Synergistic engineering of buried interfaces for high-efficiency and stable perovskite solar cells

Abstract

To address the challenges of interfacial defects and energy-level alignment in perovskite solar cells (PSCs), this study introduces p-toluenethiol as a molecular modifier for the SnO₂/perovskite buried interface. The thiol groups (–SH) coordinate with oxygen vacancies and hydroxyl groups on the SnO₂ surface, effectively passivating deep-level trap states and suppressing the formation of PbI₂ secondary phases and lattice defects. The aromatic benzene ring induces interfacial dipole moments via π-conjugation, optimizing energy-level alignment between SnO₂ and perovskite to reduce electron transport barriers, and its hydrophobicity also enhances the device's environmental stability. Experimental results show that PSCs achieve a power conversion efficiency (PCE) of 25.53%, while flexible devices exhibit a PCE of 23.27%. Stability tests demonstrate significantly improved performance retention under continuous illumination and environmental exposure. This work synergistically optimizes device efficiency and stability through molecular-scale inter facial engineering, providing a foundation for the application of perovskite solar cells.

---

Lin Song

Lin Song received his PhD degree in the physics department from Technische Universität München (Garching, Germany) in 2017. He continued to work as a postdoc at the physics department of Technische Universität München. Now he is a professor at Frontiers Science Center for Flexible Electronics (FSCFE) and Institute of Flexible Electronics (IFE) of Northwestern Polytechnical University (NPU) in Xi'an, China. His research interest is mainly on the study of hybrid photovoltaics using advanced X-ray and neutron scattering techniques.

---

Introduction

Perovskite solar cells (PSCs) have emerged as a next-generation photovoltaic technology, achieving a certified efficiency of 27%, demonstrating a great potential to rival conventional silicon-based photovoltaics.^1–5 However, their practical implementation is challenged by two persistent issues: non-radiative recombination caused by interfacial defects that limits power conversion efficiency,⁶ and irreversible degradation during operation due to chemical instability at the interfaces.^7–10 These issues collectively underscore the pivotal role of the buried interface—serving as a critical nexus for charge transport, crystallization regulation, and stress buffering—in determining the efficiency ceiling and operational stability of devices.^11–13

At the buried interface, intrinsic oxygen vacancies (V_O) and chemisorbed non-lattice oxygen create deep-level trap states, significantly impairing charge extraction efficiency and triggering phase separation.^13–16 Hydroxyl (–OH) groups further catalyze perovskite lattice degradation, generating electrically inert impurities such as PbI₂ and δ-FAPbI₃, which disrupt interfacial energy level alignment.^17–20 Therefore, developing advanced buried interface engineering strategies that simultaneously address defect passivation and energy-level optimization, is essential for overcoming the performance bottlenecks of PSCs.²¹

To address this, researchers have developed various buried interface engineering strategies through molecular engineering and material design.^22–24 Zhao et al. introduced aminomethyl phosphonic acid (AMPA) into SnO₂ ETLs, passivating oxygen vacancies on the SnO₂ surface to improve carrier mobility and align energy levels with perovskite. Consequently, the PCE was boosted from 19.91% to 24.22%;²⁵ Li et al. used bisphenol S (BPS) to crosslink SnO₂ nanoparticles. This strategy, on the one hand, passivated defects and formed a network for uniform films. On the other hand, the coordination between sulfone groups and Pb²⁺ enabled the regulation of perovskite crystallization;²⁶ Chen et al. modified SnO₂ with sodium thiosulfate, which enhanced film uniformity and conductivity through strong chemical interaction, and promoted perovskite crystallization to reduce trap density and improve charge transport.²⁷

In this study, we introduce p-toluenethiol as a molecular modifier for the SnO₂/perovskite buried interface. This molecule enhances interfacial quality through multiple functions, including selective passivation of interfacial defects, modulation of charge transport energy levels, and suppression of ion migration. With this approach, the modified PSCs achieve a champion PCE of 25.53%, featuring a remarkable open-circuit voltage (V_OC = 1.18 V), short-circuit current density (J_SC = 26.01 mA cm⁻²), and fill factor (FF = 83.23%). And the flexible perovskite devices exhibit a PCE of 23.27% with a V_OC of 1.17 V, a J_SC of 24.84 mA cm⁻², and a FF of 79.75%. Moreover, these devices combine high performance with high stability, validating the approach effectiveness in balancing efficiency and stability.

Results and discussion

The chemical structural formula of p-toluenethiol is shown in Fig. S1,† which is deposited on top of SnO₂ layer to modify the buried interface of perovskite films. The multifunctional efficacy of p-toluenethiol as a buried interfacial modifier stem from its distinct molecular architecture. The electron-donating methyl (–CH₃) group ortho to the sulfhydryl (–SH) moiety augments the proton-donating capability of the thiol group, facilitating deprotonation reactions with hydroxyl groups on the SnO₂ surface to passivate oxygen vacancies.^28,29 The aromatic benzene ring exhibits dual functionality: its π-conjugated system enables electron delocalization to stabilize interfacial charge transfer states, while the inherent hydrophobicity forms a moisture-blocking barrier that inhibits halide ion migration.^30,31 The spatial arrangement of these functional groups generates an asymmetric electron density distribution, inducing interfacial dipole moments that align energy levels between the SnO₂ and perovskite layers.³² This synergistic combination of chemical passivation, environmental shielding, and electronic structure modulation suppresses interfacial recombination and enhances carrier extraction efficiency.

To elucidate the interaction between p-toluenethiol and SnO₂, X-ray photoelectron spectroscopy (XPS) analysis is performed. Fig. 1a–c present the high-resolution spectra of the O 1s and Sn 3d orbitals. Fig. 1a displays the high-resolution O 1s spectrum of the pristine SnO₂ film, showing two distinct peaks at 531 eV and 532.2 eV. The former peak corresponds to lattice oxygen (Sn–O) on the SnO₂ surface, whereas the latter one originates from oxygen vacancies (V_O). Fig. 1b shows the O 1s spectrum of the p-toluenethiol-modified SnO₂ film. Compared with the pristine SnO₂, the lattice oxygen (Sn–O) peak moves to 531.1 eV and the oxygen vacancy (V_O) peak shifts to 532.4 eV. The concurrent increase in lattice oxygen (Sn–O) content and decrease in oxygen vacancy (V_O) concentration demonstrate effective suppression of oxygen vacancy defects. Additionally, the integral area ratio of oxygen vacancies (V_O) to lattice oxygen (Sn–O) decreases from 0.394 to 0.347, indicating partial occupation of oxygen vacancies by the thiol groups of p-toluenethiol.

Fig. 1 XPS spectra of O 1s for (a) the control and (b) modified SnO₂ layers. (c) Sn 3d and (d) Pb 4f for the buried interfaces of control and modified samples.

In Fig. 1c, the Sn 3d_5/2 and Sn 3d_3/2 peaks of the pristine SnO₂ film are observed at 487.3 eV and 495.7 eV, respectively. For the p-toluenethiol modified SnO₂ film, these peaks shift toward lower binding energies, reading 487.1 eV and 495.6 eV, respectively. The binding energy shift confirms the interaction between SnO₂ and the p-toluenethiol interfacial layer, indicative of increased electron cloud density around Sn atoms.

The shift of Sn 3d and O 1s peaks arises from Lewis acid–base interactions between Sn atoms and thiol groups. These interactions promote passivation of predominant SnO₂ defects, including tin interstitials (Sn_i) and oxygen vacancies (V_o). Concurrently, the thiol groups may occupy oxygen vacancy sites in SnO₂, further contributing to V_o defect passivation.

To investigate the interaction between p-toluenethiol and the perovskite layer, XPS analysis are performed on the buried interfaces of both unmodified and p-toluenethiol-modified perovskite films. Fig. 1d presents the Pb 4f core-level spectra, where the Pb 4f_7/2 and 4f_5/2 peaks in the modified film exhibit distinct shifts toward higher binding energies compared to the control sample. Specifically, the 4f_7/2 peak shifts from 138.3 eV (control) to 138.4 eV (modified), and the 4f_5/2 peak moves from 143.1 eV to 143.2 eV. This higher binding energy shift indicates that thiol (–SH) groups interact with coordination-deficient Pb²⁺ ions, effectively reducing their concentration and decreasing the electron density around these ions. The observed peak displacement demonstrates that p-toluenethiol passivates uncoordinated Pb²⁺ sites at the interface.

¹H nuclear magnetic resonance (¹H NMR) is also employed to investigate the interactions between p-toluenethiol, perovskite, and SnO₂. As depicted in Fig. S2a,† upon mixing p-toluenethiol with SnO₂, the sulfhydryl proton shifts from 5.18 ppm to 5 ppm, indicating an interaction between the –SH group and tin oxide. Fig. S2b† shows the result of mixing p-toluenethiol with PbI₂, in which the sulfhydryl proton chemically shifts from 5.18 ppm to 5.2 ppm. This observation confirms the interaction between p-toluenethiol and uncoordinated Pb²⁺.

To investigate the regulatory effect of the modifier on the microstructure of perovskite films, the buried interface morphologies of control and modified devices are first comparatively analyzed via scanning electron microscopy (SEM) measurements. The control sample exhibited distinct surface roughness and discrete PbI₂ aggregates at the interface, a phase segregation phenomenon attributed to the disordered accumulation of unreacted PbI₂ during perovskite crystallization (Fig. 2a). In contrast, the modified sample demonstrated significantly optimized interfacial morphology (Fig. 2b), with a complete elimination of PbI₂ aggregates, indicating that the modifier optimizes the crystallization through coordination effects.

Fig. 2 SEM images of buried interfaces for (a) the control and (b) modified perovskite films. (c) XRD spectra of the control and modified perovskite films measured from buried interface side. SEM images of surfaces for (d) the control and (e) modified perovskite films. (f) XRD spectra of the control and modified perovskite films measured from surface side.

X-ray diffraction (XRD) analysis of the buried interface reveals notable structural differences (Fig. 2c). The modified sample exhibits enhanced characteristic peak intensity for perovskite (e.g., at 14.1°) compared to the control group, accompanied by the absence of the PbI₂ diffraction peak. These results agree well with the SEM observation, collectively confirming that the modification suppresses PbI₂ secondary phase formation to enhance perovskite phase purity while promoting preferred orientation growth of perovskite grains.

The surface morphology of perovskite films is probed with SEM measurements as well. The control perovskite films contained pinholes (Fig. 2d), which could induce charge recombination. In contrast, the treated perovskite films demonstrated enhanced crystallinity and improved crystal quality, manifested in increased grain size and reduced grain boundary (GB) density (Fig. 2e). Fig. 2f compares the XRD patterns of perovskite films before and after modification. After the p-toluenethiol treatment, the intensities of the (100) and (200) diffraction peaks increased, confirming the improved perovskite crystallinity achieved through the interfacial incorporation of p-toluenethiol.

Fig. S3† compares the water contact angles of pristine SnO₂ and p-toluenethiol-treated SnO₂. The contact angle on the untreated SnO₂ surface measures approximately 51.3°, whereas the p-toluenethiol-modified surface exhibits a significantly increased contact angle of 64.7°, demonstrating enhanced hydrophobicity. This surface property modification is critical for moisture resistance, as the hydrophobic benzene ring of p-toluenethiol forms a protective barrier to inhibit water penetration and halide ion migration at the buried interface, thereby improving device long-term stability.

Atomic force microscopy (AFM) characterization results (Fig. S4†) reveal that p-toluenethiol modification significantly reduces the surface roughness (RMS) of the SnO₂ electron transport layer (ETL) from 13.36 nm to 7.89 nm. This morphological optimization promoted uniform nucleation and lateral grain growth of the perovskite layer by mitigating heterogeneity in the wettability of the precursor solution during spin-coating, thereby reducing grain boundary density. The improved interfacial contact between the flattened SnO₂ and perovskite layers effectively suppresses leakage currents caused by localized voids or microcracks. Concurrently, the modified SnO₂ interface induces a reduction in the perovskite surface roughness from 47.83 nm to 31.90 nm, which minimizes optical scattering losses to enhance photon absorption efficiency while providing an ideal substrate for uniform deposition of subsequent hole transport layers, thereby mitigating interfacial carrier recombination and transport barriers.³³

Conductive atomic force microscopy (C-AFM) measurements are performed on the buried interface of perovskite films. Fig. S5a† displays the C-AFM current map of the pristine buried interface. The distinct high-current bright spots are observed, which is ascribed to the presence of conductive defect sites at the buried interface that facilitates leakage current pathways. In contrast, the area and intensity of these high-current bright spots are significantly reduced in the modified sample (Fig. S5b†), implying improved interfacial coverage. Fig. S5c† presents the line scan profiles obtained from Fig. S5a, S5b and S5d† shows the data statistics of Fig. S5c.† The pristine sample exhibits a pronounced tail in the high-current region, whereas the modified sample shows a current distribution in the low-current range. These results collectively demonstrate that p-toluenethiol modification reduces voids and pinholes while enhancing interfacial uniformity at the perovskite buried interface, synergistically suppressing leakage current.³⁴

To systematically analyze the energy level alignment between the SnO₂ electron transport layer and the perovskite layer, this study employs ultraviolet-visible absorption spectroscopy (UV-vis) and ultraviolet photoelectron spectroscopy (UPS) for interfacial energy level engineering characterization. The UV-vis absorption spectra of perovskite films are shown in the Fig. 3a, revealing no obvious difference in the absorption characteristics between the modified and original perovskite films across the spectral range, with their bandgaps both remaining at 1.57 eV (Fig. S6†), indicating that the modification does not alter the bulk optical properties of the materials.

Fig. 3 (a) UV/vis absorption spectra data of the control and modified perovskite films. (b) Steady-state PL and (c) TRPL of the control and modified samples. (d) SCLC data based on the electron-only devices without and with the p-toluenethiol treatment. (e) V_OC of PSCs at various illumination intensities with linear fits to the data. (f) J–V characteristics of the devices sweep from 1 to −1 V in dark.

UPS measurements disclose significant modulation of the SnO₂ electronic structure that the work function (W_F) of SnO₂ decreases from 4.36 eV to 4.23 eV after modification (Fig. S7a and b†), while the valence band maximum (VBM) and conduction band minimum (CBM) shift upward from −7.41 eV/−4.05 eV to −7.38 eV/−4.02 eV with the calculation methods in the ESI.† This energy level adjustment is attributed to the n-type doping effect of p-toluenethiol, which promotes the upward shift of the CBM, forming better energy level alignment with the perovskite layer.

The UPS characterization results for the surfaces and buried interfaces of perovskite films are shown in Fig. S8.† The CBM at the buried interface of perovskite shifts from −2.76 eV (control) to −2.93 eV (modified), indicating that the energy barrier between the perovskite and SnO₂ is reduced after modification. Consequently, the interfacial charge transfer can be improved. This alignment reduces electron transport resistance at the heterojunction, providing a structural basis for efficient carrier extraction.³⁵

To investigate how the SnO₂ surface passivation modulates electron extraction and carrier dynamics in perovskite layers, we analyze photoluminescence (PL) and time-resolved photoluminescence (TRPL) characteristics. The PL spectra (Fig. 3b) demonstrate stronger fluorescence quenching in p-toluenethiol-modified SnO₂ compared to bare SnO₂, confirming enhanced electron transfer efficiency from the perovskite layer to SnO₂via surface modification.³⁶ TRPL decay curves (Fig. 3c) are fitted using a biexponential function.

The extracted τ₁ (fast decay) corresponds to charge extraction at the ETL. Key fitting parameters (see ESI Table S1†) show that τ₁ drops from 275.67 to 53.89 ns, which represents an acceleration of the charge extraction at the perovskite/ETL interface after modification. Moreover, the τ_avg decreases from 1327.65 ns (bare SnO₂) to 532.89 ns (SnO₂/p-toluenethiol) as well, demonstrating accelerated charge extraction kinetics in the modified system.³⁷

However, the fluorescence quenching phenomenon in the PL signals may also arise from stronger non-radiative recombination caused by more trap states.³⁸ To investigate the changes in the density of trap states before and after modification, the space-charge-limited current (SCLC) measurements are carried out (Fig. 3d). At low bias voltage, the device exhibits a linear ohmic response. When the applied voltage exceeds the transition point, the current increases significantly due to the filling of trap states by charge carriers. Therefore, the trap density N_trap can be calculated from the trap-filled limit voltage V_TFL using the following equation:

where L and e are the thickness of the film and the elementary charge, while ε₀ and ε represent the vacuum permittivity and the relative permittivity, respectively.^39,40

Compared to SnO₂/perovskite (N_trap = 1.50 × 10¹⁵ cm⁻³), SnO₂/p-toluenethiol/perovskite exhibits a lower trap density (N_trap = 1.92 × 10¹⁵ cm⁻³). The reduced trap density contributes to the enhanced optoelectronic performance of the device.

Electrochemical impedance spectroscopy (EIS) measurements are performed on the control and modified PSCs, and the obtained Nyquist curves and their related fitting are displayed in Fig. S9.† The low- and high-frequency regions correspond to carrier recombination (R_rec) and charge transport processes (R_tr), respectively.⁴¹ Fitting results reveal that the charge transport resistance (R_tr) decreases to 1.09 × 10⁴ Ω for the modified device compared with the reference device (1.43 × 10⁴ Ω). While the carrier recombination resistance (R_rec) increases from 4.57 × 10⁴ Ω to 4.91 × 10⁴ Ω. The results indicate that p-toluenethiol modification effectively suppresses carrier recombination and enhances charge transport.⁴²

The ideality factor (n_id) can be obtained from the linear fit in Fig. 3e, according to the equation:⁴³

The n_id of the modified PSCs is 1.71 compared to 1.83 for the control PSCs, indicating reduced trap-assisted recombination within the device under open-circuit conditions. The reduction in n_id is mainly attributed to the buried heterogeneous interface optimized by the introduction of p-toluenethiol.

Dark current characterization shown in Fig. 3f reveals the improvement in carrier transport properties after modification. The modified devices exhibit a substantial reduction in dark current density under −1–1 V bias. The suppression of leakage current indicates improved interface charge selectivity and a reduction in defect-mediated recombination pathways.

Fig. S10a† presents the dark current–voltage (I–V) characteristics of devices with the structure of ITO/SnO₂ or SnO₂/p-toluenethiol/Ag. Linear sweep voltammetry is employed to qualitatively estimate the electrical conductivity of SnO₂. It is observed that the electrical conductivity increases after the p-toluenethiol modification. The corresponding electron mobility is calculated via the Mott–Gurney law equation (Fig. S10b†):

Here, μ represents the electron mobility, V is the voltage, J denotes the current density, ε₀ is the vacuum permittivity, ε stands for the dielectric constant of SnO₂, and L is the thickness of the SnO₂ film.⁴⁴ Through calculation, the electron mobility of the SnO₂/p-toluenethiol thin film is determined to be 4.81 × 10⁻⁵ cm² V⁻¹ s⁻¹, which is higher than that of the pristine SnO₂ film (2.75 × 10⁻⁵ cm² V⁻¹ s⁻¹). The incorporation of p-toluenethiol significantly enhances the electron mobility. This improvement can be attributed to the effective passivation of defect states in the SnO₂ by p-toluenethiol, thereby facilitating electron transport.⁴¹

Perovskite devices with the structure of indium tin oxide (ITO)/electron transport layer (ETL)/perovskite/hole transport layer (HTL)/silver (Ag) are fabricated to evaluate their photovoltaic performance. The device structure is shown in Fig. 4a, and cross-sectional scanning electron microscope (SEM) images of devices with p-toluenethiol treatment are shown in Fig. 4b, the perovskite layer has a thickness of ∼728 nm. The current density–voltage (J–V) characteristics of the best-performing PSCs under simulated AM 1.5 G solar radiation (100 mW cm⁻²) are shown in Fig. 4c, including both forward and reverse scan records. The corresponding performance parameters are summarized in Table 1. The modified device exhibits a J_SC of 26.01 mA cm⁻², an V_OC of 1.18 V, a FF of 83.23%, yielding a final PCE of 25.53%. This performance surpasses that of the control device, which shows a PCE of 24.16% with a J_SC of 25.65 mA cm⁻², a V_OC of 1.15 V, and a FF of 81.83%. Concurrently, the device exhibits significantly reduced hysteretic behavior, with the hysteresis index (HI) decreasing from 0.05 to 0.03. The photovoltaic parameters all are improved, which is attributed to the reduction of non-radiative recombination centers after structural optimization. Moreover, the average photovoltaic parameters with standard deviation are listed in Table S2.† The average PCE is (23.87 ± 0.27) % for the control devices and (25.33 ± 0.18) % for the modified solar cells. The larger PCE and smaller deviation indicate a generally superior photovoltaic performance and repeatability. Furthermore, the photovoltaic parameters of PSCs with different p-toluenethiol concentrations are compared in Fig. S11.† The results show that the modification effect is optimal at the concentration of 4 mg mL⁻¹.

Fig. 4 (a) Schematic diagram of an implemented device. (b) Cross-sectional SEM image of the PSCs. (c) J–V characteristic curves of the best-performing PSCs before and after p-toluenethiol modification. (d) EQE spectra. (e) Steady-state J_SC at the maximum power point for the devices with and without p-toluenethiol modification. (f) J–V curves of the flexible PSC. (g) PCE tracking at MPP under the continuous illumination of the simulated light (100 mW cm⁻²) for 300 h. Environmental stability of PSCs under (h) nitrogen atmosphere and (i) ambient air (RH = 40%, 25 °C).

Table 1 Photovoltaic parameters of the best-performing devices before and after p-toluenethiol modification

Sample		V OC (V)	J SC (mA cm^−2)	FF (%)	PCE (%)
Modified	Forward	1.18	26.01	83.23	25.53
	Reverse	1.16	25.83	82.85	24.76
Control	Forward	1.15	25.65	81.83	24.16
	Reverse	1.15	24.88	80.33	22.92

---

The external quantum efficiency (EQE) spectra of the corresponding devices are shown in Fig. 4d. The integrated J_SC is 25.33 mA cm⁻² (control device) and 24.50 mA cm⁻² (modified device), which match well with the values obtained from the J–V measurements.

In addition, the modified device exhibits superior J_SC stability (24.53 mA cm⁻²) compared to the control device during 300 s operation (Fig. 4e), with the maximum power point voltage (V_max) of 1 V and 1.02 V for the control and modified PSCs, respectively. Under light immersion, the stable photogenerated carrier output in the modified device also demonstrates that the destructive photocatalytic effect at the heterojunction is significantly suppressed.

The current density–voltage (J–V) characteristics of flexible perovskite solar cells are measured under simulated AM 1.5 G illumination (100 mW cm⁻²) (Fig. 4f), and the corresponding performance parameters are summarized in Table 2. The flexible device exhibits a J_SC of 24.84 mA cm⁻², a V_OC of 1.17 V, an FF of 79.75%, and a PCE of 23.27%. The statistic of the photovoltaic parameters for 10 devices is displayed in Fig. S12 and Table S2,† in which the average PCE is (22.09 ± 0.76) %. This performance demonstrates the feasibility of achieving high photovoltaic performance in flexible configurations while maintaining the mechanical adaptability required for roll-to-roll processing applications.

Table 2 Photovoltaic parameters of the best-performing flexible devices after p-toluenethiol modification

Sample		V OC (V)	J SC (mA cm^−2)	FF (%)	PCE (%)
	Forward	1.17	24.84	79.75	23.27
	Reverse	1.15	22.70	77.28	20.25

---

The long-term stability of the devices is investigated. Fig. 4g shows the continuous illumination stability of devices with and without p-toluenethiol treatment. The p-toluenethiol-modified devices maintain 93% of their initial efficiency after maximum power point tracking (MPPT) for 300 h, while the control devices retain only 39% of their initial efficiency under identical conditions.

Environmental stability is evaluated by exposing control and modified devices to nitrogen and ambient air (RH = 40%, 25 °C), as shown in Fig. 4h and i. After 2000 h in a nitrogen atmosphere, the PCE of modified devices remain stable without any decay, whereas control devices exhibit significant degradation, dropping to 73% of their initial efficiency. When exposed to ambient air for 2000 h, the modified photovoltaic cells retain 86% of their initial performance, while the control group suffers substantial efficiency loss and retains only 53% of its initial efficiency. Furthermore, the thermal stability is assessed by continuous aging the PSCs at 65 °C for 630 h (Fig. S13†). The results demonstrate that the p-toluenethiol modified devices retain 80% of their initial PCE, while the control devices maintain only 49%. This observation highlights the superior tolerance of the PSC to heat stress after the p-toluenethiol modification.

Conclusions

This study achieves synergistic regulation of interfacial defect passivation, energy-level alignment, and carrier dynamics in perovskite photovoltaics through molecular modification of the SnO₂/perovskite buried interface with p-toluenethiol. The thiol (–SH) groups of the modifier specifically bind to oxygen vacancies and hydroxyl groups on the SnO₂ surface via coordination reactions, forming a chemical passivation layer that effectively suppresses the formation of deep-level trap states. Concurrently, these groups establish strong coordination bonds with uncoordinated Pb²⁺ ions at the perovskite interface, reducing the precipitation of PbI₂ secondary phases and lattice defects. The aromatic benzene ring in the molecular structure induces interfacial dipole moments through π-conjugation, lowering the electron transport barrier between SnO₂ and perovskite to facilitate efficient charge extraction at the heterojunction. Furthermore, the hydrophobic nature of the benzene ring creates a moisture-resistant barrier at the interface, inhibiting halide ion migration and water penetration to mitigate ion migration and phase segregation under humid conditions.

With this strategy, the PSCs achieve a PCE of 25.53% and the related flexible devices deliver a PCE of 23.27%. Stability tests reveal that the modified devices preserve 93% of their initial PCE after continuous illumination for 300 h, remain nearly unchanged in performance after 2000 h storage in a nitrogen atmosphere, and exhibit only a 14% efficiency loss in air, outperforming unmodified comparable devices.

By integrating molecular-scale interfacial engineering, this work resolves the intrinsic trade-offs among efficiency and stability in perovskite devices, establishing both theoretical and technical foundations for their applications in high-performance optoelectronic systems and integrated energy solutions.

Data availability

The data of this manuscript are available upon reasonable request from the authors.

Author contributions

Yikun Hua: writing, visualization, methodology. Lei Zhao: methodology, formal analysis. Xinyue Song: methodology. Chao Wu: visualization, methodology. Jie Zhang: visualization, methodology. Weiyuan Chen: software, methodology. Lin Song: review & editing, conceptualization.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work is financially supported by the Fundamental Research Funds for the Central Universities and Key Research and Development Program of Shaanxi (Program No. 2023GXLH-091).

References

1. W. Hui, Y. G. Yang, Q. Xu, H. Gu, S. L. Feng, Z. H. Su, M. R. Zhang, J. O. Wang, X. D. Li, J. F. Fang, F. Xia, Y. D. Xia, Y. H. Chen, X. Y. Gao and W. Huang, Adv. Mater., 2020, 32, 9.
2. S. D. Stranks, G. E. Eperon, G. Grancini, C. Menelaou, M. J. P. Alcocer, T. Leijtens, L. M. Herz, A. Petrozza and H. J. Snaith, Science, 2013, 342, 341–344.
3. J. S. Manser, J. A. Christians and P. V. Kamat, Chem. Rev., 2016, 116, 12956–13008.
4. J. J. Yoo, G. Seo, M. R. Chua, T. G. Park, Y. L. Lu, F. Rotermund, Y. K. Kim, C. S. Moon, N. J. Jeon, J. P. Correa-Baena, V. Bulovic, S. S. Shin, M. G. Bawendi and J. Seo, Nature, 2021, 590, 10.
5. A. Lan, H. Lu, B. Huang, F. Chen, Z. K. Chen, J. Wang, L. Q. Li and H. Do, ACS Appl. Mater. Interfaces, 2024, 16, 64825–64833.
6. M. L. Zhang, Z. Q. Ying, X. Li, S. Li, L. Chen, X. C. Guo, L. H. Liu, Y. H. Sun, J. Wu, Y. H. Zeng, C. X. Xiao, J. Wu, X. Yang and J. C. Ye, Adv. Mater., 2025, 37, 11.
7. J. Zhou and M. Zhong, Acta Mater. Compositae Sin., 2022, 39, 1937–1955.
8. Z. W. Gao, Y. Wang and W. C. H. Choy, Adv. Energy Mater., 2022, 12, 20.
9. S. F. Wu, J. Zhang, Z. Li, D. J. Liu, M. C. Qin, S. H. Cheung, X. H. Lu, D. Y. Lei, S. K. So, Z. L. Zhu and A. K. Y. Jen, Joule, 2020, 4, 1248–1262.
10. B. Yu, K. Wang, Y. P. Sun and H. Z. Yu, Adv. Mater., 2025, 37, 12.
11. B. B. Liu, H. Bi, D. M. He, L. Bai, W. Q. Wang, H. K. Yuan, Q. L. Song, P. Y. Su, Z. G. Zang, T. W. Zhou and J. Z. Chen, ACS Energy Lett., 2021, 6, 2526–2538.
12. Q. Y. Li, H. Liu, C. H. Hou, H. M. Yan, S. D. Li, P. Chen, H. Y. Xu, W. Y. Yu, Y. P. Zhao, Y. P. Sui, Q. X. Zhong, Y. Q. Ji, J. J. Shyue, S. Jia, B. Yang, P. Y. Tang, Q. H. Gong, L. C. Zhao and R. Zhu, Nat. Energy, 2024, 9, 13.
13. B. Du, Y. X. Lin, J. T. Ma, W. D. Gu, F. Liu, Y. J. Yao and L. Song, Chem. Sci., 2025, 16, 1876–1884.
14. Z. Y. Wu, J. Z. Su, N. Y. Chai, S. Y. Cheng, X. Y. Wang, Z. L. Zhang, X. L. Liu, H. Zhong, J. F. Yang, Z. P. Wang, J. B. Liu, X. Li and H. Lin, Adv. Sci., 2023, 10, 9.
15. G. Q. Tong, L. K. Ono, Y. Q. Liu, H. Zhang, T. L. Bu and Y. B. Qi, Nano-Micro Lett., 2021, 13, 14.
16. Y. N. Zhang, B. Yu, Y. P. Sun, J. K. Zhang, Z. Su and H. Z. Yu, Angew. Chem., Int. Ed., 2024, 63, 11.
17. Q. Q. Zhao, B. Q. Zhang, W. Hui, Z. H. Su, H. Wang, Q. Zhang, K. Gao, X. X. Zhang, B. H. Li, X. Y. Gao, X. Wang, S. De Wolf, K. Wang and S. P. Pang, J. Am. Chem. Soc., 2024, 146, 19108–19117.
18. Y. S. Jeon, D. H. Kang, J. H. Kim and N. G. Park, J. Mater. Chem. A, 2023, 11, 3673–3681.
19. Z. Xiong, X. Chen, B. Zhang, G. O. Odunmbaku, Z. P. Ou, B. Guo, K. Yang, Z. P. Kan, S. R. Lu, S. S. Chen, N. A. N. Ouedraogo, Y. Cho, C. Yang, J. Z. Chen and K. Sun, Adv. Mater., 2022, 34, 10.
20. Y. Pu, H. J. Su, C. C. Liu, M. Guo, L. Liu and H. Z. Fu, Energies, 2023, 16, 30.
21. B. Zhou, C. Shang, C. Wang, D. Qu, J. Qiao, X. Zhang, W. Zhao, R. Han, S. Dong, Y. Xue, Y. Ke, F. Ye, X. Yang, Y. Tu and W. Huang, Research, 2024, 2024.
22. Y. Xiao, X. Y. Yang, R. Zhu and H. J. Snaith, Science, 2024, 384, 846–848.
23. M. Ebic, Nanotechnology, 2025, 36, 12.
24. W. C. Xiang, Y. H. Gao, B. B. Yuan, S. P. Xiao, R. Wu, Y. R. Wan, Z. Q. Liu, L. Ma, X. B. Chen, W. J. Ke, G. J. Fang and P. L. Qin, Energy Environ. Sci., 2025, 18, 406–417.
25. Y. J. Gao, W. Y. Gong, Z. Q. Zhang, J. N. Guo, J. Y. Ma, X. Li, Y. L. Zeng and M. X. Wu, Angew. Chem., Int. Ed., 2025, 13,  DOI:10.1002/anie.202424479.
26. J. He, J. Y. Zhang, Y. Zhang, J. M. Xu, Z. Liang, P. D. Zhu, W. B. Peng, G. P. Qu, X. Pan, X. Z. Wang and B. M. Xu, Angew. Chem., Int. Ed., 2025, 64, 10.
27. T. Y. Xia, Y. F. Ouyang, C. Wang, Y. Pan, Q. Gao, X. Chen, B. Zhang, K. Chen, Z. J. He, X. B. Yuan, C. X. Shen, B. Guo, Y. H. Deng, S. J. Chen, T. M. Jiang and K. Sun, J. Phys. Chem. Lett., 2024, 15, 5854–5861.
28. W. C. Xiang, Y. H. Gao, B. B. Yuan, S. P. Xiao, R. Wu, Y. R. Wan, Z. Q. Liu, L. Ma, X. B. Chen, W. J. Ke, G. J. Fang and P. L. Qin, Energy Environ. Sci., 2025, 18, 406–417.
29. J. Cao, J. Yin, S. F. Yuan, Y. Zhao, J. Li and N. F. Zheng, Nanoscale, 2015, 7, 9443–9447.
30. P. M. Gschwend, F. M. Schenk, A. Gogos and S. E. Pratsinis, Materials, 2021, 14, 13.
31. M. Tang, S. L. Zhu, Z. T. Liu, C. Jiang, Y. C. Wu, H. Y. Li, B. Wang, E. J. Wang, J. Ma and C. L. Wang, Chem, 2018, 4, 2600–2614.
32. Y. Li, L. Dong, Y. Cai, Y. Li, D. Xu, H. Lei, N. Li, Z. Fan, J. Tan, R. Sun, B. Wang, J. Gong, Z. Lin, K. Guo, X. He and Z. Liu, Angew. Chem., 2025, e202504902,  DOI:10.1002/anie.202504902.
33. S. P. Koiry, P. Jha, C. Sridevi, D. Gupta, V. Putta and A. K. Chauhan, Mater. Today Commun., 2023, 36, 11.
34. S. C. Liu, H. Y. Lin, S. E. Hsu, D. T. Wu, S. Sathasivam, M. Daboczi, H. J. Hsieh, C. S. Zeng, T. G. Hsu, S. Eslava, T. J. Macdonald and C. T. Lin, J. Mater. Chem. A, 2024, 12, 2856–2866.
35. J. M. Qiu, X. Y. Mei, M. X. Zhang, G. L. Wang, S. W. Zou, L. Wen, J. M. Huang, Y. Hua and X. L. Zhang, Angew. Chem., Int. Ed., 2024, 63, 11.
36. M. Stolterfoht, V. M. Le Corre, M. Feuerstein, P. Caprioglio, L. J. A. Koster and D. Neher, ACS Energy Lett., 2019, 4, 2887–2892.
37. P. C. Zhu, S. Gu, X. Luo, Y. Gao, S. L. Li, J. Zhu and H. R. Tan, Adv. Energy Mater., 2020, 10, 7.
38. P. Caprioglio, M. Stolterfoht, C. M. Wolff, T. Unold, B. Rech, S. Albrecht and D. Neher, Adv. Energy Mater., 2019, 9, 1901631.
39. D. Shi, V. Adinolfi, R. Comin, M. J. Yuan, E. Alarousu, A. Buin, Y. Chen, S. Hoogland, A. Rothenberger, K. Katsiev, Y. Losovyj, X. Zhang, P. A. Dowben, O. F. Mohammed, E. H. Sargent and O. M. Bakr, Science, 2015, 347, 519–522.
40. Z. Y. Wu, X. L. Liu, H. Zhong, Z. H. Wu, H. Chen, J. Z. Su, Y. C. Xu, X. Y. Wang, X. Li and H. Lin, Small Methods, 2022, 6, 10.
41. J. S. Xie, K. Huang, X. G. Yu, Z. R. Yang, K. Xiao, Y. P. Qiang, X. D. Zhu, L. B. Xu, P. Wang, C. Cui and D. R. Yang, ACS Nano, 2017, 11, 9176–9182.
42. A. R. C. Bredar, A. L. Chown, A. R. Burton and B. H. Farnum, ACS Appl. Energ. Mater., 2020, 3, 66–98.
43. Z. Liang, H. F. Xu, Y. Zhang, G. Z. Liu, S. L. Chu, Y. L. Tao, X. X. Xu, S. D. Xu, L. Y. Zhang, X. J. Chen, B. M. Xu, Z. G. Xiao, X. Pan and J. J. Ye, Adv. Mater., 2022, 34, 12.
44. P. N. Murgatroyd, J. Phys. D: Appl. Phys., 1970, 3, 151.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta03804a

‡ These authors contributed equally.

---