Introduction

Perovskite solar cells have become a hot topic due to their rapid increase in power conversion efficiency (PCE), and their certified efficiency has reached 26.7%1,2,3,4,5,6,7,8,9,10,11,12,13,14. In light of the current development of perovskite solar cells (PSCs), researchers have high expectations for their commercialization. PSCs can be fabricated by one-step or two-step fabrication methods. Compared with the one-step method, the two-step method has a commercial advantage because it does not require an anti-solvent process8,15,16,17,18. In the two-step method, a PbI₂ layer reacts with subsequently deposited organic halide salts to produce the perovskite layer. Ion interdiffusion and the formation of intermediate phases, which ultimately convert into perovskites during the thermal annealing process, determine the crystallization process19. Notably, controlled moisture exposure has been shown to promote intermediate hydrate phases and regulate ion diffusion kinetics, leading to improved crystallinity and film morphology20,21,22. This moisture-assisted transformation has proven crucial for achieving high-performance perovskite films under ambient fabrication conditions via the two-step method23,24. However, to mitigate the asynchronous crystallization caused by uncontrolled ion diffusion, most high-efficiency devices still rely on low-humidity and low-oxygen glovebox environments, raising significant concerns regarding scalability and manufacturing cost25,26.

In addition to crystallization control, defect passivation is equally vital for achieving highly efficient PSCs. Usually, undercoordinated Pb2+ ions at surfaces and grain boundaries act as nonradiative recombination centers, severely limiting device PSCs27,28,29. The complexation of Pb2+ ions is crucial for fabricating PSCs via the two-stage method fully in the atmosphere. Conventional passivation strategies, such as bulk additive incorporation and surface treatments with ammonium ligands30,31, often introduce side effects. For example, an insulating organic layer can impede charge transport, and charge-extraction barriers have been demonstrated due to ligands in alky or aromatic spacers modified PSCs32,33,34. Furthermore, these ligands bind to the perovskite through a single active site, which leads to a resistive barrier once densely packed8,35,36,37,38,39,40. Although Sargent et al. passivated undercoordinated Pb2+ at the surfaces and grain boundaries of the perovskite using dual-site-binding ligands41, a more robust solution—finding multi-site-binding (≥3 sites) ligands—is of urgent need to simultaneously achieve deep trap passivation and low interfacial resistance for efficient charge extraction.

In this article, we develop an antimony chloride-N,N-dimethyl selenourea complex, formulated as Sb(SU)₂Cl₃, as a multi-site passivator for Pb²⁺ defects. Such a complex can bind four adjacent sites of perovskite via two Se and two Cl atoms and form an extended hydrogen-bonding network through three NH-Cl bonds and dual intramolecular/intermolecular hydrogen bonds. Detailed characterizations and analyses reveal that the Sb(SU)₂Cl₃ enhances crystallinity, suppresses defects, and improves charge transport across interfaces. Consequently, two-step fully air-processed PSCs achieve a champion PCE of 25.03%—among the highest reported for ambient-fabricated devices. Unencapsulated devices retain 98.98% of their initial PCE after 1584 h storage in dark conditions (20–40% RH, 25 °C), projecting a T₈₀ lifetime of 23,325 h, rendering them as one of the most stable PSCs to date.

Results

Synthesis and characterization of multi-site binding ligands

As shown in Fig. 1a, antimony chloride reacts with N,N-dimethylselenourea (SU) in dichloromethane to form a Sb(SU)₂Cl₃ complex, synthesized following previously reported procedures42. The synthesized complex is soluble in polar solvents and can form various hydrogen bonds between the amine nitrogen atoms and chloride ions. These can be categorized as either intramolecular or intermolecular hydrogen bonds, which play a critical role in promoting crystal nucleation and growth. To further probe the complex’s electronic characteristics, density functional theory (DFT) calculations were conducted to generate the electrostatic potential (ESP) map (Fig. 1b). The electron-deficient (positively charged) regions are mainly localized around the amino and methyl groups, favoring the formation of hydrogen bonds with I⁻ anions. In contrast, the chloride and selenide atoms exhibit high electron density (negatively charged regions). Efficient orbital coupling between molecular units facilitates electron transfer from donor to acceptor sites, thereby enhancing coordination or binding interactions between Cl and Se atoms within the complex and undercoordinated Pb²⁺ defects in the perovskite lattice. In addition, the complex exhibits an elevated highest occupied molecular orbital (HOMO) energy level (Fig. 1b), which supports efficient hole transport. The hydrophobic methyl groups, combined with the oxygen-repelling effect of chloride ions, also contribute to improved moisture and oxidation resistance of the resulting perovskite solar cells under operational conditions.

Fig. 1: Synthesis and characterization of Sb(SU)2Cl3.
Fig. 1: Synthesis and characterization of Sb(SU)2Cl3.
Full size image

a Schematic of the synthesis for the antimony chloride-N,N-dimethyl selenourea complex (Sb(SU)2Cl3). b Top view and side view for ESP and energy levels of Sb(SU)2Cl3. c FTIR spectrum and (d) XRD pattern of Sb(SU)2Cl3. e UV-vis absorption spectra of pure selenourea in ethyl acetate solution (SU-EAaq), Sb(SU)2Cl3 in ethyl acetate solution (Sb(SU)2Cl3-EAaq), and Sb(SU)2Cl3 in the solid-state thin film (Sb(SU)2Cl3-S).

Fourier transform infrared (FTIR) spectroscopy, ultraviolet–visible (UV–vis) spectroscopy, and X-ray diffraction (XRD) were employed to verify the formation of Sb(SU)2Cl3. As shown in Fig. 1c, the FTIR spectrum of Sb(SU)2Cl3 aligns closely with previously reported data42. Two broad absorption bands at ~3300 cm−1 and ~3200 cm−1 correspond to N–H stretching vibrations, indicating that the N,N-dimethylselenourea ligands retain their hydrogen-bond donor characteristics within the complex. A strong absorption peak at 1650 cm−1 is attributed to N–H bending, further confirming the presence of intramolecular or intermolecular hydrogen bonding. In addition, a moderate band between 1000–800 cm-1 is assigned to C–Se stretching, supporting successful coordination of selenium with antimony. It also highlights a characteristic Se–Sb vibrational band at 350–300 cm-1, serving as direct evidence of complex formation. XRD analysis (Fig. 1d) reveals multiple diffraction peaks between 10° and 50°, consistent with a crystalline phase. Prominent peaks at 15° and 30° suggest high structural symmetry and an ordered arrangement within the lattice of Sb(SU)₂Cl₃. These features underscore the periodicity and crystallinity of the complex. UV–vis absorption spectra (Fig. 1e) show distinct differences among three sample types. Pure SU in ethyl acetate (SU-EAaq) exhibits major absorption between 250–300 nm. In contrast, Sb(SU)2Cl3 in solution (Sb(SU)2Cl3-EAaq) displays enhanced intensity in the same range. Notably, the solid-state film (Sb(SU)2Cl3-S) shows a red-shift of ~ 10 nm and significantly increased absorption intensity. This shift is likely driven by enhanced intermolecular hydrogen bonding in the solid state, which strengthens electronic coupling. Additional characterization data are provided in Supplementary Figs. 1–3.

Interaction between ligands and perovskite

To evaluate the interaction between the multidentate ligand Sb(SU)₂Cl₃ and the perovskite surface, we modeled four distinct adsorption configurations: single-site binding via Se, single-site via Cl, dual-site via Se–Cl, and quadruple-site via 2Se–2Cl (Fig. 2a–d). Charge transfer analysis at the Sb(SU)2Cl3/PbI2-terminated surface interface revealed substantial electron accumulation (Supplementary Fig. 4), indicating enhanced interfacial bonding and defect passivation. Notably, as the number of binding sites increases, the extent of charge transfer rises while the adsorption energy decreases (Fig. 2e), suggesting stronger, more stable binding. In the most favorable configuration, Se and Cl atoms from Sb(SU)2Cl3 coordinate simultaneously with four neighboring undercoordinated Pb2+ sites, forming four bonds within a single perovskite lattice unit. Meanwhile, the square composed of Se and Cl atoms can well match the FAPbI3 lattice (Supplementary Fig. 5)43. This configuration exhibits the strongest charge transfer and the most stable adsorption. For clarity, pristine FAPbI3 and Sb(SU)2Cl3-modified FAPbI3 are hereafter referred to as the control and target samples, respectively. In addition, the terminal hydrogen atom in Sb(SU)2Cl3 forms a hydrogen bond with an iodine atom on the PbI2-terminated surface, further stabilizing the interface via both chemical and hydrogen bonding. We next examined point defect formation at the PbI2-terminated surface. Supplementary Fig. 6 illustrates four typical surface defects (top view), and the corresponding defect formation energies are shown in Fig. 2f. The presence of Sb(SU)2Cl3 increases the formation energies of three key defect types—iodine vacancies (VI), lead vacancies (VPb), and anti-site defects (IPb)—due to strong interfacial bonding that suppresses vacancy generation and atomic substitution. Among these, the I vacancy has the lowest formation energy (0.82 eV), which is 2.34 eV lower than VPb, 3.06 eV lower than IPb, and 0.76 eV lower than the PbI anti-site defect, consistent with prior reports identifying VI as the most common defect in perovskite films. Importantly, treatment with Sb(SU)2Cl3 in the 2Se + 2Cl binding configuration significantly increases the formation energy of VI, VPb, and IPb, effectively suppressing these defects. However, the formation energy of the PbI defect decreases slightly due to the attractive interaction between the Pb atom (substituting for an I atom) and the Cl atoms of Sb(SU)2Cl3, which facilitates the formation of PbI anti-site defects. Notably, the presence of three relatively isolated Cl atoms in Sb(SU)2Cl3 will hold a high chance to fill the iodine vacancies for defect passivation, with a favorable binding energy of –2.03 eV, thereby enhancing the overall structural stability (Supplementary Fig. 7). Given the diverse functional groups in Sb(SU)2Cl3, we systematically investigated its interactions across various binding sites and structural configurations. The analysis revealed that the bonding to perovskite via Cl atoms is the most favorable one (Supplementary Figs. 8–10). Comparisons between the monomeric and calixarene forms of the interacting species with the perovskite surface further confirmed that the complexation process enables stronger, multi-site binding (Supplementary Fig. 10). To probe the interaction between Sb(SU)2Cl3 and perovskite constituents (FA⁺ and Pb²⁺), we performed FTIR and nuclear magnetic resonance (NMR) spectroscopy. These analyses focused on non-covalent interactions, proton coupling, polarization, and coordination effects. The FTIR spectra (Fig. 2g) revealed enhanced N–H stretching signals in the 3600–2800 cm−1 range, indicative of strengthened hydrogen bonding between Sb(SU)2Cl3 and FA⁺ ions. In addition, vibrational bands in the 1650–1500 cm−1 range, corresponding to C=O and C=N stretching, confirmed molecular interactions between the ligand and functional groups within the perovskite lattice. In the lower frequency region (1100–900 cm−1), spectral changes suggest the formation of Pb–N and Pb–Se coordination bonds, which likely contribute to improved charge transport and enhanced electron mobility within the perovskite. Furthermore, UV–vis absorption spectra (Fig. 2h) show distinct shifts upon the incorporation of Sb(SU)₂Cl₃ into the perovskite matrix, further supporting the formation of strong electronic interactions between the ligand and the perovskite. The observed red shift in absorption indicates strong interactions between Sb(SU)2Cl3 and PbI2, reflecting significant changes in the material’s electronic structure. This effect likely originates from coordination between the complex and Pb2+ ions, wherein the nitrogen and selenium atoms in the selenourea ligands form Pb–N and Pb–Se bonds. These interactions modify the local chemical environment of PbI2, inducing lattice distortion and reorganization, which in turn reduces the energy required for electronic transitions. This mechanism is analogous to coordination effects observed in systems such as PbI2–dimethyl sulfoxide (DMSO) complexes, where solvent molecules coordinate with Pb2+ ions, altering the Pb–I network and modulating the material’s optical and electronic properties. Such interactions promote new charge-transfer transitions and confirm the formation of stable coordination complexes in solution, which benefit the subsequent crystallization and film quality of perovskite. NMR spectroscopy further reveals the impact of Sb(SU)2Cl3 on the chemical environments of FA⁺ and Pb²⁺ ions. In the one-dimensional ¹H NMR spectrum (Fig. 2i), proton coupling and polarization effects induced by the complex result in an upfield shift and splitting of the FA⁺ proton signal (originally at 8.83 ppm). This shift is also observed in the perovskite precursor solution, indicating strong hydrogen bonding between FA⁺ and the complex, which enhances structural stability. Two-dimensional ¹H–¹H COZY spectra (Fig. 2j, k) show increased coupling between FA⁺ protons, further confirming that the complex reduces defect states and stabilizes the ionic environment within the perovskite lattice. Raw NMR data are provided in Supplementary Figs. 11–16. Taken together, FTIR and NMR results demonstrate that the multidentate ligand Sb(SU)₂Cl₃ enhances both the structural stability and optoelectronic properties of perovskite materials. These improvements are achieved through a combination of mechanisms, including strengthened hydrogen bonding, proton polarization, and Pb2+ coordination, which jointly optimize charge transport and light absorption. Comparative studies of SbCl3, SU, and the Sb(SU)2Cl3 complex with the perovskite precursor (Supplementary Figs. 17 and 18) revealed that Sb(SU)2Cl3 exhibits the strongest binding affinity. Importantly, the complex is not a simple physical mixture of SbCl3 and SU, but rather a chemically integrated structure in which halide and SU ligands synergistically stabilize the PbI2 coordination environment and enhance interactions with FA⁺ ions. The detailed explanation can be found in Supplementary Figs. 17 and 18 of the Supporting Information.

Fig. 2: Interaction between Sb(SU)2Cl3 and perovskite based on density function theory (DFT) and experimental characterization.
Fig. 2: Interaction between Sb(SU)2Cl3 and perovskite based on density function theory (DFT) and experimental characterization.
Full size image

a–d Adsorption structures, and (e) adsorption energies of Sb(SU)2Cl3 on the surface of PbI2 termination with one binding site (Se), one binding site (Cl), two binding sites (Se + Cl), and four binding sites (2Se + 2Cl). f Formation energies of four possible intrinsic neutral point defects for pure FAPbI3 (control) and Sb(SU)2Cl3 modified FAPbI3 with four binding sites (2Se + 2Cl) (Target). g FTIR spectra of FAI and perovskite without and with Sb(SU)₂Cl₃. The blue bands denote the N-H stretching peaks. h UV-Vis absorption spectra of FAPbI3, Sb(SU)₂Cl₃, and Sb(SU)₂Cl₃ treated FAPbI3. i 1H NMR spectra of FAI and perovskite without and with adsorption of Sb(SU)₂Cl₃. j, k 2D ¹H-¹H COZY spectra of FAPbI3 with and without Sb(SU)₂Cl₃, dissolved in DMSO-d6, with a small amount of DMF added to improve solubility.

To investigate the influence of the multidentate ligand Sb(SU)2Cl3 on the nucleation and growth of perovskite crystals, we fabricated perovskite films using a two-step method. In the control process (Supplementary Fig. 19a), a PbI2 solution in DMF/DMSO was spin-coated onto the substrate, followed by spin-coating of an FAI solution in IPA. The films were then annealed at 150 °C for 10 min to form perovskite layers. In the modified process (Supplementary Fig. 19b), a defined amount of Sb(SU)2Cl3 was added to the PbI2 solution, while the remaining steps were identical to those of the control. Surface morphology and crystallographic structure of the resulting films—prepared under ambient conditions (20–40% relative humidity)—were characterized using scanning electron microscopy (SEM) and X-ray diffraction (XRD) (Supplementary Figs. 20 and 21). Compared to the control, the target films displayed a denser and more uniform morphology, accompanied by diminished PbI₂ diffraction peaks and a reduced full width at half maximum (FWHM), indicating enhanced film quality and improved crystallinity resulting from the incorporation of Sb(SU)₂Cl₃. The dependence of film morphology and crystallinity on Sb(SU)2Cl3 concentration is presented in Supplementary Figs. 22 and 23. To further assess crystallization dynamics, in situ UV–vis absorption spectroscopy was performed during the two-step film formation (Fig. 3a, b). The crystallization of the target film proceeded more slowly during thermal annealing, with the transition to the α-FAPbI3 phase completed at 2.98 s, in contrast to 1.29 s for the control film. This delayed and more ordered crystallization process, driven by interactions between Sb(SU)2Cl3 and FAPbI3, led to enhanced film crystallinity and suppressed defect formation. In situ grazing-incidence wide-angle X-ray scattering (GIWAXS) measurements (Supplementary Fig. 24) confirmed the evolution of the α-FAPbI3 phase, with a characteristic diffraction signal appearing at q ≈ 1.0 Å⁻¹. The α-phase emerged more immediately in the target film compared to the control, suggesting that Sb(SU)2Cl3 accelerates the nucleation of the desired perovskite phase, thereby improving overall film quality.

Fig. 3: Effect of multi-site binding ligands (Sb(SU)2Cl3) on the perovskite layers.
Fig. 3: Effect of multi-site binding ligands (Sb(SU)2Cl3) on the perovskite layers.
Full size image

a, b In situ UV-Vis absorption spectroscopy for control (a) and target (b) groups, (c, d) Grazing-incidence X-ray diffraction (GIXRD) profiles of perovskite thin films (012) crystallographic plane. High-resolution Pb 4 f XPS peaks of the control (e) and target (f) perovskite film. g, h GIWAXS images of the control and target samples at an incidence angle of 0.8°. i, j Pole figure plots from the (100) lamellar diffraction as a function of incidence angles of the control and target samples. k, l Corresponding out-of-plane line cuts of the GIWAXS images as a function of incidence angles of the control and target samples.

To visualize the distribution of residual strain within perovskite films, grazing-incidence X-ray diffraction (GIXRD) was performed (Fig. 3c, d), revealing depth-dependent variations in residual stress and microstrain along the (012) plane. Upon incorporation of Sb(SU)2Cl3, the absolute residual stress decreased from 16.3 MPa (control) to 8.6 MPa (target), indicating partial stress relaxation in the perovskite lattice (Supplementary Fig. 25). X-ray photoelectron spectroscopy (XPS) was conducted to examine the chemical states of lead (Fig. 3e, f). The signal corresponding to metallic Pb⁰ was significantly suppressed, and the Pb 4 f binding energy exhibited a + 0.25 eV shift upon incorporation of Sb(SU)₂Cl₃. These observations indicate reduced residual PbI₂ and more complete conversion to the perovskite phase, as well as strong coordination interactions between Pb²⁺ and the introduced ligands. To further investigate the effects on crystallinity and crystal orientation, grazing-incidence wide-angle X-ray scattering (GIWAXS) was employed. The 2D GIWAXS patterns at an incident angle of 0.800° (Fig. 3g, h) show diffraction signals corresponding to α-FAPbI3, δ-FAPbI3, and PbI2 phases in both control and target films. After Sb(SU)2Cl3 incorporation, the intensities of δ-phase and PbI2 peaks decreased markedly, while the α-phase exhibited stronger, more uniform orientation, indicating enhanced phase purity and improved crystallinity. To quantify crystallographic orientation, the azimuthal angle (χ) dependence of the integrated diffraction intensity was analyzed. The orientation fraction was derived using the ratio Axy/(Axy + Az), representing the proportion of face-on to edge-on crystallites. Pole figure analysis of the (110) diffraction from α-FAPbI3 (Fig. 3i, j) showed an increased out-of-plane orientation in the target film (53.06%) compared to the control (43.40%), suggesting that Sb(SU)2Cl3 promotes more vertically aligned crystal domains. In addition, the variation in the out-of-plane diffraction peak with incidence angle reflects changes in d-spacing from the surface to the bulk of the film. The Sb(SU)2Cl3-modified films exhibited a greater d-spacing gradient (Fig. 3k, l), which may arise from the ligand-induced lattice modulation and preferential out-of-plane crystal growth.

Effect of ligands on the device performance

PSCs with the device structure FTO/SnO₂/perovskite/phenethylammonium iodide (PEAI)/spiro-OMeTAD/Au were fabricated to investigate the effect of multi-site binding ligands (Sb(SU)₂Cl₃) on device performance (Supplementary Fig. 26). Figure 4a shows the current density–voltage (J–V) curves of devices with and without the incorporation of Sb(SU)₂Cl₃. The control device exhibited a short-circuit current density (JSC) of 24.94 mA cm−², an open-circuit voltage (VOC) of 1.14 V, and a fill factor (FF) of 81.6%, corresponding to a power conversion efficiency (PCE) of 23.19%. Upon introducing the multi-site binding ligands, JSC, VOC, and FF increased to 25.66 mA cm−², 1.18 V, and 82.7%, respectively, yielding a PCE of 25.03%. The modified device also showed a steady-state PCE of 24.83% (inset of Fig. 4a), consistent with the value obtained from the J–V measurement. The corresponding external quantum efficiency (EQE) spectra of the PSCs are shown in Fig. 4b. By integrating the overlap between the EQE spectra and the standard AM 1.5 G solar photon flux, current densities of 24.43 mA cm⁻² and 24.81 mA cm⁻² were obtained for the control and target devices, respectively. These values are in good agreement with the JSC values measured from the J–V curves. The champion device was certified at an accredited photovoltaic certification laboratory (Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences (SIMIT)), confirming the reliability of the PCE (Supplementary Fig. 27). A reduced hysteresis was also observed after introducing Sb(SU)₂Cl₃ (Supplementary Fig. 28). The additive is also effective in the one-step fabrication process (Supplementary Fig. 29). Compared with the control cells, the target cells exhibited a more concentrated performance distribution (Supplementary Fig. 30). To investigate the origin of the enhanced device performance, various characterizations were conducted. The short-circuit current density (JSC) is jointly determined by light-harvesting efficiency (LHE), electron injection yield, and charge collection efficiency. Comparable LHE was confirmed by the nearly identical absorption spectra (Fig. 4c). Nyquist plots obtained from electrochemical impedance spectroscopy (EIS) revealed reduced series resistance (Rs) and transport resistance (Rtrans), along with an increased recombination resistance (Rrec), indicating suppressed charge recombination and enhanced charge transfer efficiency (i.e., improved electron injection yield and charge collection efficiency) upon incorporation of Sb(SU)₂Cl₃ (Fig. 4d).

Fig. 4: Device performance.
Fig. 4: Device performance.
Full size image

a J–V characteristics under standard illumination conditions (100 mW cm−2, AM 1.5 G) and the stabilized PCE (the inset), (b) EQE spectra, (c) UV-vis absorption spectra, (d) Nyquist plots under dark condition, (e) steady-state photoluminescence spectra, and (f) carrier lifetime of the control and target samples, (g) Dependence of dark current on temperature of the control and target perovskite solar cells, (h) Thermally stimulated current (TSC) spectra of the control and target samples, (i) Space charge limiting current (SCLC) curves of the control and target samples.

The fitting parameters based on the equivalent circuit shown in the inset are summarized in Supplementary Table 1. Charge transport within the hole transport material (HTM) and at the perovskite/HTM interface is represented by the semicircles in the high- and low-frequency regions, respectively44. The incorporation of Sb(SU)₂Cl₃ improves perovskite film quality and reduces defect density, as confirmed by the increased photoluminescence (PL) intensity (Fig. 4e). To further investigate carrier dynamics, time-resolved photoluminescence (TRPL) spectroscopy was performed, with the resulting spectra fitted using a biexponential function (Fig. 4f)45.

$$y={A}_{1}*{\exp }^{\left(-x/{t}_{1}\right)}+{A}_{2}*{\exp }^{(-x/{t}_{2})}+{y}_{0}$$
(1)

where the longer decay time (t1) corresponds to radiative recombination, and the shorter decay time (t2) corresponds to defect-induced nonradiative recombination46,47. The significantly reduced defect-assisted recombination was evidenced by the increased carrier lifetime after incorporation of Sb(SU)₂Cl₃. Temperature-dependent current–voltage measurements were used to determine the average activation energy of trapped carriers, calculated using the Richardson–Dushman equation48:

$$J\propto {{{\rm{e}}}}^{-\Delta E/{kT}}$$
(2)

where ΔE, k, and T represent the electron activation energy, Boltzmann constant, and absolute temperature, respectively. Compared with the control sample, the target sample exhibited a lower trap activation energy, indicating shallower traps in the perovskite films upon addition of the multi-site binding ligand Sb(SU)₂Cl₃ (Fig. 4g and Supplementary Fig. 31). The thermally stimulated current (TSC) curves and space-charge-limited current (SCLC) measurements further confirmed a reduction in defect density (Fig. 4h, i). Therefore, we conclude that the incorporation of Sb(SU)₂Cl₃ enhances the crystalline quality and reduces defect density, leading to improved device performance.

Time-resolved X-ray diffraction (XRD) measurements were performed at 110 °C and 70% relative humidity (R.H.) to assess the phase stability of the control and target samples (Fig. 5a, b). Compared with the control sample, the target sample exhibited a slower degradation rate. In the fresh control sample, a relatively strong PbI₂ diffraction peak—attributed to fabrication under ambient conditions (20–40% R.H.)—was significantly suppressed after incorporating Sb(SU)₂Cl₃. Unencapsulated devices were stored in ambient air (20–40% R.H.) to evaluate dark shelf stability. The target cell retained 98.98% of its initial PCE after 1584 h of storage, corresponding to a projected T80 lifetime of 23,325 h by linear extrapolation, whereas the control cell’s PCE decreased to 80.74% after 1008 h (Fig. 5c, d).

Fig. 5: Phase stability and device stability.
Fig. 5: Phase stability and device stability.
Full size image

a, b XRD patterns of control and target perovskite films aged in ambient conditions (70% humidity) and at 110 °C. The samples were prepared in the atmosphere (20–40% humidity) and at room temperature. c, d Dark shelf stability of unencapsulated control and target PSCs aged in ambient conditions at room temperature with R.H. around 20–40% and the corresponding linear extrapolation. e, f Evolution of the PCEs tracked under continuous heating at about 85 °C following the ISOS-D-2 protocol and the corresponding linear extrapolation. g, h Evolution of the PCEs tracked under continuous one sun light soaking under the maximum power point and the corresponding linear extrapolation.

Thermal stability was tested by storing the devices at 85°C following the ISOS-D-2 protocol. While Kim et al. previously provided mechanistic insights into de-doping effects49, we employed Poly [bis(4-phenyl)(2,4,6-trimethylphenyl)amine] (PTAA) as the hole transport layer (HTL) to better isolate the intrinsic thermal stability of the perovskite layer. Under these conditions, the target cells maintained 91.46% of their initial PCE after 1012 h of storage, corresponding to a projected T80 lifetime of 5004 h, whereas the control cells retained only 76.73% of their initial PCE after 528 h (Fig. 5e, f). Device operational stability was evaluated by maximum power point tracking (MPPT). The target cell retained 99.98% of its initial PCE after 969 h, corresponding to a projected T80 lifetime of 5209 h (Fig. 5g, h), while the control cell retained 55.67% after 489 h. These results demonstrate that the device is among the most stable and efficient perovskite solar cells fabricated under ambient conditions reported to date (Supplementary Tables 2–4). To elucidate the origin of the enhanced stability, first-principles calculations were performed to evaluate the adsorption energies of O₂ and H₂O molecules on the FAPbI₃ (100) surface without and with Sb(SU)₂Cl₃ (Supplementary Figs. 32 and 33; Supplementary Tables 5 and 6). The increased adsorption energies of O₂ and H₂O indicate significantly improved resistance to moisture upon addition of Sb(SU)₂Cl₃, consistent with the observed enhancements in thermal, ambient, and operational stability.

Discussion

This study was initiated with the understanding that most ammonium-based ligand additives and surface passivators interact with the perovskite lattice through a single active binding site, often resulting in dense ligand packing that introduces unwanted resistive barriers. We hypothesized that this limitation could be overcome by employing ligands capable of binding at multiple active sites on the perovskite surface. Such multi-site coordination with undercoordinated Pb2+ ions effectively suppresses defect formation and promotes enhanced crystallinity. In particular, the Sb(SU)2Cl3 complex, which forms both chemical and hydrogen bonds with the perovskite lattice, significantly improves structural stability and moisture resistance. Perovskite solar cells fabricated entirely in ambient air using a two-step method and incorporating these multi-site binding ligands achieved a power conversion efficiency (PCE) of 25.03%—among the highest reported for air-processed devices. These devices also exhibited exceptional long-term stability, retaining approximately 99% of their initial performance after 1584 h of storage in ambient air (20–40% relative humidity), with a projected T80 of 23,325 h under dark shelf conditions. This work highlights the potential of multi-site binding ligands as a promising strategy for simultaneously enhancing both the efficiency and durability of perovskite solar cells.

Methods

Synthesis of the Sb(SU)2Cl3 complex

The Sb(SU)₂Cl₃ complex was synthesized following a previously reported procedure42. Briefly, 200 mg (0.877 mmol) of SbCl₃ was dissolved in 60 mL of dichloromethane to form solution A. Separately, 265 mg (1.754 mmol) of N,N-dimethyl selenourea was dissolved in 60 mL of dichloromethane to form solution B. Solution A was added dropwise to solution B over 30 minutes, and the resulting mixture was stirred under a nitrogen-protected atmosphere for 2 h. A yellow precipitate formed, which was collected by filtration through Whatman filter paper and washed several times with dichloromethane to yield the final product, the Sb(SU)₂Cl₃ complex. The yield was 95%. 1H NMR (400 MHz, DMSO-d6) δ 7.69 (2H), 3.18 (6H). 13C NMR (400 MHz, DMSO-d6) δ 177.42, 55.38.

Device fabrication

The SnO₂ solution was diluted with deionized water at a volume ratio of 1:3. The diluted solution was spin-coated onto FTO glass at 3000 rpm for 30 s, followed by annealing at 150 °C for 30 min. To prepare the perovskite film, 691.5 mg (1.5 M) of PbI₂ was dissolved in a mixed solvent of DMF (0.9 mL) and DMSO (0.1 mL). The resulting solution was spin-coated onto the substrate at 1500 rpm for 30 s and annealed at 70 °C for 60 s. The PbI₂ film was then cooled at room temperature for 60 s. For the modified samples, a Sb(SU)₂Cl₃ solution was prepared by dissolving 0.5 mg of Sb(SU)₂Cl₃ in 1 mL of DMF. Various volumes (0, 0.5, 1, and 2 μL) of the Sb(SU)₂Cl₃ solution were added to 100 μL of the PbI₂ solution, respectively. Separately, 90 mg of FAI, 9 mg of MACl, and 6.39 mg of MAI were dissolved in 1 mL of IPA to form the organic halide salt solution. This solution was spin-coated onto the PbI₂ film at 2000 rpm for 30 s. Finally, the films were annealed at 150 °C for 15 min to complete the perovskite film formation. For surface passivation, 4.98 mg of PEAI was dissolved in 1 mL of IPA and spin-coated onto the samples at 5000 rpm for 30 s. To prepare the hole transport layer solution, 260 mg of bis(trifluoromethane)sulfonimide lithium salt (LiTFSI) was dissolved in 1 mL of acetonitrile to form the lithium salt solution, and 375 mg of FK209 Co(III) TFSI salt (Lumtec) was dissolved in 1 mL of acetonitrile to form the cobalt salt solution. Subsequently, 72.3 mg of spiro-OMeTAD, 35 μL of the lithium salt solution, 30 μL of 4-tert-butylpyridine, and the cobalt salt solution were dissolved in 1 mL of chlorobenzene. The hole transport layer solution was spin-coated at 3000 rpm for 30 s. A 70 nm gold layer was thermally evaporated onto the samples to form the metal electrode.

Stability testing

For maximum power point tracking (MPPT) measurements, the PEAI layer was omitted50. A hole transport layer solution was prepared by dissolving 72.3 mg of Spiro-OMeTAD, 7 mg of PTAA, 35 μL of Li-TFSI, and 28.8 μL of 4-tert-butylpyridine in 1 mL of chlorobenzene. This solution was spin-coated onto the samples at 3000 rpm for 20 s. For thermal stability tests, 20 mg of PTAA and 2.25 mg of DPI-TPFB were dissolved in 1 mL of chlorobenzene to form the hole transport layer solution. This solution was spin-coated at 1500 rpm for 30 s and annealed at 70 °C for 5 min51. All other procedures followed the standard device fabrication process described in the Device Fabrication section.

Material and device characterization

An X-ray diffractometer (D/MAX2500V/PC, Rigaku, Japan), UV–vis spectrophotometer (Jasco V-780), field-emission scanning electron microscope (Hitachi S-4800, Hitachi High-Technologies), and Thermo Scientific Flash 2000 analyzer were used to characterize the fundamental properties of the materials. A solar simulator (Newport, Oriel Class A, 91195 A) coupled with a source meter (Keithley 2400) was used to measure the J–V curves. The light intensity was calibrated using a standard reference cell. An internal quantum efficiency system (Oriel, IQE 200B) was used to measure the EQE spectra. Electrochemical impedance spectroscopy (EIS) was performed using an AUTOLAB (AUT302N) system under a bias voltage of 0.9 V and a frequency range from 100 kHz to 0.1 Hz. A FluoTime 300 spectrometer (PicoQuant) was used to acquire steady-state and time-resolved photoluminescence (PL) spectra. A Fourier transform infrared spectrometer (Nicolet iS50) and a 400 MHz NMR spectrometer (Bruker) were used to measure FTIR and NMR spectra, respectively. GIWAXS measurements were carried out at the Pohang Accelerator Laboratory in South Korea.

Theoretical calculations

Vienna ab-initio simulation package (VASP) was used to perform the density function theory (DFT) calculations52,53,54. The exchange-correlation functional was carried out using the Perdew-Burke-Ernzerhof (PBE) with generalized gradient approximation (GGA)55. The FAPbI3 (100) surfaces were modeled using 3 × 3 slabs. A plane wave basis with the cutoff energy of 400 eV was chosen in this work. A 3 × 3 × 1 k-point mesh was used in these calculations. We consider the van der Waals (vdW) interaction using the DFT-D3 method56. A vacuum layer along the out-plane direction of 15 Å was constructed to restrain the interactions between adjacent slabs. The residual force and energy convergence thresholds were set to 0.02 eV Å−1 and 10−4 eV, respectively.

The definition formula of charge density difference is:

$$\Delta \rho=\rho {{\rm{total}}}-\rho {{\rm{FAPbI}}}_3-\rho {{\rm{mol}}}$$
(3)

Where ρtotal is the total charge density of the adsorption interface, ρFAPbI3 and ρmol are the charge densities of the FAPbI3 (100) surface and the molecular, respectively.

The adsorption energy is defined as:

$${E}_{{{\rm{ads}}}}={E}_{{{\rm{total}}}}-{E}_{{{\rm{FAPbI}}}_{3}}-{E}_{{{\rm{O}}}_{2}/{{\rm{H}}}_2{{\rm{O}}}}$$
(4)

Where Etotal is the total energy of the O2/H2O adsorption interface, EFAPbI3 and EO2/H2O are the energies of FAPbI3 (100) surface and O2/H2O, respectively.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.