Synthesis of stable iodoplumbate and perovskite for efficient annealing‐free device and long‐term storage

As a next‐generation photovoltaic device, perovskite solar cells are rapidly emerging. Nevertheless, both solution and device stability pose challenges for commercialization due to chemical degradation caused by internal and external factors. Especially, the decomposition of iodoplumbate in a perovskite solution hinders the long‐term use of perovskite solutions. Moreover, the synthesis of stable perovskites at low temperature is important for stable devices and wide applications (flexible devices and high reproducibility). Herein, the critical composition of perovskite is found to obtain high stabilities of both iodoplumbate and perovskite crystals by utilizing CsPbBr3 and FAPbI3, exhibiting high device performance and long‐term solution storage. The novel composition of CsPbBr3‐alloyed FAPbI3 not only crystallizes under annealing‐free conditions but also demonstrates excellent iodoplumbate stability for 100 days (∼3000 h) without any degradation. Furthermore, high device stabilities are achieved over 2000 and 3000 h under extreme conditions of A.M. 1.5 and 85°C/85% relative humidity, respectively. Overall, the device exhibited a high power conversion efficiency of 23.4%, and furthermore, CsPbBr3‐alloyed FAPbI3 was devoted to widen the applications in both flexible and carbon‐electrode devices, thereby addressing both scientific depths and potential commercial materials.


INTRODUCTION
2][3][4][5] Due The fabrication of PSCs mostly relied on the solution process, and the formation of iodoplumbates occurs inside the perovskite solution.A perovskite solution is known to be ionized; however, perovskite colloids can be formed, and various iodoplumbates (PbI 2 ) 0 -(PbI 6 ) 4− grow inside of the solution.When the iodoplumbates are in (PbI 4 ) 2− , iodoplumbate is favorable for perovskite synthesis, while excess or deficiency of iodide in lead could cause defective films and materials. 27,28Iodoplumbate is an unstable colloidal phase that easily reacts in ambient conditions, and this property makes it difficult to utilize in large-scale production and storage in a market.Therefore, the ionic and solvent behavior of perovskites should be carefully understood to develop stable perovskites.
0][31][32][33][34][35] These bonding features determine the crystallization temperature, where organic-cation-based perovskite requires 100 • C-150 • C, and inorganic-cationbased perovskite requires 150 • C-300 • C. [36][37][38] Not only the share of different bonding nature in perovskite but also the nucleophilicity and the distortion pattern of the perovskite also influences their crystallization.Lowering the synthesis temperature is important in PSCs, as it reduces fabrication cost and increases reproducibility by minimizing thermal discrepancy between multiple layers in a device.Most importantly, the perovskite solution must be stable, and the perovskite film should be stable at low temperature for commercial use.
The concept of annealing-free perovskite was previously introduced, but the PCE and stability also have to be satisfied for commercialization. 391][42][43][44] However, a high concentration of substituted ions increases the bandgap, resulting in inferior current generation.Along with that, an excess amount of ions is thermodynamically unstable due to the metastable behavior of perovskite.In the ABX 3 of perovskite, charge transport mainly depends on the X site and charge generation depends on the B site.6][47][48][49] Therefore, achieving both stability and device performance requires complex compositional engineering.
Herein, we developed perovskite compositions using CsPbBr 3 and FAPbI 3 to control stable iodoplumbate and annealing-free film.The prepared solution was stable for 100 days without any degradation; therefore, changes in iodoplumbate and device performance were not observed.
By controlling the thermodynamic barrier of perovskite, we discovered an annealing-free composite that exhibits a PCE of 23.4% and robust long-term stability (maximum power point tracking [MPPT] 2000 h and 85 • C/85% relative humidity [RH] 3000 h, both encapsulated).Both density-functional theory (DFT) calculations and experiments were conducted to investigate the novel perovskite composition that satisfies both annealing-free and highly performing devices.Annealing-free perovskite is further introduced into a flexible device, which is sustained for 2000 cycles under a 4-mm bending curvature without any PCE loss.This indicates that the material and interfaces are well aligned by the low-temperature process.Beyond that, the device was further fabricated with carbon electrode and achieved a PCE of 15%.It is worth noting that carbon device is fabricated in annealing and vacuum-free process, including n-type, perovskite, p-type, and electrode, paving the way to complete annealing-and vacuum-free devices.

DISCUSSION
2][3][4][5][6][7][8] However, the development of an ideal perovskite composition to obtain both solution and material stability has not been clearly studied.1][52][53][54] The complex obtains Pb as a central metal and I attached to Pb. [55][56][57] In a perovskite solution, dimethyl sulfoxide (DMSO) is used to form a complex and supports ideal crystallization of perovskite.However, recent studies have claimed that the DMSO complex can be changed based on the condition of iodoplumbate. 57As the solution gets aged, the binding site for DMSO is replaced by I (PbI 4 2− is favorable).][60] In Figure 1A, a schematic illustration explains the phenomenon that perovskite requires continuous PbI 6 iodoplumbate complexes are generally related to colloidal sizes.In colloidal-scale adhesion, both cation and halide take place.2][63] Figure 1C shows the absorption spectrum of the perovskite solution, indicating that the addition of CsPbBr 3 in FAPbI 3 perfectly prevents iodine attachment, whereas FAPbI 3 solution produced iodine-rich iodoplumbate.Since the ions in CsPbBr 3 are less nucleophilic than the ions in FAPbI 3 , bromine takes place as an inhibitor of iodine attachment.Ionic interaction and hydrogen interaction would co-exist; however, the reaction speed would increase rapidly.By adhesion of iodine and bromine, distortion as well as organic adhe-sion might take advantage of the reaction in the precursor and crystal growth.With the optimized solution (10% of CsPbBr 3 -alloyed FAPbI 3 ), the solution remained stable for more than 100 days with room storage, while regular FAPbI 3 solution turned red due to excess I replacement into iodoplumbate (Figures 1D and S1).
A schematic illustration of film fabrication using FAPbI 3 and CsPbBr 3 is shown in Figure 2A.Due to the similar ionic radii, bromine can easily substitute iodide in lattice, resulting in a change from a trigonal to a cubic structure with increasing bromine concentrations in FAbased perovskite.Generally, FAPbI 3 has black trigonal α-FAPbI 3 and yellow hexagonal δ-FAPbI 3 .5][66]  temperature.Interestingly, the addition of CsPbBr 3 dramatically converts FAPbI 3 into a black phase in an annealing-free condition.It should be noted that DMSO in iodoplumbate is removed rapidly and crystallized properly after antisolvent treatment.Diethyl ether (DE) was chosen as an antisolvent due to its low evaporation heat property, which can effectively support crystallization and remove residual solvent since residual solvent in perovskite film would lead to film degradation.DE could vaporize easily at room temperature, especially without thermal annealing.Right after spin-coating, semi-transparent brown film was observed, and it became a mirror-like black film as time goes by (about several minutes).It can be taken together that color and surface property changes imply that the residual antisolvent was evaporated, and crystallization was completed (Figure S2).In a single-junction PSC, MAPbI 3 or FAPbI 3 is widely used as a light absorber with bandgaps of 1.59 and 1.51 eV, respectively (Figure S3). 7,41igure 2B shows X-ray diffraction (XRD) result for different concentrations of CsPbBr 3 -alloyed FAPbI 3 .The result indicates that only hexagonal perovskite was observed in pure FAPbI 3 , while 10%, 20%, and 30% CsPbBr 3 -alloyed FAPbI 3 showed a clear trigonal phase.Even though the fresh films with 20% and 30% CsPbBr 3 -alloyed FAPbI 3 produced clear trigonal phase, they degraded in 30 days due to the phase segregation of I and Br (Figure 2C) (40%-50% RH and 25 • C). [67][68][69] There could be many deciphering in these phenomena, but the magnified XRD result in Figure 2D indicates that CsPbBr 3 was well alloyed in the lattice at the beginning.However, with increasing CsPbBr 3 , the major peaks of trigonal perovskite (<001> and <022>) decreased, and <111> peak disappeared at 30% CsPbBr 3alloyed FAPbI 3 .We have confirmed that the tetragonal phase turned into a cubic phase at 30% CsPbBr 3 -alloyed FAPbI 3 in Figure 2E.According to DFT calculations and previous works, excess amount of Br, ion migration, and phase segregation can occur. 40,68,70Therefore, excessive bromine in perovskite not only increases the bandgap but also reduces the phase stability of the film.As to say, our results are not only well corresponded but are also practical, making them appealing as ideal photovoltaic materials.
Formation enthalpy calculations were performed for FA/Cs cation and I/Br anion substitutions on FAPbI 3 to investigate the thermodynamic stability of CsPbBr 3alloyed FAPbI 3 solid solution.The formation enthalpies were considered for two FAPbI 3 polymorphs.One is a black trigonal phase, in which PbI 6 octahedra share the corners, forming a three-dimensional arrangement.The other is a yellow hexagonal phase, in which octahedra share the faces forming the linear chains perpendicular to the c axis.Due to anisotropicity of FA molecules, structure relaxation for variable configurations was conducted in both phases.As shown in Figure 3A,B, the FA molecules of the trigonal phase were perpendicular to the c axis, and those of the hexagonal phase were parallel to the c axis.In the trigonal phase for cation substitution, the 1 × 1 × 2 supercells were used, and six different cation sites were generated.All 26 possible configurations were reduced to 36 irreducible configurations by considering the symmetry of crystal structure using site-occupancy disorder (SOD) codes. 71In the same manner, for the hexagonal phase of cation substitution, eight different cation sites were generated using the 2 × 2 × 1 supercells, and all 28 possible configurations were reduced to 34 irreducible configurations.While the conventional cells were used for both phases of anion substitutions, nine and six different anion sites and the total configurations were reduced 120 and 13 irreducible anion substitution configurations.Figure 3C,D shows formation enthalpies for the trigonal and hexagonal phases of FA/Cs cation substitution and I/Br anion substitution, respectively.The stabilization of the trigonal phase on anion substitution occurred with 40%-80% I/Br substitution amount.However, significant trigonal phase stabilization upon cation substitution was demonstrated at low doping amounts (∼16%, highlighted by blue circles).This implies that a low concentration of Cs substitution may be a clue that 10% CsPbBr 3 -alloyed FAPbI 3 forms the preferred photoactive black phase without the need for annealing.To examine how low concentrations of Cs substitution and Br substitution were coupled to stabilize the black trigonal phase, the formation enthalpies of (FAPbI 3 ) 5/6 (CsPbBr 3 ) 1/6 were calculated.The 1 × 1 × 2 supercells were used, and 12 different anion sites were generated.All the possible 12 C 3 (=220) configurations were reduced to 83 irreducible configurations.The range of formation enthalpies of (FAPbI 3 ) 5/6 (CsPbBr 3 ) 1/6 is −0.62 to −0.59 eV/f.u.(Figure S4).In the group of high enthalpies, the substituted Cs and Br ions made a cluster, resulting in distortion of solid solution.On the other hand, in the group of low enthalpies, the substituted Cs and Br ions are dispersed well in the solid solution.
Phase analysis from Figures 2 and 3 indicates that CsPbBr 3 is proper in 5%-15% region in FAPbI 3 to sustain high phase stability.The result from Figure 4A indicates that the bandgap of CsPbBr 3 -alloyed FAPbI 3 (5%, 10%, and 15%) ranges from 1.53 to 1.56 eV, which is well qualified for use in single junction PSCs.2][63] Also, the absorption coefficient is the highest in 10% of CsPbBr 3alloyed FAPbI 3 , indicating that this composition produced the most favorable current generation.To identify impurities and phases, XRD was conducted (Figure 4B).The 5% CsPbBr 3 -alloyed FAPbI 3 showed a hexagonal phase dominantly; however, the 10% CsPbBr 3 -alloyed FAPbI 3 clearly formed a trigonal phase without any high-bandgap impurities (hexagonal perovskite and PbI 2 ).On the other hand, 15% CsPbBr 3 -alloyed FAPbI 3 produced a small amount of PbI 2 phase, indicating that the excess CsPbBr 3 is thermodynamically unfavorable in ambient crystallization due to its high crystallization temperature above 300 • C, while FAPbI 3 crystallizes at 150 • C. 72,73 Therefore, excess CsPbBr 3 may retard crystallization at low temperature.However, the Gibbs-phase diagram of the binary phase indicates that the Gibbs energy decreases at certain alloy compositions, and the 10% CsPbBr 3 -alloyed FAPbI 3 could lower the Gibbs energy down to room conditions. 40The resulting grain size is around 200-300 nm (Figure 4C), showing a highly dense film.
To further confirm the interaction behavior of CsPbBr 3alloyed FAPbI 3 , X-ray photoelectron spectroscopy (XPS) was further conducted, as shown in Figure 4D.XPS shows that the binding energy of perovskite is increased at the 10% CsPbBr 3 -alloyed FAPbI 3 compared with 5% CsPbBr 3alloyed FAPbI 3 .However, 15% CsPbBr 3 -alloyed FAPbI 3 maintained the same binding energy as 10% CsPbBr 3 -FAPbI 3 , suggesting that an optimum binding nature of the individual ions exists and limits the energy distribution.The result is well matched with the XRD results in Figure 4B, where the energy sharing of CsPbBr 3alloyed FAPbI 3 is optimized under ambient conditions.In Figure 4E, atomic force microscopy was also performed.Based on the XRD results in Figure 4B, 5% and 15% CsPbBr 3 -alloyed FAPbI 3 produced two phases (additional hexagonal phase or PbI 2 ), which may have led to a rough surface.However, the single-phased 10% CsPbBr 3 -alloyed FAPbI 3 resulted in a smooth film with the lowest rootmean-square value (film properties of annealed FAPbI 3 [0% CsPbBr 3 ]), as listed in Figure S5.The formation of grains and films is also related to the surface energy of the bottom film and neighboring materials, and 10% CsPbBr 3alloyed FAPbI 3 well qualified the minimized disruption in grain formation. 73,74nce 10% CsPbBr 3 -alloyed FAPbI 3 was confirmed as an effective material for solution and film, optoelectronic analysis was further conducted.Therefore, charge extraction behavior was further measured with steady-state photoluminescence (PL) in Figure 5A and time-resolved photoluminescence (TRPL) spectroscopy in Figure S6.As shown in the absorption data in Figure 4A, the bandgap ranged from 1.53 to 1.56 with increasing CsPbBr 3 ratio.This could be taken together that PL shift is derived from precursor composition differentiation.Since 5% and 10% of CsPbBr 3 did not exhibit a blue shift, however, 15% of CsPbBr 3 shows a noticeable blue shift.This could be explained by the fact that compared to 15%, 10% CsPbBr 3 exhibits improved crystallization as well as reduced bandgap. 75Moreover, PL intensity is clue of effective charge carrier extraction, which underpins that 10% CsPbBr 3 extracts charge carriers more effectively.The measurement was performed on the device architecture, and the 10% CsPbBr 3 -alloyed FAPbI 3 exhibited the fastest charge transport compared to the 5% and 15% CsPbBr 3alloyed FAPbI 3 .Additionally, trap analysis through electrochemical impedance spectroscopy further confirmed the lowest defect density at 10% CsPbBr 3 -alloyed FAPbI 3 , as shown in Figure 5B.It should be mentioned that both trap density and abovementioned PL spectra were measured by a full device.Since lower PL intensity in a full device implies reduced carrier recombination, the trap density data of 10% CsPbBr 3 -alloyed FAPbI 3 device are in good agreement with the lowest PL intensity.The major advantage of annealing-free perovskite is the possibility of depositing perovskite on underlying polymer layers, which paves the way to flexible devices.As a result, PCE of 18% was achieved from flexible device of 10% CsPbBr 3alloyed FAPbI 3 (Figure 5C).For comparison with the annealed device, the FAPbI 3 device on a flexible substrate was annealed at 150 • C; however, the PCE was below 5% and drastically lagged behind that of the glass substrate (Figure S7).FAPbI 3 requires an annealing temperature of about 160 • C; however, a flexible substrate cannot tolerate that high temperature.Since flexible substrate and indium-doped tin oxide (ITO) have different heat expansion indices, underlying heat would induce expansion, which would lead to a notorious split within layers.Therefore, relatively lower performance of annealed FAPbI 3 was a predictable result.Additionally, a bending test was measured in Figure S8, and a device with 10% CsPbBr 3alloyed FAPbI 3 sustained its initial PCE over 2000 cycles (bending radius: 4 mm) due to the annealing-free step improving the adhesion of perovskite with neighboring layers.In contrast, the device with heated FAPbI 3 could not sustain the initial PCE after 500 cycles because of the thermal discrepancy of layers.With a glass substrate, a PCE of up to 23.4% was achieved with 10% CsPbBr 3alloyed FAPbI 3 (Figure 5D), and the steady-state current and external quantum efficiency (EQE) in Figure 5E,F also underpinned the PCE result as well (device performance of annealed FAPbI 3 [0% CsPbBr 3 ] are listed in Figure S9).We conducted an annealing process on 10% CsPbBr 3 -alloyed perovskite; however, we noticed PCE drop (Figure S10).The PCE drop was reasonable since additional thermal treatment would cause thermal expansion, which would result in interface damage.It should be noted that the optimized concentration exhibited the superior reproducibility (Figure S11), indicating that the high reproducibility was due not only to the annealing step, but also to the stable phase.In general, glass substrate exhibits a higher PCE than flexible substrate due to the crystal quality of ITO, as the crystallinity and conductivity of electrode are important factors in device performance.
The annealing-free perovskite has another advantage of being cost effective.This is due to the reduced fabrication time and energy needed for crystallization.Annealingand vacuum-free PSCs (n-type and p-type are also fabricated without heat) using a carbon electrode were further fabricated and exhibited the PCE of ∼15% (as shown in Figure 5G).While the conductivities of the bottom and top electrodes are crucial and a carbon top electrode is less conductive than Au, the resulting PCE is not high compared to the device using glass/ITO/Au.Despite this, the device remains very challenging for applications because of its heat-and vacuum-free fabrication process.We have further measured light and thermal stability of the optimized device using 10% CsPbBr 3 -alloyed FAPbI 3 (glass encapsulated), as shown in Figure 5H,I (between 1000 and 1400 h has a major constancy issue; in detail, humidity and atmosphere were temporarily disturbed.Once the issue had been taken care, data were back in track).It should be noted that there is some PCE decay during the encapsulation process due to epoxy, contact, and charge transport problems.The light stability of the device sustained 80% of initial PCE (η = 17.5%) for up to 2000 h, and the heat stability (85 • C/85% RH) sustained 80% of initial PCE (η = 18.6%) up to 3000 h (actual PCE is listed in Figure S12).The device was also stored in a dark room and sustained for more than 1 year (glass encapsulated, as shown in Figure S13) (40%-50% RH and 25 • C), indicating that the 10% CsPbBr 3alloyed FAPbI 3 is an extremely stable material.The device performance and robust stability are one of the highest values in the field of annealing-free perovskite.

CONCLUSION
Despite various efforts to bring PCSs into practical use, relatively low PCE and long-term instability have kept PCSs in a nascent stage.Another challenge is the solution stability of perovskite materials, which is critical for continuous process in large-scale manufacturing.Therefore, an alternative strategy should be suggested to overcome the current limitations, from the solution preparation to the device performance.In this research, CsPbBr 3 was alloyed with FAPbI 3 to achieve high PCE and stability without annealing steps.With DFT calculations and experiments, we have found the optimum composition range.
The optimized device exhibited a high PCE of 23.4% and was sustained for 2000 h under illumination (A.M. 1.5) and 3000 h under 85 • C/85% RH.Also, the device demonstrated stability for 1 year when stored in the dark and encapsulated environment.The annealing-free step itself can be a strong advantage in low-cost and reproducible manufacturing and has wide applications, such as in a flexible device.Furthermore, the alloying of CsPbBr 3 with FAPbI 3 largely suppressed the formation of improper iodoplumbates in perovskite solution, resulting in superior solution stability for 100 days.These achievements can be widely utilized in both research and industry-scale PSCs.

Perovskite preparation
1.8 M of FAI, PbI 2 , PbBr 2 , and CsBr were dissolved in 1 mL of mixed solvent (0.85 mL of DMF and 0.15 mL of DMSO) with different ratios of 5%, 10%, and 15% CsPbBr 3 .Then, about 15 min of sonication step was followed.Every composition of perovskite precursor was fabricated under normal atmosphere.

Device fabrication
ITO glass (AMG, 7 Ω) substrates were cleaned by a sonicator with acetone, IPA, and deionized water for 15 min.Substrates were dried with N 2 gas after cleaning and treated for 15 min with a UV-ozone cleaner.As an electron transport layer (ETL), SnO 2 solution was prepared (with a ratio of 5:1 for deionized water and SnO 2 ) and spin coated on the ITO substrates at 3000 rpm for 30 s.Additional annealing step was held at 120 • C for 30 min to remove residual solvent.After the annealing step, substrates were cleaned for 15 min by UV-ozone cleaner.
After depositing the process, the perovskite precursor was spin coated on the substrates.An amount of 95 μL perovskite precursor was dropped before the first step started and spun at 500 rpm for 5 s and accelerated to the second step (5000 rpm for 20 s).Then, 1 mL of DE was used as an antisolvent and dropped 14 s after the second step started.No additional annealing step was required for every type of sample.
spiro-OMeTAD solution was prepared by dissolving spiro-OMeTAD at 72.3 mg mL −1 in a chlorobenzene solution that contained 28.8 μL tBP, 17.5 μL Li-TFSI salt (520 mg mL −1 in acetonitrile), and 21.9 μL cobalt solvent (384 mg mL −1 in acetonitrile) in a N 2 -filled glovebox.An amount of 95 μL of spiro-OMeTAD solution (as a hole transporting layer) was spin coated at 3000 rpm for 30 s.For the PTAA hole transport layer, solution was fabricated by dissolving 25 mg of PTAA in 1 mL of chlorobenzene, with the addition of 9 μL of 4-tert-butylpyridine and 6 μL of bis(trifluoromethane)sulfonimide lithium salt solution.
Finally, gold as a counter electrode was thermally evaporated on the spiro-OMeTAD or PTAA layer and the whole device fabrication process was finalized.For stability measurement, the encapsulated devices were sealed with cover glass using UV-curable epoxy resin (Nagase).
For annealing and vacuum-free PSCs, SnO 2 solution as an ETL (with the same ratio which was used in abovementioned annealing-free device) was deposited on ITO glass substrates (3000 rpm for 30 s).Substrates were cleaned for 15 min by UV-ozone cleaner, and the perovskite precursor was spin coated (same process and condition as the previously mentioned annealing-free device).Prepared hole-transporting layer (HTL) solution (same solution that was used in annealing-free device) was spin coated on prepared perovskite film, and no annealing process was conducted either.Finally, carbon electrode as a counter electrode was placed on HTL with doctor blading.

4.4
Measurements and characterization XPS (Sigma Probe; Thermo VG Scientific) with a photon energy of hν = 21.22 eV was used to study the surface and interface composition.An X-ray diffractometer (new D-8 ADVANCE; Bruker) was used to observe the crystal structure and impurities in HTL-deposited perovskite films.The optical properties of the films were analyzed by UV−vis spectroscopy (V-770; JASCO).A field-emission scanning electron microscope (Merlin Compact; Zeiss) was utilized to understand the vertical structures of devices and the surface morphology.Raman spectroscopy (FEX, NOST) was conducted to observe strengthened perovskite-bonding properties.Steady-state PL spectra were obtained via LabRAM HV Evolution, Horiba, using lasers of λ excitation = 325 and 532 nm.TRPL spectra (FluoTime 300; Picoquant) were obtained using λ excitation = 398 nm laser.The surface topographies of HTL-deposited perovskite films were characterized by atomic force microscope (NX10; Park System).Trap density was measured via AC voltage perturbation with an amplitude of 0 mV (Zive SP-1, WonATech).The photocurrent-voltage (J-V) curves of the solar cells were measured by solar cell measurement system (K3000; McScience, AM 1.5G, 100 mW cm −2 ), with an active area of 0.09 cm 2 (with mask) and 100 mV s −1 voltage scan rate (reverse scan, from 1200 to 0 eV).The device was measured at 25 • C with 30%-50% RH, and device was initially stabilized for 1 min under illumination.The crystalline solar cell (device ID: S301-K105) was firstly calibrated, and then PSCs were measured three times to see the any J-V drops or rises.

Stability measurement
For the stability test, to avoid invasion of unintentional factors, completed solar cells were encapsulated with cover glass with aid of UV-curable epoxy resin.An 85 • C/85% test was conducted to evaluate device stability under harsh conditions.The encapsulated devices were stored in a chamber with specific conditions of 85 • C and 85% RH.Devices were cooled down for about 30 min before measurements, and then measured in room conditions (25 • C, 50% RH).The initial PCE was measured before the device were moved into a chamber.
To scrutinize long-term stability under continuous illumination, MPPT was conducted.Encapsulated devices were stored under 1 sun illumination and then measured every 24 h under illumination conditions without relaxation process.Initial PCE was measured right after (within 1 min) the device was moved into an MPPT chamber.

DFT calculations
DFT calculations were performed using the projector augmented wave (PAW) method as implemented in the Vienna Ab initio Simulation Package (VASP). 76,77The generalized gradient approximation Perdew-Burke-Ernzerhof exchange-correlation functional revised for solids 78 and Γcentered grid k-meshes with k-spacing of 0.3 A −1 were employed for structural relaxations.The plane-wave cutoff energy was set to 400 eV and the energy and force convergence criteria were set to 10 −6 eV and 0.01 eV Å −6 , respectively.The PAW potentials used in this work are Br, I, Pb, C, N, H, and Cs_sv from VASP distribution.
To investigate the formation enthalpies of CsPbBr 3alloyed FAPbI 3 solid solutions, we reduced the number of configurations by considering the crystal symmetry using SOD codes.FA/Cs substitutions on cation site were conducted over 1 × 1 × 2 and 2 × 2 × 1 supercells for black trigonal phase and yellow hexagonal phase, respectively.The total numbers of configurations for these systems are 2 6 and 2 8 , but symmetry is reduced by 36 and 34 irreducible cation configurations, respectively.I/Br substitutions on halogen site were conducted over conventional cells for both black trigonal phase and yellow hexagonal phase.The total numbers of configurations for these systems are 2 9 and 2 6 , but symmetry reduced 120 and 13 irreducible anion configurations, respectively.The formation enthalpies were calculated through the following equation: +  Cs ( CsI +  PbI 2 ) +  Br ( FABr +  PbBr 2 ) } where χ A is the fraction of A atom, H A is the total enthalpy of A precursor, and  (FA1− CsCs  Cs )Pb(I1− BrBr  Br )3 is the total enthalpy per formula units of the supercell with substitution.

C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors declare no conflict of interest.

4 −
octahedron after crystallization, but these continuous crystals are favored with PbI 4 2− having two DMSO attachments.After crystallization, DMSO evaporates from iodoplumbates and halides in organic/inorganic cations (formamidinium iodide [FAI] or cesium bromide [CsBr]) attach to the void sites (created by DMSO) to finalize the conversion.However, excess iodine, such as PbI 5 3− or PbI 6 4− , retards the effect of DMSO, lowering the device performance.Therefore, we further measured the absorption of perovskite solutions with different concentrations of CsPbBr 3 in Figure 1B.The exact composition of complexes is ill-defined, but CsPbBr 3 surely hinders iodine attachment in iodoplumbate.It is well known that F I G U R E 1 Solution stability analysis.(A) Schematics for the formation of proper iodoplumbates by adding CsPbBr 3 .(B) Effect of PbI 5 3− reduction on fresh solution, and (C) aged solution by adding CsPbBr 3 using UV-vis spectroscopy.(d) Device performance using aged precursor solution.

F
I G U R E 3 Density-functional theory (DFT) calculation of CsPbBr 3 -alloyed FAPbI 3 .Relaxed crystal structure of (A) black trigonal phase and (B) yellow hexagonal phase.Calculated formation enthalpies for trigonal and hexagonal phases of (C) FA/Cs cation substitution and (D) I/Br anion substitution.
Δ f =  (FA1− CsCs  Cs )Pb(I1− BrBr  Br )3 − { (1 −  Cs −  Br )( FAI +  PbI 2 ) This work was supported by the National Research Foundation of Korea grant funded by the Korea Government (MSIT) (no.RS-2023-00212110) and the Technology Development Program (RS-2023-00225289) funded by the Ministry of SMEs and Startups (Korea).This study was supported by the research grant of the University of Suwon in 2022, and Advanced Materials Analysis Center, University of Suwon.Research at Tokyo Tech was supported by the MEXT Element Strategy Initiative Program.