Improved photovoltaic performance in perovskite solar cells based on CH₃NH₃PbI₃ films fabricated under controlled relative humidity

Abstract

Because organic–inorganic perovskite solar cells are known to be unstable in the presence of moisture, most of the reported high-efficiency perovskite solar cells are fabricated under an inert atmosphere in a glove box. This requirement impedes mass production of organic–inorganic perovskite solar cells (PVSCs). Fabricating dense uniform perovskite thin films with high surface coverage via a single one-step solution process is also a challenge in achieving high-efficiency PVSCs. In this work, we successfully develop a facile, controllable one-step solution processing method to obtain high-quality hybrid-perovskite thin films in ambient atmosphere. The entire preparation process for CH₃NH₃PbI₃ films is conducted in ambient air to investigate the effect of humidity on the molecular structure and crystallization of the hybrid perovskite. It is found that relative humidity (RH) and solvent are crucial factors in determining the final morphology of CH₃NH₃PbI₃ and the photovoltaic performance. The best device efficiency achieved for a solar cell fabricated in ambient atmosphere under a RH of 28% is 16.15%; this PCE value is comparable to that of glove box-based PVSCs. This work puts forth a possible method for the easy mass production of high-performance PVSCs under ambient conditions.

---

Introduction

Organic–inorganic hybrid perovskites such as CH₃NH₃PbI₃ have interesting optical and electronic properties such as remarkable optical absorptions, excellent charge carrier mobility, high dielectric constant, and low effective masses.^1–3 Power conversion efficiencies (PCEs) of photovoltaic (PV) devices that have been fabricated using these materials have increased dramatically from around 4% to more than 20% in the last four years.⁴ Although promisingly high PCE values of perovskite solar cells (PVSCs) have been realized, these solar cells have been found to be unstable, degrading in the presence of moisture.⁵ Further, it is well known that the hygroscopic nature of CH₃NH₃⁺ makes it difficult to deposit a perovskite layer under ambient air conditions.⁶ Most of the reported high-efficiency PV devices based on perovskite films have been fabricated under an inert atmosphere in a glove box,^7–9 which impedes mass production of PVSCs.

Recently some experimental preparation of hybrid-perovskite films in ambient air have been written.^10–13 However, there is few mention of the effects of relative humidity (RH) in the entire formation process of hybrid-perovskite. On the other hand, Yang et al. have found that the participation of water molecules is not entirely bad. They fabricated planar heterojunction solar cells using MAPbI_3−xCl_x perovskite.¹⁴ The enhanced reconstruction process of the MAPbI_3−xCl_x formed in a glove box was carried out by annealing under controlled humidity conditions (30 ± 5% RH) and resulted in an increase of device PCE. Moisture-assisted crystal growth has also been recently reported by Yang et al.,¹⁵ whose hypothesis is that moisture helps to enhance grain boundary movement. In the above two studies by Yang et al., it was focused on the effect of the controlled humidity on the crystallization or re-crystallization of the formed perovskite film, but not on the formation of hybrid-perovskite molecular structure. Up to now, there is few report that has studied the effect of controlling humidity over the entire process of hybrid-perovskite film which includes the establishment of the molecular structure and crystallization.

For solution processing of perovskite films in ambient air, a two-step method generally results in high-efficiency PVSCs, because PbI₂ is not sensitive to water during the first step of lead iodide formation. The perovskite layer is formed after PbI₂ is dipped into the CH₃NH₃I solution, minimizing the influence of humidity. However, there are few methods available to fabricate such cells in ambient air via a single one-step solution process. Herein we developed high-performance planar-heterojunction-structured perovskite solar cells with a facile one-step deposition by using a dimethylacetamide (DMAC) solution of CH₃NH₃PbI₃. We previously reported that DMAC is an effective solvent for controlling the dynamics of nucleation and grain growth of CH₃NH₃PbI₃;¹⁶ it induces rapid crystallization of perovskite films, suggesting that it is more suitable to be used in humid environments.

In this study, in order to investigate the effect of humidity (water molecules in ambient air) on the molecular structure and crystallization of hybrid-perovskite and to determine the optimal moisture content, the entire preparation process for CH₃NH₃PbI₃ films was conducted in ambient air. We found that the quantity of humidity and solvent are crucial factors determining the final morphology of CH₃NH₃PbI₃ and the photovoltaic performance.

Experimental section

Materials preparation

CH₃NH₃I was synthesized by reacting 27.86 mL methylamine (40% in methanol, Junsei Chemical Co.) and 30 mL hydroiodic acid (57 wt% in water, Aldrich) in a 250 mL round-bottomed flask at 0 °C for 2 h with stirring. The precipitate was recovered by evaporation at 50 °C for 1 h. Then, the product CH₃NH₃I, was dissolved in ethanol, recrystallized from diethyl ether, and dried at 60 °C in a vacuum oven for 24 h. Further, 2,2′,7,7′-tetrakis(N,N-di-p-methoxyphenyl amine)-9,9′-spirobifluorene (spiro-MeOTAD) was purchased from Aldrich.

Solar cell fabrication

Fluorine-doped transparent, conducting SnO₂-coated (fluorine-doped tin oxide; FTO) glass substrates (OPV Tech Co., Ltd) were cleaned by ultrasonic washing in ethanol for 30 min and were subsequently treated in an oxygen plasma cleaning machine for 10 min. A 0.15 mol L⁻¹ titanium isopropoxide with 50 μL acetylacetone was slowly added dropwise under stirring to a 2 mol L⁻¹ HCl solution in 25 mL absolute ethanol. The resulting solution of TiO₂ was then coated on the substrates by spin coating at 3000 rpm for 30 s and the resultant film was annealed at 450 °C for 60 min.

The previously synthesized CH₃NH₃I powder and PbI₂ (99.5%, Alfa-Aesar) were mixed in a 1 : 1 mole ratio in DMAC at 60 °C for 12 h. The resulting concentration of CH₃NH₃PbI₃ was 1.5 mol L⁻¹. Before the CH₃NH₃PbI₃ layer was deposited over the TiO₂ layer, the RH of ambient air was controlled by employing an air blower or a dehumidifier. Further, in order to ensure the accuracy of measurements, a thermo-hygrograph capable of carrying out instantaneous temperature and humidity measurement was used. The CH₃NH₃PbI₃ solution was coated onto the TiO₂/FTO substrate by spin coating at 1000 rpm for 10 s and 6000 rpm for 60 s in a humid environment. After a specific rotation time, i.e., ∼50 s, the pale yellow CH₃NH₃PbI₃ gradually turned brown. The film was then subjected to annealing on a hot plate at 105 °C for 10 min to evaporate the residual solvent and crystallization completed. Finally, a black CH₃NH₃PbI₃ film was obtained. This facile refinement process for the preparation of CH₃NH₃PbI₃ films is depicted schematically in Fig. 1. Further, for comparison, we also prepared a reference CH₃NH₃PbI₃ film under an inert environment in an Ar-filled glove box (<10 ppm O₂ and H₂O).

Fig. 1 Schematic procedure for preparation of CH₃NH₃PbI₃ film.

Subsequently, the spiro-MeOTAD-based hole-transfer layer (i.e., 170 mg spiro-MeOTAD, 28.5 μL 4-tert-butylpyridine, and 20 mg lithium-bis(trifluoromethanesulfonyl)imide (Li-TFSI) all dissolved in 1 mL chlorobenzene) was deposited by spin coating at 4000 rpm for 30 s. Before evaporating the silver electrodes, spiro-MeOTAD was allowed to oxidize in dry air overnight at room temperature. Finally, a 150 nm-thick silver layer was deposited by vacuum evaporation at a pressure of 1.3 × 10⁻³ Pa. For the fabrication of the best-performing devices, slightly modified conditions were used. First, the TiO₂/FTO substrates were pre-heated at 65 °C on a hot plate before coating with CH₃NH₃PbI₃. Then, the samples were put in a refrigerated cabinet for 2 min after thermal annealing of the CH₃NH₃PbI₃ film.

Device characterization

The current–voltage (J–V) curves of the solar cells were measured in ambient air with a RH of 20–40% using a Keithley 2400 source meter (Keithley Instruments Inc., Cleveland, OH, USA). The cells were illuminated using a 450 W Class AAA simulator equipped with an AM1.5G filter (XES-40S1) at a calibrated intensity of 100 mW cm⁻², as determined by a standard silicon reference cell. The J–V scan rate is 150 mV s⁻¹. The effective area of the cell was set to 0.0405 cm² by using a non-reflective metal mask. The crystallographic structure of the perovskite was analyzed by X-ray diffraction (XRD) (D/MAX Ultima III, Rigaku Corporation, Tokyo, Japan) using Cu Kα radiation. The morphology was determined by scanning electron microscopy (SEM). Atomic force microscopy (AFM) images were obtained using a Digital Instrument NanoScope NS3A system to study the surface morphologies, including the surface roughness. UV-visible (UV-vis) spectra were recorded on a Hitachi U-3010 spectrophotometer (Hitachi, Ltd., Chiyoda, Tokyo, Japan). The external quantum efficiency (EQE) measurements were carried out using a system consisting of a Xe lamp (300 W) with a monochromator (Oriel 74100). The light intensity was measured with an optical power meter (OphirOptronics 70310) equipped with a calibrated thermopile head (OphirOptronics 71964). To investigate the electrical properties of the interfaces, electrochemical impedance spectroscopy (EIS) was performed using a Zahner IM6ex electrochemical workstation, in which a perturbation of 10 mV was applied and the frequency ranged from 100 Hz to 1 MHz.

Results and discussion

The composition evolution and crystal structure of the CH₃NH₃PbI₃ films were investigated for different RH levels (20–40% humidity) by employing X-ray diffraction (XRD). As shown in Fig. 2, the strong diffraction peaks at 14.12°, 28.41°, 31.85°, and 43.19° can be assigned to the (110), (220), (310), and (330) planes of the CH₃NH₃PbI₃ crystal, corresponding to a tetragonal crystal structure of halide perovskite with high crystallinity. The diffraction peak intensity increased with decreasing humidity, which reveals that the film crystallinity and thickness can be improved within a certain humidity range. Take (110) planes of the CH₃NH₃PbI₃ crystal for example. The grain size can be counted according to Debye–Scherrer formula: D = Kλ/β cos θ, where D is the grain size, K is constant (value 0.89), λ is the wavelength of the incident ray (value 0.154 nm), θ is the diffraction angle, and β is the full width at half maximum (FWHM). As shown in Table 1, we carries out a comparison between the calculation results under different conditions. These results proved that the FWHM decreased with the increase of RH, while the grain size did the reverse. In addition, the grain size of perovskite film formed in the glove box is similar to that of perovskite film formed under ∼32% humidity. Notably, a small peak at 12.65° indicates the presence of incompletely converted PbI₂ in samples prepared under ∼40% humidity. This suggests that CH₃NH₃PbI₃ decomposes under high humidity levels. Bragg peaks of the reference and the other samples at lower humidity levels are roughly similar, which suggests that the effect of moisture on the chemical composition of the perovskite film is minimal under low humidity levels.

Fig. 2 X-ray diffraction (XRD) patterns of the perovskite films on glass/FTO substrates prepared under different environmental conditions with various relative humidity (20–40%) and glove box. Solid line: CH₃NH₃PbI₃; dotted line: glass/FTO.

Table 1 The full width at half maximum (FWHM) and the grain size (D) under different conditions for CH₃NH₃PbI₃ layer treatment

Treatment	RH (20 ± 2%)	RH (24 ± 2%)	RH (28 ± 2%)	RH (32 ± 2%)	RH (36 ± 2%)	RH (40 ± 2%)	Reference
FWHM (°)	0.379	0.385	0.372	0.363	0.357	0.351	0.361
D (Å)	209	206	213	218	222	226	219

---

Fig. 3 presents the UV-vis absorption spectra of a CH₃NH₃PbI₃ film exposed to various RH levels. All the samples exhibited panchromatic absorption of light with spectra extended over the visible to near-infrared region. However, the samples prepared at ∼36% and ∼40% RH showed lower light harvesting than the samples prepared at lower humidity levels, implying that higher moisture contents may decrease the optical absorption coefficient. Interestingly, the perovskite film exposed to an ambient environment with a RH of 28 ± 2% showed enhanced absorption characteristics. Moreover, all the moisture-exposed samples exhibit stronger absorption in the wavelength range of 550–800 nm compared to the reference sample; this confirms that moisture can increase light absorption at longer wavelengths of visible light.

Fig. 3 UV-vis absorption spectra for perovskite films exposed to various relative humidities.

The morphology of the perovskite films fabricated under different environmental conditions was analyzed via SEM, and the obtained results are shown in Fig. 4. There are three major perovskite domain morphologies with humidity levels from low to high, as seen in Fig. 4(a)–(c), respectively. Obvious grain boundary is observed for the low-humidity sample. Moreover, the poor coverage of CH₃NH₃PbI₃ films prepared in the high-humidity environment indicates that increasing the moisture reduces the film density. Significant improvement in film morphology was seen when CH₃NH₃PbI₃ films were prepared in ambient air (humidity of 28% ± 4%). High-resolution SEM images in Fig. 4(e)–(g) show that the grain size increased with increasing humidity. The inset in Fig. 4(e) shows the distinct the crystal boundary. The size of crystalline grains is in the range of 50 to 300 nm and the higher crystallinity of the particles in the films cast in air (RH = 28 ± 4%) enable the films to exhibit good light absorption. By comparison, the reference sample fabricated in the Ar-filled glove box exhibits apparent pin holes (Fig. 4(d) and 1(h)). The surface roughness of the films was further monitored by high-resolution atomic force microscopy (AFM) (Fig. 5(a) and (b)). The size of the imaged surface is 10 × 10 μm². The measured average roughness (R_a) and root-mean-square roughness (R_q) are respectively 69.4 and 88.6 nm for the CH₃NH₃PbI₃ film prepared under ∼28% humidity, and 74.5 and 103 nm for the reference film prepared in the glove box. This shows that a high-quality perovskite film can be obtained if the RH can be carefully controlled.

Fig. 4 Surface SEM images of the perovskite film prepared under different moisture levels. (a and e) RH = 20 ± 4%, (b and f) RH = 28 ± 4%, (c and g) RH = 36 ± 4%, (d and h) reference.

Fig. 5 The atomic force microscopy (AFM) images of (a) ambient air with RH = 28 ± 2%; (b) reference.

In order to further elucidate the influence of humidity, a schematic illustration of the CH₃NH₃PbI₃ precursor solution system under humid conditions is shown in Fig. 6. It has been reported previously that exposure to anhydrous gases has absolutely no effect on organohalide perovskites.¹⁷ Thus, we can conclude that the combined action of the solvent, i.e., DMAC, and moisture promotes fast formation and crystal growth in the film. Moreover, since the use of the volatile DMAC solvent dramatically promotes rapid crystallization of perovskite precursor films,¹⁶ the duration in which humidity affects crystallization is short. Nevertheless, we continued to monitor the effect of humidity on CH₃NH₃PbI₃ formation since oxygen had little effect on the degradation of the perovskite.¹⁸ When RH was low, the small amount of moisture was insufficient to enhance CH₃NH₃PbI₃ diffusion after the self-assembly of CH₃NH₃I and the PbI₂. Almost all the water molecules tended to gather at grain boundaries and cause erosion of CH₃NH₃PbI₃. When exposed to high RH, CH₃NH₃I decomposed before the formation of CH₃NH₃PbI₃, forming CH₃NH₂, hydrated HI, and unreacted solid PbI₂, which has also been previously reported by Niu et al.¹⁹ However, when a specific RH (such as ∼28% humidity) was employed along with the effective solvent DMAC, the CH₃NH₃PbI₃ crystals were obtained. Since CH₃NH₃PbI₃ are extremely soluble in the aqueous precursor solution system, it is possible that the stable systems composed of CH₃NH₃PbI₃ and water molecules are established, like CH₃NH₃PbI₃·H₂O.²⁰ When a CH₃NH₃PbI₃ molecule encounters one water molecule, water will weaken the interactions between the PbI₃ and CH₃NH₃ units in CH₃NH₃PbI₃ and make CH₃NH₃PbI₃ precursor move around more freely. This is able to positively adjust the perovskite crystallization rate and facilitate their anisotropic growth. While upon heating, water molecules inside the precursor film tend to evaporate. Meanwhile, when a CH₃NH₃PbI₃ molecule encounters two water molecules, the hydrogen bonds between the PbI₃ and CH₃NH₃ units are nearly broken.²⁰ So, we believe that the right amount of water could enhance the controllability of the perovskite crystallization and encourage homogeneous nucleation. In other words, the appropriate moisture could provide an aqueous environment to boost the diffusion length of the CH₃NH₃PbI₃ and to accelerate adjacent grain agglomeration, further promoting perovskite grain growth and reducing pinhole formation. The rapid crystallization stage completed finally with the evaporation of solvent and water. Therefore, an appropriate RH in air is absolutely necessary for realizing high-quality perovskite films.

Fig. 6 Schematic illustration of the solution system of CH₃NH₃PbI₃ precursor under moisture condition.

We further elucidated the relation between the perovskite layer and photovoltaic performance by constructing FTO/compact TiO₂/CH₃NH₃PbI₃/spiro-MeOTAD/Ag devices. The J–V curves of the solar cells based on perovskite films prepared under different humidity conditions are shown in Fig. 7(a), and the inset shows device performance variation with RH. The error is measured for 12 devices for each condition and the best photovoltaic parameters are listed in Table 2. An analysis shows that the PCE first increased with the humidity level (20–40% humidity) and then decreased. The best performance achieved was 15.3% PCE for approximately 28% RH. The unsatisfactory performance characteristics can be ascribed to the poor coverage and low crystallinity of the obtained samples. Nevertheless, we can conclude that a suitable RH improves the performance of the perovskite solar cells.

Fig. 7 Device performance and device cross-section view. (a) J–V curve of the corresponding devices with the perovskite films prepared under different environment. (b) J–V curve and (c) cross-section view of the devices with the perovskite films optimized in the ambient with ∼28% humidity.

Table 2 Best cells performance of perovskite based solar cells under different conditions for CH₃NH₃PbI₃ layer treatment

Treatment	Jsc (mA cm^−2)	Voc (V)	FF	PCE (%)
RH (20 ± 2%)	18.6	1.06	0.684	13.5
RH (24 ± 2%)	18.9	1.05	0.695	13.8
RH (28 ± 2%)	21.4	1.07	0.668	15.3
RH (32 ± 2%)	20.8	1.03	0.663	14.2
RH (36 ± 2%)	17.9	1.06	0.685	13.0
RH (40 ± 2%)	17.4	1.04	0.641	11.6
Reference	19.6	1.05	0.711	14.6

---

Ultimately, by further improving the perovskite film fabrication processes, by pre-heating the substrate as reported by Hyun-Seok et al.,¹³ and by rapidly cooling after annealing, a device efficiency of 16.15% was obtained under a RH of 28%. As shown in Fig. 7(b) and Table 3, the short circuit current (J_sc) significantly increased from 21.4 mA cm⁻² to 22.95 mA cm⁻², suggesting sufficient light absorption and efficient carrier collection. Average J–V characteristics obtained by sweeping the voltage from forward to reverse and from reverse to forward bias demonstrate no obvious J–V hysteresis. Fig. 7(c) shows the cross-sectional view of the complete device with the ∼40 nm-thick, dense TiO₂ layer acting as the electron-transport layer (ETL), the ∼400 nm-thick CH₃NH₃PbI₃ functioning as the active optical absorption layer, and the ∼200 nm-thick spiro-MeOTAD serving as the hole-transport layer (HTL).

Table 3 Device performance parameters under controlled humidity conditions (28 ± 2% relative humidity) in different scan directions

Scandirection	Jsc (mA cm^−2)	Voc (V)	FF	PCE (%)
Reverse	23.2	1.02	0.689	16.3
Forward	22.7	1.03	0.684	16.0
Average	22.95	1.025	0.6865	16.15

---

Fig. 8(a) shows the EQE in the range from 300 nm to 800 nm. The results suggest that the 28% RH-based PVSC has a higher EQE value than the glove box-based PVSC in the wavelength range of 450 to 750 nm. The integrated J_sc values are 22.5 mA cm⁻² and 19.3 mA cm⁻², respectively, which agree with the J–V measurement results. The Nyquist plots with a forward bias of 1 V in the dark are given in Fig. 8(b), and the inset shows the equivalent circuit and the plots showing the recombination resistance (R_rec) variation with the applied bias. The impedance spectra are dominated by a large semicircle and no obvious arc related to HTL has been observed, as reported by Mora-Sero et al.²¹ These spectra indicate that the 28% RH-based PVSCs exhibit smaller internal series resistance (R_s), which is related to the charge-transfer resistance (R_CT) and to the TiO₂/CH₃NH₃PbI₃ and FTO/TiO₂ interfaces.²² Therefore, a small R_CT for the 28% RH-based PVSC was obtained as a result of good contact at the TiO₂/CH₃NH₃PbI₃ interface. For both devices, a decrease in R_rec is observed with increasing forward bias voltage. In particular, at a high applied bias, close to the V_oc, both the 28% RH-based PVSC and the glove box-based PVSC exhibited a similar recombination resistance, indicating similar V_oc for both devices.²³ The EIS analyses results are consistent with the J–V measurement results for the PVSCs. These results further indicate that controlled humidity can improve the quality of the CH₃NH₃PbI₃ film, and this characteristic is beneficial to significantly improve the performance of PV devices to make them comparable to the performance of the glove box-based PVSC.

Fig. 8 (a) EQE spectra and (b) Nyquist plot of the relative humidity (28 ± 2%) – based and glovebox-based perovskite solar cells.

Finally, we also fabricated a large active area photovoltaic module for the 28% RH-based PVSC, 1 × 1 cm in size, as shown in Fig. 9. The device exhibited V_oc of 0.816 V, J_sc of 16.2 mA cm⁻², FF of 43.3%, and PCE of 5.7%. The poor performance of this large-area module is concerned primarily with the limitation of simple spin-coating and the higher series resistance, which needed to be improved and perfected.

Fig. 9 (a) Photograph of the 28% RH-based PVSC (a active area of 1 cm²) and (b) J–V curve of this large-area module.

Conclusions

This is the first report to systematically investigate the effect of RH (20–40%) during the entire perovskite film fabrication process. We also examined how the water molecules participates in formation and crystallization of hybrid-perovskite, and analyzed the effect of the quantity of water molecules. Our results demonstrate that a specific RH will positively impact the preparation of the CH₃NH₃PbI₃ film. We show that a controlled humidity of about 28% can enhance the morphology and photoelectric properties of the CH₃NH₃PbI₃ film and improve the photovoltaic performance. The best device efficiency achieved by introducing substrate pre-heating and rapid cooling under a RH of 28% was 16.15%. Further comparison between the glove box-based PVSC and 28% RH-based PVSC also confirms the beneficial effects of moderate humidity. Thus, we believe that mass production and application of PVSCs under ambient conditions is possible.

Acknowledgements

This work was supported by the Privileged Development Program of Jiangsu High Education on New energy material science and engineering, the National Natural Science Foundation of China (Grant No. 51272033), the Jiangsu Province Industry-University-Research joint innovation fund (BY2013024-01), and the Natural Science Foundation of the Jiangsu Higher Education Institutions of China (Grant No. 14KJA430001, EEKJA48000).

References

1. M. Grätzel, Nat. Mater., 2014, 13, 838–842.
2. M. A. Green, A. Ho-Baillie and H. J. Snaith, Nat. Photonics, 2014, 8, 506.
3. T. Salim, S. Sun, Y. Abe, A. Krishna, A. C. Grimsdale and Y. M. Lam, J. Mater. Chem. A, 2015, 3, 8943.
4. S. D. Stranks, P. K. Nayak, W. Zhang, T. Stergiopoulos and H. J. Snaith, Angew. Chem., Int. Ed., 2015, 54, 3240–3248.
5. H. S. Kim, S. H. Im and N. G. Park, J. Phys. Chem. C, 2014, 118, 5622.
6. J. H. Kim, S. T. Williams, N. Cho, C. C. Chueh and A. K. Y. Jen, Adv. Energy Mater., 2014, 5, 1401229.
7. N. Li, H. Dong, H. Dong, J. Li, W. Li, G. Niu, X. Guo, Z. Wu and L. Wang, J. Mater. Chem. A, 2014, 2, 14974.
8. G. E. Eperon, V. M. Burlakov, P. Docampo, A. Goriely and H. J. Snaith, Adv. Funct. Mater., 2014, 24, 156.
9. K. Wojciechowski, M. Saliba, T. Leijtens, A. Abate and H. J. Snaith, Energy Environ. Sci., 2014, 7, 1146.
10. M. Seetharaman, P. Nagarjuna, P. N. Kumar, S. P. Singh, M. Deepa and M. A. G. Namboothiry, Phys. Chem. Chem. Phys., 2014, 16, 2469–24696.
11. Y. Zhou, M. Yang, W. Wu, A. L. Vasiliev, K. Zhu and N. P. Padture, J. Mater. Chem. A, 2015, 3, 8178–8184.
12. K. L. Wu, A. Kogo, N. Sakai, M. Ikegami and T. Miyasaka, Chem. Lett., 2015, 44, 321–323.
13. H. S. Ko, J. W. Leea and N. G. Park, J. Mater. Chem. A, 2015, 3, 8808.
14. H. Zhou, Q. Chen, G. Li, S. Luo, T. B. Song, H. S. Duan and Y. Yang, Science, 2014, 345, 542.
15. J. You, Y. M. Yang, Z. Hong, T. B. Song, L. Meng, Y. Liu, C. Jiang, H. Zhou, W. H. Chang, G. Li and Y. Yang, Appl. Phys. Lett., 2014, 105, 183902.
16. M. Lv, X. Dong, X. Fang, B. Lin, S. Zhang, J. Ding and N. Yuan, RSC Adv., 2015, 5, 20521.
17. K. K. Bass, R. E. McAnally, S. Zhou, P. I. Djurovich, M. E. Thompsona and B. C. Melot, Chem. Commun., 2014, 50, 15819–15822.
18. J. Yang, B. D. Siempelkamp, D. Liu and T. L. Kelly, ACS Nano, 2015, 9, 1955–1963.
19. G. Niu, W. Li, F. Meng, L. Wang, H. Dong and Y. Qiu, J. Mater. Chem. A, 2014, 2, 705–710.
20. X. Dong, X. Fang, M. Lv, S. Zhang, J. N. Ding and N. Yuan, J. Mater. Chem. A, 2015, 3, 5360–5367.
21. I. Mora-Sero, S. Gimenez, F. Fabregat-Santiago, E. Azaceta, R. Tena-Zaera and J. Bisquert, Phys. Chem. Chem. Phys., 2011, 13, 7162–7169.
22. B. J. Kim, D. H. Kim, Y. Y Lee, H. W. Shin, G. S. Han, J. S. Hong, K. Mahmood, T. K. Ahn, Y. C. Joo, K. S. Hong, N. G. Park, S. Lee and H. S. Jung, Energy Environ. Sci., 2015, 8, 916–921.
23. F. F. Santiago, G. G. Belmonte, I. M. Seŕoa and J. Bisquert, Phys. Chem. Chem. Phys., 2011, 13, 9083–9118.

---