸ on the working electrodes of the AgNWs. Although the materials mentioned above for protection of AgNWs in other components of DSSCs may also be applied to the working electrodes, such protective materials might hinder the light harvesting ability or electron transport between the AgNWs and the porous TiO₂ layer or among the nanowires. In particular, some brittle protective materials (e.g., TiO₂, SiO₂) would not be suitable in flexible devices. In this study, we used graphene oxide (GO) as a protective material to fabricate transparent conductive working electrodes of AgNWs for use in DSSCs. A small dosage of PEDOT:PSS in the GO protective layer significantly enhanced the PCEs of these DSSCs. We have also investigated the mechanism of corrosion protection and the effect of the protective layer on electron transfer.

2. Experimental

2.1 AgNWs

AgNWs were synthesized using a two-step polyol reduction method, as reported previously,³⁴ with some of the parameters slightly modified to obtain a larger aspect ratio. Briefly, 0.3 M polyvinylpyrrolidone (36 mL, Sigma-Aldrich) and 0.2 M NaCl (80 μL) were mixed in a reaction vessel and heated at 143 °C. 1 M AgNO₃ (20 μL, Sigma-Aldrich) was added into the mixture; after 5 min, a further charge of 1 M AgNO₃ (4 mL) was added slowly using a peristaltic pump. After the color of the solution had become silver-whitish (ca. 60 min), the as-prepared AgNWs were washed three times with EtOH through centrifugation. The purified AgNWs had lengths of 40–50 μm and diameters of 40–50 nm (Fig. 1).

Fig. 1 Typical (a) SEM and (b) OM images of as-prepared AgNWs.

2.2 GO–AgNW electrodes

Glass substrates were washed with EtOH and cleaned using an O₂ plasma cleaner (PDC-23G, Harrick Plasma). The AgNW solution was spin-coated on a treated substrate and then dried at room temperature; this process was repeated several times to reach a desired sheet resistance. The AgNW-coated substrate was dipped in 1 wt% polyethoxysiloxane (PES, Shin-Etsu) EtOH solution to enhance the electrical conductivity and resistance to oxidation and thermal shock for AgNW network.³⁵ The PES-coated substrate was immersed in 1.25 wt% 1,2-aminopropyltriethoxysilane (APTS, Sigma-Aldrich) in EtOH containing 0.005 M HCl for 2 min and then dried at 100 °C for 30 min. The APTS-treated substrate was dipped twice into a mixture of GO (0.3 mg mL⁻¹), synthesized as reported in our previous study (see ESI†),¹³ and PEDOT:PSS (weight ratio of PEDOT:PSS to GO: 0, 0.0625, 0.125, 0.1875, or 0.25). For comparison, a pure PEDOT:PSS layer upon the APTS-treated substrate was also prepared; it was dipped twice in the PEDOT:PSS solution (10 mg mL⁻¹). The reduced GO samples were obtained by thermally treating the GO substrates under a N₂ atmosphere at 400 °C for 30 min. A TiO₂ blocking layer was formed by dipping the substrate in a mixture of tetrabutyl titanate (TBT, 1.2 g, Fluka), acetylacetone (AcAc, 0.06 g, Aldrich), and 0.1 M HCl (2 g) in butanol (40 g total) and then drying at 100 °C for 30 min. The structure of a GO–AgNW electrode is presented schematically in Fig. 2.

Fig. 2 Schematic representation of the structure of the GO–AgNW electrode.

2.3 Assembly of DSSCs

In our experiments, the GO–AgNW electrodes were used as the working electrodes of DSSCs. To avoid the thermal damage to GO and AgNW, DSSCs were fabricated at a low-temperature process. The post-treatment on the porous TiO₂ layer³⁶ wasn't done in order to reduce the complexity of analysis and realize the characteristics of the electrodes more clearly. Briefly, a 12.5 wt% TiO₂ slurry was prepared by dispersing P25 into a mixture of aqueous HCl (pH 3) and tert-butanol (1 : 2, v/v). Using this slurry, a porous TiO₂ layer (ca. 7 μm) was prepared on the GO–AgNW electrodes at low temperature though chemical sintering.^37,38 The working electrodes were immersed in a solution of N719 (0.3 mM) at 50 °C for 3 h, rinsed with acetonitrile, and then dried at room temperature. The dyed working electrodes were assembled with Pt counter electrodes into DSSCs, using a Surlyn film (thickness: 60 μm) as a spacer. The electrolyte solution (0.6 M BMII, 0.5 M TBP, 0.1 M GuNSC, 0.2 M LiI, and 0.05 M I₂ in MeCN) was filled into the cells through the pre-drilled holes in the Pt counter electrodes.

2.4 Characterization

The morphologies of the AgNWs and the GO–AgNW electrodes were examined using an optical microscope (M835, M&T Optics) and a field-emission scanning electron microscope (JSM-7401F, JEOL). The sheet resistances and transmittances of the GO–AgNW electrodes were determined using a four-pin probe meter (Loresta-GP, Mitsubishi Chemical) with an MCP-T610 probe and a UV-Vis spectrophotometer (Lambda 850, Perkin-Elmer), respectively. The photocurrent density–voltage (J–V) characteristics were measured under irradiation (100 mW cm⁻²) from a solar simulator (MFS-PV, Hong-Ming Technology) equipped with a source meter (Keithley 2400, Keithley Instruments). Electrochemical impedance spectroscopy (EIS; frequency range: from 50 mHz to 100 kHz; potential perturbation: 10 mV) was performed using an electrochemical workstation (Zennium, Zahner). The work functions of GO and GO/PEDOT:PSS were evaluated from photoelectron spectra measured in air (PESA; AC-2, Riken Keiki).

3. Results and discussion

There is a natural trade-off when using nanowires in conductive films: when the number of AgNWs on the substrate increases, the surface resistance decreases but so too does the transmittance. Fig. 3 reveals that the performance of the AgNWs was better than or comparable with those of ITO and FTO. Note that the transmittance in Fig. 3 is based on the reference that the transmittance of 100% is for air but not for the substrate; this aspect is often confused in the literature. The iodine electrolyte can, however, react with Ag to form AgI and, thereby, destroy the nanowire structure. When the AgNWs come into contact with the electrolyte, their nanowire structure can disappear almost immediately (Fig. S1†). In this study, we used GO to protect the AgNWs from the electrolyte; we formed a layer of GO covering the AgNWs to isolate them from the electrolyte (Fig. 4). We observed no damage to the AgNWs after immersing the GO–AgNW electrode into the iodine electrolyte solution (Fig. S2a†). Okada and Shiratori²⁰ prepared counter electrodes of AgNWs for DSSCs and used PEDOT:PSS as not only a catalyst but also a protector for the AgNWs. To compare the abilities of GO and PEDOT:PSS to prevent iodine-induced corrosion, we also applied PEDOT:PSS to protect the AgNWs in the working electrodes (PEDOT:PSS–AgNW). We found that PEDOT:PSS also prevented the AgNWs from corrosion by the iodine electrolyte. Interestingly, the RGO did not protect the AgNWs from electrolytic attack (Fig. S2b†). SEM revealed no pinholes or other damage on the surface of the RGO after the thermal reduction (Fig. S3†), similar to the result for the GO (Fig. S4†). Because both GO and PEDOT:PSS feature numerous charges, but RGO does only few,^39,40 we infer that the preventative effects of GO and PEDOT:PSS against iodine-mediated corrosion arose from electrostatic repulsion between the electrolyte and the charged protectors.

Fig. 3 Variation in sheet resistance of AgNWs plotted with respect to transmittance of incident light at 550 nm.

Fig. 4 SEM image of a GO covering on the surface of AgNWs.

Fig. 5 displays the photocurrent density–voltage curves of the DSSCs incorporating various GO/PEDOT:PSS–AgNW working electrodes; Table 1 lists the corresponding characteristics. The PCE of the DSSC containing the GO–AgNW working electrode was 0.55%, whereas that containing the PEDOT:PSS–AgNW working electrode was only 0.23%. To resist electrolytic corrosion, the thickness of the PEDOT:PSS layer had to be at least 93 nm (Fig. S5†). Fig. 6 reveals that the transmittance of the PEDOT:PSS–AgNW electrode was much lower than that of the GO–AgNW electrode (or the GO/PEDOT:PSS–AgNW electrode), due to its strong light-absorption characteristics and the thick PEDOT:PSS layer. The low transmittance of light may have led to increased electrical resistance on the porous TiO₂ layer,⁴¹ thereby resulting in a low PCE. Fig. 7a presents the Nyquist plots of the electrochemical impedance spectra. For the PEDOT:PSS–AgNW working electrode (GP01), two distinct semicircles, typical for low-temperature-sintered DSSCs,^12,37 appeared in the Nyquist plot. The right semicircle (occurring at lower frequency) was large, suggesting high electrical resistance on the porous TiO₂ layer (R₃), confirming that the low transmittance resulted in the low PCE for the PEDOT:PSS–AgNW working electrode. For the GO–AgNW DSSC working electrode (GP10), however, only one semicircle appeared in the Nyquist plot, suggesting that the impedance of one interface was much larger than that of the others. The inset to Fig. 7a displays the equivalent circuit used to fit the EIS spectra. Our equivalent circuit analysis indicated that the resistance R₂ for the GO–AgNW electrodes reached as high as 849.2 Ω, representing the charge transfer resistance between the GO–AgNW working electrode and the TiO₂ layer. Because GO is a poor electrical conductor, we doped small amounts of PEDOT:PSS into it to increase its electrical conductivity. Fig. 7b displays the EIS analysis for the samples containing various PEDOT:PSS-doping contents (GP0625, GP125, GP1875, GP25). Although two distinct semicircles appeared in the Nyquist plots, in contrast to the case for typical low-temperature ITO DSSCs, the left-hand semicircle was larger that the right-hand one. Upon increasing the amount of PEDOT:PSS, the size of the left-hand semicircle decreased, suggesting that PEDOT:PSS doping decreased the charge-transfer resistance between the GO–AgNW working electrode and the TiO₂ layer; accordingly, the resistance R₂ (Table 1) decreased gradually.

Fig. 5 Photocurrent density–voltage curves of DSSCs featuring various GO/PEDOT:PSS–AgNW working electrodes.

Table 1 Photovoltaic characteristics and EIS fitting parameters for DSSCs featuring various GO/PEDOT:PSS–AgNW working electrodes

Working electrode	GO/PEDOT:PSS	Voc/V	Jsc/mA cm^−2	FF/%	PCE/%	R1/Ω	R2/Ω	R3/Ω
GP10	1 : 0 (0.3 mg mL^−1 GO)	0.65 ± 0.02	3.0 ± 0.1	28 ± 2	0.55 ± 0.06	22	849	24
GP0625	1 : 0.0625	0.73 ± 0.01	4.3 ± 0.2	36 ± 2	1.12 ± 0.02	24	138	18
GP125	1 : 0.125	0.72 ± 0.02	5.3 ± 0.1	35 ± 2	1.35 ± 0.02	17	172	21
GP1875	1 : 0.1875	0.74 ± 0.01	5.8 ± 0.1	44 ± 2	1.87 ± 0.03	25	187	25
GP25	1 : 0.25	0.74 ± 0.01	6.6 ± 0.2	47 ± 2	2.33 ± 0.04	27	271	21
GP01	0 : 33.3 (10 mg mL^−1 PEDOT:PSS)	0.39 ± 0.03	1.7 ± 0.2	35 ± 4	0.23 ± 0.03	33	294	767

---

Fig. 6 Transmission spectra of GO/PEDOT:PSS–AgNW electrodes featuring various ratios of GO to PEDOT:PSS.

Fig. 7 Nyquist plots of the electrochemical impedance spectra of DSSCs featuring various working electrodes having protective layers of (a) pure GO (GP10) and pure PEDOT:PSS (GP01) and (b) a mixture of GO and PEDOT:PSS (GP0625, GP125, GP1875, GP25). Inset to (a) equivalent circuit used to fit the electrochemical impedance spectra.

In our experiments, all of AgNW electrodes have the surface resistance of 20 ± 1 Ω sq⁻¹. However, the surface resistance increased by 0.5∼1 Ω sq⁻¹ when a layer of GO or GO/PEDOT:PSS covered the AgNWs. The increase in the impedance at the interface between the working electrode and the porous TiO₂ layer may have resulted from other causes not related to the electrical conductivity of GO. Fig. 8 presents the PESA spectra of the GO and GO/PEDOT:PSS–AgNW. The work function of GO is 5.63 eV; its work function decreased to 5.24 eV upon doping with PEDOT:PSS. Although the work function of PEDOT:PSS (5.0 eV)²⁵ is lower than that of GO, the low work function for GO/PEDOT:PSS cannot be attributed to the mixing rule, because we had doped only a small amount of PEDOT:PSS into the GO (GO/PEDOT:PSS = 1 : 0.25). The high work function for GO is much more likely due to the presence of C^δ+–O^δ– dipoles.⁴² The presence of charged PEDOT:PSS may have alleviated the dipole moment and, therefore, decreased the work function. Because the work function of GO is much larger than that of AgNWs (4.5 eV)²⁵ and the conduction band of TiO₂, an energy barrier may exist (Fig. 9)—although determining the Fermi level of GO and the exact alignment of the Fermi energies between GO and AgNWs requires further analysis. Moreover, contact between high-work-function, p-type GO and low-work-function AgNWs will create a Schottky barrier that hinders charge transfer. When we incorporated the PEDOT:PSS into the GO, the GO/PEDOT:PSS hybrid layer possessed a lower work function, thereby decreasing the Schottky barrier as illustrated in Fig. 9. Table 1 reveals that, upon incorporation of PEDOT:PSS, the PCEs of the DSSCs featuring the GO protective layer increased from 0.55 to 2.33%—an enhancement of 324%. Fig. 10 presents the variation in dark current with respect to the applied potential. A comparison of Fig. 5 and 10 reveals that, unlike typical DSSCs, a greater PCE led to a larger dark current in these devices. Because the AgNWs, porous TiO₂, dye, electrolyte, and counter electrode used in the cells were identical for all samples, which differed only in their GO/PEDOT:PSS protective layers, the increase in the dark current implies a decrease in resistance between the AgNWs and the porous TiO₂. This finding adds further credence to the inference that PEDOT:PSS doping can decrease the Schottky barrier between AgNWs and GO and, thereby, enhance the PCEs of DSSCs.

Fig. 8 PESA analysis of the GO and the GO/PEDOT:PSS hybrid GP25.

Fig. 9 Schematic representation of the energy diagram of a DSSC featuring a GO/PEDOT:PSS–AgNW working electrode.

Fig. 10 Dark current density–voltage curves of DSSCs featuring various GO/PEDOT:PSS–AgNW working electrodes.

4. Conclusion

For the first time, we have fabricated AgNW working electrodes for DSSCs. We found that the presence of GO circumvented corrosion of the AgNWs by the iodine electrolyte, evidently due to electrostatic repulsion. In contrast to the protective effect of pure PEDOT:PSS, the GO–AgNW electrodes displayed high transmittance and, thereby, led to higher PCEs. The work function of the p-type GO was higher than that of the AgNWs. The incorporation of a small amount of PEDOT:PSS into GO decreased the work function from 5.63 to 5.24 eV, probably arising from the charged PEDOT:PSS overcoming the C–O dipole moment of the GO. After incorporating 25% PEDOT:PSS into the GO, the PCE of the GO/PEDOT:PSS–AgNW DSSC increased by 324%, presumably because of a decrease in the Schottky barrier upon decreasing the work function of the protective layer. Although the PCE of the GO/PEDOT:PSS–AgNW DSSC is low (2.33%), which is much lower than that of the DSSCs made in the high-temperature process, it is comparable to that of the DSSCs made in the low-temperature process, revealing promising for the application of flexible DSSCs.

Acknowledgements

This study was supported financially by the Ministry of Science and Technology, the Republic of China (MOST 103-2221-E-224-074-MY2).

References

1. J. Gong, J. Liang and K. Sumathy, Renewable Sustainable Energy Rev., 2012, 16, 5848–5860.
2. J. Wu, Y. Li, Q. Tang, G. Yue, J. Lin, M. Huang and L. Meng, Sci. Rep., 2014, 4, 4028.
3. X.-L. He, M. Liu, G.-J. Yang, H.-L. Yao, S.-Q. Fan and C.-J. Li, J. Power Sources, 2013, 226, 173–178.
4. A. Vomiero, V. Galstyan, A. Braga, I. Concina, M. Brisotto, E. Bontempi and G. Sberveglieri, Energy Environ. Sci., 2011, 4, 3408–3413.
5. J. Lin, Y. Peng, A. R. Pascoe, F. Huang, Y.-B. Cheng, Y.-U. Heo, A. Nattestad, W. Seung, S. K. Kim, H. J. Yoon, S.-W. Kim, Y. Yamauchi, S. X. Dou and J. H. Kim, J. Mater. Chem. A, 2015, 3, 4679–4686.
6. M. J. Yun, S. I. Cha, S. H. Seo and D. Y. Lee, Sci. Rep., 2014, 4, 5322.
7. K. Yoo, J.-Y. Kim, J. A. Lee, J. S. Kim, D.-K. Lee, K. Kim, J. Y. Kim, B. Kim, H. Kim, W. M. Kim, J. H. Kim and M. J. Ko, ACS Nano, 2015, 9, 3760–3771.
8. K. Utsumi, O. Matsunaga and T. Takahata, Thin Solid Films, 1998, 334, 30–34.
9. A. Klöppel and J. Trube, US Pat. 6,849,165, 2005.
10. Z. Wu, Science, 2004, 305, 1273–1276.
11. B. T. Liu and C. H. Hsu, J. Colloid Interface Sci., 2011, 359, 423–427.
12. S. De and J. N. Coleman, ACS Nano, 2010, 4, 2713–2720.
13. B.-T. Liu and H.-L. Kuo, Carbon, 2013, 63, 390–396.
14. A. R. Madaria, A. Kumar, F. N. Ishikawa and C. W. Zhou, Nano Res., 2010, 3, 564–573.
15. X. Y. Zeng, Q. K. Zhang, R. M. Yu and C. Z. Lu, Adv. Mater., 2010, 22, 4484–4488.
16. M. Sun, W. Fu, H. Yang, Y. Sui, B. Zhao, G. Yin, Q. Li, H. Zhao and G. Zou, Electrochem. Commun., 2011, 13, 1324–1327.
17. Y. F. Wang, J. H. Zeng and Y. Li, Electrochim. Acta, 2013, 87, 256–260.
18. Y.-S. Kim, M.-H. Chang, E.-J. Lee, D.-W. Ihm and J.-Y. Kim, Synth. Met., 2014, 195, 69–74.
19. C.-H. Kim, S.-H. Cha, S. C. Kim, M. Song, J. Lee, W. S. Shin, S.-J. Moon, J. H. Bahng, N. A. Kotov and S.-H. Jin, ACS Nano, 2011, 5, 3319–3325.
20. I. Okada and S. Shiratori, ACS Appl. Mater. Interfaces, 2013, 5, 4144–4149.
21. B. E. Hardin, W. Gaynor, I. K. Ding, S.-B. Rim, P. Peumans and M. D. McGehee, Org. Electron., 2011, 12, 875–879.
22. M. Al-Mamun, J.-Y. Kim, Y.-E. Sung, J.-J. Lee and S.-R. Kim, Chem. Phys. Lett., 2013, 561–562, 115–119.
23. H. C. Weerasinghe, F. Huang and Y.-B. Cheng, Nano Energy, 2013, 2, 174–189.
24. L. Yang, T. Zhang, H. Zhou, S. C. Price, B. J. Wiley and W. You, ACS Appl. Mater. Interfaces, 2011, 3, 4075–4084.
25. G. Y. Margulis, M. G. Christoforo, D. Lam, Z. M. Beiley, A. R. Bowring, C. D. Bailie, A. Salleo and M. D. McGehee, Adv. Energy Mater., 2013, 3, 1657–1663.
26. M. Al-Mamun, J.-Y. Kim, K.-J. Lee, Y.-H. Ko, J.-H. Lee, I.-S. In, J.-W. Lee, Y.-E. Sung and S.-R. Kim, Synth. Met., 2013, 177, 77–81.
27. K. Guo, M. Li, X. Fang, X. Liu, Y. Zhu, Z. Hu and X. Zhao, J. Mater. Chem. A, 2013, 1, 7229–7234.
28. H. Jung, B. Koo, J.-Y. Kim, T. Kim, H. J. Son, B. Kim, J. Y. Kim, D.-K. Lee, H. Kim, J. Cho and M. J. Ko, ACS Appl. Mater. Interfaces, 2014, 6, 19191–19200.
29. H.-S. Kim, C.-R. Lee, J.-H. Im, K.-B. Lee, T. Moehl, A. Marchioro, S.-J. Moon, R. Humphry-Baker, J.-H. Yum, J. E. Moser, M. Gratzel and N.-G. Park, Sci. Rep., 2012, 2, 591.
30. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643–647.
31. H. Zhou, Q. Chen, G. Li, S. Luo, T.-b. Song, H.-S. Duan, Z. Hong, J. You, Y. Liu and Y. Yang, Science, 2014, 345, 542–546.
32. T. Leijtens, G. E. Eperon, S. Pathak, A. Abate, M. M. Lee and H. J. Snaith, Nat. Commun., 2013, 4, 2885.
33. J. A. Christians, R. C. M. Fung and P. V. Kamat, J. Am. Chem. Soc., 2014, 136, 758–764.
34. B.-T. Liu, S.-X. Huang, M.-F. Lai and Z.-H. Wei, RSC Adv., 2015, 5, 1684–1689.
35. B.-T. Liu and S.-X. Huang, RSC Adv., 2014, 4, 59226–59232.
36. F. Huang, D. Chen, Q. Li, R. A. Caruso and Y.-B. Cheng, Appl. Phys. Lett., 2012, 100, 123102.
37. H. C. Weerasinghe, G. V. Franks, J. D. Plessis, G. P. Simon and Y.-B. Cheng, J. Mater. Chem., 2010, 20, 9954–9961.
38. T. Miyasaka, M. Ikegami and Y. Kijitori, J. Electrochem. Soc., 2007, 154, A455–A461.
39. X. Gao, J. Jang and S. Nagase, J. Phys. Chem. C, 2010, 114, 832–842.
40. S. Kirchmeyer and K. Reuter, J. Mater. Chem., 2005, 15, 2077–2088.
41. M. Adachi, M. Sakamoto, J. Jiu, Y. Ogata and S. Isoda, J. Phys. Chem. B, 2006, 110, 13872–13880.
42. S.-S. Li, K.-H. Tu, C.-C. Lin, C.-W. Chen and M. Chhowalla, ACS Nano, 2010, 4, 3169–3174.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra03756a

---