Fabrication of Porous Lead Bromide Films by Introducing Indium Tribromide for Efficient Inorganic CsPbBr3 Perovskite Solar Cells

In the process of preparing CsPbBr3 films by two-step or multi-step methods, due to the low solubility of CsBr in organic solvents, the prepared perovskite films often have a large number of holes, which is definitely not conducive to the performance of CsPbBr3 perovskite solar cells (PSCs). In response to this problem, this article proposed a method of introducing InBr3 into the PbBr2 precursor to prepare a porous PbBr2 film to increase the reaction efficiency between CsBr and PbBr2 and achieve the purpose of In (Ⅲ) incorporation, which not only optimized the morphology of the produced CsPbBr3 film but also enhanced the charge extraction and transport capabilities, which was ascribed to the reduction of the trap state density and impurity phases in the perovskite films, improving the performance of CsPbBr3 PSCs. At the optimal InBr3 concentration of 0.21 M, the InBr3:CsPbBr3 perovskite solar cell exhibited a power conversion efficiency of 6.48%, which was significantly higher than that of the pristine device.


Introduction
After more than ten years of rapid development, lead halide-based perovskite solar cells have made remarkable achievements, but they seem to be in a vicious circle where high efficiency and high stability are contradictory to each other. Although the power conversion efficiency (PCE) of organic-inorganic hybrid perovskite solar cells (PSCs) has increased from the initial 3.8% to more than 25% at present [1][2][3][4][5], yet due to the strong volatility of common A-site organic cations, such as organic methylammonium (MA + ) and formamidinium (FA + ), the organic components disappear under thermal stress [6,7]. In terms of thermal stability, the all-inorganic cesium-lead halide perovskite CsPbX 3 (X: iodine or bromine), which is formed by using more stable inorganic cesium ions (Cs + ) to completely replace organic cations, usually performs better stability [8][9][10][11] and is not prone to degradation at temperatures above 400 • C [8,12]. This provides the necessary conditions for the long-term stable use of CsPbX 3 PSCs [13][14][15][16]. The key factor, which influences the stability of CsPbX 3 , is the moisture in the air. The presence of humidity changes the phase of the perovskite and reduces the stability of the photovoltaic device, but this does not directly cause the decomposition of CsPbX 3 (mainly I-rich CsPbX 3 ) and the lack of components [8,17,18]. Of course, this phase change is reversible when heated [19].

Device Fabrication
All the following processes were carried out in a fume hood environment, without artificial control of the temperature, humidity, and airflow rate of the surrounding environment. The fluorine-doped tin-oxide-coated glasses were patterned by laser etching and cleaned by ultrasonic with acetone, isopropanol, ethanol, and deionized water. After being dried by high purity nitrogen, the FTO were further cleaned by an ultrasound treatment for 15 min and washed with ethanol. Afterward, the pre-conditioned FTO were spin-coated with 0.15 M titanium diisopropoxide bis(acetylacetonate) in 1-butanol at 5000 rpm for 20 s and were heated at 125 • C for 5 min. After these substrates returned to room temperature, the above procedure was repeated twice with 0.3 M titanium diisopropoxide bis(acetylacetonate) in 1-butanol and compact TiO 2 (c-TiO 2 ) was obtained. After that, the mesoporous TiO 2 (m-TiO 2 ) films were deposited on the above cooling c-TiO 2 by spincoating at 5000 rpm for 30 s by means of TiO 2 paste diluted with ethanol. Further, the obtained layers were dried at 125 • C for 5 min followed by the muffle furnace at 500 • C for 30 min. After the muffle furnace was lowered to room temperature, pre-coated substrates were acquired.
Perovskite films were synthesized by a multistep solution-processing method. 0.03, 0.09, 0.15, 0.21, and 0.27 mmol InBr 3 were added into 1 mL DMF of PbBr 2 (1 M) and stirred. After the InBr 3 was completely dissolved, the DMF mixed solution was spin-coated on pre-coated substrates at 2000 rpm for 30 s and then heated to 90 • C for 30 min. Afterward, the methanol solution of CsBr (0.07 M) was spin-coated on the InBr 3 :PbBr 2 film at 5000 rpm for 30 s and heated to 250 • C for 5 min, and this step was repeated five times. Next, the prepared sample was placed in isopropanol and soaked for 30 min and annealed at 250 • C for 15 min to remove excess CsBr. Finally, the carbon paste was deposited coated on the perovskite films by using the doctor blade coating method and dried at 100 • C for 10 min. The effective area of the back electrode was 3 mm × 3 mm, which defined the active area of each device.

Characterization
The morphologies of the synthesized films and energy-dispersive X-ray spectroscopy (EDS) mapping images were observed by a scanning electron microscope (SEM, FEI MAG-ELLAN 400, FEI, Hillsboro, OR, USA). The crystal structure of the synthesized sample was determined by means of X-ray diffraction (XRD, Cu Kα radiation, λ = 1.5418 Å, Rigaku D/max2500, Tokyo, Japan). The steady-state photoluminescence (PL) spectra of perovskite films were performed using a Renishaw InVia micro-Raman spectroscopy system (Renishaw, Wotton-under-Edge, UK) with a 473 nm excitation source. Ultraviolet photoelectron (UPS) and X-ray photoelectron spectroscopy (XPS) were carried out by an X-ray photoelectron spectrometer (EscaLab Xi+, Thermofisher, Waltham, MA, USA). UV-Vis spectrometer (UV-3600, Shimadzu, Kyoto, Japan) was employed to measure the absorption spectrum in the range of 200 nm to 800 nm. The current-voltage (J-V) characteristics and the external quantum efficiency (EQE) of the fabricated solar devices were measured by a solar cells test system (XP3000, Sanyou, Beijing, China) and an EQE measured system (QTest Station 1000A, CROWNTECH, Inc., Macungie, PA, USA), respectively. The impedance was executed at 10 −1~1 0 7 Hz by using an impedance analyzer in a dark environment (Solartron 1260 coupled to the dielectric interface 1296, Farnborough, UK).

Results and Discussion
In the two-step or multi-step method of preparation of perovskite, the quality of the PbBr 2 film determined the morphology of the following perovskite film. Figure 1 shows the top-view SEM images of PbBr 2 films by introducing different concentrations of InBr 3 . When there was no InBr 3 in the PbBr 2 precursor solution, as shown in Figure 1a, the surface of the obtained sample was rough, and the PbBr 2 grain distribution was extremely uneven, and a large area of exposed m-TiO 2 could be directly observed. When the PbBr 2 precursor solution was introduced into 0.03 M InBr 3 , as shown in Figure 1b, the surface of the PbBr 2 film was flat, and the coverage of the m-TiO 2 film was increased, and the observable exposed m-TiO 2 area was significantly reduced. With the gradual increase in the concentration of introduced InBr 3 (Figure 1c-f), the PbBr 2 film appeared porous, but the number of pores decreased as the concentration of InBr 3 increased. Meanwhile, the porosity volume increased as the concentration of InBr 3 increased. From the cross-sectional view of PbBr 2 shown in Figure S1 (Supplementary Materials), we could clearly see that the pure PbBr 2 film has a flat surface and a uniform thickness of about 50~60 nm, and the m-TiO 2 was filled with PbBr 2 . As the concentration of InBr 3 introduced gradually increased, the thickness of the PbBr 2 film also gradually increased (about 70 nm, 90 nm, 100 nm, 120 nm, 160 nm), and the film roughness increased. The above data could clearly demonstrate that the introduction of InBr 3 could effectively affect the morphology of the PbBr 2 film. The increase in the porosity volume, roughness, and thickness of the PbBr 2 film facilitated the diffusion of the subsequent CsBr solution, increased the reaction efficiency with CsBr, and then achieved the full growth of CsPbBr 3 grains [29,30].
In the two-step or multi-step method of preparation of perovskite, the quality of the PbBr2 film determined the morphology of the following perovskite film. Figure 1 shows the top-view SEM images of PbBr2 films by introducing different concentrations of InBr3. When there was no InBr3 in the PbBr2 precursor solution, as shown in Figure 1a, the surface of the obtained sample was rough, and the PbBr2 grain distribution was extremely uneven, and a large area of exposed m-TiO2 could be directly observed. When the PbBr2 precursor solution was introduced into 0.03 M InBr3, as shown in Figure 1b, the surface of the PbBr2 film was flat, and the coverage of the m-TiO2 film was increased, and the observable exposed m-TiO2 area was significantly reduced. With the gradual increase in the concentration of introduced InBr3 (Figure 1c-f), the PbBr2 film appeared porous, but the number of pores decreased as the concentration of InBr3 increased. Meanwhile, the porosity volume increased as the concentration of InBr3 increased. From the cross-sectional view of PbBr2 shown in Figure S1 (Supplementary Materials), we could clearly see that the pure PbBr2 film has a flat surface and a uniform thickness of about 50~60 nm, and the m-TiO2 was filled with PbBr2. As the concentration of InBr3 introduced gradually increased, the thickness of the PbBr2 film also gradually increased (about 70 nm, 90 nm, 100 nm, 120 nm, 160 nm), and the film roughness increased. The above data could clearly demonstrate that the introduction of InBr3 could effectively affect the morphology of the PbBr2 film. The increase in the porosity volume, roughness, and thickness of the PbBr2 film facilitated the diffusion of the subsequent CsBr solution, increased the reaction efficiency with CsBr, and then achieved the full growth of CsPbBr3 grains [29,30]. To investigate the influence of InBr3 on the structure of PbBr2, XRD patterns of InBr3:PbBr2 films are shown in Figure 2a. It could be seen that Pure PbBr2 was in the orthorhombic phase crystal structure (PDF#84-1181) [23]. When InBr3 was introduced, for all concentrations of InBr3 used, two new diffraction peaks of (011) and (200) crystallographic planes of PbBr2 could be found at 2θ of 20.94° and 22.05°, but no diffraction peak belonging to InBr3 or other protobromides were found. Since InBr3 did not exist in the form of simple In 3+ and Brin the DMF solution, it was self-ionized and formed various complexes [31][32][33]. Therefore, we speculated that in the process of PbBr2 crystal growth, In (Ⅲ) could be in the form of free In 3+ to replace a part of the Pb vacancy or exchange it with Pb, or the In cluster was directly bound to host lattice constituents [33][34][35]. Meanwhile, the PbBr2 crystal was made to grow along multiple ordered crystal orientations. When using the XPS technique to prove the presence of In in PbBr2 films, not surprisingly, characteristic peaks belonging to Br 3d, Pb 4f, and In 3d were found in the XPS spectra for the pure PbBr2 and InBr3:PbBr2 films, as shown in Figure 2b. According to Figure 2c-e, the core level In3d5/2 and 3d3/2 were located at 445.4 eV and 452.9 eV, respectively, and the Pb To investigate the influence of InBr 3 on the structure of PbBr 2 , XRD patterns of InBr 3 :PbBr 2 films are shown in Figure 2a. It could be seen that Pure PbBr 2 was in the orthorhombic phase crystal structure (PDF#84-1181) [23]. When InBr 3 was introduced, for all concentrations of InBr 3 used, two new diffraction peaks of (011) and (200) crystallographic planes of PbBr 2 could be found at 2θ of 20.94 • and 22.05 • , but no diffraction peak belonging to InBr 3 or other protobromides were found. Since InBr 3 did not exist in the form of simple In 3+ and Brin the DMF solution, it was self-ionized and formed various complexes [31][32][33]. Therefore, we speculated that in the process of PbBr 2 crystal growth, In (III) could be in the form of free In 3+ to replace a part of the Pb vacancy or exchange it with Pb, or the In cluster was directly bound to host lattice constituents [33][34][35]. Meanwhile, the PbBr 2 crystal was made to grow along multiple ordered crystal orientations. When using the XPS technique to prove the presence of In in PbBr 2 films, not surprisingly, characteristic peaks belonging to Br 3d, Pb 4f, and In 3d were found in the XPS spectra for the pure PbBr 2 and InBr 3 :PbBr 2 films, as shown in Figure 2b. According to Figure 2c-e, the core level In3d 5/2 and 3d 3/2 were located at 445.4 eV and 452.9 eV, respectively, and the Pb 4f 5/2 and 4f 7/2 peaks in Pb 4f spectrum and Br 3d 3/2 and 3d 5/2 peaks in Br 3d spectrum all moved towards higher binding energies, which showed that Pb-Br interactions were enhanced after In 3+ or In cluster incorporation [36]. Additionally, the EDS mapping was also utilized to confirm the presence of In in the InBr 3 :PbBr 2 films. Figure S2 (Supplementary Materials) demonstrated that all elements were uniformly distributed in the corresponding film, especially, there was no aggregation of In elements. Nanomaterials 2021, 11, x FOR PEER REVIEW 5 of 12 4f5/2 and 4f7/2 peaks in Pb 4f spectrum and Br 3d3/2 and 3d5/2 peaks in Br 3d spectrum all moved towards higher binding energies, which showed that Pb-Br interactions were enhanced after In 3+ or In cluster incorporation [36]. Additionally, the EDS mapping was also utilized to confirm the presence of In in the InBr3:PbBr2 films. Figure S2 (Supplementary Materials) demonstrated that all elements were uniformly distributed in the corresponding film, especially, there was no aggregation of In elements.  Figure 3 depicts the top-view SEM images of perovskite films without and with InBr3 with the corresponding cross-section SEM images inserted in the inset. The size of the crystal grain of the pure CsPbBr3 was quite different, the film uniformity and coverage were also bad, and the bare m-TiO2 could be clearly seen. As the concentration of introduced InBr3 gradually increased (0.03~0.21 M), the coverage of m-TiO2 by CsPbBr3 films also gradually increased, and the size and number of pores in each film showed a downward trend. This morphological change was conducive to the performance of the perovskite cells. However, when the concentration of InBr3 was further increased by 0.27 M, there were again obvious holes in the CsPbBr3 film. This result indicated that the quality and surface CsPbBr3 film depended on the morphology of the corresponding porous InBr3:PbBr2 film greatly that was, the morphology of CsPbBr3 film could be modified by changing InBr3 concentration.  Figure 3 depicts the top-view SEM images of perovskite films without and with InBr 3 with the corresponding cross-section SEM images inserted in the inset. The size of the crystal grain of the pure CsPbBr 3 was quite different, the film uniformity and coverage were also bad, and the bare m-TiO 2 could be clearly seen. As the concentration of introduced InBr 3 gradually increased (0.03~0.21 M), the coverage of m-TiO 2 by CsPbBr 3 films also gradually increased, and the size and number of pores in each film showed a downward trend. This morphological change was conducive to the performance of the perovskite cells. However, when the concentration of InBr 3 was further increased by 0.27 M, there were again obvious holes in the CsPbBr 3 film. This result indicated that the quality and surface CsPbBr 3 film depended on the morphology of the corresponding porous InBr 3 :PbBr 2 film greatly that was, the morphology of CsPbBr 3 film could be modified by changing InBr 3 concentration.
The XRD patterns shown in Figure 4a revealed that all CsPbBr 3 films had a cubic structure (PDF#54-0752) [23], and the positions of the diffraction peaks were not significantly shifted to high or low angles, which demonstrated that, although In cluster could promote growth along multiple ordered crystal orientations, it could not change the phase of CsPbBr 3 . When the concentration of the introduced InBr 3 was 0.00 M and 0.03 M, there existed two peaks located at 11.7 • and 29.4 • , respectively, which belonged to (002) and (213) lattice planes of the CsPb 2 Br 5 phase [37]. As the concentration of InBr 3 was further increased (0.09~0.27 M), no obvious impurity peak belonging to CsPb 2 Br 5 or Cs 4 PbBr 6 phase could be observed. In fact, the control of the reaction rate between CsBr and PbBr 2 was a necessary condition for preparing CsPbBr 3 films with a high purity phase and high coverage. Based on Figure 1, the appropriate concentration of InBr 3 could make the PbBr 2 film have higher porosity, which provided more effective diffusion paths for the diffusion of CsBr methanol solution in the PbBr 2 film, and appropriately increased the contact area between CsBr and PbBr 2 . That could also ensure the full growth of CsPbBr 3 crystal grains and, at the same time, could prevent the formation of the impurity phase due to excessive PbBr 2 or CsBr. However, if the concentration of the InBr 3 introduced into PbBr 2 was too low or too high, it was not conducive to controlling the reaction rate of CsBr and PbBr 2 . In the process of the reaction, due to the incomplete reaction of the precursors or the excessive growth of crystal grains, the morphology of the CsPbBr 3 film was easily deteriorated, accompanied by the formation of byproducts. Further XPS was employed to certify the presence of the incorporated In 3+ in the InBr 3 :CsPbBr 3 film. Figure S3 (Supplementary Materials) exhibited the XPS of Cs 3d, Pb 4f, Br 3d, and In (Ш) 3d for the CsPbBr 3 and the InBr 3 :CsPbBr 3 films, respectively. As seen in Figure 4b, compared with CsPbBr 3 film, two In signals corresponding to In 3d 5/2 and 3d 3/2 core levels were detected in InBr 3 :CsPbBr 3 film, and Cs 3d, Pb 4f, and Br 3d all moved towards higher values, which means that the chemical state of the [PbBr 6 ] 4octahedral was altered and Pb-Br and Cs-Br interactions were enhanced after replacing Pb 2+ (1.7497 Å) with In 3+ (1.6590 Å) with a smaller ion radius accompanied by the size of the [PbBr 6 ] 4octahedral and the voids decreased [35]. The contraction of lattice and the enhancement of the spatial symmetry of the crystal structure caused by the incorporation of In 3+ or In cluster could result in an efficient charge transport along with multiple directions, which perhaps was one of the important factors to improve the performance of CsPbBr 3 cells [35,38]. The EDS mapping was used to characterize the cross-sectional of InBr 3 :CsPbBr 3 , and it was confirmed that In was evenly distributed inside the perovskite, which indicated the successful incorporation of CsPbBr 3 by In ( Figure S4   The XRD patterns shown in Figure 4a revealed that all CsPbBr3 films had a cubic structure (PDF#54-0752) [23], and the positions of the diffraction peaks were not significantly shifted to high or low angles, which demonstrated that, although In cluster could promote growth along multiple ordered crystal orientations, it could not change the phase of CsPbBr3. When the concentration of the introduced InBr3 was 0.00 M and 0.03 M, there existed two peaks located at 11.7° and 29.4°, respectively, which belonged to (002) and (213) lattice planes of the CsPb2Br5 phase [37]. As the concentration of InBr3 was further increased (0.09~0.27 M), no obvious impurity peak belonging to CsPb2Br5 or Cs4PbBr6 phase could be observed. In fact, the control of the reaction rate between CsBr and PbBr2 was a necessary condition for preparing CsPbBr3 films with a high purity phase and high coverage. Based on Figure 1, the appropriate concentration of InBr3 could make the PbBr2 film have higher porosity, which provided more effective diffusion paths for the diffusion of CsBr methanol solution in the PbBr2 film, and appropriately increased the contact area Subsequently, UV-vis Spectrometer, UPS, PL were used to characterize the cells with the FTO/c-TiO 2 /m-TiO 2 /CsPbBr 3 structure. Figure S5a (Supplementary Materials) shows the absorption spectra of the CsPbBr 3 with different concentrations of InBr 3 . The absorption edge of each perovskite film was at approximately 530 nm within the visible region, which revealed that the concentration change of the introduced InBr 3 did not significantly affect the light absorption range of CsPbBr 3 . Correspondingly, the calculated bandgaps (2.34 eV) did not reveal obvious and meaningful changes ( Figure S5b, Supplementary Materials). As the concentration of InBr 3 increased, so did the film's capacity to absorb visible light. This was mainly attributed to the phase-purity of the perovskite film and the full growth of crystal grains, which was beneficial to improve the short current density (J SC ) of the cells. The mechanism of this phenomenon was mainly attributed to the partial substitution of Pb 2+ by In 3+ or In cluster [34,35]. Figure 5a,b present the UPS spectra of the pristine and InBr 3 (0.21M):CsPbBr 3 films. By formula valence band maximum E VBM = 21.22 eV − (E cutoff − E onset ) [39,40], it could be calculated that the valence band (E V ) of CsPbBr 3 and InBr 3 :CsPbBr 3 were −5.60 and −5.28 eV, respectively, which was ascribed to the rearrangement of electrons outside the Cs, Pb, and Br atoms after the incorporation of In 3+ or In cluster [34,35]. Combined with Figure S5b, the corresponding calculated conduction band (E C ) was −3.26 and −2.94 eV, and the energy band diagram of isolated semiconductors of the PSCs using carbon electrodes is plotted in Figure 5c [39,41]. For HTLfree PSCs, Ev of the perovskite should be deeper than the work function (W F ) of the carbon electrode [39] so as to facilitate the extraction of photogenerated holes and reduce the energy loss of the holes during the transmission process [39]. Obviously, the incorporation of In 3+ or In cluster effectively reduced the difference in interface energy levels, thereby facilitating the charge extraction and transfer and enhancing the photovoltaic performance of PSCs. In addition, the PL was conducted to analyze the carrier transfer behavior of CsPbBr 3 and InBr 3 :CsPbBr 3 films. As shown in Figure 5d, all perovskite films showed the typical emission band around 523 nm. InBr 3 (0.21 M):CsPbBr 3 film showed a strong quenching in contrast with the pristine and other CsPbBr 3 films introduced with InBr 3 , which indicated that InBr 3 could effectively inhibit the carrier recombination and enhance the charge extraction ability. The main reason behind this was that the defect density caused by the pinholes of CsPbBr 3 films surface, and the impurity phase of CsPbBr 3 films are obviously improved by adding InBr 3 , and the enhancement of the spatial symmetry of the crystal structure caused by partial substitution of Pb 2+ by In 3+ or In cluster [34,35,39,42].  Subsequently, UV-vis Spectrometer, UPS, PL were used to characterize the cells with the FTO/c-TiO2/m-TiO2/CsPbBr3 structure. Figure S5a (Supplementary Materials) shows the absorption spectra of the CsPbBr3 with different concentrations of InBr3. The absorption edge of each perovskite film was at approximately 530 nm within the visible region, which revealed that the concentration change of the introduced InBr3 did not significantly affect the light absorption range of CsPbBr3. Correspondingly, the calculated bandgaps (2.34 eV) did not reveal obvious and meaningful changes ( Figure S5b, Supplementary Materials). As the concentration of InBr3 increased, so did the film's capacity to absorb visible light. This was mainly attributed to the phase-purity of the perovskite film and the full growth of crystal grains, which was beneficial to improve the short current density (JSC) of the cells. The mechanism of this phenomenon was mainly attributed to the partial substitution of Pb 2+ by In 3+ or In cluster [34,35]. Figure 5a,b present the UPS spectra of the pristine and InBr3 (0.21M):CsPbBr3 films. By formula valence band maximum EVBM = 21.22 eV − (Ecutoff − Eonset) [39,40], it could be calculated that the valence band (EV) of CsPbBr3 and InBr3:CsPbBr3 were −5.60 and −5.28 eV, respectively, which was ascribed to the rearrangement of electrons outside the Cs, Pb, and Br atoms after the incorporation of In 3+ or In cluster [34,35]. Combined with Figure S5b, the corresponding calculated conduction band The HTL-free PSCs were synthesized based on the standard mesoscopic architecture of c-TiO 2 /m-TiO 2 /InBr 3 :CsPbBr 3 /carbon, and the cross-section of the complete device is given in Figure 6a. The current J-V characteristics of relevant devices under reverse scanning are presented in Figure 6b, the corresponding forward scanning curve is shown in Figure S6 (Supplementary Materials), and the key parameters including Jsc, Voc, FF, PCE, and hysteresis index (HI) are summarized in Table 1. The PCE of all devices with InBr 3 introduced were better than that of the pristine ones, and all parameters showed a regular trend of first increasing and then decreasing with an increase of the concentration of InBr 3 introduced. When the concentration of InBr 3 was 0.21 M, the corresponding device exhibited the best performance. Compared with the pristine device, the PCE of InBr 3 (0.21 M):CsPbBr 3 device was significantly improved from 3.29% to 6.48% with the continuously increased J SC of 4.21 and 6.52 mA/cm 2 , Voc of 1.28 and 1.38 V, FF of 0.61 and 0.72, and HI of 0.25 and 0.03. When the InBr 3 concentration was further increased to 0.27 M, the J SC of the device dropped by about 0.5 mA/cm 2 , while the Voc and FF did not change significantly, which was due to the deterioration of the InBr 3 :CsPbBr 3 film morphology. The PEC of PSCs was determined by a variety of complex factors. According to the experimental results of SEM, UV-vis, UPS, and PL, the improvements of Voc and FF were due to the rise in the energy difference between the perovskite conduction band and electron transport layer, thereby reducing the energy loss of the holes in the transmission process. The improved Jsc was not only ascribed to the quality of the InBr 3 :CsPbBr 3 film morphology or the increase in film coverage to absorb more photons to generate more electrons but also reduced vacancy defects in the optimized CsPbBr 3 films to improve the charge extraction and transfer process after the incorporation of In 3+ or In cluster. In addition, the smaller hysteresis of the InBr 3 :CsPbBr 3 cells performed than that of the pristine device might be enabled by the passivation function of InBr 3 to diminish the defects of Pb 2+ and Br − [34,35]. Figure 6c shows the external quantum efficiency (EQE) spectrum. The highest EQE value of 84% was achieved at the InBr 3 concentration of 0.21 M, whereas the reference devices with less or excessive InBr 3 concentration displayed lower EQE responses. This regular change was consistent with the results of J-V characteristics. Additionally, the integrated current density calculated by the EQE curve of each device was very close to the J SC , and the mismatch was less than 5%. Figure 6d demonstrates the Nyquist plots of pristine CsPbBr 3 and InBr 3 (0.21 M):CsPbBr 3 devices measured at a reverse potential of 1.0 V and the corresponding equivalent circuit model. Table S1 (Supplementary Materials) also provides a list of the fitting values of the series resistance (R s ) and the charge recombination resistance (R rec ). After the introduction of InBr 3 , R rec increased from 765 to 1152 Ω, which showed that the incorporation of In had efficient repression of carrier recombination due to the significantly improved film formation quality of perovskites, thus reducing the trap state density and improving carriers mobility [39]. The HTL-free PSCs were synthesized based on the standard mesoscopic architecture of c-TiO2/m-TiO2/InBr3:CsPbBr3/carbon, and the cross-section of the complete device is given in Figure 6a. The current J-V characteristics of relevant devices under reverse scanning are presented in Figure 6b, the corresponding forward scanning curve is shown in Figure S6 (Supplementary Materials), and the key parameters including Jsc, Voc, FF, PCE, and hysteresis index (HI) are summarized in Table 1. The PCE of all devices with InBr3 tine CsPbBr3 and InBr3 (0.21 M):CsPbBr3 devices measured at a reverse potential of 1.0 V and the corresponding equivalent circuit model. Table S1 (Supplementary Materials) also provides a list of the fitting values of the series resistance (Rs) and the charge recombination resistance (Rrec). After the introduction of InBr3, Rrec increased from 765 to 1152 Ω, which showed that the incorporation of In had efficient repression of carrier recombination due to the significantly improved film formation quality of perovskites, thus reducing the trap state density and improving carriers mobility [39].

Conclusions
In the process of preparing the CsPbBr3 film by the multi-step method, we introduced InBr3 into the PbBr2 precursor, so that the PbBr2 film was transformed from a flat membrane to a porous membrane, which was beneficial to improve the reaction efficiency of CsBr and PbBr2, reduced the impurity in CsPbBr3, and optimized the surface morphology,

Conclusions
In the process of preparing the CsPbBr 3 film by the multi-step method, we introduced InBr 3 into the PbBr 2 precursor, so that the PbBr 2 film was transformed from a flat membrane to a porous membrane, which was beneficial to improve the reaction efficiency of CsBr and PbBr 2 , reduced the impurity in CsPbBr 3 , and optimized the surface morphology, and, finally, enabled the performance of CsPbBr 3 PSCs to be significantly improved. When combined with host lattices, the In 3+ or In cluster could effectively suppress the carrier recombination in the CsPbBr 3 film and shift up the Ev of CsPbBr 3 , thereby enhancing the charge extraction and transportation capabilities. When the InBr 3 concentration in the PbBr 2 precursor solution was 0.21 M, the InBr 3 :CsPbBr 3 device presented the best photovoltaic performance with a PCE of 6.48% and, especially the V OC significantly increased by 100 mV compared with the pristine CsPbBr 3 . These research results confirmed that InBr 3 has solid potentials for improving the performance of CsPbBr 3 PSCs and also provided a reference for InBr 3 or some other metal bromide applications in the inorganic CsPbI 3 PSCs field and developmental direction.  Table S1: Electrochemical Impedance Spectroscopy parameters of the pristine and InBr 3 (0.21 M):CsPbBr 3 modules.
Author Contributions: K.C., H.Y. and W.F. conceived the idea; X.M. performed research, analyzed data, and wrote the paper; K.C., Q.L., B.F., P.Z., H.W. and T.G. provided assistance for data acquisition and data analysis. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.