Enhanced performance of perovskite solar cells with solution-processed n-doping of the PCBM interlayer

Abstract

Here, we report a facile and efficient sequential n-doping method to increase the device performance of planar-type organic/inorganic perovskite solar cells. Compared to the pristine PCBM-based electron transport layer (ETL), power conversion efficiencies of cells with a doped ETL are dramatically enhanced up to 12.53%.

---

Organic/inorganic hybrid perovskite solar cells (PSCs) based on lead halide (CH₃NH₃PbX_aY_(3−a), X, Y = Cl, Br, or I) have recently attracted considerable attention as promising next generation photovoltaics because of their cost-efficiency, solution-processability, and high photovoltaic performance.^1–4 Several studies have reported that perovskite materials have unique properties such as high absorption coefficient, long carrier diffusion lengths, and efficient exciton dissociation.^3–8 By utilizing the above distinguishing properties, several groups have reported high performance PSCs typically by depositing a perovskite material onto a mesoporous TiO₂ scaffold acting as an efficient electron transport layer (ETL);^9–11 the power conversion efficiency (PCE) of PSCs of over 20% was thus recently realized.¹² Followed by a successful demonstration of mesostructured TiO₂-based PSCs, a planar heterojunction (PHJ)-based PSC with a configuration of indium tin oxide (ITO)/interfacial layer/perovskite/interfacial layer/top electrode has also been actively studied.^8,10,13 In particular, considering that a high-temperature sintering step is necessary to make the TiO₂ scaffold, the advantages of the simple and low-temperature manufacturing of the PHJ PSC make it more attractive for practical application, especially in flexible architectures.¹⁴

Perovskite film morphology plays an important factor in PHJ PSCs for improving photovoltaic parameters such as open circuit voltage (V_oc), fill factor (FF), and short circuit current density (J_sc). To date, most of the works have therefore focused on controlling perovskite film morphology through sequential deposition, additive inclusion, solvent engineering, etc.^2,8,14–18 Aside from controlling the film morphology, it is well-known that hole and electron interfacial layers, which prevent the perovskite photoactive layer from directly contacting electrodes, have an influence on the device performance.^14,18 In addition, the energy level, carrier mobility, and electrical property of interfacial materials are critical for obtaining high-efficiency PSCs.^19,20 An n-type semiconducting material, phenyl-C₆₁-butyric acid methylester (PCBM) has been widely used as an ETL in PHJ PSCs due to its high electron accepting property, suitable energy level, and orthogonal solvent solubility.^8,13 In PCBM-based PHJ PSCs, despite the high solubility of PCBM in organic non-polar solvent, a low viscosity of PCBM solution can form small cracks in the film,^14,18 functioning as a shunting pass, eventually leading to a decrease in V_oc and FF. Furthermore, carrier recombination arising from an imperfect PCBM film morphology can lead to a decreased J_sc.^14,18 To this end, a PCBM layer having enough thickness to fully cover the underlying perovskite and to avoid direct contact between the perovskite and electrode has been normally used. However, the thickness of the PCBM layer could be quite limited because a layer that is too thick can also lead to an increased series resistance (R_s) ascribed to the low electrical property of PCBM, thus reducing the overall efficiency.^14,18 Therefore, the electrical properties of PCBM-based ETL need to be modified to realize high-performance PHJ PSCs.

One strategy for modifying the electrical conductivity of PCBM is to use a suitable n-type dopant. While in general, many n-dopants can be used, the challenge of n-doping an electron transporting semiconductor is that the n-type dopant governed by a one-electron transfer doping pathway has a low ionization energy, which causes air-instability.^21,22 In previous reports, 1,3-dimethyl-2-phenyl-2,3-dihydro-1H-benzoimidazole (DMBI) was reported as an efficient and air-stable n-type dopant for fullerene derivatives through the hydride radical transfer pathway irreversibly.^21,22 From these reports, the conductivity of doped PCBM film could be more than 4 orders of magnitude higher than that of un-doped PCBM film.²² In particular, since the DMBI is a highly alcohol soluble, a direct solution-based deposition on top of the PCBM film is possible without the dissolution of either the underlying PCBM or the perovskite film. This feature can eliminate and fill unintentionally formed cracks in the PCBM film by DMBI deposition, thus eventually increasing the shunt resistance (R_sh) and FF values. However, despite the aforementioned advantages of the DMBI, to date few approaches have been available for using the DMBI as an n-type doping agent of fullerene derivatives in optoelectronics.^22–24 In particular, considering that studies on increasing electrical conductivity of a hole transport layer using new p-doping agents were carried out to enhance device performances in the field of PSCs,^25–27 it is urgently needed to develop a new n-dopant for fullerene derivatives and to investigate the n-doping effect on the cell-performances.

In this study, we introduced a facile solution-processed DMBI as an n-type doping agent for PCBM ETL, and the effects of the DMBI-treated PCBM on the device performances in planar heterojunction perovskite solar cells were investigated. From the conductivity and work-function investigations, it is revealed that PCBM is efficiently doped by the DMBI molecules. In particular, dramatically increased performance of PSCs was demonstrated by introducing the DMBI as a PCBM dopant to the PSCs (PCE of 12.53%; V_oc of 0.96 V; J_sc of 16.42 mA cm⁻²; FF of 79.12%), which is attributed mostly to the enhanced conductivity of PCBM film and the decreased work-function of PCBM by the sequential DMBI-coating. These results indicate that solution-processable DMBI is a promising n-type PCBM dopant for low-cost and efficient PSCs.

To evaluate the effects of PCBM-based ETLs with and without n-type doping using the DMBI molecule on cell-performance in CH₃NH₃PbI₃-based PHJ PSCs (see Fig. 1a and b), different types of PSCs were fabricated: with a pristine PCBM layer as a reference ETL; with PCBM layers each doped with 0.4, 0.7, and 1 wt% DMBI. Additionally, two fabrication methods were used for making doped PCBM films: (1) spin-casted from a pre-mixed solution of PCBM and DMBI in chlorobenzene and (2) sequential casting of different amounts of DMBI solution of 0.4–1 wt% in isopropyl alcohol onto the pre-formed PCBM layer. The detailed experimental methods and procedures are presented in the ESI.† Fig. 2a and b exhibit the representative current density–voltage (J–V) curves, and the corresponding device parameters are summarized in Table 1. For the cells with doped PCBM ETLs made from a pre-mixed solution of PCBM and DMBI, the photovoltaic performances were slightly reduced compared to those of un-doped PCBM-based cells, as shown in Fig. 2a. On the other hand, compared to the un-doped PCBM-based PSCs, all photovoltaic parameters were improved by the sequential doping of the PCBM layer, regardless of the doping levels up to 1 wt%, as shown in Fig. 2b–d, Tables 1 and S1.† The best performance was obtained in cells doped sequentially with 0.7 wt% DMBI. In detail, with an increase in the amount of DMBI dopant from 0.4 to 0.7 wt%, efficiencies gradually improved up to ∼12.5%; however, higher doping than 0.7 wt% led to a slightly decreased PCE of 9.47%. Since the sequential doping method showed more improved cell-performances than those of the pre-mixed case, hereafter we further focus on the sequential deposition method and the effects of the DMBI-treated PCBM on PSC-performances. More importantly, these results indicate that a simple introduction of a doped PCBM layer by the sequential method into PSCs can be a very effective way to increase device performances.

Fig. 1 (a) Chemical structures of DMBI and DMBI-doped PCBM. (b) Device configuration of the planar heterojunction (PHJ) perovskite solar cells and doping methods used in this study.

Fig. 2 Representative J–V curves of cells processed by (a) the pre-mixed method and (b) the sequential deposition method. Photovoltaic performances (c) V_oc, J_sc, (d) FF, PCE of PSCs with un-doped and DMBI-doped PCBM films.

Table 1 Photovoltaic parameters of CH₃NH₃PbI₃-based PSCs with different DMBI doping methods and concentrations

	ETL	Voc (V)	Jsc (mA cm^−2)	FF (%)	PCE (%)
Sequential deposition method	Pristine	0.86	12.75	74.51	8.15
	0.4 wt%	0.99	14.33	75.21	10.72
	0.7 wt%	0.96	16.42	79.12	12.53
	1 wt%	0.93	13.1	77.71	9.47
Pre-mixed method	0.4 wt%	0.77	12.96	76.48	7.65
	0.7 wt%	0.79	13.07	74.83	7.73
	1 wt%	0.79	11.48	72.54	6.58

---

From the above J–V analysis, we verified that a doped form of PCBM processed by a sequential deposition method is better suited to ETL; however, the reasons for the performance deviations among the PCBM-based ETLs without and with doping treatments remain unclear. In particular, comparing the pristine PCBM with the doped PCBM-based PSCs, we observed increased cell-parameters such as FF, V_oc, and J_sc. It is well-known that FF is closely related to the R_s and R_sh values,²⁸ and it has been shown that a reduced R_s resulting from the doping-induced conductivity enhancement of an interfacial layer can lead to improved FF and J_sc in photovoltaic cells,^29,30 while the V_oc can be enhanced with a favorable work-function (WF) change by n- or p-type doping.^31–33 We therefore attempted to find doping evidence for DMBI-treated PCBM film and the effects of DMBI-treated PCBM on cell-parameters.

PCBM-doping evidence could be obtained by modifying the electrical property according to previous reports;^21,34 we thus conducted electrical conductivity using a two-terminal probe station, as can be seen in Fig. 3a. The electrical conductivity of the doped and un-doped PCBM films was measured using device configuration of the top gate/bottom contact where DMBI was inserted between source/drain electrodes and the PCBM film. With DMBI, the electrical conductivity of the pristine PCBM film was dramatically increased from 1.06 × 10⁻⁹ to 4.15 × 10⁻⁵ S cm⁻¹, which is 4 orders of magnitude higher than that of the pristine PCBM film, demonstrating that DMBI doping treatment can dramatically increase the electrical property of PCBM film. Considering that FF and J_sc are closely related to R_s, having a strong influence on material conductivity,^29,30 and considering that the electrical conductivity of PCBM film was shown to be improved by DMBI doping, the improved FF and J_sc are attributed to a reduced R_s by doping PCBM with DMBI, which can explain the reason for the improved FF and J_sc in the J–V curves of Fig. 2b. Further representative evidence for PCBM doping could be explained by the WF changes of PCBM;^21,35 thus, we also studied the changes in the PCBM WFs induced by DMBI materials measured using scanning kelvin probe microscopy (SKPM). As shown in Fig. 3b, the measured WF of un-doped PCBM film was 4.48 eV, while with increasing DMBI concentrations, the WF value was up-shifted to 4.10 eV, indicating that PCBM is n-doped by DMBI.²¹ In addition, as shown in Fig. 3c, the CN⁺ peak related to the ion mass for DMBI dopant is gradually decreased from top to bottom PCBM, which can support that the DMBI doping could not be occurred only in a distinct surface of PCBM. Furthermore, the DMBI layer on top of PCBM layer can also function as an additional buffer layer to prevent the direct contact between the metal and the perovskite,^36,37 which could also be responsible for the improved PSC-performances shown in Fig. 2b–d. Notably, considering that the potential energy-loss at the perovskite/ETL interface can be highly reduced due to the decreased WF of the PCBM by DMBI-doping (which can improve the built-in potential and electron extraction, thereby improving the V_oc, FF, and J_sc),^31–33 the reduced WF could also be responsible for the enhanced PSC-parameters shown in the DMBI-doped PCBM based PSCs. Although the origin of the overall increased cell performances can be mostly explained by doping-induced conductivity enhancement and the decreased WF of DMBI-doped PCBM, the overall cell parameters did not improve further with increasing DMBI concentrations beyond 0.7 wt%, suggesting that other factors are involved in the changed device efficiency besides the doping-induced conductivity and WF improvement.

Fig. 3 (a) Electrical properties of un-doped and DMBI-doped PCBM films. (b) Work function of pristine PCBM, 0.4 wt% DMBI, 0.7 wt% DMBI, and 1 wt% DMBI. Inset: energy level diagram of the components in PSC with PCBM or PCBM/DMBI ETL. (c) TOF-SIMS analysis of the cell with DMBI treatment.

Typically, an interfacial film property has a key role in transporting a carrier from perovskite film to electrodes;³⁸ thus, we conducted morphological analysis by SEM and AFM measurement, as shown in Fig. 4. As shown in Fig. 4, the pristine PCBM showed a relatively smooth and flat film morphology and a low rms roughness of 1.472 nm, but with increasing doping concentration of DMBI from 0.4 to 1 wt%, there is a tendency of larger aggregation of DMBI and rougher surface morphology. Considering that an interfacial layer with rough and non-uniform morphology can cause a carrier recombination to produce decreased R_sh, V_oc, and FF,^14,18,19 and the higher DMBI concentration can produce a rougher surface and a thicker DMBI to restrict electron flows from perovskite to the cathode (thus resulting in overall reduction in R_s and FF),^16,39–41 the slightly decreased PCE shown beyond 0.7 wt% DMBI-doping could be attributed to the larger DMBI aggregation and rougher surface shown in 1 wt% DMBI concentration, as can be confirmed from the observed R_s (pristine: 3.023 Ω cm², 0.4 wt%: 2.83 Ω cm², 0.7 wt%: 2.48 Ω cm², and 1 wt%: 3.72 Ω cm²) and R_sh (pristine: 2.505 × 10³ Ω cm², 0.4 wt%: 3.162 × 10³ Ω cm², 0.7 wt%: 4.793 × 10³ Ω cm², and 1 wt%: 4.197 × 10³ Ω cm²). As a result, the highest efficiency was obtained in PSCs doped with 0.7 wt% DMBI since as the DMBI concentration increases, there is a trade-off between more enhanced doping effects such as increased conductivity and decreased WF shown in Fig. 3a and b to produce better cell-performances and poorer film morphology with rougher surface shown in Fig. 4 to induce worse cell-performances.

Fig. 4 SEM top-view images (a–d) of pristine PCBM film and DMBI-doped PCBM film: (a) pristine PCBM, (b) 0.4 wt% DMBI, (c) 0.7 wt% DMBI, and (d) 1 wt% DMBI. Topographic AFM images (e–h) of PCBM film and DMBI-doped PCBM film (5 μm × 5 μm): (e) pristine PCBM, rms = 1.472 nm; (f) 0.4 wt% DMBI-doped PCBM, rms = 4.435 nm; (g) 0.7 wt% DMBI-doped PCBM, rms = 7.454 nm; and (h) 1 wt% DMBI-doped PCBM, rms = 17.551 nm.

Similarly, the reason for a reduced cell efficiency of pre-mixed PCBM:DMBI-based cells may be explained; cells with pre-mixed PCBM:DMBI layer showed a reduced V_oc, J_sc, and FF compared to cells with sequential deposition method due to the fact that DMBI aggregates in solution state (see Fig. S1†) can cause inhomogeneous and rough surface morphology acting as charge recombination/trapping sites and leakage pathways, thus resulting in poor J_sc, V_oc, FF, and PCEs, compared to those of cells prepared by the sequential deposition method.

As photocurrent–voltage hysteresis is a major issue in the accurate evaluation of device performance in PSCs,⁴² we measured the J–V curves for cells with un-doped PCBM and 0.7 wt% DMBI-doped PCBM as a reference and the best performing photovoltaic cell, respectively, by changing sweep directions and scan rates. Fig. 5 presents the J–V curves of pristine and 0.7 wt% DMBI-doped PCBM based cells measured using different scan directions and rates. When measured by scanning from a positive (+1.2 V) to a negative (−0.4 V) direction and from a negative (−0.4 V) to a positive (+1.2 V) direction, a slightly larger J–V hysteresis was observed in DMBI-doped cells in comparison with pristine cells (see Fig. 5a). In addition, with decreasing scan time from 500 to 0 ms, a slightly increasing J–V hysteresis emerged in DMBI-doped cells. The origin of the scan-direction and rate dependent J–V hysteresis in both PSCs could be explained as follows; when the perovskite and PCBM penetrate efficiently into one another, mobile ions in perovskite and PCBM molecules can form fullerene halide radical anions, thus resulting in the negligible hysteresis in J–V curves.³⁶ In the case of doped PCBM, it was believed that it is more difficult to form fullerene halide radical anion probably due to binding up both rich halide antisites and unincorporated halides with the pristine case.^36,42 Thus the ions in perovskite will be freer to move around with changing scan-direction and rate, resulting in a slight hysteresis in DMBI-doped cells, as presented in a previous report.^36,42 We also measured the external quantum efficiency (EQE) and internal quantum efficiency (IQE) of both cells. As shown in Fig. 5c, the J_sc calculated from the EQE data shown in the DMBI-doped PCBM device was 15.83 mA cm⁻², which is very similar to the J_sc obtained from the J–V power curve. More importantly, as shown in Fig. 5c, the DMBI-doped PCBM device produced higher EQE and IQE compared to those of the pristine PCBM device. The EQE is increased by ∼35% for DMBI-doped PCBM relative to pristine PCBM, and the degree of the EQE enhancement is similar to that of the IQE increment, indicating that the enhanced IQE is responsible mainly for the increased EQE, and thus we can also confirm that the enhanced EQE and J_sc shown in the DMBI-doped PCBM PSCs could be attributed mainly to the improved charge-transport and charge-collection efficiency from the perovskite layer to electrodes, resulting from the use of DMBI-doped PCBM.^43,44

Fig. 5 J–V curves of the pristine and 0.7 wt% DMBI-doped PCBM based cells measured with (a) different scan directions and (b) different scan rates. (c) EQE and IQE of PSCs with PCBM layer without and with doping of 0.7 wt% DMBI.

Conclusions

In conclusion, we demonstrated that solution-processable DMBI can be used as an efficient PCBM dopant for efficient PSCs. PCBM properties were dramatically modified by DMBI doping, and most of the cell-parameters with DMBI-treated PCBM were improved. The best performing PSC shows a PCE of 12.53%, V_oc of 0.96 V, J_sc of 16.42 mA cm⁻², and FF of 79.12%, which are higher values than those of the un-doped PCBM-based cell (PCE = 8.15%; V_oc = 0.86 V; J_sc = 12.75 mA cm⁻²; FF = 74.51%). Compared to cells with un-doped PCBM ETLs, it is revealed that the overall improved device characteristics shown in doped PCBM-based cells originate mainly from the increased electrical conductivity and decreased work-function of PCBM film by DMBI n-doping. These results suggest that a simple introduction of a doped PCBM layer by the DMBI treatment into PSCs can be a promising strategy to increase the device performances of PHJ PSCs.

Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (No. 2012M2A2A6013183).

Notes and references

1. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643.
2. J. Burschka, N. Pellet, S. J. Moon, R. H. Baker, P. Gao, M. K. Nazeeruddin and M. Grätzel, Nature, 2013, 499, 316.
3. J.-H. Im, C.-R. Lee, J.-W. Lee, S.-W. Park and N.-G. Park, Nanoscale, 2011, 3, 4088.
4. Q. Wang, Y. Shao, Q. Dong, Z. Xiao, Y. Yuan and J. Huang, Energy Environ. Sci., 2014, 7, 2359.
5. H. J. Snaith, J. Phys. Chem. Lett., 2013, 4, 3623.
6. N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu and S. I. Seok, Nat. Mater., 2014, 13, 897.
7. C. C. Homes, T. Vogt, S. M. Shapiro, S. Wakimoto and A. P. Ramirez, Science, 2001, 293, 673.
8. G. Xing, N. Mathews, S. Sun, S. S. Lim, Y. M. Lam, M. Grätzel, S. Mhaisalkar and T. C. Sum, Science, 2013, 342, 344.
9. B. Yang, O. Dyck, J. Poplawsky, J. Keum, A. Puretzky, S. Das, I. Ivanov, C. Rouleau, G. Duscher, D. Geohegan and K. Xiao, J. Am. Chem. Soc., 2015, 137, 9210.
10. N. J. Jeon, J. H. Noh, W. S. Yang, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Nature, 2015, 517, 476.
11. K. Wojciechowski, M. Saliba, T. Leijtens, A. Abate and H. J. Snaith, Energy Environ. Sci., 2014, 7, 1142.
12. W. S. Yang, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Science, 2015, 348, 1234.
13. S. D. Stranks, G. E. Eperon, G. Grancini, C. Menelaou, M. J. P. Alcocer, T. Leijtens, L. M. Herz, A. Petrozza and H. J. Snaith, Science, 2013, 342, 341.
14. J. You, Z. Hong, Y. Yang, Q. Chen, M. Cai, T.-B. Song, C.-C. Chen, S. Lu, Y. Liu, H. Zhou and Y. Yang, ACS Nano, 2014, 8, 1674.
15. D. Bi, S.-J. Moon, L. Häggman, G. Boschloo, L. Yang, E. M. J. Johansson, M. K. Nazeeruddin, M. Grätzel and A. Hagfeldt, RSC Adv., 2013, 3, 18762.
16. Y.-J. Jeon, S. Lee, R. Kang, J.-E. Kim, J.-S. Yeo, S.-H. Lee, S.-S. Kim, J.-M. Yun and D.-Y. Kim, Sci. Rep., 2014, 4, 6953.
17. 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.
18. Y. Bai, H. Yu, Z. Zhu, K. Jiang, T. Zhang, N. Zhao, S. Yang and H. Yan, J. Mater. Chem. A, 2015, 3, 9098.
19. J. Kim, G. Kim, T. K. Kim, S. Kwon, H. Back, J. Lee, S. H. Lee, H. Kang and K. Lee, J. Mater. Chem. A, 2014, 2, 17291.
20. C. Sun, Q. Xue, Z. Hu, Z. Chen, F. Huang, H.-L. Yip and Y. Cao, Small, 2015, 11, 3344.
21. B. D. Naab, S. Gao, S. Olthof, E. G. B. Evans, P. Wei, G. L. Millhauser, A. Kahn, S. Barlow, S. R. Marder and Z. Bao, J. Am. Chem. Soc., 2013, 135, 15018.
22. P. Wei, J. H. Oh, G. Dong and Z. Bao, J. Am. Chem. Soc., 2010, 132, 8852.
23. S. Shao, Z. Chen, H.-H. Fang, G. H. ten Brink, D. Bartesaghi, S. Adjokatse, L. J. A. Koster, B. J. Kooi, A. Facchetti and M. A. Loi, J. Mater. Chem. A, 2016, 4, 2419.
24. S. S. Kim, S. Bae and W. H. Jo, Chem. Commun., 2015, 51, 17413.
25. A. Abate, D. J. Hollman, J. Teuscher, S. Pathak, R. Avolio, G. D'Errico, G. Vitiello, S. Fantacci and H. J. Snaith, J. Am. Chem. Soc., 2013, 135, 13538.
26. J. H. Noh, N. J. Jeon, Y. C. Choi, M. K. Nazeeruddin, M. Grätzel and S. I. Seok, J. Mater. Chem. A, 2013, 1, 11842.
27. K. Neumann and M. Thelakkat, RSC Adv., 2014, 4, 43550.
28. M.-S. Kim, B.-G. Kim and J. Kim, ACS Appl. Mater. Interfaces, 2009, 1, 1264.
29. H. Ma, H.-L. Yip, F. Huang and A. K.-Y. Jen, Adv. Funct. Mater., 2010, 20, 1371.
30. C. J. Brabec, V. Dyakonov, J. S. Parisi and N. S. Sariciftci, Organic Photovoltaics: Concepts and Realization, Springer, Berlin, 2003.
31. K.-G. Lim, H.-B. Kim, J. Jeong, H. Kim, J. Y. Kim and T.-W. Lee, Adv. Mater., 2014, 26, 6461.
32. S.-H. Kim, C.-H. Lee, J.-M. Yun, Y.-J. Noh, S.-S. Kim, S. Lee, S. M. Jo, H.-I. Joh and S.-I. Na, Nanoscale, 2014, 6, 7183.
33. F. Zhang, M. Ceder and O. Inganäs, Adv. Mater., 2007, 19, 1835.
34. S. Fabiano, S. Braun, X. Liu, E. Weverberghs, P. Gerbaux, M. Fahlman, M. Berggren and X. Crispin, Adv. Mater., 2014, 26, 6000.
35. E.-S. Choi, Y.-J. Jeon, S.-S. Kim, T.-W. Kim, Y.-J. Noh, S.-N. Kwon and S.-I. Na, Appl. Phys. Lett., 2015, 107, 023301.
36. L. Meng, J. You, T.-F. Guo and Y. Yang, Acc. Chem. Res., 2016, 49, 155.
37. C.-C. Chueh, C.-Z. Li and A. K.-Y. Jen, Energy Environ. Sci., 2015, 8, 1160.
38. H. Kim, K.-G. Lim and T.-W. Lee, Energy Environ. Sci., 2015, 9, 12.
39. R. Hamilton, J. Smith, S. Ogier, M. Heeney, J. E. Anthony, I. McCulloch, J. Veres, D. D. C. Bradley and T. D. Anthopoulos, Adv. Mater., 2009, 21, 1166.
40. C.-H. Chiang, Z.-L. Tseng and C.-G. Wu, J. Mater. Chem. A, 2014, 2, 15897.
41. J.-W. Lee, D.-J. Seol, A.-N. Cho and N.-G. Park, Adv. Mater., 2014, 26, 4991.
42. J. Xu, A. Buin, A. H. Ip, W. Li, O. Voznyy, R. Comin, M. J. Yuan, S. Jeon, Z. J. Ning, J. J. McDowell, P. Kanjanaboos, J. P. Sun, X. Z. Lan, L. N. Quan, D. H. Kim, I. G. Hill, P. Maksymovych and E. H. Sargent, Nat. Commun., 2015, 6, 7081.
43. J.-S. Yeo, J.-M. Yun, D.-Y. Kim, S.-S. Kim and S.-I. Na, Sol. Energy Mater. Sol. Cells, 2013, 114, 104.
44. M. Hu, C. Bi, Y. Yuan, Z. Xiao, Q. Dong, Y. Shao and J. Huang, Small, 2015, 11, 2164.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra13082h

---