Perovskite materials have demonstrated remarkable efficacy, facilitating substantial advancements in the efficiency of PSCs. In the last ten years, certified power conversion efficiencies (PCEs) have surged to 27.0%^[1], primarily through advancements in component engineering^{[2, 3]}, solvent engineering^[4], crystal growth control^[5−7], and surface passivation^[8]. Despite these breakthroughs, commercialization of PSCs remains hindered by critical challenges, most notably the long-term stability. The instability of perovskite films is primarily associated with the defects at grain boundaries, in the bulk, and at interfaces with charge transport layers^[9]. These defects serve as recombination centers, ion migration pathways, and degradation initiators, collectively limiting device efficiency and operational lifetime^[10].

Interface defects, especially on the upper surface of perovskite films, are especially problematic due to incomplete coverage, lattice mismatches, and chemical instability during processing. Key defect species include undercoordinated Pb²⁺ ions, iodine vacancies (V_I), and organic cation vacancies, which induce non-radiative recombination, reduce charge extraction efficiency, and accelerate moisture-induced degradation^[11]. For example, unpassivated Pb²⁺ sites create deep-level traps that reduce open-circuit voltage (V_OC), while V_I promotes iodide migration, leading to phase segregation and hysteresis^{[12, 13]}. The interplay between these defects and environmental stressors underscores the urgent need for robust interfacial passivation strategies^[14].

To address defect-mediated losses, extensive research has focused on interface engineering. Early approaches employed Lewis base ligands (e.g., phenethylammonium bromide, PEABr) to passivate Pb²⁺ defects via coordination interactions^[15]. Subsequent studies introduced bimolecular systems, such as thiol-based and ammonium-based molecules, to simultaneously target iodine vacancies and Pb-related defects^[16]. Cross-linked 1D/3D heterostructures have also been explored to passivate grain boundaries and suppress ion migration^[17]. Despite these advancements, single-component passivation strategies often struggle to address the complexity of interfacial defects. Small molecules may fail to neutralize deep-level traps or introduce interfacial energy barriers, while 2D capping layers can impede charge transport due to insulating organic spacers^[18].

Due to these limitations, recent studies have shifted toward multifunctional passivation schemes. Competitive adsorption strategies, such as dual-molecule systems combining 3-bromopropylphosphonic ammonium (PPAABr) and PEABr, have achieved record efficiencies by synergistically targeting multiple defect types^[19]. Similarly, molecular conformational engineering using planar aromatic ketones has optimized dipole moments to enhance defect passivation and energy alignment^[20]. However, these approaches often involve complex synthesis protocols or unpredictable interfacial dynamics, necessitating simpler yet effective solutions.

In this study, we introduce a bilayer passivation strategy using phenethylammonium iodide (PEAI) and n-octylammonium iodide (OAI) to concurrently suppress non-radiative recombination and enhance environmental stability. PEAI targets undercoordinated Pb²⁺ ions at grain boundaries and surfaces, eliminating deep-level traps, while OAI forms a hydrophobic layer via its long alkyl chains to inhibit moisture penetration. This dual-layer approach leverages complementary functionalities, chemical passivation and physical protection, to address interfacial defects comprehensively. The resulting devices exhibit enhanced optoelectronic performance and operational stability, achieving a maximum PCE of 24.48%.

ITO and formamidinium iodide (FAI, 99.9%) were obtained from LiaoNing Advanced Election. Additionally, a variety of other chemicals were purchased from TCI, including lead(II) iodide (PbI₂, 98%), cesium iodide (CsI, 99%), lead(II) chloride (PbCl₂, 99%), N-methylpyrrolidone (NMP, 99%), isopropyl alcohol (IPA, 99.9%), and (4-(7H-dibenzo[c,g]carbazol-7-yl)butyl)phosphonic acid (4PADCB). Furthermore, N,N-dimethylformamide (DMF, 99.8%) was obtained from Sigma-Aldrich. Other materials, including lead acetate (PbAc₂, 99%), n-octylammonium iodide (OAI, 99%), fullerene (C₆₀), and bathocuproine (BCP) were acquired from Xi'an Yuri Solar Co., Ltd. For additional components, phenylethylammonium iodide (PEAI) was obtained from GreatCell Solar Materials, while ethanol was purchased from Macklin. Lastly, silver (Ag) was obtained from ZhongNuo Advanced Material Technology.

Perovskite precursor solution: 2 M FA_0.83Cs_0.17PbI₃ perovskite solution: 171.97 mg FAI, 461.09 mg PbI₂, 259.81 mg CsI, 1 mg PbAc₂, 278.11 mg PbCl₂ were dissolved in 500 μL DMF and 96 μL NMP solution. The solution was stirred for 2 h in N₂-filled glovebox before use.

Passivation layer precursor solution: 2 mg Phenylethylammonium iodide (PEAI) and 2 mg n-Octylammonium Iodide (OAI) were dissolved in 1 mL IPA, respectively.

Hole transport layer precursor solution: 0.5 mg 4PADCB was dissolved in 1 mL ethanol as HTL solution.

ITO/glass substrates were subjected to a cleaning procedure that involved a sequence of detergent, deionized water, and ethanol, utilizing ultra-sonication for 20 min. Afterward, the substrates were dried with a N₂ flow. Subsequently, an ultraviolet ozone cleaner was used to treat the substrates for 30 min prior to their application. The SAM solution, prepared at a concentration of 0.5 mg/mL, was subjected to spin-coating at 4000 rpm for 30 s, followed by an annealing process at 100 °C for 10 min. Subsequently, a 2 M perovskite precursor solution was applied to the substrates via spin-coating at 5000 rpm for 50 s without anti-solvent dripped, and underwent an annealing step at 150 °C for 10 min. For the application of a single passivation layer, a solution of either 2 mg/mL PEAI or OAI was spin-coated onto the perovskite layer at 5000 rpm for 30 s, and then subjected to annealing at 100 °C for 10 min. As for the bilayer passivation, PEAI solution (2 mg/mL) and OAI (2 mg/mL) solutions were sequentially deposited on perovskite films, followed by an annealing step at 150 °C for 10 min. Finally, C₆₀ (25 nm), BCP (5 nm), and Ag (100 nm) were thermally deposited within a high-vacuum chamber (9 × 10⁻⁵ Pa), respectively.

J−V measurements were performed utilizing a xenon lamp solar simulator (Enlitech, SS-F5-3A, Class AAA) paired with a source meter (Keysight B2901A, USA). The AM 1.5G simulated irradiation (100 mW/cm²) was calibrated against a standard silicon cell (certified by NREL, SRC-2020) inside a N₂ glove box, kept at a temperature of 25 ± 2 °C. A metal mask precisely defined the device area of 0.09 cm² to ensure accurate delineation of the active region. The crystalline structures of perovskite were examined through XRD within a 2θ range of 5° to 60° on a Rigaku Smart Lab X-ray diffractometer (9 kW, Cu Kα radiation, λ = 1.540593 Å). Top-view SEM images of the perovskite films were captured via a Gemini SEM 500 (Zeiss, Germany). UV−vis absorption spectra were obtained with a UV−vis spectrophotometer (U-3900H, Hitachi, Japan). XPS analysis was conducted using a ThermoFisher ESCALAB Xi⁺ system under a high vacuum pressure of 8 × 10⁻¹⁰ Pa, utilizing Al Kα (1486.6 eV) as the excitation source.

XRD analysis of perovskite films treated with OAI, PEAI, and PEAI + OAI bilayer revealed distinct impacts on structural quality (Fig. 1(a)). For the unmodified control sample, prominent peaks at approximately 12.7° and 14°, corresponding to residual PbI₂ and the perovskite (100) planes, respectively, indicate incomplete perovskite conversion and PbI₂ residuals^{[21, 22]}. The presence of numerous small white spots (highlighted by red circles in Fig. 1(b)) observed in SEM image of the control sample corresponds to residual PbI₂. This phenomenon is associated with the addition of excess PbCl₂ in perovskite components, leading to incomplete incorporation of PbI₂ within the crystal lattice of the perovskite during crystallization process, resulting in its segregation as unreacted residues. Such residual PbI₂ imposes dual penalties on device performance: (1) introducing defect states at grain boundaries and interfaces that enhance non-radiative recombination and impede charge transport^[23]; (2) accelerating moisture-induced degradation due to its hygroscopic nature.

To address these challenges, passivation strategies were implemented to improve phase purity through optimized stoichiometry or surface engineering. The OAI-treated sample showed reduced PbI₂ peak intensity compared to the control, suggesting partial suppression of Pb²⁺ through coordination with ammonium groups of OAI. Furthermore, PEAI passivation has been shown to reduce PbI₂ peak intensity, which is attributed to phenethylammonium ligands preferentially binding to undercoordinated Pb²⁺, thereby effectively passivating defects and inhibiting PbI₂ formation. The PEAI + OAI bilayer configuration demonstrated the most positive effect, nearly eliminating PbI₂ signatures through synergistic passivation mechanisms.

Surface morphologies of perovskite films were analyzed by SEM, as illustrated in Figs. 1(b)−1(e). The detailed analysis indicated that the sample without passivation showed the presence of white grains on its surface, indicative of PbI₂ precipitation (highlighted by red circles in Fig. 1(b)). In contrast, the samples treated with OAI, PEAI, and PEAI + OAI exhibited a notable decrease in the number of white grains. This reduction was likely due to the reaction between OA⁺ and PEA⁺ cations with PbI₂, which suppresses residual PbI₂ formation. Furthermore, the PEAI + OAI bilayer passivation strategy produced a uniform surface morphology in perovskite film, optimizing separation and extraction of charge at perovskite/ETL interface.

UV−vis absorption spectroscopy was employed to examine the optical characteristics of perovskite films. As shown in Fig. 2(a), following passivation, a significant increase in absorption intensity was detected, suggesting a decrease in surface defects and an overall enhancement in the quality of the film^[24]. The PEAI + OAI bilayer passivation exhibited synergistic effects, achieving the highest absorbance across the entire spectral range by effectively suppressing both surface and bulk defects. Moreover, the corresponding Tauc plots presented in Fig. 2(b) for the perovskite films revealed that every sample exhibited a consistent bandgap of 1.57 eV, suggesting the addition of passivation molecules at the interface did not substantially influence the perovskite intrinsic optical characteristics. These treatments enhanced light harvesting by minimizing sub-bandgap absorption losses while maintaining the integrity of the perovskite electronic framework, with bilayer passivation providing comprehensive performance optimization^[19].

Steady-state photoluminescence (PL) spectra were illustrated in Fig. 2(c). Among the various films analyzed, the PEAI + OAI bilayer film exhibited the highest PL intensity, followed by PEAI, OAI, and the control films, indicating improved perovskite film quality, which was consistent with the XRD and SEM analysis. The maximum PL intensity of the PEAI + OAI passivated film further confirmed that non-radiative recombination was effectively inhibited by synergistic passivation effect. TRPL results were presented in Fig. 2(d), along with the calculation of the average carrier lifetime. The PEAI + OAI passivated film exhibited the longest carrier lifetime of 2781.59 ns. Both PEAI (τ_ave = 931.99 ns) and OAI (τ_ave = 796.46 ns) passivation showed significantly increased lifetimes compared to the control film (τ_ave = 76.44 ns). These findings indicated that the PEAI + OAI bilayer passivation is superior in reducing surface defects compared to treatments with a single layer. The prolonged carrier lifetime in the PEAI + OAI film indicated efficient defect passivation^{[25, 26]}.

The surface modification would have an impact on interface energy band alignment and the n-type semiconducting contact could promote the electron transfer and extraction at the interface. The spectra of secondary electron cutoff and valence band regions are shown in Figs. 3(a)−3(d), and the resulting schematic diagram of energy level alignment is presented in Fig. 3(e). Compared to the control group and modification techniques, the unmodified perovskite surface exhibited the electron affinity (3.33 eV) and difference between the Fermi level (E_f) and the conduction band minimum (CBM) (0.6 eV). After PEAI + OAI modification, these values changed to 3.49 and 0.45 eV, respectively. An energy offset between the CBM of the perovskite film and that of C₆₀ can lead to the accumulation of electrons and holes at this interface after charge separation, resulting in extensive interfacial recombination. The unmodified perovskite showed a difference of 0.67 eV between its CBM and the CBM of C₆₀, indicating poor photoelectric performance. However, after modification with PEAI + OAI, these values changed to 0.51 eV, indicating that the energy level alignment between the perovskite and C₆₀ has been improved due to the synergistic effect of two molecules.

XPS was employed to investigate the interaction of PEAI/OAI with the perovskite, illustrated in Fig. 4(a). Detailed analysis of the Pb 4f core levels revealed distinct shifts in binding energy (E_B) for perovskite films passivated with OAI, PEAI, and the combined PEAI + OAI methods. In the control film, the Pb 4f peaks (4f_7/2 and 4f_5/2) were centered at 138.4 and 143.2 eV, respectively^{[27, 28]}. Both PEAI and OAI treatments resulted in a slight shift to lower E_B compared to the control. Notably, OAI-only passivation induced a more pronounced negative shift in the binding energy of Pb 4f compared to PEAI-based treatments. This observation indicated stronger chemical interaction between the long-chain alkyl ammonium groups in OAI and the perovskite lattice, which preferentially coordinates with undercoordinated Pb²⁺ ions and Pb⁰ defects at grain boundaries and surfaces^[21].

Remarkably, the PEAI + OAI bilayer configuration produced the largest E_B reduction (~0.25 eV), suggesting a synergistic defect passivation mechanism arising from the complementary functionalities of PEAI and OAI molecules. Quantitative analysis of Pb 4f spectral components further revealed progressive elimination of metallic Pb⁰ species (~136.7 eV) following passivation treatments, with the bilayer configuration achieving complete suppression of Pb⁰ signals. This trend aligns with XRD and SEM observations of reduced PbI₂ residuals, indicating enhanced phase purity in passivated films. The improved chemical homogeneity achieved through bilayer passivation is attributed to dual-functional defect mitigation: (1) PEAI's phenyl-ethyl groups provide steric stabilization and defect capping; (2) OAI's long alkyl chains facilitate deep-level defect passivation via strong coordinate bonding. These combined effects effectively reduce non-radiative recombination centers and stabilize Pb²⁺ oxidation states, thereby improving charge carrier dynamics and operational stability.

XPS analysis of the I 3d core levels revealed slight shifts in E_B of perovskite films treated with OAI, PEAI, and PEAI + OAI, as illustrated in Fig. 4(b). For the control sample, the I 3d_3/2 and I 3d_5/2 peaks were centered at 630.3 and 619.0 eV, respectively^{[29, 30]}. Both PEAI and OAI treatments alone exhibited a slight shift to lower E_B compared to the control, likely due to the bonding of PEA⁺ and OA⁺ with undercoordinated I⁻ ions at the grain boundaries, which stabilized the I−Pb framework and reduced surface defects. The PEAI + OAI bilayer passivation synergistically integrated these effects, resulting in a more pronounced negative E_B shift (~0.2 eV), indicating synergistic suppression of iodine vacancy formation and ion migration. These variations of E_B correlated with enhanced device stability, as reduced surface defects and optimized I−Pb−I bonding networks mitigated non-radiative recombination and halide segregation, both essential for the enduring efficacy of PSCs^{[31, 32]}.

Fig. 5(a) presented the J−V curves for the highest-performing PSCs obtained from control, OAI, PEAI, and PEAI + OAI treated devices. The control PSC exhibited a PCE of 22.01%. Moreover, the PSC based on OAI reached a maximum PCE of 22.97%, demonstrating improved V_OC and J_SC values of 1.14 V and 24.85 mA/cm², respectively, under identical testing conditions. Similarly, the PEAI-based PSC reached a maximum PCE of 23.40%, with enhancements in V_OC (1.15 V) and FF (82.91%). The PSC utilizing both PEAI and OAI exhibited the highest PCE of 24.48%, alongside further improvements in V_OC (1.16 V), J_SC (25.27 mA/cm²), and FF (83.02%). The photovoltaic parameters for each passivation method in the champion cells are summarized in Fig. 5(a).

The experimental results demonstrated significant enhancements in the optoelectronic performance of PSCs achieved through OAI, PEAI, and PEAI + OAI passivation strategies compared to the control group. Statistical distributions of photovoltaic parameters for the PSCs (20 devices) were presented in Figs. 5(b)−5(e). Improvements in V_OC were most pronounced in the PEAI + OAI configuration (1.16 V), attributed to the effective passivation of non-radiative recombination at the interfaces. The FF exhibited a progressive improvement, indicating robust charge transport efficiency and reduced electrical losses. The J_SC observed in the PEAI + OAI configuration reached a peak value of 25 mA/cm², likely due to the synergistic suppression of bulk and interfacial defects, this suppression enhanced light absorption and charge extraction. These results demonstrated the superiority of bilayer passivation in holistically optimizing charge dynamics and mitigating defects, thereby establishing it as a viable approach for advancing high-efficiency PSCs.

The analysis of dark J−V curves (Fig. 6(a)) further corroborates these findings. The dark current behavior of PSCs subjected to control, OAI, PEAI, and PEAI + OAI bilayer passivation indicated distinct mechanisms for defect mitigation. The control devices exhibited the highest dark current density (1.23 × 10⁻⁴ mA/cm²), primarily due to significant defect-assisted recombination occurring at grain boundaries and interfaces. In contrast, the PEAI + OAI bilayer passivation demonstrated synergistic effects, resulting in a low dark current of 4.33 × 10⁻⁵ mA/cm² through complementary defect passivation. The PEAI + OAI device showed a reduction of nearly one order of magnitude compared to control, which directly correlates with suppressed leakage currents in the J−V characteristics. These results further validate the pivotal role of hierarchical defect passivation in optimizing charge dynamics for high-performance perovskite photovoltaics.

The PEAI + OAI bilayer passivated PSCs exhibited exceptional performance across a range of light intensities (from 1.0 to 0.1 sun), as analyzed through J−V curves and light-dependent V_OC/J_SC (Fig. 6(b)). To further investigate the influence of bilayer treatment on trap-assisted surface recombination, measurements of V_OC and J_SC that depend on light intensity were conducted (Figs. 6(c) and 6(d)). Fig. 6(c) illustrated the correlation between V_OC and light intensity across four distinct types of devices. The gradient of the light-dependent V_OC curves indicated the extent of non-radiative recombination; a deviation of 1 kT/q suggests the potential for trap-assisted recombination under open-circuit conditions^{[33, 34]}. Notably, the calculated slopes (which correlate with the ideality factor n) for the control, OAI, PEAI, and PEAI + OAI groups were measured as 1.98, 1.33, 1.42, and 1.19 kT/q, respectively. The ideality factor (n) indicated that the passivation process employing PEAI + OAI has effectively diminished defects, which in turn alleviated charge trapping caused by these defects and supporting the observed increase in V_OC. Additionally, the relationship illustrated by the double logarithmic plot of J_SC versus light intensity adheres to the correlation J_SC∝Iα, where α is the exponent^{[35, 36]}. In comparison with the control devices, the exponent α for devices treated with PEAI + OAI was closer to 1 (Fig. 6(d)), indicating reduced bi-molecular recombination.

Furthermore, XRD patterns of both the control and PEAI + OAI bilayer modified perovskite films were analyzed after storage under ambient conditions (25 °C, 30 ± 10% relative humidity (RH)) for 14 days (Fig. 7(a)). The XRD patterns, combined with contact angle measurements, revealed significant differences in stability and surface properties between the unpassivated control and the PEAI + OAI bilayer passivated perovskite films^{[37, 38]}. In the XRD analysis, the control film exhibited distinct PbI₂ decomposition peaks at approximately 12.7°, along with diminished α-phase perovskite peaks at 14.05° and 28.1°, and the emergence of prominent δ-phase perovskite signals at 11.9° and 26.3°. These findings indicate severe moisture-induced degradation and a phase transition from α-phase to δ-phase perovskite^[39]. In contrast, the PEAI + OAI bilayer passivated film maintained strong α-perovskite peaks, with suppressed PbI₂ and δ-perovskite intensities, demonstrating enhanced structural stability. This suggests that the dual passivation layer mitigates ion migration and halide segregation by stabilizing Pb²⁺ ions and passivating I⁻ vacancies, thereby inhibiting hydrolysis and phase transitions through synergistic defect suppression^[40].

As shown in Figs. 7(b) and 7(c), the perovskite films (unpassivated and PEAI + OAI passivated) exhibited distinct color changes upon light exposure, indicating film degradation. After aging for 14 days under ambient conditions (25 °C, 30 ± 10% RH), optical photographs of perovskite films passivated by the two methods revealed notable differences in degradation behavior and stability. The control sample demonstrated significant color fading from an initial light dark to lighter yellow, indicating severe instability due to inadequate protection against environmental factors such as moisture and oxygen. In contrast, the PEAI + OAI passivated film retained a dark film, suggesting improved stability through surface defect passivation.

The variations in water contact-angle correlated with surface hydrophobicity and passivation efficacy. The control film exhibited a contact angle of 52.7°, indicating weak hydrophobicity and rendering it vulnerable to moisture infiltration. The passivation with PEAI + OAI significantly increased the contact angle to 68.6°, demonstrating enhanced hydrophobicity. This improvement was attributed to the synergistic densification and crosslinking of the surface: PEAI facilitated the formation of a quasi-2D perovskite phase, while the long alkyl chains of OAI suppressed water adsorption by reducing surface energy^[41]. These findings highlighted the multifunctional role of bilayer passivation in stabilizing bulk crystallinity and modifying surface wettability, enabling air-stable operation of perovskite photovoltaics under ambient conditions. Notably, the optimized unencapsulated film demonstrated good stability under ambient conditions of 25 °C and 30 ± 10% RH.

In our work, we proposed a bilayer modification strategy aimed at overcoming the efficiency limitations of inverted PSCs. Specifically, the top surface of the perovskite was sequentially treated with PEAI and OAI to minimize recombination losses at the interface and enhance charge extraction. PEAI effectively eliminated deep-level traps by strongly binding to undercoordinated Pb²⁺ ions at grain boundaries or surfaces, thereby passivating defects and mitigating non-radiative recombination. The long alkyl chains of OAI adsorbed onto the perovskite surface, forming a compact hydrophobic barrier that suppressed moisture infiltration while maintaining efficient charge transport. Ultimately, the bilayer modified devices achieved a champion PCE of 24.48%. Notably, the optimized unencapsulated film demonstrated good stability under ambient conditions of 25 °C and 30 ± 10% RH. This advancement provides new perspectives into the industrialization of inverted perovskite devices.