Dual‐Site‐Binding Ligand‐Driven (100) Orientation for Efficient and Stable Perovskite Solar Cells

ABSTRACT

Regulating crystal orientation is a powerful strategy for enhancing the optoelectronic performance of perovskite solar cells (PSCs). However, most existing approaches rely on single‐site ligand binding, which not only forms insulating layers that hinder charge transport but also involves weak coordination that limits orientation control. Here, we report a holistic strategy based on a dual‐site‐binding ligand, dithiopyr, that simultaneously modulates nucleation kinetics, crystal orientation, defect passivation, and charge transport in perovskite films. By coordinating its dual thioester functionalities with adjacent Pb²⁺ defect sites, dithiopyr effectively relieves lattice strain and promotes a preferential (100) orientation, thereby facilitating efficient charge transport. As a result, the inverted PSCs achieve power conversion efficiencies of 26.90% (certified 26.14%) (1.55 eV), 22.02% (1.25 eV), and 23.11% (1.68 eV), respectively. Notably, large‐area modules (30 × 30 cm²) deliver an efficiency of 21.28%, which is among the highest values reported for inverted perovskite solar cells. Unencapsulated devices retain 93.1% of their initial efficiency after 3600 h under ambient conditions and maintain over 95% of their initial performance after 1100 h of maximum power point tracking. This work establishes dithiopyr as a versatile and robust platform for precise crystal orientation control toward high‐performance perovskite photovoltaics.

Dithiopyr directs oriented crystallization of FAPbI₃ perovskites through dual‐site coordination and facet‐selective stabilization. By moderating precursor‐to‐perovskite conversion, suppressing defect formation, and reducing nonradiative recombination, this fluorinated ligand promotes highly ordered (100)‐oriented films. The resulting devices exhibit improved efficiency and operational stability, highlighting molecularly programmed crystallization as an effective route toward high‐performance perovskite photovoltaics.

1. Introduction

The rapid evolution of metal halide perovskite solar cells (PSCs) has fundamentally redefined the landscape of third‐generation photovoltaics. Within little more than a decade, their PCEs have surged from single digits to certified values exceeding 27%, rivaling the performance of established silicon‐based technologies [1, 2, 3]. These achievements stem from a unique combination of superior optoelectronic properties, including steep absorption onsets and defect‐tolerant charge transport, alongside the promise of cost‐effective, low‐temperature solution manufacturing [4, 5, 6, 7]. However, despite these strides, bridging the gap between laboratory efficiency and the theoretical Shockley‐Queisser limit, while simultaneously ensuring multi‐decadal operational longevity, remains a formidable challenge [8, 9, 10]. Crucially, controlling crystallographic orientation can significantly improve device performance. It is generally observed that the (111) crystal plane offers moisture resistance [11] and inhibits ion migration, favoring environmental stability. The (100) crystal plane exhibits the superior carrier mobility required for high‐efficiency devices [12, 13], yet maintaining this orientation typically comes at the cost of stability [14].

Numerous strategies have been proposed to regulate the perovskite crystal orientation. Lin et al. leveraged dimethyl sulfoxide (DMSO) to form a complex with Pb²⁺, retarding the crystallization rate to favour (100) growth, thereby reducing surface energy and defect density [15]. Hou et al. employed a self‐guided crystallization strategy via vapor deposition to promote preferential (100) alignment through an intermediate phase [16]. Alternatively, Ding et al. promoted (110) exposure by introducing 1‐butyl‐4‐methylpyridinium chloride (4MBPCl), which selectively reduced the surface energy of that specific plane [17]. Similarly, Fang et al. added potassium tetrakis(pentafluorophenyl)borate (KTFB) to the antisolvent and optimized the grain boundary growth by adjusting the (110) crystal plane, thereby successfully preparing a film with a grain size of more than 2 microns [18].

However, most existing regulation strategies rely on ligands that bind to a single site within the perovskite lattice. On the one hand, while such approaches can promote vertically oriented molecular packing, they often lead to the formation of insulating layers that impede interfacial charge transport. Although an ordered crystallographic orientation can be achieved, it is frequently accompanied by compromised defect passivation and reduced charge transport efficiency [19, 20, 21, 22]. In addition, these ligands may induce the formation of voids or secondary phases [23] or fail to effectively direct crystal growth [24, 25, 26]. On the other hand, single‐site ligand binding involves dynamic and relatively weak coordination, primarily regulating nucleation and crystallization kinetics, with limited influence on crystal orientation and typically resulting in random or weak preferential alignment. In contrast, dual‐ or multi‐site‐binding ligands establish strong multipoint chelating or bridging interactions with the lattice, enabling selective passivation of specific crystal facets, thereby effectively inducing oriented crystallization and pronounced preferential orientation [27, 28]. Therefore, developing a holistic strategy that employs dual‐ or multi‐site‐binding ligands to simultaneously regulate nucleation dynamics, control crystal orientation, suppress defects, and preserve efficient charge transport is crucial for fully realizing the potential of inverted perovskite solar cells.

In this work, we simultaneously enhance device efficiency and stability by employing a dual‐site‐binding ligand to regulate nucleation dynamics and control crystal orientation of perovskites. Initial analysis via high‐resolution transmission electron microscopy (HR‐TEM) and scanning electron microscopy (SEM) revealed that perovskite films suffer from random polycrystalline orientation and disordered grain boundaries. By introducing dithiopyr, a molecule equipped with multi‐functional groups, we successfully induced preferential (100) orientation, significantly improving carrier transport and crystallinity. This alignment is attributed to the geometric matching between the active sites of dithiopyr and the perovskite (100) lattice. Furthermore, the fluorinated moieties of dithiopyr establish a hydrophobic barrier at the grain boundaries. The unencapsulated device retained 93.1% of its initial efficiency after 3600 h of storage in ambient air. Notably, this strategy proves universally effective across perovskite active layers with different band gaps, underscoring the commercial potential of dithiopyr as a versatile additive for high‐performance photovoltaics.

2. Results and Discussion

We established a theoretical model of three crystal planes (100), (110), and (111). By calculating the surface energy (based on density functional theory using the CP2K package with the GGA‐PBE functional and Grimme's DFT‐D3 dispersion correction), it is found that FAPbI₃ tends to form the (100) crystal plane (Figure 1a and Note S1) [29, 30]. Theoretically, the energy of FAPbI₃ (100) crystal plane is the lowest, but the crystallization kinetics, experimental methods, system temperature, and other factors lead to no obvious preferred orientation of crystallization. To elucidate the crystallization dynamics of the pristine perovskite films, we employed in situ photoluminescence (PL) spectroscopy. As shown in Figure 1b,c, the control film displays a weak spectral signal during both the spin coating and annealing phases. Specifically, the PL intensity initially rises but subsequently decays after antisolvent addition, suggesting that rapid, uncontrolled nucleation leads to particle agglomeration and a high density of defects, which consequently act as non‐radiative recombination centers. The structural consequences of this kinetic instability were further characterized by grazing‐incidence wide‐angle X‐ray scattering (GIWAXS). The control samples exhibit broad, ring‐shaped scattering features, indicating a polycrystalline morphology with random orientation (Figure 1d). This disorder is corroborated by HR‐TEM throughout the film. Analysis of lattice fringes from multiple regions reveals a chaotic mixture of crystallographic planes, confirming the lack of preferential orientation (Figure 1e,f) [31]. Furthermore, diffraction patterns extracted from the grain interiors show inconsistency (Figure 1g). The corresponding SEM analysis of the surface and bottom interface shows that the surface is unevenly distributed with grains (average size ≈ 0.576 µm), and there are many impurity phases in the bottom interface in addition to pinholes. (Figure 1h; Figure S1) [32]. We attribute these poor morphological features to a kinetic imbalance during nucleation and growth, which inevitably increases defect density and compromises device performance.

FIGURE 1.

Crystallinity and crystal orientation of the control films. (a) Theoretical model and surface energy of each crystal plane of FAPbI₃. In situ PL of the control perovskite film during (b) spin coating and (c) annealing. (d) The GIWAXS of the control was a diffusion ring. (e) The cross‐sectional HR‐TEM of the control perovskite film. (f) The lattice fringes of the serial numbered position in the cross‐section of the perovskite film. (g) The nanobeam diffraction (NBD) pattern extracted from the cross‐sectional HR‐TEM image of the perovskite film in the control. (h) SEM images of the surface and bottom of the perovskite film in the control.

In order to solve the problem of disordered orientation of perovskite crystals, we introduced dithiopyr as a multifunctional modulator to passivate undercoordinated sites via a host‐guest mechanism (Figure S2) [33, 34]. Density functional theory (DFT) simulations reveal the molecule's distinct polarity (Figure 2a), with thioester atoms exhibiting a strong negative electrostatic potential, identifying them as potent Lewis bases that can coordinate with and neutralize cationic Pb²⁺ traps [35, 36]. Conversely, the fluorinated periphery displays a neutral‐to‐positive potential, creating a low‐surface‐energy barrier that repels moisture [37, 38]. Frontier molecular orbital (FMO) analysis (Figure S3) shows the highest occupied molecular orbital (HOMO) is delocalized over the core, while the lowest unoccupied molecular orbital (LUMO) is confined to the sulfonyl anchoring groups. Figure 2b shows that dithiopyr interacts with the lattice in the perovskite layer and induces the reconstruction of [PbI₆]⁴⁻ from a disordered arrangement to a high‐order continuous lattice, thereby achieving high‐quality crystallization of the perovskite film. When the anti‐solvent was added, the PL peak of the target film continued, indicating that the crystallization kinetics slowed down. (Figure 2c) [39]. This enhancement suggests that dithiopyr effectively passivates surface states in operando and optimizes the crystal nucleation process [40, 41]. More critically, to quantitatively analyze the crystallization kinetics without visual ambiguity, we extracted 1D time‐slice profiles of the PL peak position and intensity from the raw in situ spectra (Figure S4). The extracted temporal profiles explicitly reveal that Target film exhibits a higher and more stable PL intensity following anti‐solvent application, demonstrating that dithiopyr effectively suppresses defect generation and reduces grain boundary disorder during early nucleation (Figure S4a). During the annealing stage, the extracted temporal profiles reveal that the second‐stage PL enhancement of the Target film is delayed compared to the Control, and it maintains a higher PL intensity toward the end of the annealing process (Figure S4b). This definitively indicates that dithiopyr successfully decelerates the rapid crystallization and secondary recrystallization processes, enabling a slower, more ordered, and highly controlled formation of the perovskite phase (Figure 2d) [42]. In addition, we performed in situ XRD measurements (Figure S5). The structural evolution indicates the direct formation of the photoactive perovskite phase in the target film, without obvious δ phase. Our findings suggest that the target molecule regulates crystallization mainly through the synergistic effects of thermodynamic surface‐energy optimization and kinetic deceleration. We monitored in situ PL for samples with various dithiopyr concentrations and found that the optimal condition is 0.2 mg/mL additive (Figures S6 and S7), and the optimal dithiopyr‐to‐Pb molar ratio is calculated to be approximately 0.033mol%. Herein, the perovskite film without dithiopyr is defined as the control sample, whereas the film incorporating 0.2 mg/mL dithiopyr is denoted as the target sample. The time‐dependent PL spectra of the control and target samples are shown in Figure S8. The strong intensity indicates the improvement of crystal quality and the decrease of defect density in the target film. In addition, the 1D time‐slice absorbance profiles (Figure S10) extracted from the in situ UV ‐visible absorption spectroscopy (Figure S9) confirm that dithiopyr extends the transition window to 3.61 s, slowing the conversion kinetics. This delay promotes the growth of larger, (100)‐oriented domains while suppressing defects, optimizing grain growth and resulting in thermodynamically stable orientations [43]. This kinetic control induces a drastic microstructural evolution. GIWAXS measurements of the target film reveal a concentration of scattering signals into intense, discrete Bragg reflections (Figure 2e). Furthermore, varying dithiopyr concentrations confirmed that 0.1 mg/mL exhibited a relatively uniform diffuse ring, whereas the 0.2 mg/mL treatment yielded the most intense and discrete Bragg reflections (Figure S7).

FIGURE 2.

Crystallinity and crystallographic orientation of the dithiopyr‐treated films. (a) Local electron nucleophilic energy of dithiopyr calculated via DFT. (b) Schematic illustration of the perovskite film formation process upon the addition of dithiopyr. In situ PL spectra of the dithiopyr‐treated film during (c) spin‐coating and (d) annealing. (e) GIWAXS pattern of the dithiopyr‐treated film. (f) Cross‐sectional HR‐TEM image of the dithiopyr‐modified perovskite film, where the blue boxed region highlights the grain boundaries. (g) Lattice fringes at the numbered locations within the film, with spacings corresponding to the measured distances of the indicated crystal planes. (h) Diffraction patterns extracted from the cross‐sectional HR‐TEM image of the target perovskite film. (i) Top‐view and buried‐interface SEM images of the target perovskite film.

Cross‐sectional HR‐TEM further validates this improvement, showing continuous, pore‐free crystals (Figure 2f). Lattice fringe analysis indicates a uniform spacing of 0.32 nm, corresponding to the (100) plane (Figure 2g), while diffraction patterns confirm the (100) orientation across the grain cross‐section (Figure 2h) [31]. Macroscopically, SEM analysis shows that the dithiopyr‐ treated films are highly crystalline and the buried interface is flat and pinhole‐free (Figure 2i). The average grain size increases from 0.576 to 0.671 µm, with reduced grain boundary density. Dithiopyr's nucleation inhibition effect promotes larger grain growth, reducing interface defects [44, 45, 46].

We further elucidated the adsorption energetics via DFT simulations on various perovskite crystal facets (Figure 3a; Figure S11). To determine the preferred binding site, we compared the energetics of oxygen (O)‐ versus sulfur (S)‐anchoring on the (100) surface. Calculations reveal that the carbonyl oxygen anchoring configuration (E_ads = −1.679 eV) is thermodynamically favored over the sulfide anchoring group (E_ads = −1.420 eV). This energy difference of 0.26 eV confirms the formation of a robust Pb─O coordinate bond. Similarly, the (110) and (111) surfaces also exhibit a preference for O‐anchoring. To establish a quantitative structural basis for the proposed dual‐site bidentate mechanism, we compared the unrelaxed FAPbI₃ (100) surface with the optimized dithiopyr configuration. The nearest‐neighbor distance between two undercoordinated Pb²⁺ sites on the (100) surface is approximately 6.32 Å (Figure S12), close to the calculated O─O distance of 6.38 Å between the two carbonyl oxygen atoms in dithiopyr. This small spatial mismatch provides quantitative support for the proposed dual‐site anchoring configuration. Moreover, control experiments using rationally selected monodentate and bidentate molecules (Figure S13–S16, Table S1) further support the importance of bidentate coordination in optimizing crystal orientation and defect passivation [42]. This strong interaction indicates chemisorption [47], which passivates undercoordinated Pb²⁺ defects [40, 48]. Comparing adsorption energies across different facets, dithiopyr exhibits a stronger affinity for the (100) surface [17]. This thermodynamic preference aligns perfectly with the experimentally observed (100) orientation discussed earlier. Furthermore, the density of states (DOS) analysis shows that the deep traps in the band gap are significantly suppressed. (Figure 3b). The clean DOS profile near E_F eliminates the major non‐ radiative recombination pathways, which are responsible for the enhanced material stability under prolonged illumination and thermal stress. To elucidate how molecular interactions modulate the lattice structure, we analyzed the average surface Pb─I bond lengths on different facets before and after passivation (Figure 3c). The pristine bare (100) facet shows severe structural contraction caused by undercoordinated Pb²⁺, with an average Pb─I bond length (∼3.167 Å) much shorter than those of the (110) and (111) facets. After introducing dithiopyr, strong Lewis acid‐base coordination triggers local electron redistribution and weakens adjacent Pb─I interactions, significantly elongating the Pb─I bond on the (100) surface to ∼3.195 Å and releasing surface strain. Combined with the molecular intrinsic bond length (Figure S17), the (100)‐O configuration exhibits the narrowest bond‐length distribution, indicating a highly uniform local structure. The pronounced elongation and homogenization of Pb–I bonds verify that dithiopyr effectively relaxes the (100) surface toward a more stable, ideal bulk‐like geometry.

FIGURE 3.

Interactions between dithiopyr and the perovskite lattice. (a) Adsorption configurations and binding energies of dithiopyr on different perovskite crystal facets, revealing a thermodynamic preference for coordination on the (100) surface. (b) Density of states (DOS) of dithiopyr on the corresponding crystal planes. (c) Calculated bond lengths for the interaction between dithiopyr and the perovskite. (d) FTIR spectra showing redshifts in C═O and Pb─I vibrations, which confirms the formation of chemical coordination. (e) XPS spectra of the Pb 4f core levels.

The chemical nature of this interaction was corroborated by Fourier transform infrared spectroscopy (FTIR) analysis (Figure 3d). FTIR spectroscopy reveals interfacial chemical interactions. The C═O stretching peak shifts to a higher wavenumber upon contact with the perovskite, indicating a changed local chemical environment around the carbonyl group. Although metal coordination may weaken this bond, the observed blueshift may be related to the disruption of intermolecular interactions among pristine target molecules after anchoring to the perovskite lattice [49]. The Pb─I vibration peak also shifts to a higher frequency, suggesting that molecular coordination modifies the local Pb–I environment and contributes to residual surface‐strain modulation [35, 36]. Finally, X‐ray photoelectron spectroscopy (XPS) provided direct evidence of the interfacial electronic modulation (Figure 3e; Figure S18). Upon incorporating dithiopyr, the Pb 4f core‐level peaks exhibit a uniform shift toward lower binding energies (from 138.19 to 138.00 eV) accompanied by a similar downward shift in the I 3d spectra. After C 1s calibration, the unchanged C 1s peak excludes obvious charging artifacts. The parallel shifts of Pb 4f and I 3d toward lower binding energies suggest that dithiopyr modulates the perovskite surface electronic structure while interacting with Pb─I related species. This effect may arise from local coordination with undercoordinated Pb sites and surface dipole formation induced by coordinated dithiopyr molecules and outward‐oriented fluorinated groups, consistent with the KPFM and Mott–Schottky results.

The macroscopic structural effects of this kinetic control were further evaluated using X‐ray diffraction (XRD). Dithiopyr‐processed film displays a significantly enhanced crystallographic texture, characterized by sharper and more intense (100) and (200) reflections (Figure 4a). This observation supports our hypothesis that extending the growth window promotes the formation of high‐quality grains. To quantify the (100) orientation, we calculated the intensity ratio of the (100) diffraction peak to other reflections (Table S2) [50], finding that the (100)/(110) ratio was 3.1 times higher for the dithiopyr‐modified film than the control. At the same time, XRD (Figure S19) measured at the perovskite buried interface shows that the dithiopyr‐modified film maintains this preferred (100) orientation (Table S3). Furthermore, full‐pattern Rietveld refinement of the high‐resolution XRD data decouples structural texture effects and confirms a substantial reduction in the mass fraction of residual PbI₂ from 4.16% to 0.22% (Figure S20) [46, 51]. Since the mechanism relies on interacting with the ubiquitous [PbI₆]⁴⁻ inorganic sublattice rather than specific organic cations, we postulated that dithiopyr serves as a broad‐spectrum modulator applicable to a wide range of perovskite formulations: narrow‐bandgap system (1.25 eV, Figure 4b), wide‐bandgap system (Figure 4c; Figure S21), MAPbI₃ (Figure 4d), and FAPbI₃ (Figure S22). In all cases, we observed a uniform improvement in crystallinity, manifested by stronger and narrower diffraction peaks. This broad‐spectrum efficacy validates that the interaction is driven by the common lead‐halide backbone. This establishes dithiopyr as a versatile and robust strategy for high‐performance perovskite photovoltaics [37, 47, 52]. We further analyzed the lattice stress using the 2θ‐sin²φ method [53, 54, 55]. For the control film, the 2θ position gradually shifted toward smaller angles with increasing sample tilt (Figure 4e). The resulting negative slope in the fitting curve (Figure 4g) indicates the presence of tensile strain in the perovskite film. After the addition of dithiopyr, the grazing incidence X‐ray diffraction (GIXRD) peaks showed the opposite trend (Figure 4f); the fitting curve showed a positive slope, confirming that the tensile strain was successfully relieved and transitioned to compressive strain, which prevents excessive volume shrinkage during desolvation by promoting grain accumulation and modifying the lattice stress.

FIGURE 4.

Dithiopyr‐induced enhancement in crystalline, strain relaxation, and surface electronic modulation. (a–d) The XRD patterns of the control and dithiopyr‐treated films show significantly enhanced (100) and (200) diffraction peaks in the four formulations. GIXRD spectra of (e) control and (f) target perovskite films obtained at various grazing incidence angles (10°–50°). (g) Linear fitting of 2θ versus sin²ψ for the control and target films. (h–j) KPFM surface potential maps and corresponding CPD histograms. (k) TOF‐SIMS depth profiles confirming the uniform vertical distribution of dithiopyr throughout the perovskite film.

Kelvin probe force microscopy (KPFM) measurements (Figure 4h–j) revealed a shift in the contact potential difference (CPD) from +0.2 to −0.25 V upon dithiopyr treatment, corresponding to the formation of an interfacial dipole layer [56]. This electronic modulation optimizes energy level alignment between the perovskite surface and the electron transport layer (ETL), improving electron extraction and reducing interfacial recombination, which is vital for maximizing the device's fill factor (FF) and open‐circuit voltage (V _OC) [47, 57, 58]. The bandgap of the active layer was also checked using UV‐visible absorption spectroscopy (Figure S23). The introduction of dithiopyr does not significantly change the absorption of the perovskite active layer. To translate this molecular logic to the macroscopic scale, we probed the vertical distribution of dithiopyr using time‐of‐flight secondary ion mass spectrometry (TOF‐SIMS) depth profiling (Figure 4k). A pervasive challenge with large molecular additives is their tendency toward surface segregation, which can inadvertently form insulating barriers. However, the continuous depth profile traces of the fluorine fragment (F⁻) as a unique molecular fingerprint provide reliable qualitative evidence. Compared with the F signal of the Control thin film perovskite layer (Figure S24), it shows that the dithiopyr additive effectively penetrates the overall grain boundary network [47, 59, 60].

Prior to full device integration, we systematically evaluated the fundamental physical properties of the thin films to assess their potential in high‐performance perovskite solar cell applications. Additional spectroscopic characterization was performed to validate the passivation mechanism. Steady‐state PL (Figure 5a) exhibits a significant increase in emission intensity for the modified film, signaling reduced non‐radiative recombination [55]. More quantitatively, time‐resolved photoluminescence (TRPL) (Figure 5b; Table S4) reveals a dramatic enhancement in carrier lifetime, which surged from 391.22 ns in the control to 1005.49 ns with dithiopyr. This microsecond‐scale lifetime confirms that the sulfonyl groups effectively anchor to lattice defects, significantly suppressing deep‐level traps. This is further corroborated by Trap Density of States (tDOS) measurements (Figure 5c), the deep traps are significantly suppressed while the shallow traps are also effectively passivated [61, 62]. This passivation effect significantly prolongs carrier diffusion lengths, which is critical for maximizing the V _OC and FF of the device [47, 63, 64]. To further investigate the recombination kinetics, we analyzed the dependence of V _OC on light intensity (Figure 5d) [65]. In contrast, the target device achieves a substantially reduced ideality factor of n = 1.38 (compared to n = 1.858 for the control device). This shift confirms that dithiopyr modification effectively suppresses deep‐level trap‐assisted Shockley‐Read‐Hall (SRH) recombination. However, since the ideality factor remains above the band‐to‐band limit of 1.0, it suggests that SRH recombination is not entirely eliminated. The residual non‐radiative V_OC loss is likely dominated by remaining shallow‐level defects and recombination pathways at unpassivated regions of the charge transport interfaces [42, 66]. According to the Shockley‐Queisser limit, the intrinsic radiative recombination loss of perovskite solar cells is about 0.28 V, and the non‐radiative voltage loss is precisely calculated to be only 0.063 V. This value was directly extracted from the electroluminescence external quantum efficiency (EQE‐EL) measurements (Figure 5e), where the dithiopyr‐passivated device exhibited a substantially enhanced peak EQE‐EL of ∼8.7% compared to the control (∼4.1%). Such a low non‐ radiative loss indicates that the density of defect states in the film is effectively suppressed, and the non‐radiative recombination is significantly reduced [67].

FIGURE 5.

Optoelectronic characterization of dithiopyr‐Modified Films. (a, b) Steady‐state PL and TRPL spectra reveal enhanced emission intensity and prolonged carrier lifetime. (c) The tDOS spectra of the control and dithiopyr‐treated. (d) Light‐intensity‐dependent V _OC yields a lower ideality factor, indicating suppressed trap‐assisted recombination. (e) Voltage‐dependent electroluminescence external quantum efficiency EQE‐EL of the control and target devices. (f) SCLC measurements confirm reduced trap density.

By minimizing these non‐radiative energy losses, the molecular treatment significantly prolongs carrier diffusion lengths and directly contributes to the substantial enhancement in open‐circuit voltage (V _OC) [57, 68]. We further quantified the trap‐state density using space‐charge‐limited current (SCLC) analysis on electron‐only devices (Figure 5f). The dithiopyr modification lowers the density of electron defect states from 1.12 × 10¹⁶ cm⁻³ to 8.62 ×10¹⁵ cm⁻³, which corresponds to a tangible reduction in bulk trap density [69, 70]. By fitting the Nyquist plots with a standard Rs‐(Rrec//CPE) equivalent circuit model, the series resistance (Rs) remained highly comparable (∼20 Ω), whereas the target device displays a markedly higher recombination resistance (Rrec, increasing from 4154 to 6597 Ω) (Figure S25 and Table S5), confirming reduced leakage pathways [71]. Additionally, the built‐in potential(V_bi) increases from 0.88 to 1.00 V (Figure S26). This potential rise, driven by the optimized surface dipole layer, strengthens the internal electric field. This enhanced field facilitates rapid carrier extraction, resulting in an increased FF [58, 72].

Dark measurements (Figure S27) further corroborate the improved film quality [73]. The leakage current under reverse bias is suppressed by nearly an order of magnitude in the target device. This reduction is indicative of minimized shunt losses [57, 69]. These results confirm that dithiopyr acts as an effective interfacial buffer, where its fluorinated moieties and grain‐boundary anchoring capability work synergistically to eliminate pinholes and block leakage pathways [37, 38].

To assess the final device performance, we fabricated inverted p‐i‐n perovskite solar cells with a configuration of ITO/NiO_x/SAM/PVK/PEACl/PCBM/BCP/Ag (Figure S28). The current density‐voltage (J‐‐V) characteristics measured under standard AM 1.5G illumination (Figure 6a) reveal a pivotal enhancement in device efficiency. The champion device incorporating the dithiopyr‐modified absorber achieved a remarkable PCE of 26.90% (certified 26.14%) (Figure S29), significantly outperforming the control device (24.34%). Their performance statistics are detailed in Figure S30. The photocurrent obtained by integrating the external quantum efficiency (EQE) spectrum (Figure S31) reached 25.88 mA cm⁻².

FIGURE 6.

Device performance and stability. (a) 1.55 eV; (b) 1.68 eV; and (c) 1.25 eV. (d) J–V characteristic of the large‐area 30 cm × 30 cm device exhibiting a PCE of 21.28%. (e) SEM images of the films after thermal aging at 100°C for 100 h. (f) XRD patterns of the perovskite films after thermal aging at 100°C for 50 h; the PbI₂ peak intensity in the control exceeds the (100) main peak, whereas the PbI₂ peak in the dithiopyr‐treated remains significantly lower than the (100) peak. (g) Long‐term stability of the unencapsulated target device in ambient air, which retains 93.1% of its initial PCE after more than 3600 h. (h) Operational stability under continuous MPPT, showing that the target device maintains 95% of its initial PCE after 1100 h of continuous illumination.

Subsequent steady‐state power output measurements at the maximum power point (MPP, Figure S32) yielded a stabilized efficiency of 26.40%, underscoring a robust and consistent power output. Furthermore, the modified devices exhibited reduced hysteresis with a mathematically quantified Hysteresis Index (HI) of 0.78% (Figure S33) [74, 75, 76]. To demonstrate the universality of the dithiopyr strategy, experiments were carried out across various bandgap compositions and scaled‐up devices. In wide‐bandgap (1.68 eV) cells (Figure 6b), the efficiency reached 23.11%, with simultaneous improvements in current density, open‐circuit voltage, and fill factor. In narrow‐bandgap (1.25 eV) devices (Figure 6c), the efficiency increased from 19.92% to 22.02%. Most notably, a large area 30 cm × 30 cm (aperture area = 610 cm²) module reached a high efficiency of 21.28% (Figure 6d). The module architecture gives a geometric fill factor (GFF) of 72.72%, with P1, P2, and P3 laser scribing widths of 30 µm, 100 µm, and 30 µm, respectively, resulting in an aperture area of 610 cm² and an aperture ratio of 89.78% (Table S6). To evaluate reproducibility, we fabricated five independent modules with efficiencies ranging from 20.70% to 21.26% (Figure S34). The module maintained a nearly constant PCE during 300 s of operation, yielding a stabilized efficiency of 20.77%. After 1200 h of continuous operation, the unencapsulated module retained approximately 95% of its initial efficiency (Figure S35).

Recognizing that long‐term stability represents the ultimate criterion for the viability of interface engineering, we subjected our devices to environmental stress tests. We first challenged the structural integrity of the perovskite films under severe thermal stress (100°C), a condition known to accelerate lattice decomposition and volatile species loss. As evidenced by the SEM images (Figure 6e) and photographs (Figure S36), the control film underwent rapid degradation after 100 h, manifesting as severe pinhole formation, grain boundary dissolution, and visible bleaching to the non‐photoactive yellow phase. In stark contrast, the dithiopyr‐modified target film retained a pristine, compact morphology and its characteristic black color. This thermal resilience was crystallographically confirmed by XRD (Figure 6f). While the control film showed a rapid decay in peak intensity, the target film preserved its structural coherence. This robust thermal stability validates our hypothesis that the strong Pb─O coordinate bonds act as molecular anchors, effectively locking the surface lattice and suppressing the thermal release of volatile organic cations. In addition, we analyzed the storage stability of the precursor solution as a function of aging time. The comparative stability data have been added to Figures S37–S41 and Table S7). These results indicate that dithiopyr stabilizes the precursor state, suppresses aging‐induced film degradation, and extends the shelf‐life of the precursor solution [77, 78]. Moreover, the stability of perovskite films upon water exposure reveals that the control sample exhibited a quicker phase transition to the yellow phase within the identical time frame (Figure S42), further corroborating the effectiveness of the aforementioned fluorine‐containing moieties serving as a hydrophobic protective barrier [37, 38, 79]. The contact angle test showed that the target group was 50.5° and the control group was 44.9°. Within 2 min of dripping, the water droplets in the control group could spread evenly, while the infiltration of the target group was poor (Figure S43). The results showed that dithiopyr doping could effectively isolate water from the environment. Environmental stability was tested in ambient air (Figure 6g), where the unencapsulated target device retained 93.1% of its initial performance after 3600 h. Operational longevity was further assessed via maximum power point tracking (MPPT) measurements (Figure 6h). While the control device degraded significantly (T₇₅ ≈ 800 h) due to accelerated ion migration, the dithiopyr‐based device demonstrated superior photostability, maintaining over 95% of its initial PCE after 1100 h. This confirms that dithiopyr effectively suppresses the ion migration channels that typically plague perovskite films under illumination.

3. Conclusion

In summary, we have developed a robust crystallization modulation strategy by employing a dual‐site‐binding ligand, dithiopyr, to simultaneously regulate nucleation dynamics, control crystal orientation, suppress defects, and maintain efficient charge transport of perovskite films. By leveraging the strong coordination between thioester atoms and undercoordinated Pb²⁺ sites, dithiopyr successfully modulates the crystallization kinetics, facilitating the growth of high‐quality, preferentially (100)‐oriented perovskite films. Furthermore, the fluorinated hydrophobic periphery of the dithiopyr molecule establishes a resilient moisture barrier, while its anchoring effect effectively suppresses ion migration and thermal degradation. This structural refinement, combined with effective defect passivation and lattice strain relaxation, resulted in a champion PCE of 26.90% (certified 26.14%) and demonstrated universal efficacy across various bandgap compositions and large‐area 30 cm × 30 cm modules. Consequently, the unencapsulated devices retained 93.1% of their initial performance after 3600 h of ambient air storage and exhibited exceptional operational stability by maintaining over 95% PCE after 1100 h of continuous MPPT. This work provides a versatile and scalable approach for precisely controlling crystal texture and interfacial chemistry, offering a highly promising pathway toward the commercialization of high‐performance, durable perovskite photovoltaics.

Funding

This work was supported by the National Natural Science Foundation of China (52203359), the Fund of Basic Research Program of Jiangsu (BK20253025), the Research Fund of State Key Laboratory of Mechanics and Control for Aerospace Structures (Nanjing University of Aeronautics and astronautics) (Grant No. MCAS‐I‐0525K01), the Fund of Prospective Layout of Scientific Research for NUAA (Nanjing University of Aeronautics and Astronautics), the Open Project Program of Key Laboratory of Energy Low‐Carbon Conversion and Storage, Jiangxi Provincial Department of Education, Jiangxi University of Water Resources and Electric Power (ELCC‐25‐08) and the Natural Science foundation of Jiangxi Province (20252BAC200293).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adma74033‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.