precursor solution annealing forms cubic‐phase perovskite

FULL PAPERwww.afm-journal.de

© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1801508 (1 of 7)

Precursor Solution Annealing Forms Cubic-Phase Perovskite and Improves Humidity Resistance of Solar Cells

Petr P. Khlyabich, J. Clay Hamill Jr., and Yueh-Lin Loo*

Solar cells with light-absorbing layers comprising organometal halide perovskites have recently exceeded 22% efficiency. Despite high power-con-version efficiencies, the stability of these devices, particularly when exposed to humidity and oxygen, remains poor. In the current study, a pathway to increase the stability of methylammonium lead iodide (CH3NH3PbI3) based solar cells towards humidity is demonstrated, while maintaining the sim-plicity and solution-processability of the active layers. Thermal annealing of the precursor solution prior to deposition induces the formation of cubic-phase perovskite films in the solid state at room temperature. The experi-ments demonstrate that this improved ambient stability is correlated with the presence of the cubic phase at device operating temperatures, with the cubic phase resisting the formation of perovskite monohydrate—a pathway of degradation in conventionally processed perovskite thin films—on exposure to humidity.

DOI: 10.1002/adfm.201801508

Dr. P. P. Khlyabich, J. C. Hamill Jr., Prof. Y.-L. LooDepartment of Chemical and Biological EngineeringPrinceton UniversityPrinceton, NJ 08544, USA E-mail: [email protected]. Y.-L. LooAndlinger Center for Energy and the EnvironmentPrinceton UniversityPrinceton, NJ 08544, USA

The ORCID identification number(s) for the author(s) of this article can be found under https://doi.org/10.1002/adfm.201801508.

Despite these achievements, perovskite solar cells remain unstable when exposed to humidity, oxygen, and light.[8] Yet, the decomposition routes of both perovskite films and devices comprising them are not completely understood. Approaches to improve the stability of perovskite-containing solar cells via active-layer and interfacial engineering have recently been reported, including the substitution of cations and/or anions with bulkier coun-terparts,[8,9] the use of 2D perovskites as active layers,[10] the intermixing or stacking of 3D and 2D perovskites,[11–14] and the incorporation of passive layers atop the active layers.[15,16] In the case of humidity-induced degradation of methyl-ammonium lead iodide (CH3NH3PbI3), it is speculated that degradation occurs when water, acting as a Lewis base, coor-

dinates with the organic cation and catalyzes the decomposition of the perovskite layer to CH3NH2, HI, and PbI2.[17] We were intrigued by recent findings that demonstrated methylammo-nium lead halide perovskites containing bromide and chloride to be more resilient towards humidity exposure.[8] Interestingly, these perovskites adopt a cubic crystal structure, as opposed to the tetragonal crystal structure that is adopted by their more conventional counterpart, CH3NH3PbI3.[8] Relative to Cl- and Br-containing perovskites, CH3NH3PbI3-containing solar cells have demonstrated higher power-conversion efficiencies due to its smaller bandgap.[3] The apparent correlation between sta-bility and access to cubic-phase perovskites in CH3NH3PbBr3 and CH3NH3PbCl3 implies that we should be able to improve the stability of solar cells formed from CH3NH3I and PbI2 pre-cursors if we could access cubic-phase films across the temper-ature range relevant for device fabrication and operation.

In this study, we demonstrate a route to form cubic-phase perovskite films at temperatures below the established tetrag-onal-to-cubic transition temperature of 54 °C for phase-pure CH3NH3PbI3.[18] Once accessed, this cubic phase appears to be stable across a wide temperature window that is relevant for solar cell fabrication and operation. The preservation of the cubic phase leads to improved stability of perovskite thin films towards humidity, and accordingly, augments the operational stability of solar cells. We further correlate the enhanced sta-bility of the cubic phase to its resistance for forming perovskite monohydrate (CH3NH3PbI3·H2O) on exposure to humidity,

Cubic-Phase Perovskite

1. Introduction

Solar cells containing organometal halide perovskites as light-absorbing layers have attracted significant attention in recent years.[1–3] Perovskite films are characterized with low exciton binding energies, large free charge carrier diffusion lengths, and high mobilities;[3–5] these properties make perovskite films ideal absorbing layers for harvesting solar energy. Additionally, perovskite films can be processed from solution at ambient conditions,[6] thus preserving the simplicity of active-layer man-ufacturing. Solar cells with hybrid organic–inorganic perovs-kites as light absorbing layers have demonstrated efficiencies in excess of 22%,[1] approaching the efficiencies of silicon solar cells.[7]

Adv. Funct. Mater. 2018, 28, 1801508

www.afm-journal.dewww.advancedsciencenews.com

1801508 (2 of 7) © 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

which extends the integrity of perovskite films and improves the lifetime of solar cells comprising such active layers.

2. Results and Discussions

Films were processed from solutions comprising 1:1:1 of PbI2:CH3NH3I:dimethylsulfoxide (DMSO) in N,N-dimethylfor-mamide (DMF). Recent studies demonstrated that solar cells with active layers processed from this solution can exhibit effi-ciencies in excess of 15%.[19] In making the active layers, we either deposited the solution at room temperature, or we first annealed the solution at 150 °C for 40 min, and then cooled it down to room temperature prior to deposition. We refer to these solutions as the “standard” solution and the “annealed” solu-tion, respectively. Solar cells containing active layers processed from both standard and annealed solutions were fabricated in an inverted configuration (Scheme S1, Supporting Informa-tion). Figure S1 (Supporting Information) contains representa-tive J-V characteristics of such devices, fabricated with silver top electrodes. In both devices whose active layers are fabricated from the standard and annealed solutions, the short-circuit current densities (Jsc) exceeded 19 mA cm−2 and the fill factors (FF) were consistently above 75%, as seen in Figures S1 and S2 (Supporting Information). Both types of solar cells also demon-strate marginal hysteresis (Figure S1, Supporting Information). Relative to devices whose active layers were fabricated from the standard solution, however, the open-circuit voltage (Voc) of devices whose active layers were formed from the annealed solu-tion exhibit a 50 mV drop, the origin of which will be discussed later. As a result, the power-conversion efficiencies of solar cells processed from standard and annealed solutions are 14% and 12%, respectively (Figure S1, Supporting Information).

Morphologically, these perovskite thin films appear to be qualitatively unaffected by solution annealing. Figure S3 (Sup-porting Information) shows the scanning electron micrographs of perovskite thin films processed from both standard and annealed solutions. Both films appear continuous and pinhole-free, with the film cast from the annealed solution showing a slightly smaller average domain size. Films cast from both solu-tions are 550 ± 20 nm thick.

When exposed to humidity and oxygen, we observed that devices processed from the annealed solution retain the brown color that is characteristic of CH3NH3PbI3, while devices pro-cessed from the standard solution undergo discoloration from brown to yellow, characteristic of degradation of CH3NH3PbI3 to PbI2. This observation suggests that solar cells whose active layers are processed from the annealed solution to be more stable than those whose active layers are processed from the standard solution. To quantitatively assess this phenomenon and eliminate the convoluting effects of oxidation of the silver top electrodes, we fabricated solar cells with gold electrodes instead (Figure 1). Similar to devices with silver electrodes, devices with gold electrodes exhibit comparable Jsc and FF, independent of solution annealing (Figure 1). We observe a drop in Voc with devices whose active layers are processed from annealed solutions, comparable to what we had noted when silver, instead of gold, top contacts were used. The power-con-version efficiencies of solar cells processed from standard and annealed solutions were above 12% in both cases, as seen in Figure 1. With the oxidation of top electrodes eliminated, we proceeded to examine the operational stability on humidity exposure with this set of devices.

Figure 1c,f shows the evolution of device characteristics when these solar cells are exposed to 80% relative humidity at ambient temperature. Solar cells processed from the standard

Adv. Funct. Mater. 2018, 28, 1801508

Figure 1. J-V curves and efficiencies of solar cells processed from a,b) standard and d,e) annealed solutions. c,f) Normalized efficiencies and device parameters of perovskite solar cells with gold top electrodes as they are exposed to 80% relative humidity at ambient temperature.



solution show a noticeable decrease in efficiency after the first day, and only retain 40% of their initial efficiency after 6 days. All other solar cell parameters (Jsc, Voc, and FF) decrease simul-taneously with exposure, which contribute to the sharp drop in the power-conversion efficiency. In contrast, solar cells with active layers cast from the annealed solution retain more than 80% of their efficiency after 6 days of exposure to humidity and air. The Jsc and Voc of these devices remain almost invariant over the same period, with FF drop being the major contributor to the overall power-conversion efficiency decrease. The evolu-tion of performance characteristics of both types of devices with humidity beyond 6 days is shown in Figure S4 (Supporting Information). After 6 days, we observe continued degradation in device performance for both types of cells, though the dete-rioration in performance is substantially more rapid in devices that are made from the standard solution.

To identify the origin of this improved stability in devices pro-cessed from the annealed solution, we conducted X-ray diffrac-tion (XRD) measurements of active layers that were processed from both standard and annealed solutions. Figure 2 contains the XRD traces of thin films processed from standard and annealed solutions; reference X-ray traces corresponding to the cubic and tetragonal phases are included for comparison. We observe that the experimentally obtained diffraction traces are very similar, with one notable difference. In the highlighted 2θ range between 22° and 26°, the XRD trace of the film processed from the standard solution reveals two reflections at 23.40° and 24.48° that are consistent with the (211) and (202) reflections of the tetragonal phase of CH3NH3PbI3. The XRD trace of the film processed from the annealed solution reveals only one reflection in this range. At 2θ of 24.44°, this peak corresponds to the (111) reflection of the cubic phase. Given the similarity in XRD traces between the cubic and tetragonal phases and the proximity of the reflections, we look for the absence of the (211) reflection that is associated with its tetragonal phase to confirm the presence of the cubic phase.

We speculate that the preservation of the cubic phase at room temperature is responsible for the improved stability of solar cells (Figure 1). There is precedent in the literature that implicates the enhanced stability of the cubic phase of per-ovskites. For example, the cubic phases of CH3NH3PbBr3 and CH3NH3PbCl3 are more stable when exposed to the ambient compared with films having the tetragonal or orthorhombic phases of the same compounds.[8] Nonetheless, the ability to form the cubic phase at room temperature is surprising because the tetragonal-to-cubic transition is reported to be at 54 °C for phase-pure CH3NH3PbI3, with the cubic phase reported to only be stable above this transition.[18] Thermally annealing the precursor solution appears to favor the formation of the cubic phase at ambient conditions.

We studied the degradation pathways in films comprising tetragonal and cubic phases upon exposure to 80% humidity. Figure 3a shows the evolution of XRD traces of a CH3NH3PbI3 film in its tetragonal phase during humidity exposure. The intensities of the relevant reflections are tracked in Figure 3c. We observe a decrease in the intensity of the (110) reflection of CH3NH3PbI3 and a concurrent appearance of the (001) reflec-tion of PbI2 peak at 2θ of 12.68° for films[20] processed from the standard solution in the first several days of exposure to 80% relative humidity at ambient temperature. After 6 days of exposure to 80% relative humidity, reflections associated with CH3NH3PbI3 are no longer discernible in the diffraction trace. Only reflections associated with PbI2 and those associ-ated with perovskite monohydrate (CH3NH3PbI3·H2O; 2θ of 8.52° and 10.52°, respectively)[21,22] are detected in the thin-film X-ray traces. Our experiments indicate that perovskite monohy-drate continues to convert to PbI2 with continued exposure to humidity beyond 30 days. Figure 3b,d shows the X-ray traces and the intensity evolution as a cubic-phase film is exposed to humidity. We observe that reflections associated with the cubic phase persist for 7 weeks and reflections associated with PbI2 are only observed after 6 days of humidity exposure. Remarkably, we do not observe any reflections associated with perovskite monohydrate, even after 7 weeks of exposure. The half-life of the perovskite films, as quantified by the decrease in intensity of the (110) and (100) reflections, is 4 and 15 days for the tetragonal and cubic phases, respectively. While thin films adopting the cubic phase do eventually degrade when exposed to humidity, the rate of degradation is suppressed rela-tive to films adopting its tetragonal phase. Moreover, the deg-radation pathways for these films are different, with the cubic phase resisting the formation of perovskite monohydrate, a pathway of degradation present in conventionally processed perovskite thin films that adopt the tetragonal phase on expo-sure to humidity. With this data, we surmise that the tetragonal phase is more susceptible to humidity exposure, resulting in the formation of perovskite monohydrate that eventually leads to more rapid thin film and device degradation for perovskites processed from the standard solution.

Figure 4 and Figure S5 (Supporting Information) contain room-temperature XRD traces collected on perovskite thin films that were processed from a standard solution as well as solutions that were annealed at 150 °C for different dura-tions. Figure S5 (Supporting Information) tracks the primary reflection of the thin films with different extents of solution

Adv. Funct. Mater. 2018, 28, 1801508

Figure 2. XRD traces of perovskite films processed from standard (blue) and annealed solutions (green). Reference diffraction traces of the tetragonal (black) and cubic (red) phases with reflections identified are included.



annealing and reveals that this reflection shifts progressively from a 2θ of 28.44°, associated with the (220) reflection of its tetragonal phase,[18] to 2θ of 28.28°, associated with the (200) reflection of its cubic phase,[23] as the solution annealing time is increased from 0 to 40 min. It thus appears that increasing the annealing time of the perovskite precursor solution results increases the propensity for the cast film to adopt the cubic phase, at the expense of the tetragonal phase, with films that are formed with solutions that have been annealed for 40 min exhibiting exclusively the cubic phase. Returning to the XRD collected during humidity exposure of these films in Figure 4, we note that the extent of coexistence of the tetragonal and cubic phases depends on solution annealing time. In Figure 4b where the coexistence of the tetragonal and cubic phases is most dis-cernible, we observe that the intensity associated with the (220) reflection of the tetragonal phase at 2θ of 28.44° decreases more rapidly than that associated with the (200) reflection of the cubic phase at 2θ of 28.28°. This observation is consistent with what we had seen in the neat films of tetragonal and cubic phases and suggests decomposition in films with mixed phases to be mechanistically similar. These data further suggest that the formation of this cubic phase is correlated with the extent

of thermal annealing of the precursor solution, and the cubic phase is responsible for the increased stability of solar cell com-prising films cast from annealed solutions.

Earlier, we reported a 50 mV drop in the Voc of devices whose active layers are fabricated from the annealed solution. The origin of this decrease relative to the Voc of solar cells processed from standard solution can be correlated with pref-erential access of the cubic phase in the active layer. Earlier studies demonstrated that solar cells tested at temperatures above the tetragonal-to-cubic transition (54 °C) of phase-pure CH3NH3PbI3 are characterized with decreased Voc, while Jsc and FF remain comparable, with respect to solar cells tested at room temperature.[24,25] We believe the voltage drop stems from accelerated interfacial recombination between the active layer and charge selective contacts when we access the cubic phase.[25] While we cannot completely ignore the fact that the slightly smaller average domain size in thin films cast from annealed solutions can also facilitate recombination and ion collection at the grain boundaries,[26,27] XRD and energy-disper-sive X-ray (EDX) results rule out the formation of any I-defi-cient phases that could lead to bulk recombination. As such, we speculate that the marginally larger hysteresis and minute loss

Adv. Funct. Mater. 2018, 28, 1801508

Figure 3. Evolution of XRD traces of perovskite films processed from a) standard and b) annealed solutions during exposure to 80% relative humidity at ambient temperature. c) Quantification of the intensity evolution of the (110) reflection of tetragonal CH3NH3PbI3 (2θ = 14.10°), (100) reflection of CH3NH3PbI3·H2O (2θ = 8.52°) and (001) reflection of PbI2 (2θ = 12.68°) extracted from the traces shown in (a). d) Quantification of the intensity evolution of the (100) reflection of the cubic phase (2θ = 14.02°) and (001) reflection of PbI2 (2θ = 12.68°) extracted from the traces shown in (b). CH3NH3PbI3·H2O is absent during the degradation of films derived from the annealed solution.



in Voc observed in solar cells whose active layers are fabricated from annealed solutions with respect to those from standard solutions stem from accelerated interfacial recombination (Figure 1).

Figure 5 shows the UV–vis absorbance of the standard solu-tion and precursor solutions that had been annealed for dif-ferent times at 150 °C. These spectra were acquired after the solutions had cooled to room temperature. We observe a con-tinuous red-shift of the absorbance onset from 510 to 650 nm as annealing time for the solutions is increased from 0 to 40 min. This red-shift is manifested as a color change in the solutions from yellow to dark orange; this color change persists even after the solutions have cooled to room temperature. In the literature, this red-shift in the absorbance spectra of pre-cursor solutions have been attributed to the formation of trii-odide (I3

−), which absorbs up to 540 nm,[28] and the formation of (CH3NH3)2·Pb3I8·DMSO2 complex when the precursor solution is heated in the presence of DMSO.[29,30] To under-stand this color change on solution annealing in the context of favoring the formation of the cubic phase, we conducted

two control experiments. In one experi-ment, we added I2 to the precursor solution to intentionally form triiodide (I3

−).[31] The addition of I2 alters the solution color, as expected. But films cast from such solutions do not adopt the cubic phase. We instead access the more conventional tetragonal phase of CH3NH3PbI3 when films are cast from standard solutions with excess trii-odide. This control experiment suggests that while the formation of triiodide (I3

−) may be responsible for the color change observed on heating, it is not responsible for the forma-tion of the cubic phase in thin films.

In the second experiment, we annealed the precursor solution beyond an hour at 150 °C to induce precipitation so we can isolate and harvest the initial seeds of crystal-lization for characterization. XRD identifies the precipitate as (CH3NH3)2·Pb3I8·DMSO2 (Figure S6, Supporting Information).[19] And the presence of (CH3NH3)2·Pb3I8·DMSO2 is consistent with the color change we observe on annealing the precursor solu-tion. We thus speculate interactions between PbI2 and CH3NH3I with DMSO at elevated temperatures to result in the formation of (CH3NH3)2·Pb3I8·DMSO2, which ulti-mately nucleates the cubic phase in thin films. Though (CH3NH3)2·Pb3I8·DMSO2 is I-deficient when compared to the fully formed perovskite, XRD and EDX results rule out the formation of I-deficient phases in the perov-skite films formed from annealed solutions. To further support our assertion, we added formic acid to a precursor solution. The addi-tion of formic acid is known to catalyze the

Adv. Funct. Mater. 2018, 28, 1801508

Figure 4. Evolution of XRD traces of perovskite films during their exposure to 80% relative humidity at ambient temperature. The films were processed from a) a standard solution, and solutions that had been annealed for b) 10 min, c) 20 min, and d) 40 min prior to deposition. The peak at 2θ = 14.10° corresponds to the (110) reflection of tetragonal phase of CH3NH3PbI3, 2θ = 14.02° corresponds to (100) reflection of the cubic phase, 2θ = 12.68° corresponds to (001) reflection of PbI2, 2θ = 8.52° and 2θ = 10.52° correspond to (100) and (10-1) reflections of CH3NH3PbI3·H2O, respectively.

Figure 5. Absorbance spectra of solutions of PbI2:CH3NH3I:DMSO at 1:1:1 in DMF as a function of annealing time at 150 °C. The spectra were obtained after the solutions had cooled to room temperature.



formation of (CH3NH3)2·Pb3I8·DMSO2.[32] That we induce the formation of the cubic phase with the addition of formic acid to the standard solution is yet further correlative evidence that links (CH3NH3)2·Pb3I8·DMSO2 with the cubic phase. We caution, however, that since the presence of both triiodide (I3

−) and (CH3NH3)2·Pb3I8·DMSO2 induce a color change in the solu-tion, this color change should not be taken as a prerequisite for the formation of the cubic phase. Our findings demonstrate a route to forming cubic-phase thin films from methylammo-nium iodide and PbI2 precursor solutions and maintaining the cubic phase across a wide temperature range (Figure S7, Sup-porting Information) without sacrificing the simplicity of film formation. As a result, perovskite solar cells with improved sta-bility can be obtained.[1]

3. Conclusion

Until now, forming cubic phase perovskite films from MAI and PbI2 precursors at room temperature has been challenging and has required either physical confinement,[33] thin film decom-position and PbI2 formation,[34] or a change in the pH of the precursor solution through the addition of hydrohalic acid.[35] Our approach allows the spontaneous formation of the cubic phase with solution annealing prior to deposition, without the need to incorporate any additives. Accessing the cubic phase at room temperature via thermal annealing of the precursor solu-tion improves the stability of perovskite thin films and hence solar cells towards humidity exposure. The ability to form the cubic phase at room temperature by annealing the precursor solution can result in solar cells with markedly enhanced sta-bility with exposure to humidity.

4. Experimental SectionMaterials: All reagents from commercial sources were used without

further purification. PbI2 with 99.999% purity, bathocuproine (BCP) with 99.99% purity, DMF, DMSO, hydroiodic acid (HI) (57% in water), methylamine (CH3NH2) (33% in absolute ethanol), toluene, ethanol, and diethyl ether were purchased from Sigma Aldrich. C60 was purchased from American Dye Source.

Synthesis of Methylammonium Iodide (CH3NH3I): Fifteen milliliter of hydroiodic acid and 14 mL of methylamine were stirred in a round-bottom flask at 0 °C for 2 h. A rotary evaporator was used to extract the white powder from solution. In order to purify CH3NH3I, the white powder was redissolved in ethanol and reprecipitated in diethyl ether. A rotary evaporator was used to extract the purified white powder. The purification step was repeated three times. CH3NH3I crystals were collected and dried under a vacuum at 60 °C for 48 h.

Device Fabrication and Characterization: Prepatterned indium tin oxide (ITO)-coated glass substrates (10 Ω □−1, Thin Film Devices Inc.) were sequentially cleaned by sonication in deionized water, acetone, and isopropyl alcohol, and dried in a nitrogen stream. A thin layer of PEDOT:PSS (Clevios PH500, filtered with a 0.45 µm PVDF syringe filter from Pall Life Sciences) was first spin-coated on precleaned, prepatterned ITO-coated glass substrates and baked at 130 °C for 40 min. CH3NH3I, PbI2, and DMSO were dissolved in DMF at 1:1:1 molar ratio with a total concentration of 48 wt%[19] and stirred for 2 h in a nitrogen glovebox. This solution was referred to as the standard solution. Separately, solutions of the same composition were annealed at 150 °C for 40 min before cooling to 25 °C. This solution

was referred to as the annealed solution. The solution of choice (standard or annealed; 100 µL) was spin-coated on PEDOT:PSS-coated ITO/glass substrates at 4000 rpm for 25 s. Toluene (450 µL) was applied 4 s after the start of spin-coating as the antisolvent treatment.[19,36] These films were annealed at 70 °C for 1 min and then at 120 °C for 2 min in a nitrogen glovebox before they were transferred to a deposition chamber at pressure <3 × 10−6 Torr. C60 (30 nm), BCP (5.5 nm), and silver or gold (100 nm) were thermally evaporated atop the active layer through shadow masks to define an active area of 10 mm2.

The current density-voltage (J-V) characteristics of the photovoltaic devices were measured in a nitrogen glovebox using a Keithley 2635 source-measurement unit. A solar simulator with Xenon lamp (300 W) and an AM 1.5G filter were used as the solar simulator. A Newport reference cell (model 71582) was used for calibration. To calibrate the light intensity of the solar simulator (to 100 mW cm−2), the power of the Xenon lamp was adjusted to match the short-circuit current density (Jsc) of the reference cell under simulated sunlight to that specified by the manufacturer.

External quantum efficiency measurements were performed using a 300 W Xenon arc lamp (Newport Oriel), with filtered monochromatic light from a Cornerstone 260 1/4 M double grating monochromator (Newport 74125). A silicon photodiode (model 71580) calibrated at Newport was used as the reference cell.

UV–vis Absorption Measurements: Solutions for UV–vis absorption measurements were prepared in a manner identical to those used for fabricating solar-cell active layers. The spectra were obtained after the solutions had cooled to room temperature. UV–vis absorption spectra were obtained on an Agilent Technologies Cary 5000 spectrophotometer.

Thin-Film Fabrication: Conditions for depositing thin films for XRD measurements were identical as those used to fabricate solar-cell active layers.

XRD Measurements: XRD measurements were conducted on a Bruker D8 Discover diffractometer using Cu Kα radiation source (λ = 1.54 Å). The step size was 0.04°.

Scanning Electron Microscopy Measurements: A low-voltage scanning electron microscope (FEI Verios 460 XHR) was used to image the morphology of perovskite thin films. The accelerating voltage was kept at 5 keV to prevent beam damage to the specimens.

Film Thickness Measurements: Film thicknesses were determined using Bruker NanoMan AFM.

Humidity Studies: NaNO3 dissolved in deionized water was used to obtain 80% relative humidity. The fully dissolved salt solution in the glass dish was placed in the sealed container alongside solar cells and thin films.

Supporting InformationSupporting Information is available from the Wiley Online Library or from the author.

AcknowledgementsThe authors acknowledge the financial support from the National Science Foundation (CMMI-1537011) and Princeton Center for Complex Materials funded under NSF-MRSEC (DMR-1420541). J.C.H. was supported by the Department of Defense (DoD) by a National Defense Science and Engineering Graduate Fellowship (NDSEG).

Conflict of InterestThe authors declare no conflict of interest.

Adv. Funct. Mater. 2018, 28, 1801508


1801508 (7 of 7) © 2018 WILEY-VCH Verlag GmbH & Co. KGaA, WeinheimAdv. Funct. Mater. 2018, 28, 1801508

Keywordscubic phase, humidity exposure, perovskite hydrates, perovskites, stability

Received: February 27, 2018Revised: May 31, 2018

Published online: July 13, 2018

[1] L. K. Ono, E. J. Juárez-Pérez, Y. Qi, ACS Appl. Mater. Interfaces 2017, 9, 30197.

[2] J.-P. Correa-Baena, M. Saliba, T. Buonassisi, M. Grätzel, A. Abate, W. Tress, A. Hagfeldt, Science 2017, 358, 739.

[3] J. S. Manser, J. A. Christians, P. V. Kamat, Chem. Rev. 2016, 116, 12956.

[4] O. E. Semonin, G. A. Elbaz, D. B. Straus, T. D. Hull, D. W. Paley, A. M. van der Zande, J. C. Hone, I. Kymissis, C. R. Kagan, X. Roy, J. S. Owen, J. Phys. Chem. Lett. 2016, 7, 3510.

[5] L. M. Herz, ACS Energy Lett. 2017, 2, 1539.[6] M. L. Petrus, J. Schlipf, C. Li, T. P. Gujar, N. Giesbrecht,

P. Müller-Buschbaum, M. Thelakkat, T. Bein, S. Hüttner, P. Docampo, Adv. Energy Mater. 2017, 7, 1700264.

[7] M. A. Green, Y. Hishikawa, W. Warta, E. D. Dunlop, D. H. Levi, J. Hohl-Ebinger, A. W. H. Ho-Baillie, Prog. Photovoltaics Res. Appl. 2017, 25, 668.

[8] Z. Wang, Z. Shi, T. Li, Y. Chen, W. Huang, Angew. Chem., Int. Ed. 2017, 56, 1190.

[9] T. A. Berhe, W.-N. Su, C.-H. Chen, C.-J. Pan, J.-H. Cheng, H.-M. Chen, M.-C. Tsai, L.-Y. Chen, A. A. Dubale, B.-J. Hwang, Energy Environ. Sci. 2016, 9, 323.

[10] H. Tsai, W. Nie, J.-C. Blancon, C. C. Stoumpos, R. Asadpour, B. Harutyunyan, A. J. Neukirch, R. Verduzco, J. J. Crochet, S. Tretiak, L. Pedesseau, J. Even, M. A. Alam, G. Gupta, J. Lou, P. M. Ajayan, M. J. Bedzyk, M. G. Kanatzidis, A. D. Mohite, Nature 2016, 536, 312.

[11] Y. Hu, J. Schlipf, M. Wussler, M. L. Petrus, W. Jaegermann, T. Bein, P. Müller-Buschbaum, P. Docampo, ACS Nano 2016, 10, 5999.

[12] Y. Bai, S. Xiao, C. Hu, T. Zhang, X. Meng, H. Lin, Y. Yang, S. Yang, Adv. Energy Mater. 2017, 7, 1701038.

[13] G. Grancini, C. Roldán-Carmona, I. Zimmermann, E. Mosconi, X. Lee, D. Martineau, S. Narbey, F. Oswald, F. De Angelis, M. Graetzel, M. K. Nazeeruddin, Nat. Commun. 2017, 8, 15684.

[14] N. Li, Z. Zhu, C.-C. Chueh, H. Liu, B. Peng, A. Petrone, X. Li, L. Wang, A. K.-Y. Jen, Adv. Energy Mater. 2017, 7, 1601307.

[15] F. Wang, W. Geng, Y. Zhou, H.-H. Fang, C.-J. Tong, M. A. Loi, L.-M. Liu, N. Zhao, Adv. Mater. 2016, 28, 9986.

[16] Z. Wu, S. R. Raga, E. J. Juarez-Perez, X. Yao, Y. Jiang, L. K. Ono, Z. Ning, H. Tian, Y. B. Qi, Adv. Mater. 2018, 30, 1703670.

[17] J. M. Frost, K. T. Butler, F. Brivio, C. H. Hendon, M. Van Schilfgaarde, A. Walsh, Nano Lett. 2014, 14, 2584.

[18] A. Poglitsch, D. Weber, J. Chem. Phys. 1987, 87, 6373.[19] J. Cao, X. Jing, J. Yan, C. Hu, R. Chen, J. Yin, J. Li, N. Zheng, J. Am.

Chem. Soc. 2016, 138, 9919.[20] Y. Wu, A. Islam, X. Yang, C. Qin, J. Liu, K. Zhang, W. Peng, L. Han,

Energy Environ. Sci. 2014, 7, 2934.[21] A. M. A. Leguy, Y. Hu, M. Campoy-Quiles, M. I. Alonso, O. J. Weber,

P. Azarhoosh, M. van Schilfgaarde, M. T. Weller, T. Bein, J. Nelson, P. Docampo, P. R. F. Barnes, Chem. Mater. 2015, 27, 3397.

[22] F. Hao, C. C. Stoumpos, Z. Liu, R. P. H. Chang, M. G. Kanatzidis, J. Am. Chem. Soc. 2014, 136, 16411.

[23] S. G, P. Mahale, B. P. Kore, S. Mukherjee, M. S. Pavan, C. De, S. Ghara, A. Sundaresan, A. Pandey, T. N. Guru Row, D. D. Sarma, J. Phys. Chem. Lett. 2016, 7, 2412.

[24] T. J. Jacobsson, W. Tress, J.-P. Correa-Baena, T. Edvinsson, A. Hagfeldt, J. Phys. Chem. C 2016, 120, 11382.

[25] H. Zhang, X. Qiao, Y. Shen, T. Moehl, S. M. Zakeeruddin, M. Grätzel, M. Wang, J. Mater. Chem. A 2015, 3, 11762.

[26] T. Leijtens, G. E. Eperon, A. J. Barker, G. Grancini, W. Zhang, J. M. Ball, A. R. S. Kandada, H. J. Snaith, A. Petrozza, Energy Environ. Sci. 2016, 9, 3472.

[27] J. S. Yun, J. Seidel, J. Kim, A. M. Soufiani, S. Huang, J. Lau, N. J. Jeon, S. Il Seok, M. A. Green, A. Ho-Baillie, Adv. Energy Mater. 2016, 6, 1600330.

[28] A. D. Awtrey, R. E. Connick, J. Am. Chem. Soc. 1951, 73, 1842.[29] J. C. Hamill, J. Schwartz, Y.-L. Loo, ACS Energy Lett. 2017, 3, 92.[30] S. Rahimnejad, A. Kovalenko, S. M. Fores, C. Aranda, A. Guerrero,

ChemPhysChem 2016, 17, 2795.[31] W. S. Yang, B.-W. Park, E. H. Jung, N. J. Jeon, Y. C. Kim, D. U. Lee,

S. S. Shin, J. Seo, E. K. Kim, J. H. Noh, S. Il Seok, Science 2017, 356, 1376.

[32] N. K. Noel, M. Congiu, A. J. Ramadan, S. Fearn, D. P. McMeekin, J. B. Patel, M. B. Johnston, B. Wenger, H. J. Snaith, Joule 2017, 1, 328.

[33] Q. Wang, M. Lyu, M. Zhang, J.-H. Yun, H. Chen, L. Wang, J. Phys. Chem. Lett. 2015, 6, 4379.

[34] D. Luo, L. Yu, H. Wang, T. Zou, L. Luo, Z. Liu, Z. Lu, RSC Adv. 2015, 5, 85480.

[35] C. M. M. Soe, C. C. Stoumpos, B. Harutyunyan, E. F. Manley, L. X. Chen, M. J. Bedzyk, T. J. Marks, M. G. Kanatzidis, ChemSusChem 2016, 9, 2656.

[36] N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu, S. Il Seok, Nat. Mater. 2014, 13, 897.

precursor solution annealing forms cubic‐phase perovskite

Documents