[BMP]+[BF4]−-Modified CsPbI1.2Br1.8 Solar Cells with Improved Efficiency and Suppressed Photoinduced Phase Segregation

With the rapid progress in a power conversion efficiency reaching up to 26.1%, which is among the highest efficiency for single-junction solar cells, organic–inorganic hybrid perovskite solar cells have become a research focus in photovoltaic technology all over the world, while the instability of these perovskite solar cells, due to the decomposition of its unstable organic components, has restricted the development of all-inorganic perovskite solar cells. In recent years, Br-mixed halogen all-inorganic perovskites (CsPbI3−xBrx) have aroused great interests due to their ability to balance the band gap and phase stability of pure CsPbX3. However, the photoinduced phase segregation in lead mixed halide perovskites is still a big burden on their practical industrial production and commercialization. Here, we demonstrate inhibited photoinduced phase segregation all-inorganic CsPbI1.2Br1.8 films and their corresponding perovskite solar cells by incorporating a 1-butyl-1-methylpiperidinium tetrafluoroborate ([BMP]+[BF4]−) compound into the CsPbI1.2Br1.8 films. Then, its effect on the perovskite films and the corresponding hole transport layer-free CsPbI1.2Br1.8 solar cells with carbon electrodes under light is investigated. With a prolonged time added to the reduced phase segregation terminal, this additive shows an inhibitory effect on the photoinduced phase segregation phenomenon for perovskite films and devices with enhanced cell efficiency. Our study reveals an efficient and simple route that suppresses photoinduced phase segregation in cesium lead mixed halide perovskite solar cells with enhanced efficiency.


Introduction
Organic-inorganic hybrid perovskite solar cells have gained significant attention due to their exceptional performance, tunable bandgap, high light-absorption coefficient, low cost and versatile applications, making them highly valuable for commercial use.The certified power conversion efficiency (PCE) of perovskite solar cells has rocketed up to 26.1% at a lab-scale at present, from 3.8% in 2009, which is comparable to that of welldeveloped single-junction silicon solar cells [1][2][3][4][5].Currently, most of the perovskite solar cells with high PCEs over 20% are constructed from organic-inorganic hybrid perovskite materials listed as ABX 3 , where usually A is one or more monovalent methylamine ions such as methylamine ions (CH 3 NH 3 + , referred to as MA + ), ethylamine ions (CH 3 CH 2 NH 3 + , referred to as EA + ), formamidine cations (NH = CHNH 3 + , referred to as FA + ), etc. B is one or more of a divalent metal cation of a carbon family such as lead (Pb 2+ ), tin (Sn 2+ ), germanium (Ge 2+ ) cations), etc.And X is one or more of the halogen ions such as Cl − , Molecules 2024, 29, 1476 2 of 11 Br − , and I − [3,[6][7][8].However, the long-term stability issues of organic-inorganic hybrid perovskite materials under moisture, oxygen, thermal external force and even persistent light are still big burdens for their future application.These issues are caused by the inevitable phase segregation, ion migration, and crystal decomposition in the volatile nature of the organic components [9][10][11][12][13].
To address the issue of instability, it is suggested that the organic cations be replaced with all inorganic Cs + to create CsPbX 3 (X = I − , Br − , Cl − or their mixtures) [14][15][16][17][18]. CsPbX 3 typically has four types of crystal structures, including the cubic structure (α-, Pm3m), the tetragonal structure (β-, P4/mbm), the orthorhombic structure (γ-, Pbnm), and a nonperovskite structure (δ-, Pnma) [19].Research into CsPbX 3 perovskite solar cells started in 2014, when Choi et al. attempted to improve the performance of perovskite solar cells by removing MA + from Cs x MA 1−x PbI 3 solar cells.All inorganic δ-CsPbI 3 (E g = 2.82 eV) perovskite solar cells showed a PCE of only 0.09% [20].In 2015, Gary Hodes et al. first fully investigated whether the organic cation was necessary to obtain devices with high photovoltaic performance.They found that the fabricated inorganic CsPbBr 3 perovskite solar cells, which performed equally as well as the organic ones, were much more temperature stable than the hybrid analogues.This study conducted useful comparative studies between hybrid organic-inorganic and all-inorganic perovskite materials [21].Shortly thereafter, Snaith et al. reported on inorganic CsPbI 3 (E g = 1.73 eV) perovskite solar cells.The current density-voltage (J-V) curve showed an efficiency of up to 2.9%, which was more stable than that of the hybrid organic-inorganic perovskite solar cells [22].In 2017, Snaith's group used vacuum-based vapor deposition for alternating very thin layers of CsI and PbI 2 and obtained a stabilized PCE of 7.8 %, compared to the 4.3 % for spin-coated CsPbI 3 solar cells.The main improvement in vapor-deposited perovskite films was attributed to the much longer carrier lifetimes (>10 µs) compared to the spin-coated ones and their presumed lower trap densities [23].CsPbX 3 perovskite solar cells then entered a period of rapid development through various optimization strategies such as additive engineering, interfacial engineering and precursor solution engineering.Cs-based perovskite solar cells have achieved a high PCE of approximately 21% with a high stability against humidity and heat [24].From this progress, CsPbX 3 materials are emerging as a new research area and will play a significant role in the photovoltaic field in the near future.
Within the CsPbX 3 family, CsPbBr 3 exhibits exceptional moisture and thermal stability.However, due to its large band gap of approximately 2.3 eV, it can only absorb light from the UV region [25].Compared to other CsPbX 3 compounds, cubic CsPbI 3 has the narrowest band gap of 1.73 eV, which allows for a wider range of light absorption.However, its application is limited due to the phase instability caused by transitions between the desirable photoactive perovskite black phase (α-CsPbI 3 , cubic Pm3m) and the undesirable non-perovskite yellow phase (δ-CsPbI 3 , orthorhombic Pnma), although the transition to the yellow phase is reversible [26,27].Br-mixed halogen perovskites (CsPbI 3−x Br x ) are of great interest due to their ability to balance the band gap and phase stability of CsPbX 3 .Additionally, the band gap of CsPbI 3−x Br x can be precisely tuned from ~1.2 to 2.4 eV by adjusting the value of x.This is because the changes in the Br − and Cl − ratio can monotonically shift the absorption onset and lead to lattice transitions for perovskite, resulting an alteration in the photonic bandgaps and, thus, making them promising candidates for the sub-cells of tandem solar cells [17,[28][29][30].
Unexpectedly, some researchers have found that mixed halide perovskites suffer from intrinsic photoinduced halide segregation, with more bromide replacing iodide ions [31][32][33][34].And, numerous studies have demonstrated that the phenomenon of photoinduced phase segregation decreases the V oc of the device and significantly impacts its photoelectric performance [35,36].Eric T. Hoke et al. are the pioneers who discovered the photoinduced phase segregation in mixed halide perovskite films in 2015.They found that when 0.2 < x < 1, the photoluminescence (PL) spectra of the MAPb(Br x I 1−x ) 3 films formed a new emission peak at 1.68 eV and a redshift phenomenon occurred, indicating photoinduced phase segregation.Surprisingly, these changes in photoinduced phase segregation were completely reversible by switching between light and dark [31].In 2017, Yi-Bing Cheng et al. observed iodide-rich phase segregation near grain boundaries in nanoscale all-inorganic CsPbIBr 2 films.The authors found that iodide segregation can occur due to the immiscibility between iodine and bromine, which exacerbates ion migration and hysteresis, thereby affecting the photoelectric performance of the corresponding devices [32].However, the exact mechanism by which the photoinduced phase segregation phenomenon affects the V oc of the devices is not clear.In addition, since the photoinduced phase segregation phenomenon destabilizes the perovskite solar cells, leading to a decrease in the PCE and V oc of the devices, researchers have done a great deal of research work to inhibit and mitigate the occurrence of the photoinduced phase segregation phenomenon in perovskite solar cell devices.For example, Yue Hao et al. proposed a simple and effective strategy to inhibit the photoinduced phase segregation in CsPbIBr 2 films by modifying their crystalline grains with poly(methyl methacrylate) (PMMA).As a result, the carbon electrode-based CsPbIBr 2 PSC exhibited a more suppressed photocurrent hysteresis, coupled with an excellent PCE of 9.21% and a high V oc of 1.307 V. Their work not only provided a new avenue to address the general halide phase segregation issue of CsPbIBr 2 materials, but also provided guidance to achieve the superior performance of opto-electronic devices [37].Very recently, Wei Li et al. unveiled the impact of phase segregation in Cs 0.17 FA 0.83 Pb(I 0.80 Br 0.20 ) 3 films with a bandgap of 1.67 eV through a photoconductive atomic force microscopy.By testing the I-V curves at both grain boundaries and grain interiors with nanoscopic resolution, they identified that iodide-rich phases primarily segregated at defect-enriched grain boundaries under continuous illumination, causing a more significant local open-circuit voltage (V OC ) decrease than that occurring at grain interiors [38].
In this work, all inorganic CsPbI 1.2 Br 1.8 perovskite films with inhibited photoinduced phase segregation are modified by adding different amounts of 1-butyl-1methylpiperidinium tetrafluoroborate ([BMP] + [BF 4 ] − ) (0.5, 1, 2 and 3 mg mL −1 ) to the CsPbI 1.2 Br 1.8 precursor solution, and then its effect on the photovoltaic performance of the corresponding devices and perovskite films under light is investigated.The PCE of hole transport layer-free carbon-based perovskite solar cell is improved from 7.13% for a device with a pristine perovskite film to the highest amount of 8.44% for a device with 1 mg mL −1 of [BMP] + [BF 4 ] − -modified perovskite film.And the PCE of the device increases first with the increase in the doping concentration of [BMP] + [BF 4 ] in the perovskite film, but then decreases gradually, and the stability under light is also increased.The modified CsPbI 1.2 Br 1.8 perovskite film also shows an inhibitory effect on the photoinduced phase segregation phenomenon when compared with the pristine one, with a prolonged time attributed to the reduced phase segregation terminal (pristine: 30.42%@7 min, while the modified: 24.29%@30 min), and the reduced phase segregation rate remaining constant (pristine: k s = 1.53 × 10 −2 s −1 , while modified one: k s = 1.83 × 10 −3 s −1 ).

Preparation of SnO 2 Precursor
The purchased SnO 2 colloidal dispersion was diluted with DI water in a mass ratio of 1:5, followed by ultrasonic dispersing for over 3 h before use.

Preparation of Perovskite Precursor
A total of 259.8 mg of CsI, 46.1 mg of PbI 2 and 330.3 mg of PbBr 2 (in a molar ratio of 1:0.1:0.9) were dissolved in 1 mL of DMSO.The mixture was then placed on a stirrer and stirred continuously overnight at room temperature.Before use, the resulting transparent yellow solution was filtered through a polytetrafluoroethylene (PTFE) filter with a pore size of 0.45 µm.

Various amounts of [BMP] + [BF 4 ]
− powder were added directly to the perovskite precursors with mass concentrations of 0.5, 1, 2, and 3 mg mL −1 .The mixtures were subsequently dispersed using ultrasonic waves for a duration of over 3 h.Before use, the solutions underwent filtration through a polytetrafluoroethylene (PTFE) filter with a pore size of 0.45 µm.

Device Fabrication
The ITO substrates were carefully cleaned using a foamless eradicator and then ultrasonically cleaned with DI water, acetone and ethanol for 15 min in sequence.Before the deposition of the electron transport layers, they were all dried with high-pressure nitrogen gas and then exposed to a UV/ozone cleaner for 15 min.SnO 2 electron transport layers were then spin-coated onto the ITO substrates from a SnO 2 precursor solution.The preparation process of the SnO 2 precursor solution was listed as follows: After cooling, the samples were transferred to a nitrogen-filled glove box to deposit the perovskite layers using a spin-coating method.Briefly, 100 µL of the perovskite precursor without (pristine) or with different mass concentrations of [BMP] + [BF 4 ] − (0.5, 1, 2 or 3 mg mL −1 ) was dropped onto the SnO 2 layer and then spin-coated first at a low speed of 1500 rpm for 20 s and then at a high speed of 5000 rpm for 60 s.The wet perovskite films were then annealed at 35 • C for 15 s and then at 280 • C for 10 min.Finally, compact layers of carbon electrodes were screen printed onto the perovskite films in air, followed by annealing at 120 • C for 15 min on a hot plate.The active area of each electrode was fixed at 0.07 cm −2 via screen printing.

Characterization
The top view and cross-sectional morphologies of the samples were characterized using a Quanta 250FEG (FEI Co., Tokyo, Japan) scanning electron microscope (SEM).The crystalline properties of the perovskite films were characterized by an X-ray diffractometer (XRD, SmartLab, Rigaku, Japan) using Cu K α radiation (40 kV, 30 mA).The scanning speed was 5 • per minute at a step of 0.02 • .Absorption spectra were obtained using a JASCOV-570 (JASCO., Tokyo, Japan) UV/VIS/NIR spectrometer.J-V curves were measured in air using a Keithley 2400 source meter (Keithley Instruments, Inc., Solon, OH, USA) under simulated AM 1.5 G irradiation (100 mA cm −2 ), with the illumination source pre-calibrated using a Si reference cell system (91150V, Newport) (Newport Corporation, Irvine, CA, USA).The scan was performed in 0.02 V steps from 1.2 to −0.1 V (reverse scan) with a 0.1 s time delay between each point.

Results and Discussion
In terms of film morphology, CsPbI x Br 3−x is sensitive to deposition methods, doping and processing conditions such as the spinning speed, environmental temperature and annealing temperature, etc. Figure 1  annealing temperature, etc. Figure 1   It is speculated that the (100) orientation of the CsPbI1.2Br1.8grains is preferentially perpendicular to that of the ITO/SnO2 substrate.This orientation is highly advantageous for effective carrier transport and injection in the corresponding devices.Furthermore, the peak intensity of the perovskite film decreases as the mass concentration of the [BMP] + [BF4] − increases to 3 mg mL −1 .This suggests that higher mass concentrations may weaken the crystallinity of the modified films with increased scattering from grain boundaries or intragranular defects, which is not beneficial for the photovoltaic performance of the final devices [40].
The absorption spectra in Figure 2b indicate that the control film and the [BMP] + [BF4] − -modified CsPbI1.2Br1.8films process a similar sharp-cutting absorption edge at ~605 nm corresponding to a bandgap of ~2.05 eV, which is consistent with the values reported in the literature [39,41].However, this increase in mass concentration leads to a slight decrease in the absorption intensity in the 500-580 nm absorption region, mainly due to the addition of [BMP] + [BF4] − in the perovskite films.8 grains is preferentially perpendicular to that of the ITO/SnO 2 substrate.This orientation is highly advantageous for effective carrier transport and injection in the corresponding devices.Furthermore, the peak intensity of the perovskite film decreases as the mass concentration of the [BMP] + [BF 4 ] − increases to 3 mg mL −1 .This suggests that higher mass concentrations may weaken the crystallinity of the modified films with increased scattering from grain boundaries or intragranular defects, which is not beneficial for the photovoltaic performance of the final devices [40].
The absorption spectra in Figure 2b  annealing temperature, etc. Figure 1   It is speculated that the (100) orientation of the CsPbI1.2Br1.8grains is preferentially perpendicular to that of the ITO/SnO2 substrate.This orientation is highly advantageous for effective carrier transport and injection in the corresponding devices.Furthermore, the peak intensity of the perovskite film decreases as the mass concentration of the [BMP] + [BF4] − increases to 3 mg mL −1 .This suggests that higher mass concentrations may weaken the crystallinity of the modified films with increased scattering from grain boundaries or intragranular defects, which is not beneficial for the photovoltaic performance of the final devices [40].
The absorption spectra in Figure 2b indicate that the control film and the [BMP] + [BF4] − -modified CsPbI1.2Br1.8films process a similar sharp-cutting absorption edge at ~605 nm corresponding to a bandgap of ~2.05 eV, which is consistent with the values reported in the literature [39,41].However, this increase in mass concentration leads to a slight decrease in the absorption intensity in the 500-580 nm absorption region, mainly due to the addition of [BMP] + [BF4] − in the perovskite films.The effect of the [BMP] + [BF 4 ] − additive on the performance of CsPbI 1.2 Br 1.8 solar cell devices are also investigated.Therefore, we depict the all-inorganic device with a hole transport layer-free architecture (Figure 3a), where SnO 2 and carbon are used as the electron transport layer and counter electrode, respectively.The cross-sectional SEM image for a representative cell based on 1 mg mL −1 of a [BMP] + [BF 4 ] − (see Figure 3b for the chemical structure)-modified CsPbI 1.2 Br 1.8 film is shown in Figure 3c (note that the SnO 2 is too thin to be observed).For comparison, we also prepare pristine and 0.5, 2 and 3 mg mL −1 of [BMP] + [BF 4 ] − -modified CsPbI 1.2 Br 1.8 perovskite solar cells, of which the typical J-V curves and corresponding parameters for the best-performing target devices are shown in Figure 3d and Table 1, respectively.As we can see, the pristine device exhibited an open circuit voltage (V OC ) of 1.14 V, a short circuit current density (J SC ) of 12.06 mA cm −2 and a fill factor (FF) of 0.52, resulting in a PCE of 7.13%.With the increase in the doping concentration of [BMP] + [BF 4 ] in the perovskite film, the PCE increases first and then decreases gradually.When the doping concentration of the [BMP] + [BF 4 ] reaches 1 mg mL −1 , the corresponding CsPbI 1.2 Br 1.8 device exhibits the highest PCE, that of 8.44%, with a V OC of 1.21 V, a J SC of 12.32 mA cm −2 and an FF of 0.57.In addition, the additions of [BMP] + [BF 4 ] − to the perovskite light absorber all provide a higher performance compared to that of the pristine one, mainly in terms of the increase in the V OC and FF.The optimized addition mass concentration of the [BMP] + [BF 4 ] − can be determined to be 1 mg mL −1 .

The effect of the [BMP] + [BF4]
− additive on the performance of CsPbI1.2Br1.8solar cell devices are also investigated.Therefore, we depict the all-inorganic device with a hole transport layer-free architecture (Figure 3a), where SnO2 and carbon are used as the electron transport layer and counter electrode, respectively.The cross-sectional SEM image for a representative cell based on 1 mg mL −1 of a [BMP] + [BF4] − (see Figure 3b for the chemical structure)-modified CsPbI1.2Br1.8film is shown in Figure 3c (note that the SnO2 is too thin to be observed).For comparison, we also prepare pristine and 0.5, 2 and 3 mg mL −1 of [BMP] + [BF4] − -modified CsPbI1.2Br1.8perovskite solar cells, of which the typical J-V curves and corresponding parameters for the best-performing target devices are shown in Figure 3d and Table 1, respectively.As we can see, the pristine device exhibited an open circuit voltage (VOC) of 1.14 V, a short circuit current density (JSC) of 12.06 mA cm −2 and a fill factor (FF) of 0.52, resulting in a PCE of 7.13%.With the increase in the doping concentration of [BMP] + [BF4] in the perovskite film, the PCE increases first and then decreases gradually.When the doping concentration of the [BMP] + [BF4] reaches 1 mg mL −1 , the corresponding CsPbI1.2Br1.8device exhibits the highest PCE, that of 8.44%, with a VOC of 1.21 V, a JSC of 12.32 mA cm −2 and an FF of 0.57.In addition, the additions of [BMP] + [BF4] − to the perovskite light absorber all provide a higher performance compared to that of the pristine one, mainly in terms of the increase in the VOC and FF.The optimized addition mass concentration of the [BMP] + [BF4] − can be determined to be 1 mg mL −1 .To further understand the effect of [BMP] + [BF 4 ] − additives on the performance of CsPbI 1.2 Br 1.8 perovskite solar cells, the J-V curves of the pristine and 1 mg mL −1 of [BMP] + [BF 4 ] − -modified perovskite solar cells were tested after exposure to different light durations.Figure 4a-d shows the variations in the four normalized photovoltaic parameters (V OC , J SC , FF and PCE) of the different devices with respect to the light duration.It can be seen that the four parameters of the pristine device show a significant decrease with the increase in light exposure time, whereas the decrease for that of the 1 mg mL −1 [BMP] + [BF 4 ] − -modified device is significantly slower.For example, the PCE of the untreated device gradually decreased from an initial 6.99% to 4.04% after 30 min of light exposure, with a decrease of about 42%, while the PCE of the [BMP] + [BF 4 ] − -treated device gradually decreased from an initial 7.39% to 6.95%, with a decrease of only about 6%.It can be speculated that due to the suppressive effect of the [BMP] + [BF 4 ] − additive on the photoinduced phase segregation phenomenon of the CsPbI 1.2 Br 1.8 films, the stability of the perovskite solar cells under light is increased.
To further understand the effect of [BMP] + [BF4] − additives on the performance of CsPbI1.2Br1.8perovskite solar cells, the J-V curves of the pristine and 1 mg mL −1 of [BMP] + [BF4] − -modified perovskite solar cells were tested after exposure to different light durations.Figure 4a-d shows the variations in the four normalized photovoltaic parameters (VOC, JSC, FF and PCE) of the different devices with respect to the light duration.It can be seen that the four parameters of the pristine device show a significant decrease with the increase in light exposure time, whereas the decrease for that of the 1 mg mL −1 [BMP] + [BF4] − -modified device is significantly slower.For example, the PCE of the untreated device gradually decreased from an initial 6.99% to 4.04% after 30 min of light exposure, with a decrease of about 42%, while the PCE of the [BMP] + [BF4] − -treated device gradually decreased from an initial 7.39% to 6.95%, with a decrease of only about 6%.It can be speculated that due to the suppressive effect of the [BMP] + [BF4] − additive on the photoinduced phase segregation phenomenon of the CsPbI1.2Br1.8films, the stability of the perovskite solar cells under light is increased.Upon irradiation, Br-mixed halogen perovskites will undergo the internal segregation of halogen ions, resulting in the formation of structural domains that are enriched in either bromides or iodides.To investigate the photoinduced phase separation phenomenon of CsPbI1.2Br1.8thin films, we prepared pristine and 1 mg mL −1 of [BMP] + [BF4] − -modified CsPbI1.2Br1.8perovskite films.Both films were then subjected to UV-Vis absorption spectroscopy after irradiation for different times using simulated sunlight at 400 mW cm - 2 AM1.5G.Under this continuous irradiation, the temperature of the films is approximately 40 °C.From the UV-Vis absorption spectra (Figure 5a) of the pristine perovskite film, it can be found that when the light duration was between 0 and 7 min, the intensity Upon irradiation, Br-mixed halogen perovskites will undergo the internal segregation of halogen ions, resulting in the formation of structural domains that are enriched in either bromides or iodides.To investigate the photoinduced phase separation phenomenon of CsPbI 1.2 Br 1.8 thin films, we prepared pristine and 1 mg mL −1 of [BMP] + [BF4] − -modified CsPbI 1.2 Br 1.8 perovskite films.Both films were then subjected to UV-Vis absorption spectroscopy after irradiation for different times using simulated sunlight at 400 mW cm -2 AM1.5G.Under this continuous irradiation, the temperature of the films is approximately 40 • C. From the UV-Vis absorption spectra (Figure 5a) of the pristine perovskite film, it can be found that when the light duration was between 0 and 7 min, the intensity of the absorption peak at the wavelength of 584 nm gradually weakened with the increase in light duration, indicating a decrease in the mixed halide content in the perovskite films.The intensities of the absorption peaks at 560 and 620 nm gradually increase, indicating the formation of Br-rich and I-rich phases.As the illumination time is further increased (10 min and 15 min), the intensity of the absorption peaks at 584 nm gradually increases, while the intensity of the absorption peaks at 560 and 620 nm gradually decreases (Figure 5a).This suggests that after the film reaches the phase segregation terminal (30.42%@7 min, Figure 5b,e), the continued strong illumination leads to the phase recovery phenomenon of the perovskite film portion, i.e., the light-induced self-healing phenomenon, which is consistent with a previous report [17].Similarly, from Figure 5c-e, it can be seen that under continuous light irradiation, the CsPbI 1.2 Br 1.8 (treated with 1 mg mL −1 of [BMP] + [BF 4 ] − ) film initially showed a photoinduced phase segregation phenomenon, and the continuous light irradiation led to a photoinduced self-healing phenomenon when the duration time exceeded the limit of the phase segregation terminal (24.29%@30 min, Figure 3d,e).In addition, the curves representing the phase segregation rates were analyzed using a monoexponential decay function, as shown in Figure 5f.The phase segregation rate constant k s of the [BMP] + [BF 4 ] − -modified CsPbI 1.2 Br 1.8 is 1.83 × 10 −3 s −1 , while that of the pristine one is 1.53 × 10 −2 s −1 .There is an order of magnitude difference between the former and the latter, suggesting that the addition of [BMP] + [BF 4 ] − could reduce the rate of phase segregation.Thus, we can conclude that under continuous light irradiation, the CsPbI 1.2 Br 1.8 films will undergo photoinduced phase separation followed by the light-induced selfhealing phenomenon upon reaching the limit of the phase separation.And, the addition of [BMP] + [BF 4 ] − can inhibit the photoinduced phase separation of CsPbI 1.2 Br 1.8 films.
of the absorption peak at the wavelength of 584 nm gradually weakened with the increase in light duration, indicating a decrease in the mixed halide content in the perovskite films.The intensities of the absorption peaks at 560 and 620 nm gradually increase, indicating the formation of Br-rich and I-rich phases.As the illumination time is further increased (10 min and 15 min), the intensity of the absorption peaks at 584 nm gradually increases, while the intensity of the absorption peaks at 560 and 620 nm gradually decreases (Figure 5a).This suggests that after the film reaches the phase segregation terminal (30.42%@7 min, Figure 5b,e), the continued strong illumination leads to the phase recovery phenomenon of the perovskite film portion, i.e., the light-induced self-healing phenomenon, which is consistent with a previous report [17].Similarly, from Figure 5c-e, it can be seen that under continuous light irradiation, the CsPbI1.2Br1.8(treated with 1 mg mL −1 of [BMP] + [BF4] − ) film initially showed a photoinduced phase segregation phenomenon, and the continuous light irradiation led to a photoinduced self-healing phenomenon when the duration time exceeded the limit of the phase segregation terminal (24.29%@30 min, Figure 3d,e).In addition, the curves representing the phase segregation rates were analyzed using a monoexponential decay function, as shown in Figure 5f.The phase segregation rate constant ks of the [BMP] + [BF4] − -modified CsPbI1.2Br1.8 is 1.83 × 10 −3 s −1 , while that of the pristine one is 1.53 × 10 −2 s −1 .There is an order of magnitude difference between the former and the latter, suggesting that the addition of [BMP] + [BF4] − could reduce the rate of phase segregation.Thus, we can conclude that under continuous light irradiation, the CsPbI1.2Br1.8films will undergo photoinduced phase separation followed by the light-induced self-healing phenomenon upon reaching the limit of the phase separation.And, the addition of [BMP] + [BF4] − can inhibit the photoinduced phase separation of CsPbI1.2Br1.8films.

Conclusions
Our study mainly focuses on the effect of [BMP] + [BF 4 ] − additives on the performance of CsPbI 1.2 Br 1.8 solar cells and their impact on the phenomenon of the photoinduced phase segregation of CsPbI 1.2 Br 1.8 thin films, paving the way for their use as efficient and light irradiation stable photovoltaic devices.The effects of [BMP] + [BF 4 ] − additives on the photovoltaic properties and stability of the devices were investigated by preparing hole transport layer-free carbon-based perovskite solar cells, with pristine or 0.5, 1, 2 and 3 mg mL −1 of [BMP] + [BF 4 ] − -modified CsPbI 1.2 Br 1.8 films as the light harvester layers for each device.
shows the top view SEM images of the pristine CsPbI 1.2 Br 1.8 and CsPbI 1.2 Br 1.8 films doped with different mass concentrations (0.5, 1, 2 and 3 mg mL −1 ) of [BMP] + [BF 4 ] − .It is evident that the surface morphologies of the CsPbI 1.2 Br 1.8 films doped with different concentrations of [BMP] + [BF 4 ] − do not differ significantly from the undoped film.All films have a flat surface without any pinholes.
shows the top view SEM images of the pristine CsPbI1.2Br1.8 and CsPbI1.2Br1.8films doped with different mass concentrations (0.5, 1, 2 and 3 mg mL −1 ) of [BMP] + [BF4] − .It is evident that the surface morphologies of the CsPbI1.2Br1.8films doped with different concentrations of [BMP] + [BF4] − do not differ significantly from the undoped film.All films have a flat surface without any pinholes.

Figure 1 .
Figure 1.Top view SEM images of pristine CsPbI1.2Br1.8 and CsPbI1.2Br1.8films doped with different mass concentrations (0.5, 1, 2 and 3 mg mL −1 ) of [BMP] + [BF4] − .To establish the impact of the [BMP] + [BF4] − additive on the crystallinity and photon absorption of the perovskite thin films, an X-ray diffractometer and UV/VIS spectrometer characterization of the pristine and [BMP] + [BF4] − -modified thin films is carried out.The XRD patterns of the pristine and [BMP] + [BF4] − -modified films all exhibit typical perovskite peaks at 2θ = 15.45°,20.41° and 30.63°, which are assigned to the (100), (110), and (200) lattice planes of the CsPbI1.2Br1.8perovskite crystal, respectively [39], as shown Figure 2a.And there are no impurity phases in any of the films.However, the normalized XRD spectra in Figure S1 (Supporting Information) show that the crystalline grains in the [BMP] + [BF4] − -modified CsPbI1.2Br1.8film all have a much more preferred (100) orientation.It is speculated that the (100) orientation of the CsPbI1.2Br1.8grains is preferentially perpendicular to that of the ITO/SnO2 substrate.This orientation is highly advantageous for effective carrier transport and injection in the corresponding devices.Furthermore, the peak intensity of the perovskite film decreases as the mass concentration of the [BMP] + [BF4] − increases to 3 mg mL −1 .This suggests that higher mass concentrations may weaken the crystallinity of the modified films with increased scattering from grain boundaries or intragranular defects, which is not beneficial for the photovoltaic performance of the final devices[40].The absorption spectra in Figure2bindicate that the control film and the [BMP] + [BF4] − -modified CsPbI1.2Br1.8films process a similar sharp-cutting absorption edge at ~605 nm corresponding to a bandgap of ~2.05 eV, which is consistent with the values reported in the literature [39,41].However, this increase in mass concentration leads to a slight decrease in the absorption intensity in the 500-580 nm absorption region, mainly due to the addition of [BMP] + [BF4] − in the perovskite films.

Figure 1 .
Figure 1.Top view SEM images of pristine CsPbI 1.2 Br 1.8 and CsPbI 1.2 Br 1.8 films doped with different mass concentrations (0.5, 1, 2 and 3 mg mL −1 ) of [BMP] + [BF 4 ] − .To establish the impact of the [BMP] + [BF 4 ] − additive on the crystallinity and photon absorption of the perovskite thin films, an X-ray diffractometer and UV/VIS spectrometer characterization of the pristine and [BMP] + [BF 4 ] − -modified thin films is carried out.The XRD patterns of the pristine and [BMP] + [BF 4 ] − -modified films all exhibit typical perovskite peaks at 2θ = 15.45 • , 20.41 • and 30.63 • , which are assigned to the (100), (110), and (200) lattice planes of the CsPbI 1.2 Br 1.8 perovskite crystal, respectively [39], as shown Figure 2a.And there are no impurity phases in any of the films.However, the normalized XRD spectra in Figure S1 (Supporting Information) show that the crystalline grains in the [BMP] + [BF 4 ] −modified CsPbI 1.2 Br 1.8 film all have a much more preferred (100) orientation.It is speculated that the (100) orientation of the CsPbI 1.2 Br 1.8 grains is preferentially perpendicular to that of the ITO/SnO 2 substrate.This orientation is highly advantageous for effective carrier transport and injection in the corresponding devices.Furthermore, the peak intensity of the perovskite film decreases as the mass concentration of the [BMP] + [BF 4 ] − increases to 3 mg mL −1 .This suggests that higher mass concentrations may weaken the crystallinity of the modified films with increased scattering from grain boundaries or intragranular defects, which is not beneficial for the photovoltaic performance of the final devices[40].The absorption spectra in Figure2bindicate that the control film and the [BMP] + [BF 4 ] −modified CsPbI 1.2 Br 1.8 films process a similar sharp-cutting absorption edge at ~605 nm corresponding to a bandgap of ~2.05 eV, which is consistent with the values reported in the literature [39,41].However, this increase in mass concentration leads to a slight decrease in the absorption intensity in the 500-580 nm absorption region, mainly due to the addition of [BMP] + [BF 4 ] − in the perovskite films.
indicate that the control film and the [BMP] + [BF 4 ] −modified CsPbI 1.2 Br 1.8 films process a similar sharp-cutting absorption edge at ~605 nm corresponding to a bandgap of ~2.05 eV, which is consistent with the values reported in the literature [39,41].However, this increase in mass concentration leads to a slight decrease in the absorption intensity in the 500-580 nm absorption region, mainly due to the addition of [BMP] + [BF 4 ] − in the perovskite films.
shows the top view SEM images of the pristine CsPbI1.2Br1.8 and CsPbI1.2Br1.8films doped with different mass concentrations (0.5, 1, 2 and 3 mg mL −1 ) of [BMP] + [BF4] − .It is evident that the surface morphologies of the CsPbI1.2Br1.8films doped with different concentrations of [BMP] + [BF4] − do not differ significantly from the undoped film.All films have a flat surface without any pinholes.

Figure 1 .
Figure 1.Top view SEM images of pristine CsPbI1.2Br1.8 and CsPbI1.2Br1.8films doped with different mass concentrations (0.5, 1, 2 and 3 mg mL −1 ) of [BMP] + [BF4] − .To establish the impact of the [BMP] + [BF4] − additive on the crystallinity and photon absorption of the perovskite thin films, an X-ray diffractometer and UV/VIS spectrometer characterization of the pristine and [BMP] + [BF4] − -modified thin films is carried out.The XRD patterns of the pristine and [BMP] + [BF4] − -modified films all exhibit typical perovskite peaks at 2θ = 15.45°,20.41° and 30.63°, which are assigned to the (100), (110), and (200) lattice planes of the CsPbI1.2Br1.8perovskite crystal, respectively [39], as shown Figure 2a.And there are no impurity phases in any of the films.However, the normalized XRD spectra in Figure S1 (Supporting Information) show that the crystalline grains in the [BMP] + [BF4] − -modified CsPbI1.2Br1.8film all have a much more preferred (100) orientation.It is speculated that the (100) orientation of the CsPbI1.2Br1.8grains is preferentially perpendicular to that of the ITO/SnO2 substrate.This orientation is highly advantageous for effective carrier transport and injection in the corresponding devices.Furthermore, the peak intensity of the perovskite film decreases as the mass concentration of the [BMP] + [BF4] − increases to 3 mg mL −1 .This suggests that higher mass concentrations may weaken the crystallinity of the modified films with increased scattering from grain boundaries or intragranular defects, which is not beneficial for the photovoltaic performance of the final devices[40].The absorption spectra in Figure2bindicate that the control film and the [BMP] + [BF4] − -modified CsPbI1.2Br1.8films process a similar sharp-cutting absorption edge at ~605 nm corresponding to a bandgap of ~2.05 eV, which is consistent with the values reported in the literature [39,41].However, this increase in mass concentration leads to a slight decrease in the absorption intensity in the 500-580 nm absorption region, mainly due to the addition of [BMP] + [BF4] − in the perovskite films.

Figure 4 .
Figure 4. Evolution of the four normalized parameters obtained from the J-V curves for the representative pristine and 1 mg mL −1 of [BMP] + [BF4] − -modified CsPbI1.2Br1.8perovskite solar cells after exposure to different light durations, aged under a simulated air mass and 1.5 sunlight (400 mW cm −2 ) at room temperature in ambient air: (a) VOC, (b) JSC, (c) FF and (d) PCE.

Figure 4 .
Figure 4. Evolution of the four normalized parameters obtained from the J-V curves for the representative pristine and 1 mg mL −1 of [BMP] + [BF 4 ] − -modified CsPbI 1.2 Br 1.8 perovskite solar cells after exposure to different light durations, aged under a simulated air mass and 1.5 sunlight (400 mW cm −2 ) at room temperature in ambient air: (a) V OC , (b) J SC , (c) FF and (d) PCE.

Figure 5 .
Figure 5. UV-Vis absorption spectra of (a) pristine and (c) 1 mg mL −1 of [BMP] + [BF4] − -modified CsPbI1.2Br1.8perovskite films recorded after different light exposure duration times (AM 1.5G 400 mW cm −2 ).(b,d) Spectra of ΔA values obtained from (a,c), respectively.(e) The phase segregation terminal and (f) spectra of ΔA values as a function of the light exposure duration curves obtained from (a,c).The dashed, doted lines in (f) are linear fits to the data using monoexponential decay.
Our study mainly focuses on the effect of [BMP] + [BF4] − additives on the performance of CsPbI1.2Br1.8solar cells and their impact on the phenomenon of the photoinduced phase segregation of CsPbI1.2Br1.8thin films, paving the way for their use as efficient and light irradiation stable photovoltaic devices.The effects of [BMP] + [BF4] − additives on the photovoltaic properties and stability of the devices were investigated by preparing hole

Figure 5 .
Figure 5. UV-Vis absorption spectra of (a) pristine and (c) 1 mg mL −1 of [BMP] + [BF 4 ] − -modified CsPbI 1.2 Br 1.8 perovskite films recorded after different light exposure duration times 1.5G 400 mW cm −2 ).(b,d) Spectra of ∆A values obtained from (a,c), respectively.(e) The phase segregation terminal and (f) spectra of ∆A values as a function of the light exposure duration curves obtained from (a,c).The dashed, doted lines in (f) are linear fits to the data using monoexponential decay.

Table 1 .
J-V parameters of the representative devices according to Figure3d.

Table 1 .
J-V parameters of the representative devices according to Figure3d.