Efficient planar heterojunction solar cell employing ... - IOPscience

REGULAR PAPER

Japanese Journal of Applied Physics 54, 092301 (2015) http://dx.doi.org/10.7567/JJAP.54.092301

Efficient planar heterojunction solar cell employing CH3NH3PbI2+xCl1%x mixed halide perovskite utilizing modified sequential deposition Chunfu Zhang1,2*, Shi Tang2, Jing Yan1, Zhizhe Wang2, He Xi2, Dazheng Chen2, Haifeng Yang2, Jincheng Zhang2, Genquan Han1,2, Yan Liu1*, and Yue Hao2 1 Key Laboratory of Optoelectronic Technology and Systems of the Ministry of Education, College of Optoelectronic Engineering, Chongqing University, Chongqing 400044, China 2 State Key Discipline Laboratory of Wide Band Gap Semiconductor Technology, School of Microelectronics, Xidian University, Xi’an 710071, China E-mail: [email protected]; [email protected]

Received March 20, 2015; accepted June 20, 2015; published online August 24, 2015 In this work, a modified method of sequential deposition of a CH3NH3PbI2+xCl1%x organometal halide perovskite material with a low chlorine fraction and easy morphology control is developed. The as-prepared mixed halide perovskite CH3NH3PbI2+xCl1%x thin film shows an intense adsorption from 400–800 nm. Scanning electron microscopy images show that almost the entire compact TiO2 surface is covered with CH3NH3PbI2+xCl1%x mixed halide perovskite with grain sizes of 250–350 nm. The planar heterojunction solar cell employing CH3NH3PbI2+xCl1%x mixed halide perovskite shows a twofold enhancement of photovoltaic performance, compared with a planar CH3NH3PbI3 heterojunction solar cell, achieving a 9.53% power conversion efficiency (PCE) with a VOC of 0.96 V, a JSC of 18.86 mA/cm2, and a fill factor (FF) of 0.53. © 2015 The Japan Society of Applied Physics

1.

Introduction

The continuously growing demand for clean energy has increasingly directed research attention to solar energy. Thus, solar cells that convert sunlight into electricity become increasingly important. Over the past few years, many emerging solar cells, such as organic solar cells, dye-sensitized solar cells, and organometal halide perovskite solar cells, have achieved impressive power conversion efficiencies (PCEs).1–8) Among them, the development of organometal halide perovskite solar cells is the most rapid and has stunned almost everyone involved in solar cell research. The organometal halide perovskite material in the form of ABX3 (A = CH3NH3+; B = Pb2+ or Sn2+; and X = Cl−, Br−, or I−) is one of the most promising materials on earth that absorb sunlight owing to its suitable band structure and convenient deposition process. This material was first synthesized by Weber9) and further developed by Liang et al.10) and Lee et al.11) utilizing a solution process that enables its application to the fabrication of solar cells by spin-coating, ink-printing, screen-printing, and other large-scale low-cost thin-film preparation methods. According to the Shockley–Queisser limit, the material with a direct bandgap approaching 1.5 eV would be the optimum choice, and CH3NH3PbI3 with a bandgap of 1.55 eV meets this requirement quite well.12) Although CH3NH3PbI3 has the optimum bandgap, it suffers from a smaller diffusion length than that of another organometal halide perovskite, CH3NH3PbI3−xClx. The carrier diffusion length of CH3NH3PbI3−xClx could exceed 1 µm; however, the carrier diffusion length of CH3NH3PbI3 is on a scale of 100 nm, which is only 1=10 of that of CH3NH3PbI3−xClx.13) With a small diffusion length, electrons and holes cannot be efficiently collected at the cathode and anode, respectively. Hence, several attempts have been made to incorporate chlorine into CH3NH3PbI3.14–17) Unfortunately, these attempts encounter two constraints. Firstly, the incorporation of a high chlorine fraction into CH3NH3PbI3 would increase the material bandgap, resulting in the deviation from the optimum bandgap and ultimately the degradation of the device performance.13–15) Both theoretical and experimental results

suggest that the bandgap of CH3NH3PbX3 perovskite materials increases with the halide elements alternating in the order of I, Br, and Cl.18–20) When iodine atoms are completely replaced by chlorine atoms, the bandgap of the perovskite material will greatly deviate from the optimum bandgap. Secondly, the morphology of the present prevailing mixed halide perovskite CH3NH3PbI3−xClx is very sensitive to the deposition process.16,17) Thus, the method of incorporating chlorine into CH3NH3PbI3 while maintaining an appropriate bandgap and improving the morphology is crucial in both the deposition of a mixed halide perovskite thin film and the fabrication of high-performance planar heterojunction perovskite solar cells. Initially, continuous organometal halide perovskite films with a mesoporous structure were obtained by only two methods: the one-step deposition method and the sequential deposition method. In the one-step deposition method, as shown in Fig. 1(a), the film morphology cannot be easily controlled since it is sensitive to the thickness of the compact TiO2 layer, annealing time, annealing temperature, and the thickness of the perovskite layer.21,22) Accordingly, a sequential deposition method was developed. As shown in Fig. 1(b), a lead halide film is firstly prepared by the spin-coating method, and then it is dipped into CH3NH3I or CH3NH3Cl solution to complete the preparation of the organometal halide perovskite. To further optimize the morphology of organometal halide perovskite films, a mesoporous TiO2 layer and a modified sequential deposition have been employed and the solar cells achieved PCEs of approximately 10%.23–25) However, the additional process of fabricating the mesoporous layer will inevitably increase the fabrication difficulty and cost of perovskite solar cells, compared with the planar process. Further studies focusing on the carrier transport characteristic of perovskite materials suggest ambipolar and nonexcitonic characteristics of the perovskite materials.26,27) On the basis of these characteristics of the perovskite materials, the planar heterojunction structure could be employed in perovskite solar cells.28) However, it is still very difficult to obtain a continuous CH3NH3PbI3 or CH3NH3PbI3−xClx film on planar surfaces by using the sequential deposition method.29) Hence,

092301-1

© 2015 The Japan Society of Applied Physics

Jpn. J. Appl. Phys. 54, 092301 (2015)

C. Zhang et al.

Fig. 1. (Color online) (a) One-step deposition on mesoporous TiO2, (b) sequential deposition on mesoporous TiO2, and (c) modified sequential deposition on compact TiO2.

Fig. 2. (Color online) (a) Structure of CH3NH3PbI2+xCl1−x devices. (b) Energy diagram of CH3NH3PbI2+xCl1−x devices.

to obtain planar heterojunction perovskite solar cells by a method that provides a bandgap close to the optimum bandgap and an improved morphology of the organometal halide perovskite film, a modified deposition method for organometal halide perovskite has been developed, as illustrated in Fig. 1(c) by sequentially depositing PbI2 and PbI2:PbCl2 solutions to incorporate a small fraction of chorine into CH3NH3PbI3. Because PbI2 and PbI2:PbCl2 solutions were employed and no matter how the PbCl2 fraction changes, the chlorine fraction in the solution could not exceed 1, the molar ratio of chlorine is 1 − x, and the molar ratio of iodine is 2 + x. The as-prepared mixed halide perovskite CH3NH3PbI2+xCl1−x shows an intense adsorption from 400–800 nm. Scanning electron microscopy (SEM) images show that almost the entire compact TiO2 surface is covered with CH3NH3PbI2+xCl1−x mixed halide perovskite with grain sizes of 250–350 nm. Atomic force microscopy (AFM) images show a root mean squared (RMS) roughness of 50.3 nm, and after spin-coating the hole transport material (HTM) 2,2A,7,7A-tetrakis[N,N-bis(4-methoxyphenyl)amino]9,9A-spirobifluorene (Spiro-MeOTAD) this value is tuned to 4 nm. Using this material, planar heterojunction perovskite

solar cells were fabricated and the devices demonstrated the highest PCE of 9.53% under the AM 1.5 standard condition. This is much higher than that of the CH3NH3PbI3-based reference perovskite solar cell with the highest PCE of 5.5%. 2.

Experimental methods

As illustrated in Fig. 2(a), the planar heterojunction perovskite solar cell with a CH3NH3PbI2+xCl1−x absorber was fabricated on the glass=FTO substrate. The energy diagram of the planar heterojunction perovskite solar cell was presented in Fig. 2(b). A low chlorine fraction in the absorber was achieved by spin-coating PbI2 and PbI2:PbCl2 mixed solutions (PbI2 : PbCl2 ¼ 0:85 : 0:15 in molar ratio) sequentially on the compact TiO2 film. The resultant film was denoted as PbI1+xCl1−x. After that, the substrate was dipped in CH3NH3I solution to form the CH3NH3PbI2+xCl1−x absorber. The SpiroMeOTAD hole transport layer was obtained on the absorber by spin-coating Spiro-MeOTAD chlorobenzene solution. Finally, a 100-nm-thick Au anode was deposited to finish the device fabrication. For comparison, two kinds of solar cells with the same structure were fabricated, except for the absorber prepared employing CH3NH3PbI3−xClx or CH3NH3-

092301-2


Jpn. J. Appl. Phys. 54, 092301 (2015)

C. Zhang et al.

PbI3. To clarify the structure and morphology of thin films in each step of our planar CH3NH3PbI2+xCl1−x perovskite solar cell fabrication, FTO=TiO2, FTO=TiO2=PbI1+xCl1−x, FTO= TiO2=PbI1+xCl1−x=CH3NH3PbI2+xCl1−x, and FTO=TiO2= PbI1+xCl1−x=CH3NH3PbI2+xCl1−x=Spiro-MeOTAD films were prepared by the same process as that of CH3NH3PbI2+xCl1−x perovskite solar cell fabrication; for simplification, they were abbreviated to TiO2, PbI1+xCl1−x, CH3NH3PbI2+xCl1−x, and Spiro-CH3NH3PbI2+xCl1−x, respectively. Fluorine-doped tin oxide (FTO; Zhuhai Kaivo Optoelectronic Technology, 14 Ω=sq) glass was cleaned sequentially in 5% Decon-90 solution, deionized water, acetone, and ethanol under ultrasonication for 15 min each and then the FTO substrate was treated with O2 plasma for 15 min. A compact TiO2 layer on the FTO glass was prepared by spincoating 0.15 M titanium diisopropoxide bis(acetylacetonate) (75 wt % in isopropanol) solution in 1-butanol (99.8%, Sigma-Aldrich) at 4,000 rpm for 30 s and dried at 125 °C for 5 min. Then this process was repeated twice with 0.3 M titanium diisopropoxide bis(acetylacetonate) solution. Finally, the substrate was annealed at 500 °C for 15 min. After that, the resultant TiO2 film was immersed into a 40 mM TiCl4 (99%, Aladin) aqueous solution at 70 °C for 30 min, washed with deionized water and ethanol, and again annealed at 500 °C for 15 min. For the preparation of CH3NH3PbI3−xClx devices, 40 wt % CH3NH3PbI3−xClx precursor solution with a molar ratio of CH3 NH3 I : PbCl2 ¼ 2:44 : 0:88 was spin-coated onto a TiO2 thin film at 3000 rpm for 45 s and the film was annealed at 90 °C for 45 min. For CH3NH3PbI3 devices, 1 M PbI2 solution was spin-coated onto a TiO2 thin film at 3000 rpm for 45 s, and the film was annealed at 90 °C for 45 min. Then, the resultant PbI2 was dipped into 20 mg=mL CH3NH3I solution at 60 °C for 20 min to form CH3NH3PbI3. For CH3NH3PbI2+xCl1−x devices, 1 M PbI2 solution was spin-coated onto a TiO2 thin film at 3000 rpm for 45 s, and the film was annealed at 70 °C for 30 min. Then, a mixture solution of PbCl2 and PbI2 (molar ratio PbCl2 : PbI2 ¼ 0:85 : 0:15) was spin-coated at 3000 rpm for 45 s, and the film was dried at 70 °C for 30 min. After that the resultant film was dipped into a 20 mg=mL CH3NH3I solution to form CH3NH3PbI2+xCl1−x and further annealed at 90 °C for 45 min. To complete the device fabrication, HTM was then deposited by spin-coating Spiro-MeOTAD solution at 2000 rpm for 45 s. The spin-coating formulation was prepared by dissolving 72.3 mg of Spiro-MeOTAD (99%, Lumtec), 28.8 µl of 4-tert-butylpyridine (96%, SigmaAldrich), and 17.5 µl of a stock solution of 520 mg=ml lithium bis(trifluoromethylsulphonyl)imide (98%, SigmaAldrich) in acetonitrile (anhydrous, 99.8%, Sigma-Aldrich) in 1 mL chlorobenzene (99.9%, Sigma-Aldrich). Before the Au anode deposition, the devices were placed in the dark for 5–10 h to slowly evaporate the solvent. Finally, 100-nm-thick Au was thermally evaporated under vacuum as the anode. The UV–visible absorption spectra of TiO2, PbI1+xCl1−x, CH3NH3PbI2+xCl1−x, and Spiro-CH3NH3PbI2+xCl1−x were recorded with a UV–visible spectrophotometer (Perkin-Elmer Lambda 950) using precleaned FTO glass to collect blank signals. The morphologies of the perovskite layers were analyzed by SEM (Quanta ×50 FEG) and AFM (Bruker Dimension Icon). Energy-dispersive spectroscopy (EDS) was performed in the elemental mapping mode using Oxford

Fig. 3. (Color online) XRD spectra of TiO2, PbI1+xCl1−x, and CH3NH3PbI2+xCl1−x.

Instruments AZtecEnergy. X-ray diffraction (XRD) analysis was conducted on Bruker D8 Advance with the samples prepared by the same process as that for device fabrication. Photovoltaic performance characteristics were measured using a Keithley 2400 source meter under simulated sunlight from an XES-70S1 solar simulator matching the AM 1.5G standard. The system was calibrated against an NREL-certified reference solar cell and the measurements of the solar cells were performed with an active area of 0.07 cm2 in ambient atmosphere at room temperature without encapsulation. 3.

Results and discussion

The structures of TiO2, PbI1+xCl1−x, CH3NH3PbI2+xCl1−x; and Spiro-CH3NH3PbI2+xCl1−x were first studied by XRD analysis and the results are shown in Fig. 3. Note that the result for CH3NH3PbI2+xCl1−x coincides with that for Spiro-CH3NH3PbI2+xCl1−x; hence, the spectra of these two films were labeled CH3NH3PbI2+xCl1−x in Fig. 3. Two peaks at 37.90 and 51.66° corresponded to the FTO substrate. Owing to the very small thickness of the TiO2 film, two weak peaks were observed at 26.58 and 33.81°, indicating an anatase phase of the TiO2 film. When PbI1+xCl1−x was deposited onto the compact TiO2 thin film, a strong peak appeared at 12.72°, corresponding to the 〈001〉 crystal orientation of PbI1+xCl1−x. When PbI1+xCl1−x was fully transformed into CH3NH3Pb2+xCl1−x, the 〈001〉 peak of PbI1+xCl1−x disappeared and the 〈110〉 peak of CH3NH3Pb2+xCl1−x appeared at 14.15° with three extra peaks at 28.48, 31.87, and 42.60°, corresponding to 〈220〉, 〈310〉, and 〈330〉 crystal orientations, respectively. This confirms the formation of a tetragonal perovskite structure with a = b = 8.84 Å and c = 12.66 Å. The Scherrer expression suggests that the particle size of CH3NH3Pb2+xCl1−x is in the range of 250–350 nm. It can be seen in Fig. 4 that the particle size of CH3NH3PbI2+xCl1−x is in the range of 250–350 nm, which is coincident with the results of XRD analysis. There are also no observable pinholes in all the films. Although the CH3NH3PbI2+xCl1−x film is rough to some extent, the introduction of Spiro-MeOTAD makes the film smoother owing to the pore-filling capability of Spiro-MeOTAD, which makes Spiro-CH3NH3PbI2+xCl1−x acceptable for use in solar cells. As indicated in Fig. 5, the RMS roughness of TiO2 is

092301-3


Jpn. J. Appl. Phys. 54, 092301 (2015)

C. Zhang et al.

Fig. 4. SEM images of TiO2, PbI1+xCl1−x , CH3NH3PbI2+xCl1−x, and Spiro-CH3NH3PbI2+xCl1−x. Upper left: SEM photograph of TiO2 on FTO substrate with a 1 µm scale bar. Upper right: SEM photograph of PbI1+xCl1−x on FTO=TiO2 with a 1 µm scale bar. Bottom left: SEM photograph of CH3NH3PbI2+xCl1−x on FTO=TiO2 with a 1 µm scale bar. Bottom right: SEM photograph of Spiro-MeOTAD-coated CH3NH3PbI2+xCl1−x on FTO=TiO2 with a 10 µm scale bar.

(a)

(b)

(c)

(d)

Fig. 5. (Color online) AFM images of (a) TiO2, (b) PbI1+xCl1−x, (c) CH3NH3PbI2+xCl1−x, and (d) Spiro-CH3NH3PbI2+xCl1−x .

25.8 nm. When the Pb1+xCl1−x thin film was formed on the TiO2 thin film, the RMS roughness slightly increased to 25.9 nm. The formation of the CH3NH3PbI2+xCl1−x film leads to a RMS roughness of 50.6 nm. Owing to the excellent polefilling capability of Spiro-MeOTAD, the RMS roughness was reduced to 4 nm after Spiro-MeOTAD was spin-coated onto the CH3NH3PbI2+xCl1−x film. This trend of change in RMS roughness is consistent with the SEM measurement results in Fig. 4. The smooth surface of the overall film could effectively prevent the shunting in perovskite solar cells. It could be seen in Fig. 6 that the absorption of the TiO2,

PbI1+xCl1−x, and CH3NH3PbI2+xCl1−x thin films continues to increase with increasing film number, which is consistent with our intuition. Interestingly, when compared with CH3NH3PbI2+xCl1−x, a downshift was observed in the UV–visible absorption spectra after the Spiro-MeOTAD film was introduced. Intuitively, the larger the number of films, the more light should be absorbed. However, here, the introduction of the Spiro-MeOTAD film decreased the overall light absorption. It is assumed that this effect is attributed to the decrease in film roughness.17) A rough surface could reflect more incident light and this reflected incident light could not be

092301-4


Jpn. J. Appl. Phys. 54, 092301 (2015)

C. Zhang et al.

Fig. 6. (Color online) UV–visible spectra of TiO2, PbI2+xCl1−x, CH3NH3PbI2+xCl1−x, and Spiro-CH3NH3PbI2+xCl1−x. Fig. 8. (Color online) Distributions of Cl, I, and Pb elements in CH3NH3PbI2+xCl1−x thin film.

Table I. Photovoltaic performance characteristics of different perovskite solar cells. Device

received by the detector in the UV–visible spectrophotometer. The introduction of Spiro-MeOTAD smoothened the film as indicated by the results of SEM (Fig. 4) and AFM (Fig. 5) measurements. Owing to the downshift in the spectra, a 0.01 eV effective bandgap deviation was observed between CH3NH3PbI2+xCl1−x and Spiro-CH3NH3PbI2+xCl1−x films when UV–visible absorption spectra were used to estimate the effective bandgap. In Fig. 7, the effective bandgap of CH3NH3PbI2+xCl1−x is 1.57 eV, while that of Spiro-CH3NH3PbI2+xCl1−x is 1.58 eV. Although this 1.58 eV bandgap is higher than the optimum bandgap of the Shockley–Queisser limit, it is still lower than that of CH3NH3PbI3 formed by dual-source coevaporation,30) indicating successful bandgap control as a result of incorporating a very small fraction of chlorine. After the above analyses, the element distribution in CH3NH3PbI2+xCl1−x was analyzed on a FTO=TiO2=CH3NH3PbI2+xCl1−x film without capping the Spiro-MeOTAD, since the solvent of SpiroMeOTAD is chlorobenzene. As shown in Fig. 8, the EDS result suggests a uniform distribution of the Pb, I, and Cl elements in the CH3NH3PbI2+xCl1−x film, confirming good morphology control. Element mapping suggests an approximate ratio of Pb : I : Cl 1 : 2:85 : 0:15, indicating a small

Material

VOC (V)

JSC (mA=cm2)

FF

PCE (%)

A B

Planar Planar

CH3NH3PbI3−xClx CH3NH3PbI3

0.72 0.74

7.4 17.09

0.28 0.37

1.5 4.67

C

Planar

CH3NH3PbI2+xCl1−x

0.96

18.86

0.53

9.53

0.80

17.8

0.53

7.6

Ref. 9

Fig. 7. (Color online) Bandgap of CH3NH3PbI2+xCl1−x deduced from absorption spectra of CH3NH3PbI2+xCl1−x and Spiro-CH3NH3PbI2+xCl1−x.

Structure

Mesoporous CH3NH3PbI3−xClx

fraction of chlorine incorporated. This value is much smaller than that (0.26) reported elsewhere for a similar deposition process, indicating that only small amounts of byproducts containing chlorine were released during the deposition and annealing.16) This result was also consistent with the crystal parameters deduced from XRD results, since our CH3NH3PbI2+xCl1−x exhibited an identical c parameter (c = 12.66 Å) to CH3NH3PbI3, but slightly smaller a and b parameters (a = b = 8.84 Å), which may be attributed to the incorporation of a small amount of chlorine in the CH3NH3PbI3 crystal. The photovoltaic performance characteristics of planar perovskite solar cells employing the CH3NH3PbI2+xCl1−x absorber and those of reference solar cells employing the CH3NH3PbI3−xClx or CH3NH3PbI3 absorber are summarized in Table I, and the current density-to-voltage curves (J–V ) of these three kinds of solar cells are illustrated in Fig. 9. For the CH3NH3PbI3−xClx device, CH3NH3PbI3−xClx could not form a continuous film on the planar TiO2 surface in our experiment. This could easily cause the direct contact between the electron transport layer and the hole transport layer, which would lead to the partial shunting of the device and consequently the low short circuit current density (JSC) and fill factor (FF). Compared with CH3NH3PbI3−xClx, CH3NH3PbI3 was formed as a continuous film on the planar TiO2 thin film by sequential deposition. However, a low open circuit voltage (VOC) indicates an imperfect interface between CH3NH3PbI3 and Spiro-MeOTAD. The interface quality can be further improved by optimizing the morphology of

092301-5


Jpn. J. Appl. Phys. 54, 092301 (2015)

C. Zhang et al.

Fig. 9. (Color online) J–V curves of CH3NH3PbI2+xCl1−x-, CH3NH3PbI3-, and CH3NH3PbI3−xClx-based solar cells.

CH3NH3PbI3 by using CH3NH3PbI2+xCl1−x as absorber. The VOC increases to 0.96 V with a twofold FF leading to a 9.53% PCE. Compared with CH3NH3PbI3−xClx, CH3NH3PbI2+xCl1−x presents an advance control of morphology, leading to a reduction in loses of VOC and FF. 4.

Conclusions

We developed a modified method of sequential deposition of the CH3NH3PbI2+xCl1−x organometal halide perovskite material with a low chlorine fraction and easy morphology control. A mixed PbCl2 and PbI2 precursor solution was used in preparing CH3NH3PbI2+xCl1−x to ensure a low chlorine fraction and to further guarantee a bandgap approaching the optimized value. In a planar heterojunction perovskite solar cell, CH3NH3PbI2+xCl1−x shows significant increases in VOC and FF, achieved by controlling the morphology and suppressing the electron hole recombination. As a result, a 9.53% PCE has been achieved with a JSC of 18.86 mA=cm2 and a FF of 0.53. This PCE is much higher than that of the CH3NH3PbI3-based perovskite solar cell under the same fabrication process. This work has unveiled a promising absorber material for organometal halide perovskite with easy morphology and bandgap control. Acknowledgement

This study was partly financially supported by the National Natural Science Foundation of China under grant numbers 61334002 and 61106063.

1) C. C. Chen, W. H. Chang, K. Yoshimura, K. Ohya, J. You, J. Gao, Z. Hong, and Y. Yang, Adv. Mater. 26, 5670 (2014). 2) L. P. Heiniger, F. Giordano, T. Moehl, and M. Grätzel, Adv. Energy Mater. 4, 1400168 (2014). 3) Z. Wang, C. Zhang, D. Chen, J. Zhang, Q. Feng, S. Xu, X. Zhou, and Y. Hao, Jpn. J. Appl. Phys. 51, 122301 (2012). 4) Y. L. Li, S. R. Min, H. N. Cho, J. Kim, and C. W. Chung, Jpn. J. Appl. Phys. 47, 6896 (2008). 5) D. Chen, C. Zhang, T. Heng, W. Wei, Z. Wang, G. Han, Q. Feng, Y. Hao, and J. Zhang, Jpn. J. Appl. Phys. 54, 042301 (2015). 6) Z. Wang, C. Zhang, R. Gao, D. Chen, S. Tang, J. Zhang, D. Wang, X. Lu, and Y. Hao, Sol. Energy Mater. Sol. Cells 127, 193 (2014). 7) H. Zhou, Q. Chen, G. Li, S. Luo, T.-B. Song, H.-S. Duan, Z. Hong, J. You, Y. Liu, and Y. Yang, Science 345, 542 (2014). 8) A. Kojima, K. Teshima, Y. Shirai, and T. Miyasaka, J. Am. Chem. Soc. 131, 6050 (2009). 9) D. Weber, Z. Naturforsch. 33, 1443 (1978) [in German]. 10) K. Liang, D. B. Mitzi, and M. T. Prikas, Chem. Mater. 10, 403 (1998). 11) M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami, and H. J. Snaith, Science 338, 643 (2012). 12) W. Shockley and H. J. Queisser, J. Appl. Phys. 32, 510 (1961). 13) S. D. Stranks, G. E. Eperon, G. Grancini, C. Menelaou, M. J. Alcocer, T. Leijtens, L. M. Herz, A. Petrozza, and H. J. Snaith, Science 342, 341 (2013). 14) H. Yu, F. Wang, F. Xie, W. Li, J. Chen, and N. Zhao, Adv. Funct. Mater. 24, 7102 (2014). 15) E. Edri, S. Kirmayer, M. Kulbak, G. Hodes, and D. Cahen, J. Phys. Chem. Lett. 5, 429 (2014). 16) Y. Ma, L. Zheng, Y.-H. Chung, S. Chu, L. Xiao, Z. Chen, S. Wang, B. Qu, Q. Gong, and Z. Wu, Chem. Commun. 50, 12458 (2014). 17) L. Zheng, Y. Ma, S. Chu, S. Wang, B. Qu, L. Xiao, Z. Chen, Q. Gong, Z. Wu, and X. Hou, Nanoscale 6, 8171 (2014). 18) E. Mosconi, A. Amat, M. K. Nazeeruddin, M. Grätzel, and F. De Angelis, J. Phys. Chem. C 117, 13902 (2013). 19) N. Kitazawa, Y. Watanabe, and Y. Nakamura, J. Mater. Sci. 37, 3585 (2002). 20) K. Tanaka, T. Takahashi, T. Ban, T. Kondo, K. Uchida, and N. Miura, Solid State Commun. 127, 619 (2003). 21) G. E. Eperon, V. M. Burlakov, P. Docampo, A. Goriely, and H. J. Snaith, Adv. Funct. Mater. 24, 151 (2014). 22) J. M. Ball, M. M. Lee, A. Hey, and H. J. Snaith, Energy Environ. Sci. 6, 1739 (2013). 23) J. Shi, Y. Luo, H. Wei, J. Luo, J. Dong, S. Lv, J. Xiao, Y. Xu, L. Zhu, and X. Xu, ACS Appl. Mater. Interfaces 6, 9711 (2014). 24) P. Docampo, F. 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. 4, 1400355 (2014). 25) P. Docampo, F. C. Hanusch, N. Giesbrecht, P. Angloher, A. Ivanova, and T. Bein, APL Mater. 2, 081508 (2014). 26) G. Giorgi, J.-I. Fujisawa, H. Segawa, and K. Yamashita, J. Phys. Chem. Lett. 4, 4213 (2013). 27) Q. Lin, A. Armin, R. C. R. Nagiri, P. L. Burn, and P. Meredith, Nat. Photonics 9, 106 (2015). 28) B. Conings, L. Baeten, C. De Dobbelaere, J. D’Haen, J. Manca, and H. G. Boyen, Adv. Mater. 26, 2041 (2014). 29) J. Burschka, N. Pellet, S.-J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin, and M. Grätzel, Nature 499, 316 (2013). 30) L. K. Ono, S. Wang, Y. Kato, S. R. Raga, and Y. Qi, Energy Environ. Sci. 7, 3989 (2014).

092301-6
