Enhanced charge collection with passivation of the tin ...

Journal of

Published on 06 June 2017. Downloaded by ECOLE POLYTECHNIC FED DE LAUSANNE on 15/06/2017 16:51:31.

Materials Chemistry A View Article Online

COMMUNICATION

Cite this: DOI: 10.1039/c7ta04128d

Received 12th May 2017 Accepted 6th June 2017

View Journal

Enhanced charge collection with passivation of the tin oxide layer in planar perovskite solar cells† Yonghui Lee, *a Sanghyun Paek,a Kyung Taek Cho, a Emad Oveisi, b Peng Gao,a Seunghwan Lee,c Jin-Seong Park, c Yi Zhang,a Robin Humphry-Baker,a Abdullah M. Asiri d and Mohammad Khaja Nazeeruddin *a

DOI: 10.1039/c7ta04128d rsc.li/materials-a

Tin oxide is an excellent candidate to replace mesoporous TiO2 electron transport layers (ETLs) in perovskite solar cells. Here, we introduced a SnO2 layer by a low-temperature solution process, and investigated its morphology, opto-physical and electrical properties affecting the device performance. We reveal that low-temperature processed SnO2 is self-passivating in nature, which leads to a high efficiency. To further enhance the blocking effect, we combined a compact TiO2 underlayer with the SnO2 contact layer, and found that the bi-layered ETL is superior compared to single layers. The best device shows photovoltaic values in a planar structure with a shortcircuit current density (Jsc) of 22.58 mA cm 2, an open-circuit voltage (Voc) of 1.13 V, a fill factor (FF) of 0.78, and a power conversion efficiency (PCE) of 19.80% under 1 sunlight illumination.

Introduction Perovskite solar cells have been of great interest over the past few years. Since the rst study by Kojima et al.,1 a large number of topics on nding better compositions of perovskite, interfacial engineering, electron/hole transport materials and methods for layer deposition have been investigated, which boosted the power conversion efficiency (PCE) up to 22.1%.2–9 In perovskite solar cells, inorganic TiO2 ETL based device structures, i.e. uorine doped SnO2 (FTO) glass/TiO2 (n)/ perovskite (i)/hole transport material (HTM) (p)/Au have been most widely used.10,11 The main reasons for using TiO2 (anatase) a

Group for Molecular Engineering of Functional Materials, EPFL Valais Wallis, CH-1951 Sion, Switzerland. E-mail: yonghui.lee@ep.ch; mdkhaja.nazeeruddin@ ep.ch

b Interdisciplinary Centre for Electron Microscopy, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland c Division of Materials Science and Engineering, Hanyang University, 222 Wangsimniro, Seongdong-gu, Seoul 133-791, Korea d

Center of Excellence for Advanced Materials Research (CEAMR), King Abdulaziz University, P. O. Box 80203, Jeddah 21589, Saudi Arabia † Electronic supplementary 10.1039/c7ta04128d

information

(ESI)

available.

This journal is © The Royal Society of Chemistry 2017

See

DOI:

are its favorable conduction band (CB) edge, 4.3 to 4.0 eV, in relation to several perovskites with a wide band gap of 3.2 eV as well as good hole blocking ability due to its deep valence band (VB) position.12,13 The ease in lm preparation with diverse processes and environmental stability are great merits of TiO2. However, a relatively low electron mobility that causes insufficient charge separation at the compact TiO2 (c-TiO2)/perovskite interface has oen been an issue.4,14–16 Thus, an additional layer of mesoporous TiO2 (mp-TiO2) is usually used to complement the decient charge collection. However, such complicated structures may bring other risks, namely efficiency degradation and poor reproducibility of the device by bad pore-lling with the increase of fabrication costs.8 Although organic ETLs such as phenyl-C61-butyric acid methyl ester (PCBM) have shown successful performance to replace compact and mesoporous TiO2,17,18 their application could be limited by unexpected obstacles. For example, the organic ETL could be damaged during the solvent-dropping process for the post-deposition of perovskite layers in the FTO/ETL/perovskite/HTM/Au structure. Therefore, it is necessary to explore new materials and structures that can provide both electrical and chemical excellence for planar-type perovskite solar cells. In this respect, SnO2 seems to be a promising candidate to fulll these necessities. SnO2 is an n-type semiconductor, and known to have a high conductivity with a wide bandgap ranging between 3.6 and 4.1 eV, which is suitable for a transparent electron transport layer.19 Accordingly, there has been diverse research with SnO2 ETLs in perovskite solar cells since the early work by Dong et al.20 Although Jiang et al. showed a remarkably high PCE of 20.5% and long-term stability by enhancing electron extraction with SnO2 in planar structures,21 further research should be carried out considering the great potential of SnO2 as an inorganic ETL. Here, we introduced a SnO2 layer by a low-temperature solution process to replace the mp-TiO2 structure. The morphological and opto-physical properties of the lm, and electrical properties combined with device performance were studied. We reveal that low-temperature processed SnO2 is self-

J. Mater. Chem. A

View Article Online


Journal of Materials Chemistry A

passivating in nature, which is strongly associated with the device performance. To enhance the blocking effect, we combined a thin and compact TiO2 with the SnO2 contact layer. It was found that our cells using bi-layered ETLs (BETLs) could be superior to single-layered ones.

Results and discussion SnO2 layers can be formed in various ways from diverse sources and techniques.22 In this work, we prepared a SnO2 layer by directly spin-coating SnCl4 (bp 114 C) dissolved in water on a FTO substrate while SnCl2$2H2O dissolved in ethanol or crystalline SnO2 colloids were mainly used in recent papers.23–26 The lms are post-annealed at low-temperature (180 C) to obtain improved conductivity with a good surface coverage.23 It was veried by reectance measurements that the lm has an optical bandgap of 3.96 eV (Fig. 1a), which is very close to 3.95 eV shown in the recent report using SnCl2$2H2O.27 A schematic of the cross-sectional structure and a photo of the perovskite solar cell with low-temperature solution-processed tin oxide (LTO) are shown in Fig. 1b. The reference device has a planar structure comprising FTO glass/LTO/perovskite/polytriarylamine (PTAA)/Au. The composition of the perovskite was xed as (FAPbI3)0.85(MAPbBr3)0.15 with 5.7 mol% enriched PbI2 to obtain a dense morphology and high efficiency,3,28 which was conrmed by the X-ray diffraction (XRD) pattern and absorbance as shown in Fig. S1 and S2 (ESI†). The device performance was compared by modulating the lm thickness (Fig. 1c). As seen, for a device with a thin LTO layer, a small J–V curve hysteresis in the scanning direction could be promising, but a small Voc value is also observed maybe due to a shunt caused by pinholes generated during spin-coating and postannealing. On the other hand, the other device with a thicker

Communication

LTO demonstrates a good Voc, but a relatively low ll factor (FF) is shown along with a strong J–V curve hysteresis owing to the increased thickness of LTO. We found that an 100 nm thick layer is necessary to obtain a good Voc and reproducible device performance using this method, which accords well with the previous report.20 The newly designed ETL combines both benets of higher electron mobility from SnO2 and a good hole blocking ability from TiO2. Fig. 1d shows an energy level diagram of the TiO2/LTO based perovskite solar cells. As seen, the cascade structure for smooth electron transfer is not hampered even aer insertion of the TiO2 underlayer. The bandgap edge positions of TiO2 and LTO were taken from ref. 13, 27 and 29. Prior to the investigation on the inuence of c-TiO2 underlayer, we explored the LTO layer with transmission electron microscopy (TEM) and scanning electron microscopy (SEM). Fig. 2a and b show the TEM images of the heat-treated LTO lm at 100 C and 180 C for 1 h, respectively. In Fig. 2a, we nd that the layer has a mixed phase of crystalline particles surrounded by an amorphous phase, which is very similar to the lm morphology prepared by chemical bath deposition (CBD) with SnCl4 solution.30 Using high-resolution TEM (HR-TEM), we can see crystalline regions of 2 to 3 nm (in diameter), that are supposed to be SnO2 crystals as displayed in Fig. 2a and S3 (ESI†). The similarity of the lm nature with that of the CBD lm can be explained by tracing the probable chemical reactions. Anhydrous SnCl4 dissolved in water was used as a precursor for crystalline SnO2, otherwise amorphous SnOCl2 could be easily produced in the solution and on the FTO substrate during spin-coating as indicated by the following equations:31 SnCl4 + 2H2O / SnO2 + 4HCl

(1)

Fig. 1 Planar perovskite solar cells with low-temperature processed tin oxide (LTO) layers. (a) Optical bandgap of the LTO film calculated by reflectance measurements. (b) Schematic of the perovskite cell, and a photo of the real devices. (c) J–V curve hysteresis according to LTO film thickness. (d) Energy bandgap diagram of the TiO2/LTO based perovskite solar cells.

J. Mater. Chem. A


View Article Online


Communication


Fig. 2 Microstructure of the passivated tin oxide (PTO) film and the perovskite cell. (a) HR-TEM image of the LTO film dried at 100 C for 1 h. (b) TEM images and SAED pattern of the LTO film. The samples are heat-treated at 180 C for 1 h. (c) SEM cross-sectional image of the FTO/PTO film heat-treated at 180 C for 1 h. (d) SEM top-view images of bare FTO and the FTO/PTO substrate, respectively. (e) SEM cross-sectional image of the perovskite solar cell comprising FTO/PTO/perovskite/PTAA/Au.

SnCl4 + 4H2O / Sn(OH)4 + 4HCl Sn(OH)4 / H2SnO3 + H2O H2SnO3 + 2HCl / SnOCl2 + 2H2O

(2)

In general, as-prepared LTO lms prepared with chlorine precursors require an additional heat treatment to improve their electrical properties.23,27 In this step, we can expect that further conversion of SnOCl2 to SnO2 accompanied by the loss of chlorine as gas occurs. Due to the crystal growth at 180 C, slightly bigger particles are formed in the amorphous matrix (Fig. 2b) that contain Sn, O, and Cl elements, as conrmed by energy dispersive X-ray (EDX) data shown in Fig. S4 (ESI†). The selected area electron diffraction (SAED) pattern (Fig. 2b inset) shows diffused rings, indicating the short-range order structure of the lm. We conrmed by XRD analysis that the lm has an amorphous nature as shown in Fig. S5 (ESI†). On the basis of our observation, we can conclude that the layer is composed of crystalline SnO2 passivated by SnOCl2, and hereaer the layer is better described as passivated tin oxide (PTO). In the lower magnication TEM image (Fig. 2b), the lm displays a layer form, and such a morphology is similarly shown in the SEM cross-sectional and top-view images in Fig. 2c and d, respectively. It appears that a 20 nm thick PTO layer with excellent surface coverage is formed over the FTO substrate. However, the use of a low concentration precursor solution and the postannealing process involving volume shrinkage by the loss of SnOCl2 may increase the chance for the formation of surface defects. The cross-sectional SEM image of the complete device is shown in Fig. 2e. For the complete cells, it is seen that a 600 nm thick perovskite layer is formed between the PTO and PTAA without any visible pinholes, which allowed us to


focus on ETLs and the ETL/perovskite interface by excluding the pin-hole issues of the perovskite layer.8 The inuence of the ETLs on the device performance is demonstrated by the J–V curves in Fig. 3a. As discussed in our previous study, unbalanced (usually slow) charge separation at the c-TiO2/perovskite interface seems to be a crucial factor that lowers the device efficiency.4,16 A relatively high Voc of 1.1 V, but a low Jsc and FF are obtained with the 20 nm thick c-TiO2 lm. To decrease the series resistance (Rs), we tested a thinner cTiO2 substrate, but it was veried to decrease the Voc due to the decreased shunt resistance (Rsh). A similar trend is observed

Fig. 3 The bi-layered electron transport layer (BETL) of c-TiO2/PTO of the perovskite solar cells. (a) J–V curves of the perovskite solar cells with different substrates. (b) Cyclic voltammograms. The measurement was carried out with a Pt wire as a reference electrode at a scan rate of 20 mV s 1. The electrolyte solution was 0.5 10 3 M K4Fe(CN)6 + 0.5 10 3 M K3Fe(CN)6 in aqueous 0.5 M KCl, pH 2.5. (c) Photoluminescence emission spectra. (d) Transmittance.

J. Mater. Chem. A

View Article Online



with PTO lms as discussed in Fig. 1c. An amazing enhancement of PCE is observed from the cell with the optimized BETL comprising a bi-layer of 5 nm thick TiO2 and 20 nm thick PTO. The average PCEs of the same batch is summarized in Table S1 (ESI†). The inuence of the differently conditioned ETLs on the device performance is distinctly shown by cyclic voltammetry (CV) measurement.32 According to the results in Fig. 3b, the highest Jsc is measured from the bare FTO substrate, and 20 nm thick c-TiO2 shows the lowest value of the Jsc, which indicates that the lm is pin-hole free and thus has good holeblocking ability. Compared to the thicker layer, 5 nm thick TiO2 clearly shows a small current leakage. Bad hole-blocking properties are detected for the 20 nm thick PTO lm, and it still shows current leakage even with the 100 nm thick PTO layer. Remarkably a good hole-blocking effect comparable to that of the 20 nm thick TiO2 lm is found from the BETL. We see that the result matches well with the trend shown by the J–V curve measurement. The superiority of the BETL is also demonstrated by the photoluminescence (PL) emission results shown in Fig. 3c. Due to the enhanced electron quenching rate at the PTO/perovskite interface, the PL intensity decreased 30% compared to that of the 20 nm thick TiO2 lm. It is notable that the 5 nm thick c-TiO2 underlayer does not hinder charge transfer from PTO to FTO as mirrored by the emission intensity change. Fig. 3d shows the optical transmittance of the lms tested in our work. Similar to other reports, enhanced lm transmittance is observed from PTO coated lms in our work, while any introduction of TiO2 lms increased the optical loss.23 The BETL retains better transmittance than FTO and TiO2 substrates, which leads to improved light harvesting. As discussed by Ke et al., one of the main reasons to adopt a low-temperature process, normally below 200 C, in perovskite solar cells is to obtain a good surface coverage.33 In their work, it was demonstrated that the lms heat-treated at high temperature (500 C) undergo a huge volume shrinkage that results in a bad surface coverage generating a shunting pathway between FTO and the perovskite. To minimize the inuence of the morphology change, we annealed the lm at 300 C. SEM top-view images taken from FTO and the SnO2 lms with different post-annealing temperatures are shown in Fig. S6 (ESI†). Now, unlike lms post-annealed at 500 C, a smooth and conformal surface morphology is still observed in the 300 C annealed lm. However, we found a huge PCE drop for the 300 C annealed lm as shown in Fig. 4a. Preliminary tests showed that it is very hard to obtain a high Voc although we use thicker TiO2 underlayers and SnO2 layers once the lms are annealed at high temperature. This is clear evidence that shows the limit of crystallization of PTO and that it should be carefully controlled for normal charge collection. It is notable that most LTO lms for perovskite solar cells are prepared below 200 C. We infer that the decomposition of SnOCl2 to form SnO2 would be accelerated rapidly above 200 C from the thermogravimetric analysis (TGA) data shown in ref. 34. A small and gradual decrease of optical transmittance also supports the crystallization of the PTO layer (Fig. S7, ESI†). Fig. 4b shows a schematic of the newly formed energy level diagram corresponding to the device structure with the c-TiO2/crystalline SnO2.35 Regarding

J. Mater. Chem. A

Communication

Fig. 4 Recombination pathways in the SnO2 based perovskite solar cells. (a) J–V curves of the PTO and TiO2/PTO based perovskite solar cells. The ETLs are annealed at 100 C and 300 C for 1 h, respectively. (b) Energy bandgap diagram of the TiO2/SnO2 based perovskite solar cells. (c) Possible main pathways in the TiO2/SnO2 ETL. (d) Proposed structure of the c-TiO2/PTO BETL in this work.

charge transfer, we can consider two recombination routes as depicted in Fig. 4c. The CB mismatch between TiO2 and SnO2 could be considered the rst route, as described in Fig. 4c. However, we reached a convincing conclusion that the shi of the CB is not the main reason for the efficiency drop. Recently Kavan et al. revealed that an amorphous SnO2 lm prepared by ALD can retain its compact morphological nature even aer sintering at 450 C to form crystalline SnO2.36 They also disclosed that amorphous and crystalline ALD SnO2 lms differ substantially in terms of their atband and CB positions, with a 0.5 V downward shi aer crystallization, as well as severe degradation of the hole-blocking ability.36 In our previous work, we already observed optical bandgap changes assumed to be due to the Burstein–Moss effect for the ALD SnO2 lms deposited above 250 C.22 All of these results imply that the electrical characteristic of the SnO2 lm can be changed radically during the crystallization. To conrm this, we fabricated cells with ALD SnO2 layers, and obtained a very analogous result with solution process devices as demonstrated in Fig. S8 (ESI†). It is important to note that a single LTO layer was used without a c-TiO2 underlayer with a very similar result. Therefore, we nd that a reason for the metal-like nature of SnO2 is surface defects. It is well-known that SnO2 is a typical degenerate semiconductor, which can easily have donor-type defects and interstitial (Sni) and oxygen vacancies (Vo) that are responsible for the disorder in the lattice structure and generate free electrons which provide high conductivity.37,38 Due to this reason, SnO2 has been favored for conductive oxide lms and sensors that require high conductivity.39 Now we can see that perovskite solar cells should undergo a severe Voc drop due to the lack of hole-blocking ability of the deciently passivated SnO2 layer as shown in Fig. 4a and c. Consequently, the surface defects of SnO2 should be properly passivated by amorphous SnOCl2, in this work, for photovoltaic applications as illustrated in Fig. 4d.


View Article Online

Communication



Acknowledgements The authors acknowledge the SNSF NRP 70 project, number: 407040_154056, European Commission H2020-ICT-2014-1, SOLEDLIGHT project, grant agreement No.: 643791, Swiss State Secretariat for Education, Research and Innovation (SERI), and CTI 15864.2 PFNM-NM, Solaronix, Aubonne, Switzerland. This research was supported by the Global Frontier Program through the Global Frontier Hybrid Interface Materials (GFHIM) of the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (2013M3A6B1078870). Y. L. acknowledges support from special funding for energy research, managed by Prof. Andreas Z¨ uttel, Fund No. 563074. Fig. 5 Performance of the best cell with the TiO2/PTO BETL. (a) J–V curves under 1 sunlight illumination and in the dark. The curve was obtained by backward scanning. (b) EQE and the calculated Jsc. (c) J–V curve hysteresis. (d) Long-term stability test without encapsulation (relative humidity, 20%). The device was measured under 1 sunlight illumination, and kept in the dark until the next measurement.

The J–V curve of the champion cell is shown in Fig. 5a. The best device shows promising photovoltaic values in a planar structure with a Jsc of 22.58 mA cm 2, a Voc of 1.13 V, a FF of 0.78, and a PCE of 19.80% under 1 sunlight illumination conditions when scanned backward. In the EQE measurement, we conrmed that the Jsc value from the solar simulator (1 sun, Xe lamp) agrees well with the integrated Jsc as seen in Fig. 5b. A small hysteresis is still shown from the cell as similarly reported in other papers, but we believe that it can be reduced by further modication of the perovskite layer.40,41 The device shows a good performance in the stabilized condition measurement (Fig. S9, ESI†). We also found that the BETL based perovskite cells have an excellent long-term stability for 1200 hours which is similar to those of devices with a single SnO2 layer21 when the cells are stored in a drawer without encapsulation (relative humidity, 20%), maybe due to the better interfacial conditions and self-healing effect in the dark.21,42 But further studies on the PTO/perovskite interface seem to be necessary for a better understanding.

Conclusions We investigated low-temperature processed SnO2 lms based on material, morphological and opto-physical aspects. It was found that the LTO layer is composed of crystalline SnO2 passivated by amorphous SnOCl2. Due to the naturally generated donor-type defects of crystalline SnO2, the surface passivation of SnO2 was proven to be essential. Crystallization of the ETL is benecial for improving its conductivity, but we clarify that the crystallization of PTO must be well controlled in a limited range owing to the above reasons. The BETL combining the c-TiO2 underlayer and PTO was found to further enhance charge collection due to the better hole-blocking ability of the TiO2 underlayer. Our results clearly show that surface passivation would be the hottest issue for SnO2 based PVs in future work.


References 1 A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc., 2009, 131, 6050–6051. 2 C. C. Stoumpos, C. D. Malliakas and M. G. Kanatzidis, Inorg. Chem., 2013, 52, 9019–9038. 3 N. J. Jeon, J. H. Noh, W. S. Yang, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Nature, 2015, 517, 476–480. 4 J. H. Heo, S. H. Im, J. H. Noh, T. N. Mandal, C.-S. Lim, J. A. Chang, Y. H. Lee, H.-j. Kim, A. Sarkar, M. K. Nazeeruddin, M. Gr¨ atzel and S. I. Seok, Nat. Photonics, 2013, 7, 486–491. 5 J. Burschka, N. Pellet, S.-J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin and M. Gr¨ atzel, Nature, 2013, 499, 316– 320. 6 Y. H. Lee, J. Luo, R. Humphry-Baker, P. Gao, M. Gr¨ atzel and M. K. Nazeeruddin, Adv. Funct. Mater., 2015, 25, 3925–3933. 7 M. Liu, M. B. Johnston and H. J. Snaith, Nature, 2013, 501, 395–398. 8 N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu and S. I. Seok, Nat. Mater., 2014, 13, 1–903. 9 NREL Best Research-Cell Efficiencies, http://www.nrel.gov/ ncpv/images/efficiency_chart.jpg, Accessed Oct 2, 2016. 10 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. Gr¨ atzel and N.-G. Park, Sci. Rep., 2012, 2, 591. 11 E. Edri, S. Kirmayer, S. Mukhopadhyay, K. Gartsman, G. Hodes and D. Cahen, Nat. Commun., 2014, 5, 3461. 12 M. Gr¨ atzel, Nature, 2001, 414, 338–344. 13 I. Chung, B. Lee, J. He, R. P. Chang and M. G. Kanatzidis, Nature, 2012, 485, 486–489. 14 E. Edri, S. Kirmayer, A. Henning, S. Mukhopadhyay, K. Gartsman, Y. Rosenwaks, G. Hodes and D. Cahen, Nano Lett., 2014, 14, 1000–1004. 15 G. Xing, B. Wu, S. Chen, J. Chua, N. Yantara, S. Mhaisalkar, N. Mathews and T. C. Sum, Small, 2015, 11, 3606–3613. 16 Y. H. Lee, J. Luo, M.-K. Son, P. Gao, K. T. Cho, J. Seo, S. M. Zakeeruddin, M. Gr¨ atzel and M. K. Nazeeruddin, Adv. Mater., 2016, 28, 3966–3972. 17 J. Y. Jeng, Y. F. Chiang, M. H. Lee, S. R. Peng, T. F. Guo, P. Chen and T. C. Wen, Adv. Mater., 2013, 25, 3727–3732. J. Mater. Chem. A

View Article Online



18 O. Malinkiewicz, A. Yella, Y. H. Lee, G. M. M. Espallargas, M. Gr¨ atzel, M. K. Nazeeruddin and H. J. Bolink, Nat. Photonics, 2014, 8, 128–132. 19 E. Fortunato, D. Ginley, H. Hosono and D. C. Paine, MRS Bull., 2007, 32, 242–247. 20 Q. Dong, Y. Shi, K. Wang, Y. Li, S. Wang, H. Zhang, Y. Xing, Y. Du, X. Bai and T. Ma, J. Phys. Chem. C, 2015, 119, 10212– 10217. 21 Q. Jiang, L. Zhang, H. Wang, X. Yang, J. Meng, H. Liu, Z. Yin, J. Wu, X. Zhang and J. You, Nat. Energy, 2016, 2, 16177– 16183. 22 D.-w. Choi and J.-S. Park, Surf. Coat. Technol., 2014, 259, 238– 243. 23 W. Ke, G. Fang, Q. Liu, L. Xiong, P. Qin, H. Tao, J. Wang, H. Lei, B. Li, J. Wan, G. Yang and Y. Yan, J. Am. Chem. Soc., 2015, 137, 6730–6733. 24 J. X. Song, E. Q. Zheng, J. Bian, X. F. Wang, W. J. Tian, Y. Sanehira and T. Miyasaka, J. Mater. Chem. A, 2015, 3, 10837–10844. 25 H.-S. Rao, B.-X. Chen, W.-G. Li, Y.-F. Xu, H.-Y. Chen, D.-B. Kuang and C.-Y. Su, Adv. Funct. Mater., 2015, 25, 7200–7207. 26 Z. Zhu, Y. Bai, X. Liu, C.-C. Chueh, S. Yang and A. K.-Y. Jen, Adv. Mater., 2016, 28, 6478–6484. 27 M. Park, J.-Y. Kim, H. J. Son, C.-H. Lee, S. S. Jang and M. J. Ko, Nano Energy, 2016, 26, 208–215. 28 Y. C. Kim, N. J. Jeon, J. H. Noh, W. S. Yang, J. Seo, J. S. Yun, A. Ho-Baillie, S. Huang, M. A. Green, J. Seidel, T. K. Ahn and S. I. Seok, Adv. Energy Mater., 2016, 6, 1502104. 29 S. Trost, A. Behrendt, T. Becker, A. Polywka, P. G¨ orrn and T. Riedl, Adv. Energy Mater., 2015, 5, 1614–6840.

J. Mater. Chem. A

Communication

30 S. M. Yong, N. Tsvetkov, L. Larina, B. T. Ahn and D. K. Kim, Thin Solid Films, 2014, 556, 503–508. 31 Q. Zhong and X. Pan, Proc. SPIE, 1998, 3557, 82–84. 32 L. Kavan, N. T´ etreault, T. Moehl and M. Gr¨ atzel, J. Phys. Chem. C, 2014, 118, 16408–16418. 33 W. Ke, D. Zhao, A. Cimaroli, C. Grice, P. Qin, Q. Liu, L. Xiong, Y. Yan and G. Fang, J. Mater. Chem. A, 2015, 3, 24163–24168. 34 R. Al-Gaashani, S. Radiman, N. Tabet and A. R. Dau, J. Mater. Sci. Eng. B, 2012, 117, 462–470. 35 L. Xiong, M. Qin, G. Yang, Y. Guo, H. Lei, Q. Liu, W. Ke, H. Tao, P. Qin, S. Li, H. Yu and G. Fang, J. Mater. Chem. A, 2016, 4, 8374–8383. 36 L. Kavan, L. Steier and M. Gr¨ atzel, J. Phys. Chem. C, 2017, 123, 342–350. 37 C. Kilic and A. Zunger, Phys. Rev. Lett., 2002, 88, 955011– 955014. 38 S. Bansal, S. C. Kashyap and D. K. Pandya, J. Alloys Compd., 2015, 646, 483–489. 39 J.-H. Lee, Sens. Actuators, B, 2009, 140, 319–336. 40 W. Ke, C. Xiao, C. Wang, B. Saparov, H. S. Duan, D. Zhao, Z. Xiao, P. Schulz, S. P. Harvey, W. Liao, W. Meng, Y. Yu, A. J. Cimaroli, C. S. Jiang, K. Zhu, M. Al-Jassim, G. Fang, D. B. Mitzi and Y. Yan, Adv. Mater., 2016, 28, 5214–5221. 41 K. T. Cho, S. Paek, G. Grancini, C. Roldan-Carmona, P. Gao, Y. Lee and M. K. Nazeeruddin, Energy Environ. Sci., 2017, 10, 621–627. 42 E. H. Anaraki, A. Kermanpur, L. Steier, K. Domanski, T. Matsui, W. Tress, M. Saliba, A. Abate, M. Gr¨ atzel, A. Hagfeldt and J.-P. Correa-Baena, Energy Environ. Sci., 2016, 9, 3128–3134.
