High quality perovskite films fabricated from Lewis acid–base adduct through molecular exchange

Abstract

We describe a facile method to fabricate high quality CH₃NH₃PbI₃ films with smooth surface and without residual PbI₂ through molecular exchange, rather than ionic intercalation. It forms a Lewis acid–base adduct of PbI₂·xDMF when PbI₂ precipitates from DMF solution. The lattice of PbI₂ is expanded more than 30% due to the intercalation of DMF. With the lattice expansion, CH₃NH₃I (MAI) diffuses into the PbI₂·xDMF lattice easily, and the PbI₂·xDMF converts completely to CH₃NH₃PI₃ by molecular exchange between DMF and MAI. The lattice volume changes little during the molecular exchange in PbI₂·xDMF. Thus, it is easy to fabricate high quality perovskite films from the Lewis adduct of PbI₂·xDMF. The perovskite solar cells fabricated from the Lewis adduct exhibit higher photovoltaic performance than those from the PbI₂ films. This work reveals the important role of common solvent in controlling the quality of perovskite films.

---

1. Introduction

Organometallic halide perovskites, especially CH₃NH₃PbI₃ (MAPbI₃), have great potential for use as optoelectronic conversion materials in solar cells due to their high optical absorption coefficient, balance charge transport behavior and long diffusion length.^1–4 Currently, perovskite films are mainly fabricated by solution-based processing, such as a one-step method⁵ or two-step method (TSM).^6,7 The power conversion efficiency (PCE) of perovskite solar cells (PSCs) is greatly influenced by the quality of the perovskite films, including coverage, roughness, grain size, and residual PbI₂. By optimizing the device structure and perovskite layer, the PCE of the PSCs have been improved significantly from the initial 3.81% to >20%.^8–10

The typical TSM has been widely used to prepare perovskite films due to the full surface coverage and relative high PCE. In the typical TSM, PbI₂ films were first prepared by spin coating on FTO substrate from solvent of dimethylformamide (DMF) or dimethylsulfoxide (DMSO) followed by annealing. The PbI₂ films then convert to MAPbI₃ by exposing in a solution of MAI/2-propanol through ionic intercalation.^6,7 The lattice was expanded, and the unit cell volume increased from 124 ų to 248 ų when the PbI₂ convert to MAPbI₃.¹¹ Meanwhile, some grains may be extruded to surface of perovskite films due to the volume expansion, leading to rough surfaces. Furthermore, the ionic intercalation will slow down and even be passivated by MAPbI₃ at the surface of the dense PbI₂ film, resulting in residual PbI₂ in the perovskite.¹² Although, passivation effects have been reported previously,^13–17 the residual PbI₂ will lead to poor reproducibility and instability of the devices.^18,19 Many efforts have been made to eliminate the residual PbI₂, including prolonging reaction time,^14,20 elevating reaction temperature,^21,22 developing gas–solid crystallization reaction,^23–25 employing strongly coordinative solvent of DMSO as additive in solvent,²⁶ or even using pure DMSO as solvent to retard the crystallization of PbI₂.²⁷ But, it also causes some side effects in parallel to elimination of the residual PbI₂, such as dissolution or even peeling-off of the perovskite for long time reaction.²⁸

Here, we report new approach to prepare high quality perovskite films with smooth surfaces and without residual PbI₂ through molecular exchange. It forms a Lewis acid–base adduct of PbI₂·xDMF when the PbI₂/DMF solution is spin coated on the FTO substrate, which the lattice of PbI₂ is pre-expanded by the DMF molecules. There may be some DMF molecular remain in film without acting with PbI₂, so we use x (x > 1) rather than the accurate stoichiometric ratio (x = 1 theoretically)²⁹ to donate Lewis acid–base adduct in the following discussion. The MAI molecules exchange with DMF inside the Lewis adduct of PbI₂·xDMF, leading to high quality perovskite films. Compared to TSM, the devices fabricated through molecular exchange have higher performance and reproducibility.

2. Experimental

Device fabrication

FTO glass (TEC7, Pilkington) was subsequently cleaned with detergent, deionized water, acetone, 2-propanol, ethanol, deionized water, and then dried by nitrogen gas. A compact layer of TiO₂ was deposited on the FTO glass by spin coating of titanium isopropoxy solution in ethanol at 2000 rpm for 30 s, and then annealed at 500 °C for 30 min. A mesoporous TiO₂ layer was deposited on the compact TiO₂ layer by spin coating diluted TiO₂ paste (Dyesol-18NRT, Dyesol) in ethanol (2 : 7, weight ratio) at 4500 rpm for 30 s, and then annealed at 500 °C for 30 min. The mesoporous TiO₂ layer was infiltrated in PbI₂/DMF (1.4 M) solution at 100 °C and then spin coated at 6500 rpm for 30 s. Here, we fabricated perovskite layer on the mesoporous TiO₂ layer from two different routes: typical TSM through ionic intercalation and Lewis adduct reaction through molecular exchange. In the typical TSM route, the PbI₂ films spin-coated from PbI₂/DMF were first annealed at 100 °C for 60 min, and then dipped into a solution of CH₃NH₃I (MAI)/2-propanol for 60 s, followed by rinsing with 2-propanol. In the Lewis adduct reaction route, the PbI₂ films spin-coated from PbI₂/DMF were immediately dipped into the MAI/2-propanol solution for 60 s. All the perovskite films were then annealed at 100 °C for 60 min. Then, a HTM layer was deposited by spin-coating a solution prepared by dissolving 100 mg Spiro-OMeTAD, 40 μL 4-tert-butylpyridine (t-BP), 36.3 μL of a stock solution of 520 mg mL⁻¹ TFSI in acetonitrile and 60 μL of a stock solution of 300 mg mL⁻¹ FK102 dopant in acetonitrile in 1 mL chlorobenzene. Finally, a layer of gold (60 nm) was thermally evaporated on the top of HTM to form a back electrode. The active area of the electrode was fixed at 0.12 cm².

Characterization

The PbI₂ and perovskite films were characterized by field-emission scanning electron microscopy (SEM, MERLIN VP Compact), X-ray diffraction (XRD, D8-Advance), Raman spectroscopy (Lab RAM HR Evolution), UV-vis absorption spectroscopy (Cary 5000 UV-vis-NIR, Agilent Technologies) in a range from 400 nm to 900 nm, and Fourier transform infrared spectrum (VERTEX 70v). The photovoltaic properties of the perovskite solar cells were measured under one solar illumination (AM 1.5G, 100 mW cm⁻¹, 91 195, Newport) and recorded by a semiconductor characterization system (4200-SCS, Keithley) in a range of −0.5 to 1.2 V at a scan rate of 5 mV s⁻¹.

3. Results and discussion

Fig. 1 shows an illustration of the fabrication of perovskite films through molecular exchange. The PbI₂ precursor film is directly immersed in a MAI/isopropanol solution after spin coating from a solution of 1.4 M PbI₂/DMF. For comparison, we also fabricate perovskite films from the typical TSM at same reaction conditions, where the PbI₂ precursor is annealed at 100 °C before dipping in the MAI/isopropanol solution (Fig. S1†).

Fig. 1 Schematic illustration of the fabrication of perovskite films from a Lewis adduct.

We first examine the interaction between the PbI₂ and DMF in the complex compound of PbI₂·xDMF. Fig. 2a presents Fourier transform infrared (FTIR) transmittance spectra of the pure DMF solvent (black) and PbI₂·xDMF (red). A characteristic peak at 1670 cm⁻¹ is identified, corresponding to stretching vibration of C=O in DMF.^30,31 The C=O vibration downshifts to 1627 cm⁻¹ for the PbI₂·xDMF film, which indicates that bond strength between carbon and oxygen decreases duo to the formation of PbI₂·xDMF.³² It has been reported that the DMF molecules coordinated to Pb and form one-dimension structure along a-axis with a Pb–O bond length of 2.431 Å.²⁹ Thus, it forms a Lewis adduct of PbI₂·xDMF through Lewis acid–base reaction, where PbI₂ acts as Lewis acid, and DMF acts as Lewis base. In the PbI₂·xDMF, the oxygen-donor is contributed from DMF because it contains long-pair electrons on oxygen.³³

Fig. 2 (a) Fourier transform infrared spectrometer (FTIR) of DMF and Lewis adduct of PbI₂·xDMF. (b) Raman spectra of DMF and PbI₂·xDMF. (c) XRD spectra of PbI₂ powders (black), PbI₂·xDMF (red) and annealed PbI₂·xDMF (blue). (d) The evolution of the C=O stretching vibration of the Lewis adducts annealed under 70 °C.

Raman spectra also show the interaction between DMF and PbI₂. As shown in Fig. 2b, the in-plane O=C–N bending vibration of the DMF shifts from 657 cm⁻¹ to 665 cm⁻¹ when it reacts with PbI₂. The shift of the in-plane O=C–N bending vibration is similar to the co-ordination of DMF to other metal ions.^34,35 The Raman shift of the N–C=O vibration is related to the formation Pb–O bond in the complex adduct of PbI₂·xDMF.³⁶

Fig. 2c illustrates XRD spectra at low diffraction angle (8°–16°) of the PbI₂ powder and PbI₂·xDMF. The PbI₂ powder has a characteristic XRD peak of 12.7°, corresponding to (001) plane with a lattice constant of 6.98 Å. For the Lewis adduct of PbI₂·xDMF, the diffraction peak of the PbI₂ at 12.7° disappears. There are two characteristic peaks at 9.02° and 9.59° in the XRD spectrum, corresponding to (011) and (020) plane of the PbI₂·xDMF with a plane spacing of 9.82 and 9.23 Å, respectively.³⁷ The lattice of PbI₂ are expanded more than 30% due to the intercalation of DMF. Actually, the PbI₂·xDMF is not so stable due to the weak interaction between DMF and PbI₂. We characterize the evolution of the C=O stretch vibration depending on the annealing time at 70 °C, as shown in Fig. 2d. The intensity of the C=O stretching vibration weakens gradually with the increasing of the annealing time, and finally disappears after annealing for 24 hours. It indicates that the DMF escapes from the PbI₂·xDMF gradually during the annealing.

Fig. 3a presents X-ray diffraction (XRD) spectra of the perovskite films fabricated from the Lewis adduct reaction in MAI solution with various concentrations. Perovskite crystals and FTO substrate are identified in the XRD spectra. There is usually an evident peak at 12.6° in the XRD spectra of the perovskite films prepared from the PbI₂ through the typical TSM (Fig. 4a), corresponding to the (001) lattice plane of the residual PbI₂. The residual PbI₂ exists in all of the samples prepared from the typical TSM with various MAI concentrations, and even for long time (5 min) reaction (Fig. 4b). But, the PbI₂ (001) peak at 12.6° disappears in the samples prepared from the Lewis adduct, which indicates that the PbI₂·xDMF convert completely to perovskite in very short time reaction (30 s) (Fig. S2†). The conversion rate increased by more than 10 times. The rapid conversion is ascribed to the lattice expansion PbI₂·xDMF, so that the MA⁺ and I⁻ ions can diffuse into the PbI₂ lattice easily, resulting in complete conversion of PbI₂·xDMF to MAPbI₃.

Fig. 3 XRD (a) and SEM images of the perovskite films fabricated from Lewis adduct by immersing in MAI/2-proponol solutions with different MAI concentrations. (b) 10 mg mL⁻¹, (c) 20 mg mL⁻¹, (d) 30 mg mL⁻¹, respectively.

Fig. 4 XRD patterns of the perovskite films prepared from the typical TSM by dipping the PbI₂ films in a solution of MAI/2-propanol for different concentration of MAI (a) and different reaction time at 20 mg mL⁻¹ MAI (b).

The perovskite films fabricated from the Lewis adduct of PbI₂·xDMF have relative smooth surface, especially for the films prepared at a relative high MAI concentration of 20 mg mL⁻¹ (Fig. 3c). It forms a high quality (smooth, compact, uniform and PbI₂-free) perovskite film with large grains ranging from 100 nm to 500 nm (50–200 nm for typical TSM as shown in Fig. 5), which are desired for high efficiency perovskite solar cells due to decreasing interfacial recombination.³⁸ It is strange that there are cracks among the perovskite grains prepared from low MAI concentration solution of 10 mg mL⁻¹ (Fig. 3b) and 7.5 mg mL⁻¹ (Fig. S3†). Some pinholes present in the perovskite layer prepared from solution with MAI concentration of 30 mg mL⁻¹ (Fig. 3d). The photovoltaic performance of PSCs are influenced by the cracks and pinholes seriously. While for the perovskite films prepared from PbI₂ through the typical TSM, the samples usually have rough surfaces (Fig. 5). Some cub-like crystalline MAPbI₃ on the perovskite surface when the concentration of MAI is higher than 10 mg mL⁻¹, and it becomes more obvious by prolonging reaction time (Fig. S4†), the phenomenon were also observed in other groups.^12,14,39 The cub-like crystals affect for the photovoltaic performance seriously duo to increasing the roughness of Spiro-OMeTAD and Au electrodes, and even leading to short-circuit in solar cells.¹⁴

Fig. 5 SEM images of the perovskite films prepared from the different concentration of MAI via the typical TSM by reaction 60 s. (a) 10 mg mL⁻¹ (b) 20 mg mL⁻¹ (c) 30 mg mL⁻¹.

The perovskite is fabricated from Lewis adduct of PbI₂·xDMF through molecular exchange between DMF and MAI, as described in eqn (1).

PbI₂·xDMF + MAI → MAPbI₃ + xDMF (1)

The DMF molecules intercalated inside the PbI₂·xDMF are replaced by MAI because the donor of I⁻ in MAI is stronger than DMF.³² Since the lattice of PbI₂ is pre-expanded, no evident lattice volume expansion occurs in the perovskite during the molecular exchange of DMF and MAI; while the lattice volume increase about 2.1 times when PbI₂ convert to MAPbI₃ through ionic intercalation.¹¹ As a result, a smooth perovskite film with large grains (100–500 nm) is easily obtained by the Lewis adduct reaction. High quality perovskite films were also obtained by pre-expanding PbI₂ with MAI.⁴⁰ AFM topography and 3D views further show that the perovskite films fabricated from the Lewis adduct have smoother surface than those from TSM (Fig. S5†). It is noted that the conversion of the PbI₂·xDMF to perovskite completes in the molecular exchange process of DMF and MAI, according to XRD spectra (Fig. S6†) of the films before and after annealing.

Fig. 6 shows J–V curves for the best PSCs fabricated via the two different routes. The corresponding photovoltaic parameters of the solar cells are given in Table 1. Compared with the typical TSM, the photovoltaic parameters, especially J_sc, are improved significantly irrespective of the concentration of MAI (Fig. 6a and b). For the PSCs fabricated from the typical TSM and Lewis adduct reaction, the PCE increases from 7.68% to 13.20%, J_sc from 13.68 mA cm⁻² to 19.42 mA cm⁻², FF from 0.58 to 0.66, and V_oc from 0.97 V to 1.03 V. Although, the photovoltaic performances are not so high due to un-optimized fabrication parameters, they are comparable with those fabricated from the typical TSM in solution with low concentration of MAI.⁷ The increase of J_sc is ascribed to the increase of light harvesting ability in the range 400 to 800 nm (Fig. 6c). The photoluminescence peak offset slight shift for the sample fabricated via Lewis adduct reaction which might be related with the different crystals geometries.⁴¹ But, the V_oc of PSCs fabricated from the Lewis adduct decreased when the MAI concentration increases to 30 mg mL⁻¹ (Fig. 6b), which might be related to the pinholes in the perovskite layer. The observed increase in FF is decided by the change of series resistance (R_s) and shunt resistance (R_sh). The perovskite film fabricated via Lewis adduct reaction results in larger grain size and fewer number of grain boundary, leading to the decrease of R_s.⁷ Benefits from the improvement of V_oc, J_sc and FF, the PCE of perovskite solar cells fabricated via Lewis adduct reaction are improved significantly. The statistic charts of 20 PSCs also show that the PSCs fabricated from Lewis adduct reaction have higher photovoltaic performance than those from the typical TSM (Fig. 7). We believe the photovoltaic performance will be further improved by optimizing the fabrication parameters.

Fig. 6 Light J–V curves of the best perovskite solar cells fabricated via different routes from MAI/2-proponal solution with MAI concentration of (a) 20 mg mL⁻¹ and (b) 30 mg mL⁻¹, respectively. (c) UV-vis absorption and photoluminescence (PL) spectra. (d) Steady-state photocurrent and PCE at the maximum power point for the perovskite solar cells fabricated from MAI/2-proponal solution with 20 mg mL⁻¹ MAI.

Table 1 Photovoltaic performance of the best PSCs fabricated via two fabrication routes from MAI/2-propanol solutions with different MAI concentrations

Concentration of MAI	Fabrication route	Voc (V)	Jsc (mA cm^−2)	FF	PCE (%)	Rs (Ω cm^2)	Rsh (Ω cm^2)
10 mg mL^−1	Typical TSM	1.03	15.38	0.458	7.26	14.9	174
	Lewis adduct reaction	0.97	14.21	0.589	8.12	13.3	444
20 mg mL^−1	Typical TSM	0.97	13.68	0.579	7.68	15.03	180
	Lewis adduct reaction	1.03	19.42	0.660	13.20	8.49	920
30 mg mL^−1	Typical TSM	0.99	15.37	0.514	7.29	16.48	667
	Lewis adduct reaction	0.91	20.08	0.590	10.7	6.80	140

---

Fig. 7 Box chart of the photovoltaic parameters extracted from the light J–V curves of a series of PSCs fabricated from typical TSM and Lewis adduct reaction in MAI/2-proponal solutions with different MAI concentrations. (a) V_oc, (b) J_sc, (c) FF, (d) PCE.

As illustrated in Fig. 6d, we also measured the dependence of photocurrent density (J) and PCE on time by applied a bias close to the maximum output power point. The PSC fabricated from Lewis adduct reaction exhibits higher stability than that from typical TSM. It is noted that the steady state of the J and PCE are lower than those of the J–V reverse scan measurement due to hysteresis behaviours (Fig. S7†). This phenomenon was also observed in many PSCs with evident hysteresis behaviours.^42–45 A photocurrent rise time, the time taken to reach 95% of the maximum stabilized power output, is used to qualitatively evaluate the quality and trap density of the perovskite layer.⁴⁵ The rise time reduces from 66 s to 35 s for PSCs, which indicates that the perovskite film fabricated from Lewis adduct reaction has higher quality and lower trap density than that from the typical TSM.

4. Conclusions

In summary, high quality perovskite films without residual PbI₂ are fabricated from the Lewis adduct of PbI₂·xDMF in concentrated MAI/2-propanol solution. The Lewis adduct of PbI₂·xDMF convert easily to perovskite by molecular exchange between DMF and MAI. The lattice volume of the PbI₂·xDMF changes little when the DMF is replaced by MAI, resulting in a smooth surface of the perovskite film. The perovskite solar cells fabricated from the Lewis adduct exhibit higher photovoltaic performance than those from PbI₂ though the typical two-step method.

Acknowledgements

This work was supported by National Natural Science Foundation of China (51172122), Shenzhen Jiawei Photovoltaic Lighting Co., Ltd, and Tsinghua University Initiative Scientific Research Program (20161080165).

References

1. T. M. Brenner, D. A. Egger, L. Kronik, G. Hodes and D. Cahen, Nat. Rev. Mater., 2016, 1, 15007.
2. H. S. Jung, Mater. Today, 2014, 18, 65.
3. H. S. Jung and N. G. Park, Small, 2015, 11, 10.
4. A. R. M. Yusoff and M. K. Nazeeruddin, J. Phys. Chem. Lett., 2016, 7, 851.
5. 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. Grätzel and N. G. Park, Sci. Rep., 2012, 2, 591.
6. J. Burschka, N. Pellet, S. J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin and M. Grätzel, Nature, 2013, 499, 316.
7. J. H. Im, I. H. Jang, N. Pellet, M. Grätzel and N. G. Park, Nat. Nanotechnol., 2014, 9, 927.
8. N. J. Jeon, J. H. Noh, W. S. Yang, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Nature, 2015, 517, 476.
9. W. S. Yang, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Science, 2015, 348, 1234.
10. X. Li, D. Bi, C. Yi, J.-D. Décoppet, J. Luo, S. M. Zakeeruddin, A. Hagfeldt and M. Grätzel, Science, 2016, 353, 58.
11. D. Liu, M. K. Gangishetty and T. L. Kelly, J. Mater. Chem. A, 2014, 2, 19873.
12. Y. Fu, F. Meng, M. B. Rowley, B. J. Thompson, M. J. Shearer, D. Ma, R. J. Hamers, J. C. Wright and S. Jin, J. Am. Chem. Soc., 2015, 137, 5810.
13. Q. Chen, H. Zhou, T. B. Song, S. Luo, Z. Hong, H. S. Duan, L. Dou, Y. Liu and Y. Yang, Nano Lett., 2014, 14, 4158.
14. D. H. Cao, C. C. Stoumpos, C. D. Malliakas, M. J. Katz, O. K. Farha, J. T. Hupp and M. G. Kanatzidis, APL Mater., 2014, 2, 091101.
15. S. Wang, W. Dong, X. Fang, Q. Zhang, S. Zhou, Z. Deng, R. Tao, J. Shao, R. Xia, C. Song, L. Hu and J. Zhu, Nanoscale, 2016, 8, 6600.
16. L. Wang, C. McCleese, A. Kovalsky, Y. Zhao and C. Burda, J. Am. Chem. Soc., 2014, 136, 12205.
17. T. Zhang, N. Guo, G. Li, X. Qian and Y. Zhao, Nano Energy, 2016, 26, 50.
18. H. Zhang, J. Mao, H. He, D. Zhang, H. L. Zhu, F. Xie, K. S. Wong, M. Grätzel and W. C. H. Choy, Adv. Energy Mater., 2015, 5, 1501354.
19. F. Liu, Q. Dong, M. K. Wong, A. B. Djurišic, A. Ng, Z. Ren, Q. Shen, C. Surya, W. K. Chan, J. Wang, A. M. C. Ng, C. Liao, H. Li, K. Shih, C. Wei, H. Su and J. Dai, Adv. Energy Mater., 2016, 6, 1502206.
20. E. Zheng, X. F. Wang, J. Song, L. Yan, W. Tian and T. Miyasaka, ACS Appl. Mater. Interfaces, 2015, 7, 18156.
21. X. Xia, H. Li, W. Wu, Y. Li, D. Fei, C. Gao and X. Liu, ACS Appl. Mater. Interfaces, 2015, 7, 16907.
22. P. Docampo, F. C. Hanusch, S. D. Stranks, M. Döblinger, J. M. Feckl, M. Ehrensperger, N. K. Minar, M. B. Johnston, H. J. Snaith and T. Bein, Adv. Energy Mater., 2014, 4, 1400355.
23. Q. Chen, H. Zhou, Z. Hong, S. Luo, H. S. Duan, H. H. Wang, Y. Liu, G. Li and Y. Yang, J. Am. Chem. Soc., 2014, 136, 622.
24. F. Hao, C. C. Stoumpos, Z. Liu, R. P. H. Chang and M. G. Kanatzidis, J. Am. Chem. Soc., 2014, 136, 16411.
25. J. Mao, H. Zhang, H. He, H. Lu, F. Xie, D. Zhang, K. S. Wong and W. C. H. Choy, RSC Adv., 2015, 5, 73760.
26. W. Li, J. Fan, J. Li, Y. Mai and L. Wang, J. Am. Chem. Soc., 2015, 137, 10399.
27. Y. Wu, A. Islam, X. Yang, C. Qin, J. Liu, K. Zhang, W. Peng and L. Han, Energy Environ. Sci., 2014, 7, 2934.
28. K. Liang, D. B. Mitzi and M. T. Prikas, Chem. Mater., 1998, 10, 403.
29. A. Wakamiya, M. Endo, T. Sasamori, N. Tokitoh, Y. Ogomi, S. Hayase and Y. Murata, Chem. Lett., 2014, 43, 711.
30. D. Shen, X. Yu, X. Cai, M. Peng, Y. Ma, X. Su, L. Xiao and D. Zou, J. Mater. Chem. A, 2014, 2, 20454.
31. S. Bae, S. J. Han, T. J. Shin and W. H. Jo, J. Mater. Chem. A, 2015, 3, 23964.
32. N. Ahn, D. Y. Son, I. H. Jang, S. M. Kang, M. Choi and N. G. Park, J. Am. Chem. Soc., 2015, 137, 8696.
33. J. W. Lee, J. W. Lee and N. G. Park, Acc. Chem. Res., 2016, 49, 311.
34. Y. Umebayashi, K. Matsumoto, M. Watanabe and S. Ishiguro, Phys. Chem. Chem. Phys., 2001, 3, 5475.
35. Y. Umebayashi, K. Matsumoto, I. Mekata and S. Ishiguro, Phys. Chem. Chem. Phys., 2002, 4, 5599.
36. M. A. Jungbanger and C. Curran, Nature, 1964, 202, 290.
37. H. Zheng, W. Wang, S. Yang, Y. Liu and J. Sun, RSC Adv., 2016, 6, 1611.
38. H. Zhang, J. Cheng, F. Lin, H. He, J. Mao, K. S. Wong, A. K. Y. Jen and W. C. H. Choy, ACS Nano, 2016, 10, 1503.
39. B. E. Cohen, S. Gamliel and L. Etgar, APL Mater., 2014, 2, 081502.
40. T. Zhang, M. Yang, Y. Zhao and K. Zhu, Nano Lett., 2015, 15, 3959.
41. S. Yang, Y. C. Zheng, Y. Hou, X. Chen, Y. Chen, Y. Wang, H. Zhao and H. G. Yan, Chem. Mater., 2014, 26, 6705.
42. F. Huang, Y. Dkhissi, W. Huang, M. Xiao, I. Benesperi, S. Rubanov, Y. Zhu, X. Lin, L. Jiang, Y. Zhou, A. Gray-Weale, J. Etheridge, C. R. McNeill, R. A. Caruso, U. Bach, L. Spiccia and Y. B. Cheng, Nano Energy, 2014, 10, 10.
43. A. K. Jena, A. Kulkarni, M. Ikegami and T. Miyasaka, J. Power Sources, 2016, 309, 1.
44. Y. Zhang, M. Li u, G. E. Eperon, T. C. Leijtens, D. McMeekin, M. Saliba, W. Zhang, M. Bastian, A. Petrozza, L. M. Herz, M. B. Johnston, H. Lin and H. J. Snaith, Mater. Horiz., 2015, 2, 315.
45. N. K. Noel, A. Abate, S. D. Stranks, E. S. Parrott, V. M. Burlakov, A. Goriely and H. J. Snaith, ACS Nano, 2014, 8, 9815.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra15378j

---