In recent years, lead-based perovskite solar cells have achieved remarkable advancements, with their power conversion efficiency (PCE) exceeding 26%^[1]. Furthermore, the PCE of perovskite-silicon tandem solar cells has surpassed 30%^[2−3]. However, the inherent toxicity of lead in these solar cells poses significant challenges to their large-scale industrial application. To address the adverse environmental and health impacts associated with lead, researchers are increasingly focusing on the development of environmentally friendly alternatives for perovskite solar cells. One promising strategy is to substitute the divalent lead (Pb) ions with trivalent antimony (Sb) and bismuth (Bi) ions. These elements possess similar electronic structures to lead while exhibiting lower toxicity to humans. Antimony-based and bismuth-based perovskites typically exist in an A₃B₂X₉ framework, which can manifest in both zero-dimensional (0D) and two-dimensional (2D) structures, corresponding to dimeric (hexagonal crystal system) and layered (trigonal crystal system) phases, respectively. Similar to their lead-based counterparts, the bandgap and other optoelectronic properties of antimony-based and bismuth-based perovskites can be tuned by varying the components at the A, B, and X sites. Notably, alterations in the composition of A₃B₂X₉ antimony-bismuth perovskites can also induce changes in dimensionality and structural characteristics. Park et al.^[4] investigated three bismuth-based perovskite materials: Cs₃Bi₂I₉, MA₃Bi₂I₉, and MA₃Bi₂I_9−xCl_x, with bandgaps of 2.2, 2.1 and 2.4 eV, respectively. The devices fabricated from these three perovskites achieved PCE of 1%, 0.1%, and 0.003%. Lan et al.^[5] suggested that FA₃Bi₂I₉ has a smaller bandgap than MA₃Bi₂I₉ and is more suitable as a light-absorbing layer for perovskite solar cells, with devices utilizing FA₃Bi₂I₉ achieving an efficiency of 0.022%. Yu et al.^[6] introduced bromine into the Cs₃Bi₂I₉ system to improve its crystallinity and systematically studied the properties of perovskites with varying bromine content at the X site. They found that when the perovskite was Cs₃Bi₂I_9−xBr_x (x ≥ 3), the structure transitioned from a hexagonal dimeric structure to a trigonal layered structure, and the bandgap changed from an indirect to a direct bandgap. Ultimately, a device based on Cs₃Bi₂I₆Br₃ achieved an efficiency of 1.15%. Harikesh et al.^[7] proposed replacing Cs with Rb, discovering that Rb₃Sb₂I₉ is more likely to form a 2D structure, with a PCE of 0.66% for devices based on Rb₃Sb₂I₉. Jiang et al.^[8] found that the presence of chlorine in the precursor solution of MA₃Sb₂I₉ can suppress the formation of the 0D dimeric structure and promote the generation of the 2D layered structure. Devices fabricated using MACl as an additive achieved a maximum efficiency of 2.19% with minimal hysteresis. Recently, Zhang et al.^[9] discovered that the Cs₃Sb₂I_9–xCl_x perovskite films, prepared with MACl additive in the precursor solution, exhibited an intermediate phase with a mixture of 0D and 2D structures after a unique low-pressure treatment. After annealing, some I^– ions were replaced by Cl^– ions, resulting in a complete transformation to a 2D layered phase. Devices fabricated from this material achieved a PCE of 3.2%, marking the highest efficiency reported to date for Sb-based A₃B₂X₉ perovskite materials.

Currently, the efficiencies of antimony-based and bismuth-based perovskite devices remain relatively low, primarily due to intrinsic characteristics such as poor film quality, excessively wide bandgaps, and a high density of deep energy level defects^[10−11]. There has been extensive research on the effects of changing the A-site and X-site components in lead-free perovskites of the A₃B₂X₉ structure. It is evident that 2D structures are more advantageous than 0D structures in reducing non-radiative recombination and promoting charge carrier transport, leading to better optoelectronic performance. Therefore, dimensional control is one of the most effective pathways to improve device efficiency. However, there is limited reports on the effects of modifying the B-site components^[12−15]. The aforementioned studies clearly demonstrate that bismuth-based and antimony-based perovskites share a high degree of structural and property similarity. Thus, mixing antimony and bismuth elements to create a new B-site hybrid lead-free perovskite is a promising approach.

It is well known that during the spin-coating process, antimony-based and bismuth-based perovskites often experience uncontrolled film growth due to rapid crystallization rates. The presence of bromine at the X-site can effectively slow down the crystallization rate, improving film quality. Therefore, in this study, we based our work on the 2D Cs₃Bi₂I₆Br₃ perovskite and replaced bismuth with antimony to prepare a series of perovskites with varying ratios of antimony to bismuth. Through a series of characterizations, we demonstrated that the B-site alloyed perovskites with antimony and bismuth exhibit better properties compared to perovskites with a single B-site element.

Indium tin oxide glasses (ITO) were purchased from SooChow SunYo Solar Energy Technology Co., Ltd. Cesium Bromide (CsBr), [4-(3,6-Dimethyl-9H-carbazol-9-yl)butyl]phosphonic Acid (Me-4PACz), [6,6]-Phenyl C61 methyl butyrate (PC₆₁BM) and Bathocuproine (BCP) were all purchased from Xi’an Polymer Light Technology Corp. Nickel Oxide (NiO_x) nanoparticles were purchased from Shanghai MaterWin New Materials Co., Ltd. Bismuth iodide (BiI₃) were purchased from TCI (Shanghai) Chemical Industry Development Co., Ltd. Antimony iodide (SbI₃) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. Dimethyl sulfoxide (DMSO) and Isopropanol (IPA) were obtained from Sigma-Aldrich.

CsBr was dissolved in DMSO along with varying ratios of SbI₃ and BiI₃, followed by heating and stirring at 60 °C until complete dissolution was achieved. This process resulted in the preparation of a precursor solution of Cs₃(Sb_xBi_1–x)₂I₆Br₃ with a concentration of 0.225 M (For clarity, the perovskite compositions are designated as 100% Bi, 25% Sb−75% Bi, 50% Sb−50% Bi, 75% Sb−25% Bi, and 100% Sb, corresponding to x values of 0, 0.25, 0.5, 0.75, and 1, respectively). Nickel oxide (NiO_x) was selected as the hole transport layer. To fabricate this layer, 20 mg of NiO_x nanoparticles were dissolved in 1 mL of deionized water and subjected to sonication until complete dispersion was achieved. The solution was subsequently filtered through a glass fiber filter with a pore size of 0.7 μm for future applications. The self-assembled layer solution was prepared by dissolving 0.5 mg of Me-4PACz powder in 1 mL of isopropanol, followed by vigorous shaking to ensure homogeneity. PC₆₁BM was chosen as the electron transport layer material. 20 mg of PC₆₁BM powder were dissolved in 1 mL of chlorobenzene and heated to 60 °C, stirring overnight for subsequent use. For the BCP hole-blocking layer, 0.5 mg of BCP was dissolved in 1 mL of isopropanol and stirred at 60 °C overnight for subsequent application.

Initially, the ITO glass substrates were ultrasonically cleaned for 20 min using an ITO glass detergent, deionized water, acetone, and isopropanol sequentially, followed by a UV ozone treatment for 25 min. Subsequently, 45 μL of NiO_x was deposited onto the ITO glass using a spin-coating speed of 3000 rpm for 40 s, followed by annealing on a hot plate at 150 °C for 30 min. Following annealing, the device was transferred to a glove box. A 30 μL solution of Me-4PACz was applied onto the nickel oxide substrate, followed by spin-coating at 5000 rpm for 30 s. After spin-coating, the substrate was immediately placed on a hot plate at 100 °C for 10 min of annealing. The preparation process of the perovskite film is depicted in Fig. 1(a). A 50 μL perovskite precursor solution was applied to the self-assembled layer substrate, initially spinning at 500 rpm for 5 s, followed by a second spinning stage at 3500 rpm for 50 s. In the second spin-coating stage, 50−300 μL of isopropanol was added approximately 30 s (The timing and volume of isopropanol were adjusted according to the increasing bismuth content in the perovskite, as variations in elemental composition influenced the properties of the perovskite). Following spin-coating, the device was transferred to ambient conditions and annealed on a hot plate at 200 °C for 10 min. The device was subsequently transferred back to the glove box, where the PC₆₁BM electron transport layer and BCP hole-blocking layer were fabricated at spin-coating speeds of 2000 and 5000 rpm, respectively. Finally, a vacuum evaporation technique was employed to deposit 100 nm of silver as the electrode onto the BCP layer, thereby completing the device fabrication. The device architecture is illustrated in Fig. 1(b).

Following the aforementioned preparation method, we synthesized perovskite films with varying ratios of B-site elements, the crystal structure of Cs₃(Sb_xBi_1–x)₂I₆Br₃ perovskite is shown in Fig. 1(c). As illustrated in Fig. 1(d), the perovskite film with 100% Sb exhibits a yellow hue with a relatively lighter color tone. With an increasing proportion of bismuth, the film transitions progressively from yellow to orange, with its color deepening until the 100% Bi film presents a darker orange hue. The variation in the antimony-to-bismuth ratio induces a pronounced shift in the film's color, which directly correlates with alterations in the perovskite's bandgap and structure.

Subsequently, the surface morphology and crystallization quality of the perovskite films under varying conditions were characterized using field emission scanning electron microscopy (SEM), as shown in Figs. 2(a)−2(e). It was observed that the perovskite film composed entirely of Sb (100% Sb) exhibited poor quality, characterized by small grain sizes and a high density of pinholes. This phenomenon is likely attributed to the rapid nucleation and growth of the perovskite during the spin-coating process. The incorporation of bismuth effectively decelerated the crystallization rate of the perovskite, thereby enhancing the overall film quality. Consequently, a marked improvement in film quality was observed for the 75% Sb−25% Bi composition in comparison to the 100% Sb film. With increasing bismuth content, the grain size progressively increased, and the film exhibited a denser morphology. Interestingly, despite exhibiting the largest grain size, the 100% Bi perovskite film also displayed a substantial number of pinholes. Overall, the perovskite sample with a 50% Sb−50% Bi composition exhibited the highest film quality, characterized by its exceptionally smooth and compact morphology. Atomic force microscopy (AFM) was subsequently employed to further characterize the films, providing detailed insights into their surface morphology and roughness, as illustrated in Figs. 2(f)−2(j). AFM imaging revealed that the 100% Sb and 100% Bi perovskite films contained numerous pinholes, whereas films incorporating both antimony and bismuth exhibited superior quality with smoother surfaces. Notably, the 50% Sb−50% Bi film exhibited the lowest root mean square roughness (R_q), a result consistent with the observations from SEM analysis.

To gain deeper insights into the influence of varying B-site elemental ratios on perovskite properties, X-ray diffraction (XRD) analysis was performed on samples with different compositions, as illustrated in Fig. 3. All perovskite samples were found to exhibit a trigonal crystal system (space group P3¯m1), as depicted in Fig. 1(c). A comparison with the standard diffraction pattern of the trigonal crystal system Cs₃Sb₂I₉ (77-1053) revealed an overall rightward shift in the diffraction peaks of the perovskite samples. This observation suggests that the incorporation of the X-site bromine element alters the dimensionality of the perovskite structure from zero-dimensional to two-dimensional, thereby reducing the lattice constant. Furthermore, varying the ratio of B-site antimony and bismuth elements does not significantly influence the overall spatial structure of the perovskite. Notably, the 100% Bi perovskite demonstrated the highest crystallinity, as evidenced by the two prominent diffraction peaks at 17° and 26°, corresponding to the (101) and (003) planes^[6], respectively. With increasing antimony content, the diffraction peak intensity progressively decreased, accompanied by a broadening of the full width at half maximum (FWHM) and a shift in the preferential crystallization plane from (003) to (022). SEM imaging revealed that while films containing Sb exhibited a more compact morphology with smaller grain sizes, the 100% Bi perovskite film displayed the largest grain size, consistent with the observed variations in diffraction peak intensities. Additionally, the gradual incorporation of antimony induced a slight rightward shift of the diffraction peak at 26°, which can be attributed to the smaller ionic radius of antimony relative to bismuth. This incorporation causes lattice contraction and a corresponding reduction in interplanar spacing, consistent with previous reports^[12].

The absorption spectra of the films were measured using a UV−Vis spectrophotometer, and the corresponding bandgap values were calculated from the spectra, as shown in Figs. 4(a) and 4(b). It was observed that the absorption intensity of the films diminished with increasing antimony content, a trend consistent with the color variations of the film samples is presented in Fig. 1(d). With increasing antimony content, the bandgap of the perovskite gradually decreased, reaching its minimum value for the 75% Sb−25% Bi composition. However, a complete substitution of bismuth with antimony resulted in an increase in the bandgap of the perovskite. Additionally, the absorption spectrum of the 100% Bi sample exhibited a prominent shoulder peak around a wavelength of 525 nm. Interestingly, the addition of antimony led to a gradual weakening and eventual disappearance of this shoulder peak. This phenomenon may be attributed to alterations in the crystallographic orientation of the perovskite. The photoluminescence (PL) spectra of the films were also measured, as illustrated in Fig. 4(c). The emission wavelength of the films coincided with the absorption edge of their respective absorption spectra. Notably, the 50% Sb−50% Bi sample exhibited the highest emission intensity, suggesting that this composition possesses the longest carrier lifetime and minimal non-radiative recombination.

Subsequently, a series of devices with varying antimony-to-bismuth ratios were fabricated to further explore the practical implications of differences in film quality, optical absorption, and photoluminescence properties. The J−V curves and corresponding performance parameters are presented in Fig. 4(d) and Table 1, respectively. Overall, with increasing antimony content, the open-circuit voltage, short-circuit current, and power conversion efficiency (PCE) exhibited trends closely aligned with the variations observed in the SEM and PL results, following a pattern of initial improvement followed by decline. The device with a 50% Sb−50% Bi composition demonstrated the highest performance, achieving a maximum power conversion efficiency (PCE) of 0.645%.

In summary, the formation of antimony-bismuth alloyed two-dimensional perovskites, achieved by substituting bismuth with antimony at the B-site, yielded films with superior quality and enhanced optoelectronic performance compared to their single-cation counterparts containing only antimony or bismuth. The optimal antimony-to-bismuth ratio was determined to be 1 : 1, at which the best device efficiency of 0.645% was achieved. This study provides valuable insights for future research on lead-free perovskite solar cells utilizing antimony and bismuth, paving the way for higher-efficiency devices.