Electrical Stress Influences the Efficiency of CH 3 NH 3 PbI 3 Perovskite Light Emitting Devices

Transcription

1 Communication Perovskite LEDs Electrical Stress Influences the Efficiency of CH 3 NH 3 PbI 3 Perovskite Light Emitting Devices Lianfeng Zhao, Jia Gao, YunHui L. Lin, Yao-Wen Yeh, Kyung Min Lee, Nan Yao, Yueh-Lin Loo, and Barry P. Rand* Organic inorganic hybrid perovskite materials are emerging as semiconductors with potential application in optoelectronic devices. In particular, perovskites are very promising for light-emitting devices (LEDs) due to their high color purity, low nonradiative recombination rates, and tunable bandgap. Here, using pure CH 3 NH 3 PbI 3 perovskite LEDs with an external quantum efficiency (EQE) of 5.9% as a platform, it is shown that electrical stress can influence device performance significantly, increasing the EQE from an initial 5.9% to as high as 7.4%. Consistent with the enhanced device performance, both the steady-state photoluminescence (PL) intensity and the time-resolved PL decay lifetime increase after electrical stress, indicating a reduction in nonradiative recombination in the perovskite film. By investigating the temperature-dependent characteristics of the perovskite LEDs and the cross-sectional elemental depth profile, it is proposed that trap reduction and resulting device-performance enhancement is due to local ionic motion of excess ions, likely excess mobile iodide, in the perovskite film that fills vacancies and reduces interstitial defects. On the other hand, it is found that overstressed LEDs show irreversibly degraded device performance, possibly because ions initially on the perovskite lattice are displaced during extended electrical stress and create defects such as vacancies. The past 5 years have witnessed rapid progress of metal halide perovskite solar cells with power-conversion efficiencies rising from 3.8% to over 22%. [1] As an excellent photoactive material, perovskites are also being considered as light L. Zhao, Y. L. Lin, K. M. Lee, Prof. B. P. Rand Department of Electrical Engineering brand@princeton.edu Dr. J. Gao, Prof. Y.-L. Loo Department of Chemical and Biological Engineering Y.-W. Yeh, Dr. N. Yao Princeton Institute for Science and Technology of Materials Prof. Y.-L. Loo, Prof. B. P. Rand Andlinger Center for Energy and the Environment DOI: /adma emitters with tunable bandgaps, high color purity, facile solution processing, and low material cost. [2] Recently, an external quantum efficiency (EQE) of 8.5% was reported for methylammonium lead bromide (MAPbBr 3 )-based green light-emitting devices (LEDs) and 3.5% for methylammonium lead iodide (MAPbI 3 )-based near-infrared LEDs, [2c,3] which are the highest EQEs reported so far for perovskite LEDs composed of pure MAPbBr 3 and MAPbI 3 films, respectively. More recently, an EQE of 8.8% was achieved for near-infrared LEDs using layered PEA 2 (CH 3 NH 3 ) n 1 Pb n I 3n+1 (PEA = C 8 H 9 NH 3, phenylethylammonium) perovskite films [4] and an EQE of 10.4% was achieved by introducing excess largegroup ammonium halides in the precursor. [5] Despite these high efficiencies, the impacts of unique perovskite properties (such as ion migration) on LED performance have not been well studied. Trap states leading to nonradiative decay pathways are key factors that limit the performance of perovskite LEDs. The mechanisms of defect formation and their impact on perovskite device performance are not fully understood, and effective methods to reliably reduce these defects are still needed. It is suggested that the formation of metallic Pb may increase trap states and that this metallic Pb formation can be suppressed by using excess organic cations such as CH 3 NH 3 Br. [2c] A recent finding shows that illumination-induced halide redistribution may result in an order-of-magnitude reduction in trap-state density with the halides filling vacancies and reducing interstitial sites, thus increasing the photoluminescence (PL) quantum yield. [6] Trap states may also be influenced by the ambient environment, with, for example, oxygen passivating halide vacancies and enhancing PL. [7] Here, using pure MAPbI 3 perovskite LEDs with a high EQE of 5.9% as a platform, we show that electrical stress in the form of subsequent electrical scans can influence the performance of MAPbI 3 perovskite LEDs significantly, increasing the EQE from an initial 5.9% to as high as 7.4%. Consistent with the enhanced device performance, both the steady-state PL intensity and the time-resolved PL (TRPL) decay lifetime increase after electrical stress, indicating a reduction in nonradiative recombination in the perovskite film. By investigating the (1 of 6)

2 temperature-dependent characteristics of the perovskite LEDs and the cross-sectional elemental depth profile, we propose that trap reduction and resulting device-performance enhancement is due to local ionic motion of excess ions, likely excess mobile iodide, in the perovskite film that fills vacancies and reduces interstitial defects. Furthermore, the enhanced efficiency maintains >95% of its value after 3 d in N 2 and can be maintained for an extended period of time providing electrical stress is applied upon measurement. On the other hand, we found that overstressed LEDs show irreversibly degraded device performance, possibly because ions initially on the perovskite lattice are displaced during extended electrical stress, and create defects such as vacancies. Perovskite LEDs were fabricated with a device structure of ITO (150 nm)/poly-tpd (20 nm)/mapbi 3 perovskite (300 nm)/ TPBi (40 nm)/lif (1.2 nm)/al (100 nm) (ITO: indium tin oxide; poly-tpd: poly[n,n -bis(4-butylphenyl)-n,n -bis(phenyl)- benzidine]; TPBi: 2,2,2 -(1,3,5-benzinetriyl)-tris(1-phenyl-1-Hbenzimidazole)), shown schematically in Figure 1a, along with a cross-sectional scanning electron microscopy (SEM) image in Figure 1b, which confirms the film thicknesses. Figure 1c shows a schematic energy diagram of the device, with poly- TPD serving as a hole-transporting and electron-blocking layer and TPBi as an electron-transporting and hole-blocking layer. Figure 1d shows a high-magnification top-view SEM image of the prepared perovskite film that reveals a polycrystalline morphology that appears to be pinhole free. The current density voltage (J V) curve shown in Figure 1e supports the absence of pinholes as evidenced by a low leakage current of 10 5 ma cm 2. The corresponding EQE versus current density curve is shown in Figure 1f, revealing a maximum EQE of 5.9%. Roll-off at higher voltages is not significant, although devices fail at a voltage of 6.5 V due to Joule heating (Figure S1, Supporting Information). The power efficiency, current efficiency, angular-dependent electroluminescence (EL) spectra, and intensity profiles are shown in Figure S2 (Supporting Information). This device shows performance on par with efficient perovskite LEDs that have been reported thus far; [2c,3,4,5] we attribute this performance to the pinholefree film (see also the large field of view SEM image in Figure S3, Supporting Information) and efficient carrier injection and blocking in the device structure. Interestingly, we find that the EL and EQE increase with subsequent electrical scans (Figure 2a,b). As shown in Figure 2b, after 30 consecutive electrical scans, the EQE is increased from 5.9% to 7.4%, an enhancement of 26%. The EQE enhancement gradually saturates as electrical scans are repeated (see also Figure S4, Supporting Information). During subsequent electrical scans, the J V curves also change appreciably, as shown in Figure 2c. The origin of this change is not from burning out of local shorts as has been observed previously with thin film LEDs, as the leakage current is unaffected. In the range of V, we observe a steep increase of the current caused by a diffusion-dominated current. [8] The slope of the J V curves in this region increases as the device is scanned multiple times. In this region, the current can be described by the Shockley diode equation: J = J 0 (e qv/nkt 1), where J 0 denotes the saturation current density, k the Boltzmann constant, T the temperature, q the electron charge, and n the ideality factor. The ideality factor is a measure of the slope of the J V curve and can be determined via n = kt q 1 ln J V. An increase in the slope, and thus reduction of n, of the J V curves in this region indicates a reduction in trap-assisted recombination, [9] and under ideal conditions without nonradiative recombination, n equals unity. In Figure 2d, we observe that, with sequential voltage scans, n decreases dramatically, indicating a significant reduction in trap-assisted recombination. This is consistent with the increased EL intensity and EQE, all of which confirm that the reduction in trap density plays the dominant role in performance enhancement. It should be noted that the extracted n is greater than 2, likely a result of nonideal series and shunt resistances. [10] The emissive properties of the perovskite film within the device structure before and after electrical stress are investigated to further understand the origin of the Figure 1. a) Device structure of the perovskite LEDs. b) Cross-sectional SEM image of a perovskite LED. c) A schematic of the flat-band energy level diagram of the device structure. d) SEM image of the perovskite film. e) Current density and radiance versus voltage of an as-produced perovskite LED. f) EQE versus current density of an as-produced perovskite LED. Inset shows the electroluminescence spectrum (2 of 6)

3 Figure 2. a) Radiance versus voltage of an as-produced perovskite LED for subsequent voltage scans. b) EQE versus current density of an as-produced perovskite LED for subsequent voltage scans. c) Current density and d) ideality factor versus voltage of an as-produced perovskite LED for subsequent voltage scans. e) Steady-state PL of encapsulated perovskite LEDs before and after electrical stress. f) Transient PL of encapsulated perovskite LEDs before and after electrical stress. device performance enhancement. After the EQE improvement saturates, the steady-state PL intensity is increased by approximately four times (Figure 2e). The TRPL measurement shows that the average decay lifetime increases from 0.35 to 1.05 µs (Figure 2f and Table S1, Supporting Information), which supports the hypothesis that the device performance enhancement is due to the reduction of defects in the MAPbI 3 film that cause nonradiative recombination. To further investigate the mechanisms of device-performance enhancement induced by the electrical stress, we performed temperature-dependent electrical characterization of the perovskite LEDs. Figure 3 summarizes the luminance and EQE of the devices measured at four different temperatures (180, 220, 260, and 300 K), all of which are higher than the orthorhombic-tetragonal phase transition temperature of MAPbI 3 at 160 K. [11] At lower temperatures (180 and 220 K), applying multiple electrical scans produce similar luminance curves (Figure 3a,b). In contrast, at higher temperatures (260 and 300 K), increased luminance is observed upon subsequent voltage sweeps (Figure 3c,d). Figure 3e shows the Figure 3. Luminance versus voltage curves at a) 180 K, b) 220 K, c) 260 K, and d) 300 K. e) The corresponding peak EQE versus scan number at different temperatures (3 of 6)

4 Figure 4. Cross-sectional STEM images of LEDs a) before and b) after electrical stress. c) Corresponding EDS elemental depth profile of I, Pb, C, Al, and In before (solid line) and after (dashed line) electrical stress. normalized peak EQE versus voltage scan number at different temperatures. Clearly, the EQE is improved with electrical stress at 260 and 300 K, whereas no performance enhancement is observed at 180 and 220 K. It should be noted that the EQE enhancement induced by the electrical stress is 2.4 and 1.5 times at 300 and 260 K, respectively, both of which are higher than the 26% enhancement for the high-performance LED measured in N 2 at room temperature shown in Figure 2. This is because during the sample transfer process for the temperature-dependent measurements, samples are exposed to air for 1 min, which slightly degrades the initial EQE. [2h] Under this condition, multiple electrical scans have the capability to improve the performance more significantly than devices without exposure to air. Previous research has investigated the temperaturedependent hysteresis of perovskite solar cells and shown that minimal hysteresis is observed at K because of the impeded drift and diffusion of ionic species while hysteresis becomes significant at K. [12] Those observations are very similar to ours for the device-performance enhancement induced by subsequent electrical scans. It therefore appears that the electrical stress reduces trap-assisted nonradiative recombination via ionic motion and redistribution in the perovskite film. Ionic motion is widely invoked in perovskite photovoltaic cells, which can be induced by either electrical bias or illumination. [6,13] It has been reported that the trapassisted recombination is caused by quenching sites at grain boundaries [2c] and that illumination-induced halide redistribution may lead to an order-of-magnitude reduction in trap state density by filling iodine vacancies and reducing interstitial sites, thus increasing the PL significantly. [6] Based on these observations, we propose that within the as-produced perovskite film, and especially at grain boundaries, iodine vacancies (undercoordinated lead sites) with associated interstitial iodide ions generate a large trap population that serve as nonradiative recombination sites. Upon applying a voltage bias, the increased electric field aids ionic motion, filling vacancies and reducing interstitial defects. As a result, after applying the electrical stress, EL efficiency, PL intensity, and PL lifetime all increase. However, electrical stress does not influence device performance at lower temperatures (180 and 220 K) because both trap-induced nonradiative recombination and other nonradiative decay coupling to phonon modes are reduced at these temperatures, offsetting the effects of electrical stress. This is confirmed by temperature dependent steady-state PL measurements (Figure S5, Supporting Information). At 180 K, the PL intensity is 15 times higher than at room temperature, while the electrical stress treatment only leads to a fourfold enhancement in PL intensity at room temperature. It should be noted that, during the time and voltage scale of the electrical stress discussed above, the electrical stressinduced ionic motion primarily influences excess mobile ions that do not comprise the perovskite lattice sites. Under electrical stress, ionic motion is short range, enabling mobile ions to fill local traps. Figure 4 shows cross-sectional scanning transmission electron microscopy (STEM) images and corresponding energy-dispersive X-ray spectroscopy (EDS) elemental depth profiles of the devices before and after electrical stress. They show that the distribution of various elements (Pb, I, C, Al, In) is not significantly altered by the electrical stress, suggesting that ion motion is a local phenomenon. It also confirms that iodine is highly mobile, as it has diffused into both organic transport layers (TPBi and poly-tpd), even in the as-produced device before electrical stress. The diffusion of ions into the transport layers may also improve charge injection and the electron/hole balance in the device, [14] which may be another factor for the improved device performance and warrants further investigation. We further investigated the effectiveness of the device performance enhancement induced by subsequent electrical scans over extended periods of time. When stored in N 2, the elevated EQE decreases in a much slower manner compared to the timescale of the enhancement, as shown in Figure 5. Devices maintain >95% of the enhanced EQE after 3 d, 90% after 7 d, and >80% after 18 d, which were measured only once without additional electrical scans to reduce the influence of the device performance enhancement effects induced by subsequent electrical scans as much as possible. The gradually decreased EQE is possibly due to defect regeneration by ion back diffusion, and the much slower rate of efficiency reduction is because nonexcess ion back diffusion has to overcome the activation energy of ion migration in perovskites, unlike the excess mobile ion (4 of 6)

5 Figure 5. Solid square: Normalized EQE of devices with an initial electrical stress on the day of device fabrication as well as 3, 7, and 18 d after the treatment. A second electrical stress is applied after 18 d, showing that performance enhancements are possible multiple times. Open circle: Normalized EQE of devices on the day of and 22 d after device fabrication. The electrical stress is applied 22 d after device fabrication and is beneficial for device performance. concentration. [15] If electrical stress is applied again, in the form of multiple current voltage sweeps, the EQE can be enhanced again, as shown in solid square in Figure 5. Furthermore, the electrical stress can be applied at any time to enhance device performance, as shown in open circle in Figure 5, with a similar device efficiency enhancement achieved when applying electrical stress 22 d after device fabrication. This comparison suggests that electrical stress is important not only for the short term but also for long-term device performance. We report data on electrical stress primarily in the form of subsequent electrical scans as it provides more detailed information in terms of device performance, extracted from changes to the J V, L V, and EQE characteristics. Electrical stress in the form of constant electrical bias provides additional insight as it can also influence device performance, as shown in Figure 6. The EQE increases and gradually saturates after 10 min for a constant current of 3 ma cm 2. As discussed above, in this region, excess mobile ions migrate locally and fill in traps, improving device performance. However, after 15 min, EQE starts irreversibly decreasing. This is because some of the nonexcess ions initially comprising the perovskite lattice sites migrate and create more defects due to longterm electrical stressing in one direction. [13b] This suggests that the perovskite layer must be formed such that defect density is low, and the possibility of ion migration is reduced to the fullest extent. Such approaches should allow for stable and efficient LEDs. In conclusion, using pure MAPbI 3 perovskite LEDs with a high EQE of 5.9% as a platform, we studied the influence of electrical stress on the performance of perovskite LEDs. Electrical stress is shown to initially improve device efficiency, owing to excess mobile ions in the perovskite film that are able to fill local defects that cause nonradiative recombination. On the other hand, long-term electrical stressing will degrade Figure 6. EQE versus time under constant current bias. A relatively small constant current (3 ma cm 2 ), where the corresponding EQE is not at its maximum point, is used here in order to avoid any impact from Joule heating. device performance, possibly due to the migration of nonexcess ions creating more defects. Therefore, this work points to a need to develop strategies to mitigate ion migration to ensure stable device operation. Experimental Section Fabrication: Methylammonium iodide (MAI) was synthesized by mixing aqueous trimethylamine (Sigma-Aldrich) and aqueous HI (Sigma-Aldrich) at 0 C with constant stirring. MAPbI 3 perovskite LEDs were fabricated using indium-tin-oxide (ITO) patterned glass substrates with a sheet resistance of 15 Ω sq 1. Substrates were sequentially cleaned using soapy water, deionized water, acetone, and isopropyl alcohol in an ultrasonicator for 15 min each, and then treated with O 2 plasma for 5 min prior to film deposition. Poly-TPD in chlorobenzene (6 mg ml 1 ) was spin coated on top of the ITO substrates at 1500 rpm followed by thermal annealing at 150 C for 20 min. Samples were then treated with O 2 plasma for 1 s to improve wetting. PbI 2 dissolved in hot dimethylformamide (500 mg ml 1, 100 C) was cooled to 70 C for spin coating at 6000 rpm, followed by thermal annealing at 70 C for 5 min. MAI in isopropyl alcohol (50 mg ml 1 ) was then spin coated on top of PbI 2 films at the same spin speed, which was followed by sequential thermal annealing at 70 C for 10 min, 100 C for 70 min, and 60 C for 300 min. Then, samples were brought into a vacuum evaporation chamber (EvoVac, Angstrom Engineering, base pressure Torr) for thermal evaporation of TPBi (40 nm), LiF (1.2 nm), and Al (100 nm) layers on top of the perovskite film. The device area was 0.1 cm 2. Characterization: The fabricated devices were measured in an N 2 - atmosphere glovebox using a homemade motorized goniometer setup consisting of a Keithley 2400 sourcemeter unit, a calibrated Si photodiode (FDS-100-CAL, Thorlabs), a picoammeter (4140B, Agilent), and a calibrated fiber optic spectrophotometer (UVN-SR, StellarNet Inc.). Temperature dependent I L V measurements were made under vacuum (< Torr) by transferring the devices into a Lakeshore probe station (Lake Shore Cryotronics, Inc., Westerville, USA) connected with a Keithley 2400 sourcemeter unit, a calibrated Si photodiode (FDS-100-CAL, Thorlabs), and a picoammeter (4140B, Agilent). SEM measurements were made using an FEI Verios 460 XHR SEM. Cross-sectional TEM lamella samples of the devices were prepared by an FEI Helios DualBeam microscope. STEM images (5 of 6)

6 and EDS measurements were carried out in an FEI Talos (S)TEM at 200 kv. Steady-state and TRPL measurements were performed using an FLS980 spectrometer (Edinburgh Instruments). Samples were excited at 470 nm from an Xe arc lamp source for the steady-state PL measurements. Samples were excited at 635 nm by a pulsed laser diode with a 10 µs pulse period and detection wavelength of 770 nm for the TRPL measurements. Supporting Information Supporting Information is available from the Wiley Online Library or from the author. Acknowledgements This work was supported by a Defense Advanced Research Projects Agency (DARPA) Young Faculty Award (Award #D15AP00093) and an EAGER grant from National Science Foundation (NSF) (ECCS ). N.Y. and Y.-W.Y. acknowledge funding in part from the Princeton Center for Complex Materials, an MRSEC supported by NSF grant DMR Y.-L.L. and B.P.R. also acknowledge support from the Andlinger Center for Energy and the Environment. Conflict of Interest The authors declare no conflict of interest. Keywords defect reduction, ion migration, light emitting diodes, perovskites Received: October 2, 2016 Revised: March 14, 2017 Published online: April 24, 2017 [1] a) M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami, H. J. Snaith, Science 2012, 338, 643; b) M. Liu, M. B. Johnston, H. J. Snaith, Nature 2013, 501, 395; c) Z. Xiao, C. Bi, Y. Shao, Q. Dong, Q. Wang, Y. Yuan, C. Wang, Y. Gao, J. Huang, Energy Environ. Sci. 2014, 7, 2619; d) J.-H. Im, I.-H. Jang, N. Pellet, M. Grätzel, N.-G. Park, Nat. Nanotechnol. 2014, 9, 927; e) A. Kojima, K. Teshima, Y. Shirai, T. Miyasaka, J. Am. Chem. Soc. 2009, 131, 6050; f) 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, Sci. Rep. 2012, 2, 591; g) J. Burschka, N. Pellet, S.-J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin, M. Grätzel, Nature 2013, 499, 316; h) A. Mei, X. Li, L. Liu, Z. Ku, T. Liu, Y. Rong, M. Xu, M. Hu, J. Chen, Y. Yang, Science 2014, 345, 295; i) H. Zhou, Q. Chen, G. Li, S. Luo, T.-B. Song, H.-S. Duan, Z. Hong, J. You, Y. Liu, Y. Yang, Science 2014, 345, 542; j) D. Liu, T. L. Kelly, Nat. Photonics 2014, 8, 133; k) N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu, S. I. Seok, Nat. Mater. 2014, 13, 897; l) W. S. Yang, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Ryu, J. Seo, S. I. Seok, Science 2015, 348, [2] a) G. Xing, N. Mathews, S. S. Lim, N. Yantara, X. Liu, D. Sabba, M. Grätzel, S. Mhaisalkar, T. C. Sum, Nat. Mater. 2014, 13, 476; b) Z.-K. Tan, R. S. Moghaddam, M. L. Lai, P. Docampo, R. Higler, F. Deschler, M. Price, A. Sadhanala, L. M. Pazos, D. Credgington, Nat. Nanotechnol. 2014, 9, 687; c) H. Cho, S.-H. Jeong, M.-H. Park, Y.-H. Kim, C. Wolf, C.-L. Lee, J. H. Heo, A. Sadhanala, N. Myoung, S. Yoo, Science 2015, 350, 1222; d) G. Li, Z.-K. Tan, D. Di, M. L. Lai, L. Jiang, J. H.-W. Lim, R. H. Friend, N. C. Greenham, Nano Lett. 2015, 15, 2640; e) A. Sadhanala, S. Ahmad, B. Zhao, N. Giesbrecht, P. M. Pearce, F. Deschler, R. L. Hoye, K. C. Gödel, T. Bein, P. Docampo, Nano Lett. 2015, 15, 6095; f) X. Gong, Z. Yang, G. Walters, R. Comin, Z. Ning, E. Beauregard, V. Adinolfi, O. Voznyy, E. H. Sargent, Nat. Photonics 2016, 10, 253; g) Y. H. Kim, H. Cho, J. H. Heo, T. S. Kim, N. Myoung, C. L. Lee, S. H. Im, T. W. Lee, Adv. Mater. 2015, 27, 1248; h) L. Zhao, R. A. Kerner, Z. Xiao, Y. L. Lin, K. M. Lee, J. Schwartz, B. P. Rand, ACS Energy Lett. 2016, 1, 595. [3] J. Wang, N. Wang, Y. Jin, J. Si, Z. K. Tan, H. Du, L. Cheng, X. Dai, S. Bai, H. He, Adv. Mater. 2015, 27, [4] M. Yuan, L. N. Quan, R. Comin, G. Walters, R. Sabatini, O. Voznyy, S. Hoogland, Y. Zhao, E. M. Beauregard, P. Kanjanaboos, Z. Lu, D. H. Kim, E. H. Sargent, Nat. Nanotechnol. 2016, 11, 872. [5] Z. Xiao, R. A. Kerner, L. Zhao, N. L. Tran, K. M. Lee, T.-W. Koh, G. D. Scholes, B. P. Rand, Nat. Photonics 2017, 11, 108. [6] W. Zhang, V. M. Burlakov, D. J. Graham, T. Leijtens, A. Osherov, V. Bulović, H. J. Snaith, D. S. Ginger, S. D. Stranks, Nat. Commun. 2016, 7, [7] Y. Tian, M. Peter, E. Unger, M. Abdellah, K. Zheng, T. Pullerits, A. Yartsev, V. Sundström, I. G. Scheblykin, Phys. Chem. Chem. Phys. 2015, 17, [8] P. De Bruyn, A. Van Rest, G. Wetzelaer, D. M. de Leeuw, P. W. Blom, Phys. Rev. Lett. 2013, 111, [9] G. J. A. Wetzelaer, M. Scheepers, A. M. Sempere, C. Momblona, J. Ávila, H. J. Bolink, Adv. Mater. 2015, 27, [10] K. Tvingstedt, C. Deibel, Adv. Energy Mater. 2016, 6, [11] R. L. Milot, G. E. Eperon, H. J. Snaith, M. B. Johnston, L. M. Herz, Adv. Funct. Mater. 2015, 25, [12] Y. Zou, R. J. Holmes, Adv. Energy Mater. 2016, 6, [13] a) Y. Deng, Z. Xiao, J. Huang, Adv. Energy Mater. 2015, 5, ; b) Z. Xiao, Y. Yuan, Y. Shao, Q. Wang, Q. Dong, C. Bi, P. Sharma, A. Gruverman, J. Huang, Nat. Mater. 2015, 14, 193. [14] a) Y. Xia, K. Sun, J. Chang, J. Ouyang, J. Mater. Chem. A 2015, 3, 15897; b) Y. Bai, Q. Dong, Y. Shao, Y. Deng, Q. Wang, L. Shen, D. Wang, W. Wei, J. Huang, Nat. Commun. 2016, 7, [15] Y. Yuan, J. Huang, Acc. Chem. Res. 2016, 49, (6 of 6)