Inkjet‐Printing Controlled Phase Evolution Boosts the Efficiency of Hole Transport Material Free and Carbon‐Based CsPbBr3 Perovskite Solar Cells Exceeding 9%

Hole transport material free carbon‐based all‐inorganic CsPbBr3 perovskite solar cells (PSCs) are promising for commercialization due to its low‐cost, high open‐circuit voltage (Voc) and superior stability. Due to the different solubility of PbBr2 and CsBr in conventional solvents, CsPbBr3 films are mainly obtained by multi‐step spin‐coating through the phase evolution from PbBr2 to CsPb2Br5 and then to CsPbBr3. The scalable fabrication of high‐quality CsPbBr3 films has been rarely studied. Herein, an inkjet‐printing method is developed to prepare high‐quality CsPbBr3 films. The formation of long‐range crystalline CsPb2Br5 phase can effectively improve phase purity and promote regular crystal stacking of CsPbBr3. Consequently, the inkjet‐printed CsPbBr3 C‐PSCs realized PCEs up to 9.09%, 8.59% and 7.81% with active areas of 0.09, 0.25, and 1 cm2, respectively, demonstrating the upscaling potential of our fabrication method and devices. This high performance is mainly ascribed to the high purity, strong crystal orientation, reduced surface roughness and lower trap states density of the as‐printed CsPbBr3 films. This work provides insights into the relationship between the phase evolution mechanisms and crystal growth dynamics of cesium lead bromide halide films.

Hole transport material free carbon-based all-inorganic CsPbBr 3 perovskite solar cells (PSCs) are promising for commercialization due to its low-cost, high open-circuit voltage (V oc ) and superior stability.Due to the different solubility of PbBr 2 and CsBr in conventional solvents, CsPbBr 3 films are mainly obtained by multi-step spin-coating through the phase evolution from PbBr 2 to CsPb 2 Br 5 and then to CsPbBr 3 .The scalable fabrication of highquality CsPbBr 3 films has been rarely studied.Herein, an inkjet-printing method is developed to prepare high-quality CsPbBr 3 films.The formation of long-range crystalline CsPb 2 Br 5 phase can effectively improve phase purity and promote regular crystal stacking of CsPbBr 3 .Consequently, the inkjetprinted CsPbBr 3 C-PSCs realized PCEs up to 9.09%, 8.59% and 7.81% with active areas of 0.09, 0.25, and 1 cm 2 , respectively, demonstrating the upscaling potential of our fabrication method and devices.This high performance is mainly ascribed to the high purity, strong crystal orientation, reduced surface roughness and lower trap states density of the as-printed CsPbBr 3 films.This work provides insights into the relationship between the phase evolution mechanisms and crystal growth dynamics of cesium lead bromide halide films.
the ambient humidity owing to their hydrophobic properties. [11,12]][18] The crystallographic stability of perovskite materials (ABX 3 ) can be assessed by the Goldschmidt's tolerance factor (t): [19] , where R A , R B and R X are the ionic radii of the A-site cation, B-site cation and X-site anion, respectively.In general, stable lead halide perovskites phase can be formed when the t value is in the range of 0.81-1.02. [20]CsPbBr 3 perovskite has a tolerance factor of 0.92, and it can maintain stable perovskite phase over a wide temperature range. [21,22]The CsPbBr 3 nanocrystals have demonstrated an ultra-high carrier mobility of 4500 cm2 V −1 s −1 . [23]The diffusion lengths of CsPbBr 3 single crystal for electrons and holes were reported to be 10.9 and 9.5 μm, respectively. [24]Such high carrier mobilities and long diffusion lengths can ensure effective carriers transport for solar cell application.The excellent stability of CsPbBr 3 perovskite makes it an UV-filter to protect the BHJ film from UV light.In 2015, Kulbak et al. [25] first demonstrated CsPbBr 3 PSCs with a structure of FTO/meso-TiO 2 /CsPbBr 3 /spiro/Au, which achieved a PCE up to 6.2%.Later, Liang et al. fabricated HTMfree CsPbBr 3 C-PSCs with a champion PCE of 6.7%.The devices displayed negligible performance degradation after exposing in humid air (90-95% relative humidity, 25 °C) for 2640 h. [14][31] Duan et al. [16] discovered a multi-step spincoating method to fabricate high-purity CsPbBr 3 films.Cao et al. [32] developed an effective two-step spin-coating method using methanol/ H 2 O mixed solvent to obtain CsPbBr 3 films.What is more, the record PCE for CsPbBr 3 C-PSCs has achieved 11.08% with an outstanding open-circuit voltage (V oc ) of 1.702 V, further demonstrating the application potential of CsPbBr 3 C-PSCs. [33,34]In addition, CsPbBr 3 also has promising applications in tandem solar cells. [35]To further achieve practical applications of perovskite materials, the methods and understanding gained from the lab-scale results need to be transferred into a scalable fabrication process. [36]Poli et al. [37] utilized a screening coating method to prepare HTM-free CsPbBr 3 C-PSCs, realizing a PCE of 8.02%.Zhang et al. [38] reported CsPbBr 3 C-PSCs with a champion PCE of 10.22% using a water-based spray coating method.
Drop-on-demand (DOD) inkjet-printing is a facile and digital deposition technique with the advantages of low-cost, powerful patterning capacity and extremely low material consumption. [39]And it has been widely used not only in research field, but also in large scale industrial applications, such as organic-inorganic hybrid PSCs, organic solar cells, light-emitting diode and etc. [40][41][42][43] Moreover, DOD inkjet-printing can accurately jet picoliter volume of droplets onto the set position of substrates.Therefore, it can realize quantitative and uniform deposition of perovskite precursor solutions, and then manipulate chemical composition of the printed films. [44]In this work, we prepared CsPbBr 3 films by a two-step inkjet-printing method.By utilizing quantitative deposition ability of DOD inkjetprinting, we successfully realized the phase evolution from PbBr 2 to CsPb 2 Br 5 -CsPbBr 3 , and then to pure CsPbBr 3 .We found that the purity and crystal structure of as-printed CsPbBr 3 are highly associated with the quality of the intermediate CsPb 2 Br 5 phase.CsPbBr 3 films with high purity, full-coverage, large and uniform crystal grains have been printed.As a result, the optimized HTM-free CsPbBr 3 C-PSCs device achieves a champion PCE of 9.09% with a high V oc of 1.512 V for small-area device (0.09 cm 2 ), and PCEs of 8.59% (0.25 cm 2 ) and 7.81% (1.00 cm 2 ) for large-area devices, demonstrating the upscaling feasibility of our method and devices.

Results and Discussion
Cesium lead bromide halide (Cs-Pb-Br) contains three distinct phases depending on the ratio of CsBr to PbBr 2 , namely PbBr 2 -rich CsPb 2 Br 5 (CsBr:PbBr 2 = 1:2), CsPbBr 3 (CsBr:PbBr 2 = 1:1), and CsBr-rich Cs 4 PbBr 6 (CsBr:PbBr 2 = 4:1).As illustrated in Figure 1a, CsPb 2 Br 5 possesses a 2D structure with Pb 2 Br À 5 layers separated by Cs + cations. [45]sPbBr 3 presents a 3D perovskite structure with corner shared [PbBr 6 ] 4− octahedra.Cs 4 PbBr 6 has a 0D structure based on isolated [PbBr 6 ] 4− octahedra disconnected surrounded by Cs + cations. [46]The band gaps for pure CsPbBr 3 , CsPb 2 Br 5 and Cs 4 PbBr 6 crystals were measured to be 2.30, 3.1 and 3.90 eV, respectively. [45,47]Additionally, CsPb 2 Br 5 is an indirect bandgap semiconductor and exhibits a photoluminescence (PL) inactive behavior due to its symmetry-forbidden transition. [48]By contrast, CsPbBr 3 and Cs 4 PbBr 6 are direct band gap semiconductors, and exhibit green PL emission.The green PL emission of CsPbBr 3 crystals is ascribed to the direct electron-hole radiative recombination, while the green PL emission of Cs 4 PbBr 6 crystals is caused by the mid-gap state transition. [47]The three distinct phases of Cs-Pb-Br differ greatly not only in structure but also in optoelectronic properties.Liu et al. [17] reported that CsPbBr 3 films containing CsPb 2 Br 5 and Cs 4 PbBr 6 phases exhibited shorter carrier lifetimes, signifying defect density increases with CsPb 2 Br 5 and Cs 4 PbBr 6 impurity phases in CsPbBr 3 films.Zhang et al. [49] found that charge transport was blocked by excessive CsPb 2 Br 5 in CsPbBr 3 films.Evidently, the existence of CsPb 2 Br 5 and Cs 4 PbBr 6 phases in CsPbBr 3 films is not conducive to the performance of CsPbBr 3 -based solar cells. [50]n this work, we prepared high-quality CsPbBr 3 films by a DOD inkjet-printer (Figure S1a,b, Supporting Information), which can realize quantitative deposition of liquid inks.As shown in Figure S1c-e, Supporting Information, the jetted droplet volume was controlled at around 90 pL by regulating the input voltage waveform.The parameters of the input voltage waveform and deposition patterns are summarized in Table S1, Supporting Information.As shown in Figure 1b, a 0.9 M PbBr 2 solution in a mixed solvent of DMF and DMSO (9:1 volume ratio) was first printed onto the meso-TiO 2 substrates and annealed at 100 °C to promote the crystallization of PbBr 2 and form a solid meso-TiO 2 /PbBr 2 scaffold layer (Figure S2, Supporting Information).Subsequently, a 0.09 M CsBr solution in a mixed solvent of methanol and deionized water (10:1 volume ratio) was printed onto PbBr 2 layers and then annealed at 150 °C to boost the diffusion of CsBr into the PbBr 2 layers.The ratio of PbBr 2 to CsBr is precisely adjusted by controlling the total number of deposition droplets of the corresponding ink to meet the stoichiometric requirements of CsPbBr 3 .Finally, the quality of as-printed raw CsPbBr 3 films were further enhanced by the post-treatment.
Energy Environ.Mater.2024, 7, e12543 Figure S3, Supporting Information presents the cross-sectional scanning electron microscopy (SEM) images of printed CsPbBr 3 films annealed at different temperature.As depicted, printed CsPbBr 3 films display a large quantity of interface defects at low annealing temperature (150 °C).And these interface defects disappear at higher annealing temperatures (250-450 °C), showing the interface healing effect of the post-annealing treatment. [51]However, these films exhibit disorder and excessive large crystals on the upper surface, which is not conducive to the transport of carriers at the interface. [52]The reason for this runaway overgrowth of crystals will be discussed later in this paper.The above results show that the post-annealing temperature has a decisive effect on the morphology of interface and film.Based on the as-printed CsPbBr 3 films, HTM-free C-PSCs with an architecture of FTO/c-TiO 2 / meso-TiO 2 /CsPbBr 3 /Carbon were fabricated.Figure S11, Supporting Information presents the photocurrent density-voltage (J-V) curves of the printed CsPbBr 3 C-PSCs at different post-annealing temperatures, and the corresponding J-V parameters are summarized in Table S3, Supporting Information.Unfortunately, these printed CsPbBr 3 C-PSCs devices show much lower PCEs (<5%) compared with the state-of-theart records of CsPbBr 3 C-PSCs. [26,34]Liu et al. [53] fabricated highquality CsPbBr 3 films by employing porous CsPb 2 Br 5 as an intermediate.Moreover, most of the reported high-performance CsPbBr 3 PSCs are prepared by multi-step methods. [16,17,26,34]Considering CsPb 2 Br 5 film could be a promising intermediate phase for pure CsPbBr 3 film, we then improved the deposition of CsBr droplets into two consecutive steps: 1) half of the CsBr droplets were printed onto the meso-TiO 2 / PbBr 2 layer, aiming to obtain high-quality intermediate CsPb 2 Br 5 films, denoted as IJP-inter; 2) the other half of the CsBr droplets were printed onto the IJP-inter sample, represented as IJP-CsPbBr 3 .For the film that all CsBr droplets were printed onto the meso-TiO 2 /PbBr 2 layer in a single printing step, it was named as Control-CsPbBr 3 for comparison.Figure S4, Supporting Information depicts the energy-dispersive X-ray (EDX) mapping images of the printed CsPbBr 3 films, the spatial distribution of Cs, Pb and Br elements determines the CsPbBr 3 layer region.Based on the EDX spectra in Figures S5 and S6, Supporting Information, the elemental ratios of the Control (Cs:Pb:Br = 1:1.19:3.8)and IJP-CsPbBr 3 (Cs:Pb:Br = 1:1.17:3.3)show relatively ideal stoichiometry, proving accurate and controllable quantitative deposition of PbBr 2 and CsBr inks.
To study the crystal formation and growth mechanisms of Cs-Pb-Br films under stoichiometric management, the microstructures of all printed films were investigated.Figure 1c,d presents the topview and cross-sectional SEM images of different printed films, including the Control-CsPbBr 3 , IJP-inter, and IJP-CsPbBr 3 .As illustrated, most of the crystal grains in the Control-CsPbBr 3 sample are small, but there exist a few chaotic and rather large crystal grains, leading to the formation of pinhole defects and the increase in surface roughness.The IJP-inter sample, which is the intermediate film, also shows relatively poor coverage and irregular surface.Notably, the IJP-CsPbBr 3 sample exhibits uniform, full-coverage and vertical aligned crystal grains, indicative of favorable charge transport kinetics. [29]By optimizing the deposition process of CsBr, the surface and cross-sectional morphologies of printed CsPbBr 3 films were significantly improved.As shown in Figure 1d, the average thicknesses of the Control and IJP-CsPbBr 3 were calculated to be 288 AE 130 nm and 346 AE 23 nm, respectively.The IJP-CsPbBr 3 film has only one major thickness distribution range, while the Control-CsPbBr 3 exhibit two major thickness distribution ranges, indicating crystal growth process is quite different in two different printing processes.Compared with the Control-CsPbBr 3 sample, the grain size of the IJP-CsPbBr 3 sample also increases, which can be attributed to the recrystallization and coalescence of small crystal grains during the second deposition process of CsBr.
The phase evolution reactions of PbBr 2 -CsBr system can be summarized as follows: [16,17,54] According to the above phase evolution reactions, the phase evolution results are determined by the ratio of PbBr 2 to CsBr, indicating that the phase evolution mechanism is closely related to the diffusion dynamic of CsBr.To further investigate the element spatial distribution throughout the film, the SEM-EDX element line scanning was performed.As shown in Figure 1e and Figure S7, Supporting Information, Pb 2+ ions are uniformly distributed along the vertical direction in all samples, while Cs + and Br − ions exhibit distribution changes in different samples.In the Control-CsPbBr 3 sample, Cs + and Br − ions are evenly distributed on film surface, and then decrease gradually from surface to interface, suggesting composition variation along the vertical direction.
By contrast, in the IJP-CsPbBr 3 sample, all elements are regularly distributed, indicating that the composition uniformity was improved after the introduction of the IJP-inter as an intermediate during the evolution of PbBr 2 to CsPbBr 3 .
To investigate the phase evolution mechanism of two different printing progress, we have conducted X-ray diffraction (XRD) measurements.Figure S9a-c, Supporting Information presents the XRD patterns comparison of the IJP-CsPbBr 3 with the orthorhombic γphase, tetragonal β-phase, and cubic α-phase CsPbBr 3 .The main characteristic peak of the IJP-CsPbBr 3 fits best with that of the cubic αphase CsPbBr 3 and peak splitting was not observed, indicating αphase CsPbBr 3 is the dominant phase. [55]In addition, the characteristic peaks of the IJP-CsPbBr 3 shifted to a larger 2θ angle by about 0.3°(Figure S9d, Supporting Information), which can be ascribed to the experimental error, the zero error of the XRD instrument and the variation of lattice parameters of samples prepared by different methods. [56,57]Notably, the IJP-CsPbBr 3 also has a minor peak at 28.8°, belonging to orthorhombic phase CsPbBr 3 (Figure S9a-c, Supporting Information), which can be ascribed to the inevitable phase transition of CsPbBr 3 from cubic phase to orthorhombic phase during storage at room temperature. [58]As shown in Figure 2a,b, both the Control and IJP-CsPbBr 3 samples presented cubic CsPbBr 3 perovskite phase (PDF#54-0752) with the (100), ( 110), (111), and (200) lattice planes at 15.4°, 21.8°, 25.7°, and 30.9°, respectively.Besides, the diffusion peaks at 11.7°assigned to the (002) lattice plane of the tetragonal CsPb 2 Br 5 phase, and 12.8°allocated to the characteristic peak of the rhombohedral Cs 4 PbBr 6 phase, were also identified in the Control-CsPbBr 3 sample, [16,59] consistent with the previous conclusion derived from the SEM and EDX results.Chen et al. [60] reported that the formation enthalpy of Cs 4 PbBr 6 (Reaction 6, ΔH calculated = −38 kJ mol −1 ) is lower than that of CsPbBr 3 (Reaction 2, ΔH calculated = −22 kJ mol −1 ).It demonstrates that the formation of Cs 4 PbBr 6 is a thermodynamic and kinetic favorable process in CsBrrich environment.Meanwhile, the IJP-inter sample exhibits a strong peak at 12.8°and a small peak at 30.9°, corresponding to the characteristic peaks of CsPb 2 Br 5 and CsPbBr 3 phase, respectively.Remarkably, no impurity phase peak is discovered in the IJP-CsPbBr 3 sample, demonstrating high phase purity CsPbBr 3 films were formed.Furthermore, the stronger peak intensity also indicates improved crystallinity of the printed CsPbBr 3 after introducing the intermediate IJP-inter film (Figure S10a, Supporting Information).
To study the optical absorption property of the printed Cs-Pb-Br films, we have performed ultraviolet-visible (UV-vis) spectroscopy measurement.As shown in Figure 2c, all printed Cs-Pb-Br films present similar absorption edges at around 540 nm, and the corresponding Tauc plots (Figure S12, Supporting Information) indicate band gaps of 2.32 and 2.31 eV for the Control and IJP-CsPbBr 3 film, respectively, which is consistent with previous reports. [26,34,61]Moreover, the UV-vis absorption spectra implied a larger absorption intensity for the IJP-CsPbBr 3 sample, which can be attributed to the high phase purity, aligned crystal grains and larger film thickness of the IJP-CsPbBr 3 sample, as certified by the SEM and XRD results.
To study the crystal formation and growth mechanisms during the phase evolution from PbBr 2 to CsPbBr 3 in two different printing processes, 2D grazing incidence wide-angle X-ray scattering (GIWAXS), which can expose crystallinity and crystal orientation characteristics of films, was performed on the printed Cs-Pb-Br films.scattering rings of different crystal planes, where q stands for the scattering vector in reciprocal space.As expected, both the Control and IJP-CsPbBr 3 samples present strong scattering rings of cubic CsPbBr 3 perovskite crystal, which involves (100) (0.97 Å−1 < q < 1.03 Å−1 ) and (110) (1.38 Å−1 < q < 1.45 Å−1 ) crystal plane diffractions. [62]Meanwhile, the Control-CsPbBr 3 sample also illustrated scattering rings of CsPb 2 Br 5 crystal (0.74 Å−1 < q < 0.78 Å−1 ), consistent with the XRD and EDX results.In contrast, there is no diffraction feature of CsPb 2 Br 5 crystal in the IJP-CsPbBr 3 sample, while the IJP-inter sample contains strong scattering rings of CsPb 2 Br 5 crystal with weak scattering rings of CsPbBr 3 crystal.Figure 2g depicts the out-of-plane GIWAXS diffraction patterns of the Control-CsPbBr 3 , IJP-inter and IJP-CsPbBr 3 samples.Compared with other samples, the IJP-CsPbBr 3 sample shows the highest peak intensity at the (110) plane of CsPbBr 3 , suggesting higher crystallinity of CsPbBr 3 crystal in this sample.Simultaneously, the diffraction peak of the IJP-CsPbBr 3 sample shifts slightly to a higher q value, indicative of a smaller plane spacing.In the XRD results, it is also observed that the diffraction peak of the IJP-CsPbBr 3 sample shifts to a larger 2θ value (Figure S10b, Supporting Information).Zhao et al. [26] calculated that the lattice volume would expand by 2.18 times during the phase transition from the orthorhombic PbBr 2 phase to the cubic CsPbBr 3 phase.And the lattice volume would enlarge by 2.99 times from the orthorhombic PbBr 2 to the tetragonal CsPb 2 Br 5 . [48]It implies that the lattice volume is going to contract during the phase transition from tetragonal CsPb 2 Br 5 to cubic CsPbBr 3 .The co-existence of CsPb 2 Br 5 , CsPbBr 3 and Cs 4 PbBr 6 phases in the Control-CsPbBr 3 sample indicates that the phase evolution process is inconsistent in different region of the film.As a result, expansion and contraction of crystal lattices may occur in different regions of the film, which may increase residual stress and produce lattice defects, resulting in relatively irregular crystal microstructure. [26]Compared with the Control-CsPbBr 3 sample, the IJP-CsPbBr 3 sample possesses higher phase purity, better The crystal orientations of different printed films were abstracted from the 2D GIWAXS patterns.Figure 2h shows the normalized azimuth angle plots along the (100) plane (q ≈ 1 Å−1 ) of cubic CsPbBr 3 crystal for the Control and IJP-CsPbBr 3 samples.Interestingly, the IJP-CsPbBr 3 sample presents a sharp peak at the azimuth angle of around 90°, while the Control-CsPbBr 3 sample shows a relatively broad peak from the azimuth angle of about 60°to 120°.The narrower full-width half-maximum (FWHM) of the IJP-CsPbBr 3 sample indicates a stronger preferential growth of CsPbBr 3 crystal along the out-of-plane direction, which is beneficial for the charge transport of the devices. [63,64]Figure 2i illustrates the normalized azimuth angle plots along the (002) plane (q ≈ 0.76 Å−1 ) of CsPb 2 Br 5 for the Control-CsPbBr 3 and IJP-inter samples.It is found that the growth of CsPb 2 Br 5 crystal also shows relatively strong orientation preference along the out-of-plane direction.Intriguingly, compared with the Control-CsPbBr 3 sample, the peak intensity of CsPb 2 Br 5 phase in the IJP-inter sample is stronger and the FWHM is smaller (Figure 2g,i).Based on these results, the crystal stacking dynamics for the Control and IJP-CsPbBr 3 samples are proposed, as illustrated in Figure S13, Supporting Information.As the ratio of PbBr 2 to CsBr was controlled at 2:1 in the IJP-inter sample, the interconversion of CsPb 2 Br 5 and CsPbBr 3 is beneficial to the formation of the longrange CsPb 2 Br 5 phase (Reactions 1-4).As depicted in Figure 1a, CsPb 2 Br 5 presents a 2D structure consisting of Cs + sandwiched between Pb 2 Br 5 − layers.Based on the GIWAXS results, it can be inferred that the IJP-inter sample contains longer and more orderly aligned Pb 2 Br 5 − layers than that in the Control-CsPbBr 3 sample.The formed regular CsPb 2 Br 5 phase can promote the vertical diffusion orientation of CsBr, [59,65] and the vertical diffusion orientations of CsBr is favorable to the formation of long-range ordered CsPb 2 Br 5 phase.The vertical crystal growth and diffusion orientations can form a positive feedback loop, resulting in the formation of long-range vertical aligned CsPb 2 Br 5 phase in the IJPinter sample.Figure S13a,b, Supporting Information shows the atomic force microscopy (AFM) images of the Control and IJP-CsPbBr 3 samples, respectively.The root-mean-square roughness (RMS) was examined to be 56 AE 5 nm and 34 AE 3 nm for the Control and IJP-CsPbBr 3 samples, respectively.The Control-CsPbBr 3 presents much bigger Δ Height (212 AE 28 nm) compared with that of the IJP-CsPbBr 3 sample (120 AE 13 nm) (Figure S14c,d, Supporting Information).The improvement of surface morphology demonstrates that the formation of long-range crystalline CsPb 2 Br 5 is beneficial to the uniform and compact crystal stacking of CsPbBr 3 .
To study the phase composition along the vertical direction, GIWAXS measurements were performed by using different incident angles.The larger the incident angles, the larger penetration depths of X-rays, [66] allowing us to study the phase uniformity of the film along the vertical direction.Figure 3a,b shows the out-of-plane GIWAXS diffraction patterns of the Control-CsPbBr 3 and IJP-inter samples under incident angles of 0.2°, 0.4°, and 0.7°obtained from the 2D GIWAXS patterns in Figure S19, Supporting Information.In the Control-CsPbBr 3 sample, CsPbBr 3 phase is most dominant at incident angle of 0.2°, while CsPbBr 3 phase is most dominant at incident angle of 0.7°in the IJP-inter sample.As shown in Figure 3c, in the Control-CsPbBr 3 sample, the peak intensity ratio between (002) crystal plane of CsPb 2 Br 5 and (110) crystal plane of CsPbBr 3 raised from 0.59 to 2.11 as the incident angle increased from 0.2°to 0.7°, suggesting that the CsPb 2 Br 5 phase becomes dominant as the vertical depth increase.Similarly, in the IJP-inter sample, the peak intensity ratio increased from 2.74 to 4.83 as the incident angle raised from 0.2°to 0.7°, indicating that the CsPbBr 3 phase mainly exists on the upper surface.
X-ray photoemission spectroscopy (XPS) analysis was conducted to explore the surface chemistry of the Control and IJP-CsPbBr 3 films.The full survey scan proves the existence of Cs, Pb, and Br in the Control and IJP-CsPbBr 3 films (Figure S8, Supporting Information).Highresolution XPS spectra for the core level of Cs 3d, Pb 4f and Br 3d are shown in Figure 3d-f.In the spectrum of the IJP-CsPbBr 3 film, the peaks at 724.3 and 738.3 eV, 138.8 and 143.7 eV, and 68.5 and 69.5 eV can be assigned to the binding energies for Cs 3d 5/2 and Cs 3d 3/2 , Pb 4f 7/2 and Pb 4f 5/2 , and Br 3d 5/2 and Br 3d 3/2 , respectively, which is consistent with previous reports for CsPbBr 3 . [67,68]In contrast, peaks of Cs 3d 5/2 and Cs 3d 3/2 , Pb 4f 7/2 and Pb 4f 5/2 , and Br 3d 5/2 and Br 3d 3/2 were located at 724.2 and 738.2 eV, 137.9 and 142.8 eV, and 67.9 and 68.9 eV, respectively, in the spectrum of the Control-CsPbBr 3 film.The peaks of Pb 4f and Br 3d core levels of the Control-CsPbBr 3 film are similar to the reported values for Cs 4 PbBr 6 single crystal, [60,69] demonstrating our speculation that formation and growth of Cs 4 PbBr 6 crystal on the top surface of the Control-CsPbBr 3 film.Moreover, the peaks of Cs 3d, Pb 4f and Br 3d core levels of the IJP-CsPbBr 3 film shifted 0.1, 0.9 and 0.6 eV towards higher binding energy compared with the Control-CsPbBr 3 film.The changes in binding energy means that the chemical bonding properties in the formed CsPbBr 3 are different. [70]igher binding energy indicates stronger Cs-Pb, Cs-Br and Pb-Br interactions.Accordingly, the IJP-CsPbBr 3 sample has a more compact crystal structure, [71] consistent with the XRD and GIWAXS results.
Based on the above results, the diffusion dynamics of CsBr and CsPbBr 3 film growth process of the Control-CsPbBr 3 (Route 1) and IJP-CsPbBr 3 (Route 2) samples are summarized in Figure 3g.As shown in Route 1 in Figure 3g, when equal amount of CsBr was printed onto PbBr 2 layer, CsBr-rich region and PbBr 2 -rich region appeared on the surface and at the interface respectively.A CsBr-rich environment can trigger a phase transition to the Cs 4 PbBr 6 impurity phase, which can be accelerated by heating, [60,72] resulting in low-quality and rough surface of Control-CsPbBr 3 samples.As illustrated in Route 2 in Figure 2a, in the IJP-inter sample due to the PbBr 2 -rich environment, CsPb 2 Br 5 phase is predominated (Reaction 1), and a few CsPbBr 3 phase would mainly exists on the top surface (Reaction 3).In contrast, Cs + and Br − ions are homogenously distributed in the IJP-CsPbBr 3 sample, [71,72] confirming the uniform composition of the entire film.Obviously, the induced intermediate CsPb 2 Br 5 phase enhances the diffusion orientation of CsBr in the vertical direction, which can be attributed to the vertical crystal growth orientation of CsPb 2 Br 5 phase and the interconversion of CsPb 2 Br 5 and CsPbBr 3 in PbBr 2 -rich environment (Reactions 1-4).
HTM-free all-inorganic C-PSCs with an architecture of FTO/c-TiO 2 / meso-TiO 2 /CsPbBr 3 /Carbon were fabricated to evaluate the photovoltaic performance of the printed CsPbBr 3 films (Figure 4a,b).As shown in Figure S15, Supporting Information, a uniform CsPbBr 3 layer was deposited on the meso-TiO 2 substrate.The valence band (VB) and conduction band (CB) energy levels of the Control and IJP-CsPbBr 3 films were calculated by the ultraviolet photoelectron spectra and the optical bandgap (Figure S12, Supporting Information). [13]Figure 4c illustrates the energy levels of different functional layers and the transfer pathway for photogenerated carriers, indicating sufficient electrons extraction from the CsPbBr 3 conduction band to the TiO 2 conduction band and feasible hole injection from the CsPbBr 3 valence band to the carbon electrode. [16]In addition, the VB of the IJP-CsPbBr 3 film (−5.64 eV) is closer to the working function of carbon electrode than that of the Control-CsPbBr 3 film (−5.65 eV), promoting the hole Energy Environ.Mater.2024, 7, e12543 extraction at the interface. [73]Figure 4d presents the J-V curves of the champion device based on Control and IJP-CsPbBr 3 , recorded under the standard AM 1.5G illumination condition.The devices based on Control-CsPbBr 3 only obtain a best PCE of 4.67% with a short-circuit current density (J sc ) of 4.56 mA cm −2 , a V oc of 1.383 V and a fill factor (FF) of 74.1%.By developing long-range crystalline CsPb 2 Br 5 intermediate during the phase evolution process, all photovoltaic parameters show significant improvement.The champion PCE of the devices based on IJP-CsPbBr 3 boost up to 9.09% along with a J sc of 7.36 mA cm −2 , a V oc of 1.512 V and an FF of 81.7%.The key J-V parameters for a batch of 25 devices, including J sc , V oc , FF and PCE are shown in Figure S16, Supporting Information and summarized in Table 1.Obviously, the devices based on IJP-CsPbBr 3 achieved higher average J sc (7.11 AE 0.26 vs 4.45 AE 0.25 mA cm −2 ), V oc (1.49AE 0.03 vs 1.35 AE 0.04 V) and FF (80.3 AE 2.1 vs 71.3 AE 2.9%) in comparison with the devices based on Control-CsPbBr 3 .All these improvements contribute to the remarkable increase in the average PCE (8.59 AE 0.5 vs 4.27 AE 0.4%).As shown in Figure S17, Supporting Information and summarized in Table S4, Supporting Information, negligible hysteresis of the J-V characteristics was observed for both devices at reverse and forward scan directions.
The external quantum efficiency (EQE) spectra for both devices were determined to confirm their spectral response and photoelectric conversion efficiency, as illustrated in Figure 4e.Both devices present efficient charge extraction from 360 to 520 nm, and show similar onset at around 540 nm, consistent with the UV-vis absorbance results (Figure 2c).The obviously increased absorption of the devices based on  IJP-CsPbBr 3 also reveals the improvement of film quality, consistent with previous reports. [16]The J sc integrated from the IPCE spectra for devices based on Control and IJP-CsPbBr 3 was measured to be 4.30 and 7.04 mA cm −2 , in agreement with those obtained from the champion J-V curves with an error below 5%.
To evaluate the operation stability of the printed CsPbBr 3 C-PSCs, we have measured the stabilized PCE by maximum power point tracking at 1.3 V.As shown in Figure 4f, the printed CsPbBr 3 C-PSCs show no significant drop in PCE after 12 h under continuous one-sun illumination, which further confirm the outstanding operational stability of the printed CsPbBr 3 C-PSCs.
To demonstrate the upscaling potential of the printed devices, the effect of the mask sizes (0.09, 0.25, and 1 cm 2 ) on device performance was investigated.Figure 4g,h show the champion J-V curves of the devices with different metal mask sizes, and the corresponding J-V parameters variation are shown in Figure 4i and Figure S18, Supporting Information.The large-area Control-CsPbBr 3 -based devices realized PCEs of 3.79% and 3.42% for the active areas of 0.25 and 1 cm 2 , respectively.By contrast, the IJP-CsPbBr 3 -based devices can achieve PCEs of 8.59% and 7.81% when the active area is extended to 0.25 and 1 cm 2 , respectively.][76][77][78] As the active area increases from 0.09 to 1 cm 2 , the J sc , V oc , and FF declines were observed for both devices.However, the IJP-CsPbBr 3 -based device shows a reduction in J sc and FF drops (1.6% and 9.1%) compared with the Control-CsPbBr 3 -based devices (7.24% and 19.3%) (Figure S18, Supporting Information).As a result, the efficiencies of large-scale (1 cm 2 ) device achieve 73.2% and 85.9% of small-area (0.09 cm 2 ) device for the Control and IJP-CsPbBr 3 -based devices, respectively, highlighting the significance of research on scalable fabrication of perovskite films and devices.
To further explore film quality and charge recombination dynamics, steady-state photoluminescence (PL) and time-resolved photoluminescence (TRPL) measurement has been performed.As illustrated in Figure 5a, the Control and IJP-CsPbBr 3 films show similar PL peak position around 524 nm (2.36 eV), while the PL intensity of the IJP-CsPbBr 3 film was enhanced by 1.48 times compared with that of the Control-CsPbBr 3 film, proving reduction of trap states density in the IJP-CsPbBr 3 film.The gap (~0.05 eV) between PL position and optical band gap can be attributed to Stokes shift, which is similar to previous publications. [16,17]TRPL decay curves were measured to characterize the carrier lifetime in different CsPbBr 3 films.Interestingly, the IJP-CsPbBr 3 film can fit bi-exponential (rosy curves) and tri-exponential (green curves) decay model well, while the Control-CsPbBr 3 film fits the triexponential decay model obviously better (Figure 5b).Thus, the triexponential decay fitting results were selected for PL lifetime comparison.As summarized in Table 2, the carrier lifetime of the IJP-CsPbBr 3 film (33.5 ns) is nearly 4.08 times longer than that of the Control-CsPbBr 3 films (8.2 ns), which is consistent the steady state PL intensity results, indicating suppression of non-radiative recombination. [29,79]o investigate the relationship between charge recombination and device performance with different active layers, light intensity dependent V oc was measured.According to the Shockley-Reed-Hall mechanism: V oc = nkTln(I)/q + constant, where k represents the Boltzmann constant, T is the absolute temperature, q is the elementary charge and n is an ideal factor related to recombination. [80,81]The slope of V oc versus the natural logarithm of light intensity deviation kT/q indicates the existence of trap-assisted charge recombination.As shown in Figure 5c, the slope is measured to be 2.03 KT/q and 1.43 KT/q for the devices based The trap states density of different printed CsPbBr 3 films has been evaluated by the space-charge-limited current (SCLC) measurement using the device structure of FTO/TiO 2 /CsPbBr 3 /PCBM/Carbon.As depicted in Figure 5d,e, Ohmic, trap-filled limit and trap-free Child three distinct regions were observed for both devices. [29,82]When the applied bias voltage is lower than the kink-point voltage, the current increases linearly by voltage (I / V), indicating an ohmic response between the electrode and the perovskite layer. [83]voltage exceeds the kink-point voltage, the current shows a rapid nonlinear enhancement (I / V >3 ), which is ascribe to the filling of trap state by the injected charge carriers.The voltage point applied at the junction is defined as the trap-filled limit voltage (V TFL ).The trap states density (n t ) was calculated using the equation: n t = (2V TFL ε r ε 0 )/(qL 2 ), where ε r is the relative dielectric constant (measured to be 22 for CsPbBr 3 single crystal), [84] ε 0 is the vacuum permittivity, q is the elementary charge and L stands for the thickness of perovskite film.The trap states density is calculated to be 1.8 × 10 16 and 1.3 × 10 16 cm −3 for the Control and IJP-CsPbBr 3 films, respectively, proving reduction of trap states density in the IJP-CsPbBr 3 film.Based on the above results, the performance enhancement of the IJP-CsPbBr 3 C-PSCs can be attributed to longer carrier lifetime, suppression of trap-assisted recombination and reduction of trap states density.
To evaluate long-term stability of the printed CsPbBr 3 C-PSCs, we have recorded the performance of the devices based on IJP-CsPbBr 3 without encapsulation under different stress factors, including heat, humidity, and light.As shown in Figure 5f, the printed CsPbBr 3 C-PSCs remained 92.8% of initial PCE after thermal stressing at 80 °C for 1650 h.More importantly, it could maintain 91.9% of initial PCE after 984 h under continues illumination with a light intensity of 100 mW cm −2 .Nevertheless, the printed CsPbBr 3 C-PSC only remained 84.6% of initial PCE after 984 h under high humidity environment (around 75% relative humidity).The relatively poor stability of the printed CsPbBr 3 C-PSCs under humidity may ascribe to the absorption of water and oxygen molecules at vacancy sites, which would lead to a significant performance degradation. [85]Since carefully encapsulation can evade moisture and oxygen, the remarkable thermal and photo-stability of the printed CsPbBr 3 C-PSCs demonstrate high film quality and superior operational stability. [86]

Conclusion
We realized the phase evolution from PbBr 2 to CsPb 2 Br 5 -CsPbBr 3 intermediate, and then to pure CsPbBr 3 by inkjet-printing technique, providing a new scalable fabrication method for high-quality CsPbBr 3 films.The phase evolution mechanism and crystal growth dynamic of Cs-Pb-Br films have been studied with controlled stoichiometric ratio.We observed formation and overgrowth of Cs 4 PbBr 6 crystals in the Control-CsPbBr 3 film even when the deposition ratio of PbBr 2 to CsBr was controlled at 1:1, which would lead to a rough surface, and seriously disturb optical properties and carrier transport characteristics of CsPbBr 3 films.We introduced an intermediate IJP-inter film and successfully prohibited the formation of Cs 4 PbBr 6 impurity phase.More interestingly, we observed a positive feedback loop between the growth orientation of CsPb 2 Br 5 crystal and diffusion orientation of CsBr in the vertical direction, which can be ascribed to the 2D structure of the CsPb 2 Br 5 and the interconversion of CsPb 2 Br 5 and CsPbBr 3 in PbBr 2rich environment.In addition, the long-range crystalline CsPb 2 Br 5 can effectively promote the compact and uniform crystal stacking of CsPbBr 3 .Consequently, the printed CsPbBr 3 C-PSCs achieved champion PCEs of 9.09%, 8.59%, and 7.81% with active areas of 0.09, 0.25, and 1 cm 2 , respectively, demonstrating promising upscaling potential.The printed CsPbBr 3 C-PSCs show broad commercial application prospects with the advantages of easy-to-preparation, low-cost, long-term stable as well as high-performance.

Figure 1 .
Figure 1.a) Crystal structure of cesium lead bromide halides for CsPb 2 Br 5 , CsPbBr 3 and Cs 4 PbBr 6 from left to right, respectively.b) Schematic diagram of the quantitative inkjet-printing process for CsPbBr 3 films: 1) inkjet-printing of PbBr 2 ink and the number of PbBr 2 droplets deposited was controlled; 2) inkjetprinting of CsBr ink and the number of CsBr droplets deposited was controlled to meet the stoichiometric requirements of CsPbBr 3 ; 3) post-annealing treatment.c) Top-view and d) cross-sectional SEM images of the Control, IJP-inter, and IJP-CsPbBr 3 sample from left to right, respectively.The insert plot is the histogram of the thickness distribution of the corresponding sample.e) SEM-EDX line scanning profiles (yellow line) in different printed films for Br (left) and Cs (right).

Figure 3 .
Figure 3. Out-of-plane GIWAXS diffraction patterns of a) Control-CsPbBr 3 and b) IJP-inter samples under incident angles of 0.2°, 0.4°, and 0.7°.c) Peak intensity ratio between (002) crystal plane of CsPb 2 Br 5 and (110) crystal plane of CsPbBr 3 under different incident angles.High-resolution XPS spectra of d) Cs 3d, e) Pb 4f, and f) Br 3d core levels of the Control (black) and the IJP-CsPbBr 3 (red) films.The hollow circular symbol represents the raw data, and the filling region represents the corresponding fitting results.g) Schematic illustrations of the diffusion dynamics of CsBr (up) and CsPbBr 3 perovskite film growth process (down) during the phase evolution process.Route 1: formation of large Cs 4 PbBr 6 crystal due to the CsBr-rich environment on the upper surface.And CsPb 2 Br 5 phase is dominant near the interface due to the PbBr 2 -rich environment.As a result, the Control-CsPbBr 3 sample presents low purity and poor surface morphology.Route 2: the introduced IJP-inter sample prevent the formation and overgrowth of Cs 4 PbBr 6 .Therefore, the IJP-CsPbBr 3 sample shows ideal purity due to the interconversion of CsPb 2 Br 5 and CsPbBr 3 .

Figure 4 .
Figure 4. a) Schematic diagram and b) cross-sectional SEM image of a typical cell with the structure of FTO/c-TiO 2 /meso-TiO 2 /CsPbBr 3 /Carbon.c) Schematic diagram of the energy level for various functional layers of the C-PSCs based on the Control and IJP-CsPbBr 3 films.d) J-V curves of the champion devices based on Control and IJP-CsPbBr 3 films under AM 1.5G solar irradiation.e) EQE spectra and integrated J sc values calculated from the corresponding spectra.f) Stabilized power conversion efficiency of the C-PSCs based on the IJP-CsPbBr 3 measured over 12 h by maximum power point tracking.J-V curves of the large-area printed C-PSC based on g) Control and h) IJP-CsPbBr 3 films with active areas of 0.25 and 1 cm 2 .i) Active-area-dependent PCE of the device based on Control and IJP-CsPbBr 3 films.

Figure 5 .
Figure 5. a) Steady-state PL spectra, and b) TRPL spectra of Control and IJP-CsPbBr 3 films deposited on the glass substrate.c) Light intensity dependent V oc .SCLC characteristics for electron-only devices with the inserted structure for the d) Control and e) IJP-CsPbBr 3 films.f) Stability test of the printed C-PSC based on IJP-CsPbBr 3 film under different stress factors, including heat, humidity, and light.

Table 1 .
Summary of the photovoltaic parameters of different printed CsPbBr 3 C-PSCs with an active area of 0.09 cm 2 .

Table 2 .
Summary of the tri-exponential fitting results from the TRPL spectra in Figure5b.