Spatial Atomic Arrangement of Cyclohexyl‐Based Ligands for Enhanced Interface Passivation in 2D/3D Perovskite Solar Cells

Abstract

Metal halide perovskite solar cells (PSC) have emerged as a promising candidate for next‐generation photovoltaic technologies, achieving remarkable power conversion efficiencies (PCE) in polycrystalline thin‐films. Nonetheless, their PCE and long‐term stability are often limited by a high density of defects at interfaces and grain boundaries. One effective strategy to mitigate these issues is forming 2D/3D heterojunction structure by introducing a surface‐passivating interfacial layer of quasi‐2D Ruddlesden–Popper perovskite (RPP) with various functional ligand molecules. In this study, hydroxyl‐functionalized cyclohexyl ammonium hydrochloride (ACHACl) is employed as the ligand molecule, which features dual functional groups. A comparative analysis reveals that cis‐configurations of the ligands exhibit enhanced adsorption on the [PbI₆]⁴⁻ octahedra in lead halide perovskites, enabling more effective passivation of defects including metallic Pb⁰ on 3D perovskite thin‐films. Moreover, the 2D/3D structure incorporating the RPP improves hole extraction to the hole transport layer (HTL) by inducing an upward shift in the valence band edge. Consequently, the 2D/3D heterojunction PSCs achieve a PCE of 23% under 1‐Sun illumination, along with significantly improved long‐term stability, presenting the critical role of tailoring the atomic structure of the RPP ligand molecules to boost both the efficiency and long‐term stability of PSCs.

A comparative analysis of cis‐ and trans‐configured hydroxyl‐functionalized cyclohexylammonium chloride (ACHACl) ligands reveals distinct adsorption behaviors on [PbI₆]⁴⁻ octahedra during 2D/3D perovskite heterojunction formation. The cis isomer exhibits stronger surface interaction, leading to more effective defect passivation, including Pb⁰ suppression, and enabling perovskite solar cells to achieve a power conversion efficiency exceeding 23% under 1‐sun illumination.

1. Introduction

Metal halide perovskite solar cells (PSC) have emerged as a leading candidate for next‐generation photovoltaic technology, demonstrating power conversion efficiencies (PCEs) from polycrystalline thin‐film as high as 26.7% in single‐junction devices and 34.6% in perovskite‐silicon tandem devices.^{[ 1} , ² , ³ , ⁴ , ^{5 ]} Despite substantial progress in PCE, PSCs still face challenges in both efficiency and stability, primarily experiencing long‐term performance degradation due to their intrinsic structural instability and susceptibility to environmental factors including light, heat, and moisture.^{[ 6} , ⁷ , ⁸ , ^{9 ]} One of major contributors to the degradation of PSCs is the presence of metallic Pb⁰ defects, acting as deep trap states, primarily located at the grain boundaries and surfaces of polycrystalline perovskite thin‐films.^{[ 10} , ¹¹ , ^{12 ]} These metallic Pb⁰ defects originate from halide vacancies (e.g., V _Br and V _I), which facilitate the reduction of undercoordinated Pb²⁺ cations.^{[ 12} , ^{13 ]} Thus, the emergence of halide vacancies promotes the generation of Pb⁰, which negatively impacts the long‐term stability of PSCs.^{[ 10} , ¹¹ , ^{12 ]} To mitigate this degradation issue, various surface passivation strategies have been developed, including the use of conjugated polymers,^{[ 14} , ¹⁵ , ^{16 ]} self‐assembled monolayers,^{[ 17} , ¹⁸ , ¹⁹ , ^{20 ]} and Ruddlesden–Popper perovskites (RPPs).^{[ 21} , ²² , ²³ , ²⁴ , ^{25 ]}

Among these, quasi‐2D RPPs–represented by the formula of (RNH₃)₂A _{n   −  1}M _n X_{3 n  + 1}–stand out for their exceptional ability to passivate interfacial defects and suppress non‐radiative recombination by forming a 2D/3D perovskite bilayer configuration.^{[ 21} , ²² , ²³ , ²⁴ , ^{25 ]} In the RPP formula, RNH₃ represents a cationic ligand, A is a monovalent cation, M denotes a divalent metal cation, X is a halide anion, and n indicates the number of [MX₆]⁴⁻ octahedral layers sandwiched between bulky organic spacer molecules.^{[ 22 ]} The crystal orientation and formation energy of RPPs on perovskites can be engineered by changing the type and position of functional groups on the ligand (e.g., F, Cl, and OH; ortho, para, and meta), which affect interfacial interaction with adjacent layers.^{[ 26} , ²⁷ , ^{28 ]} This versatility of RPPs enabled by ligand engineering not only enables extensive tuning of their opto‐electronic characteristics, but also facilitates efficient defect passivation on perovskites, ultimately leading to enhanced performance and stability of PSCs.

RPPs are formed through hydrogen and ionic interactions between the ammonium (NH₃ ⁺) groups in the ligands and the iodide ions of corner‐sharing [MX₆]⁴⁻ octahedral within the perovskite layer.^{[ 29} , ^{30 ]} Importantly, the strength of the N─H bonds plays a crucial role in maintaining these interactions, as stronger N─H bond can suppress the release of H⁺ ions and the generation of free halide ions.^{[ 30} , ^{31 ]} These mobile halide species can diffuse into adjacent charge‐transport layers or electrodes, where they undergo irreversible chemical reactions, ultimately degrading the electrical properties and overall device performance.^{[ 31} , ³² , ^{33 ]} Conventionally, aromatic ring‐based ligands, such as phenethylammonium iodide and benzylammonium iodide, have been widely employed in RPP formation.^{[ 34} , ³⁵ , ³⁶ , ^{37 ]} However, recent studies reported that cyclohexyl‐based ligands, which lack π‐electron delocalization, exhibit stronger interactions between –NH₃ ⁺ groups and [MX₆]⁴⁻ octahedra.^{[ 21} , ³¹ , ^{38 ]} This enhanced interaction is attributed to the localized electron density and higher basicity of cyclohexylamine, which strengthens the N─H bond.^{[ 21} , ^{31 ]} As a result, cyclohexyl‐based ligands provide improved stability compared to conventional aromatic ligands.^{[ 31 ]}

Notably, even with the same ligand, variations in the tilt angle of functional groups can significantly affect steric hindrance and molecular interactions of RPPs.^{[ 39} , ⁴⁰ , ^{41 ]} The cis configuration–where functional groups reside on the same side– exhibits higher molecular polarity and a stronger dipole moment due to the uneven distribution of electron density, yielding distinct surface interaction behaviors compared to the trans form.^{[ 39} , ^{40 ]} For instance, the enhanced passivation effects of cis‐CyDAI₂ on wide‐bandgap (E _g) perovskites have been reported: the aligned –NH³⁺ groups in the cis form increase the dipole moment, stabilize surface passivation, and reduce shifts in quasi‐Fermi level splitting.^{[ 39 ]} In addition, a trans‐pyridine‐functionalized form, featuring spatially separated pyridine groups, effectively coordinates with perovskite colloids, controls crystallization, aligns interface energy levels, and suppresses Sn²⁺ oxidation.^{[ 40 ]} Hence, we hypothesize that modifying the spatial atomic arrangement of cyclohexyl‐based ligands, the passivation capability of RPP can be further enhanced, thereby improving both the efficiency and stability of PSCs.

Herein, we investigate the effects of spatial atomic modifications in RPP ligands by focusing on the cis‐ and trans‐configurations of aminocyclohexanol hydrochloride (ACHACl). To enhance the functionality of the cyclohexyl ligands, we utilize the functional groups of ACHACl to achieve complementary defect‐passivation effects, wherein the –NH₃ ⁺ and hydroxyl (–OH) groups interact with halide ions and vacancies in the perovskite, effectively passivating the interfaces.^{[ 21} , ³¹ , ⁴² , ^{43 ]} Notably, the cis ‐configuration, with its asymmetrical functional group positioning and enhanced dipole moment, facilitates the formation of a highly crystalline cis‐ACHACl‐based RPP layer on the 3D perovskite surface. Furthermore, the RPP layer can facilitate selective hole extraction to the hole transport layer (HTL) by inducing an upward shift in the valence band. As a result, PSCs passivated with cis‐ACHACl achieve a PCE exceeding 23%, a relative increase of 5.8% over the non‐passivated devices, while trans‐ACHACl treatment yields a 3.2% improvement. Furthermore, both cis‐ and trans‐ACHACl treated devices exhibit improved stability compared to the control device, demonstrating the critical role of molecular configuration of RPP ligands in 2D/3D PSCs.

2. Results and Discussion

To elucidate the impact of spatial atomic arrangement of RPP ligands on the passivation of 2D/3D heterojunction interface in PSCs, we investigated the cis‐ and trans‐configurations of 4‐amino‐cyclohexanol (ACHA). Figure 1a displays the 3D electrostatic potential (ESP) maps of the protonated forms of cis‐ACHACl and trans‐ACHACl, calculated using density functional models (ωB97X‐D/6‐31G*) in a vacuum, revealing distinct spatial configurations despite identical chemical compositions. In the cis configuration, the –OH and –NH₃ ⁺ –functional groups are oriented in the same direction, whereas in the trans configuration, they extend diagonally in opposite directions. This geometric difference profoundly affects their electrostatic potential distributions. Notably, the cis isomer features a larger dipole moment (2.54 D) than the trans isomer (1.53 D), indicating electron localization around the nitrogen and oxygen centers in the cis configuration.

Figure 1.

a) ESP maps illustrating electron‐rich (red) and electron‐deficient (blue) regions in 2D perovskites: cis‐4‐ACHA (top) and trans‐4‐ACHA (bottom). ESP calculations were performed using density functional model (ωB97X‐D/6‐31G*) in a vacuum. PL Spectra of RPP films annealed at different temperatures (without anneal to 130 °C). b) cis‐ACHA₂Cl₂PbI₂ films and c) trans‐ACHA₂Cl₂PbI₂ films. d) GIWAXS patterns of the FA film confirm the formation of α‐phase FAPbI₃. e) GIWAXS patterns of the cis‐ACHACl/FA film indicate the formation of α‐phase FAPbI₃ and the (002) orientation of cis‐ACHA₂Cl₂PbI₂. f) GIWAXS patterns of the trans‐ACHACl/FA film indicate the formation of α‐phase FAPbI₃ and the (002) orientation of trans‐ACHA₂Cl₂PbI₂.

Using the cis‐ and trans‐ACHACl ligands together with PbI₂, we synthesized 2‐µm‐thick cis‐ACHA₂Cl₂PbI₂ and trans‐ACHA₂Cl₂PbI₂ films–both adopting an n = 1 RPP phase–under various annealing temperatures ranging from room temperature to 130 °C. After annealing, distinct changes were observed in their photoluminescence (PL) spectra, as shown in Figure 1b,c. Specifically, the strongest emission peaks appeared at 506 and 536 nm, for cis‐ACHA₂Cl₂PbI₂ and trans‐ACHA₂Cl₂PbI₂ films, respectively. Increasing the annealing temperature to 85 °C resulted in enhanced PL peak intensity and narrower full‐width‐at‐half‐maximum (FWHM) values for both films. In addition, X‐ray diffraction (XRD) analysis of the cis‐ACHA₂Cl₂PbI₂ thin‐film annealed at 85 °C shows the most distinct and sharp (002) diffraction peak, as shown in Figure S1 (Supporting Information). However, beyond 85 °C, the intensities of both PL and XRD peaks diminished. Notably, the trans‐ACHA₂Cl₂PbI₂ films showed broader and less distinctive peaks compared to the cis‐ACHA₂Cl₂PbI₂, indicating the latter's superior crystallinity.

Next, we prepared the 2D/3D perovskite heterostructure by spin‐coating isopropanol solutions of cis‐ACHACl and trans‐ACHACl onto 3D FAPbI₃ perovskite thin‐films, followed by annealing at 85 °C. As shown in Figure 1d–f, grazing incidence wide‐angle X‐ray scattering (GIWAXS) measurement was carried out to examine the crystallinity of 2D/3D perovskite films, including pristine 3D perovskite (FA), cis‐ACHACl‐treated perovskite (cis‐ACHACl/FA), and trans‐ACHACl‐treated perovskite (trans‐ACHACl/FA). In all samples, the α‐FAPbI₃ phase exhibits a pronounced scattering profile along the (001) orientation at q _z = 1.00 Å⁻¹. The appearance of new peaks at q_z = 0.63 Å⁻¹ and q_z = 0.55 Å⁻¹ along the (002) orientation indicates the presence of the 2D perovskite in the cis‐ACHACl/FA and trans‐ACHACl/FA. Notably, the cis‐ACHACl/FA sample shows a relatively intense peak, while the trans‐ACHACl/FA sample's corresponding peak is barely detectable. These results suggest that the cis‐ACHACl‐based RPP layer forms more effectively on the 3D perovskite surface, achieving higher crystallinity compared to the trans‐ACHACl‐based RPP. To directly compare the reactivity of cis‐ACHACl and trans‐ACHACl for the formation of 2D perovskites, 2D/3D perovskite films were fabricated by isopropanol solutions containing both cis‐ACHACl and trans‐ACHACl at varying cis‐ACHACl/trans‐ACHACl ratios (2:1, 1:1, and 1:2). As shown in Figure S2 (Supporting Information), GIWAXS measurements reveal that the (002) orientation peak of cis‐ACHACl‐based 2D perovskites gradually decrease as the proportion of trans‐ACHACl increases. Upon converting q _z to 2θ, the (002) peaks appear at 8.91° and 7.74° for cis‐ACHACl/FA and trans‐ACHACl/FA, respectively, as shown in Figure S3 (Supporting Information). Notably, all sample treated with the mixed solutions show peaks at 8.91°, identical to the cis‐ACHACl/FA sample, indicating that cis‐ACHACl predominantly reacts with the 3D perovskite to form the 2D RPP layer on top.

To elucidate the interactions between RPP organic ligands and perovskite, Fourier‐transform infrared (FTIR) spectroscopy was performed on the RPP ligands and their complexes with PbI₂ in both thin‐film and powder forms. As shown in Figure S4 (Supporting Information), the O─H stretching range (3500–3200 cm⁻¹) exhibited broadening and weakening upon complexation with PbI₂. Specifically, for cis‐ACHACl, the peak shifted from 3267.1 to 3008.7 cm⁻¹, while for trans‐ACHACl, it shifted from 3251.7 to 3024.2 cm⁻¹. These pronounced red‐shifts result from the passivation of iodide vacancies in perovskite, which act as positively charged Lewis acid sites, by –OH groups functioning as Lewis bases.^{[ 44} , ^{45 ]} The lone pair of electrons on the oxygen atom in –OH forms a dative bond with the iodide vacancy, effectively neutralizing its positive charge.^{[ 46} , ^{47 ]} This interaction weakens the O─H bond, resulting in a reduction of its vibrational energy and a corresponding shift to lower wavenumbers.^{[ 44} , ^{45 ]} Additionally, the interaction between the –NH₃ ⁺ group and PbI₂ was investigated by analyzing shifts in the N─H bending vibrations within the range of 1650–1580 cm⁻¹. The formation of hydrogen bonds between –NH₃ ⁺ group and PbI₂ strengthened its interaction with iodide ions, leading to the elongation and weakening of the N─H bonds, which in turn reduced its vibrational energy and caused a shift to lower wavenumbers.^{[ 48} , ^{49 ]} Notably, cis‐ACHACl and trans‐ACHACl initially exhibited two distinct N─H bending peaks at 1635.5 and 1608.5 cm⁻¹ and 1627.8 and 1612.3 cm⁻¹, due to the structural asymmetry of their three N─H bonds. Upon hydrogen bonding with PbI₂, these peaks merged into a single red‐shifted peak at 1581.5 cm⁻¹ for cis‐ACHACl and 1589.2 cm⁻¹ for trans‐ACHACl. This merging occurred as the hydrogen bonds effectively restricted the otherwise orientationally disordered –NH₃ ⁺ groups, stabilizing their configuration within the perovskite structure.^{[ 50 ]} The pronounced O─H and N─H shifts observed for cis‐ACHACl indicate strong interactions with iodide vacancies and iodide ions.

Confocal PL measurements were carried out on FA, cis‐ACHACl/FA, and trans‐ACHACl/FA samples to capture the emission originating from the thin RPP layers at their surfaces. As shown in Figure S5 (Supporting Information), the spatial maps of confocal PL intensities within the 450–550 nm wavelength range (displayed in green) reveal a higher emission near 506 nm–attributed to the cis‐ACHA₂Cl₂PbI–in the cis‐ACHACl/FA sample compare to the trans‐ACHACl/FA sample. Moreover, the spatial PL maps exhibit a uniform distribution across the sample surfaces. To further investigate the spatial distribution of RPPs, atomic force microscopy (AFM) measurements were conducted on FA, cis‐ACHACl/FA, and trans‐ACHACl/FA samples, as shown in Figure S6 (Supporting Information). Both cis‐ACHACl/FA and trans‐ACHACl/FA samples feature additional surface structures, attributed to 2D perovskites, along the grain boundaries of the 3D perovskite–a feature not observed in the FA sample. The root‐mean‐square (RMS) surface roughness remains largely unchanged: 38.9 nm for FA, 39.5 nm for cis‐ACHACl/ FA, and 39.7 nm for trans‐ACHACl/FA. XRD patterns (Figure S7, Supporting Information) revealed minimal variation in the (001) orientation of α‐FAPbI₃ and a uniform d‐spacing of 6.3 Å, indicating that the introduction of RPP does not disrupt the overall crystal structure of the perovskite layer.

To investigate the surface interactions and elemental composition resulting from the integration of RPP with perovskite, X‐ray photoelectron spectroscopy (XPS) analysis was conducted. The Pb‐4f spectra of the untreated FA sample revealed distinct peaks at 138.3 and 143.2 eV, corresponding to Pb‐4f _7/2 and Pb‐4f _5/2, respectively, as shown in Figure 2a. Similarly, the I‐3d spectra of the FA sample exhibited peaks at 619.0 and 630.5 eV, assigned to I‐3d _3/2 and I‐3d _5/2, respectively, as depicted in Figure S8 (Supporting Information). To ensure accurate energy alignment, all binding energy values were calibrated relative to the C‐C peak of the C‐1s spectrum, set at 284.8 eV. Following the formation of the RPP layer on the perovskite surface, no significant shifts were observed in the Pb‐4f and I‐3d core‐level peaks. The FA film exhibited the formation of metallic Pb⁰, as evidenced by peaks at 141.5 and 136.6 eV in the Pb‐4f spectra, as commonly observed regardless of PbI₂ concentration in precursor solutions.^{[ 51 ]} By introducing ACHACl, the metallic Pb⁰ at the surface of 3D FAPbI₃ perovskite thin‐films was consumed to form ACHA₂Cl₂PbI₂, corresponding to the n = 1 phase RPP. As a result, the ratio of Pb⁰ to the total Pb content decreased from 7.4% in the FA film to 2.7% and 3.0% after treatment with cis‐ACHACl and trans‐ACHACl, respectively. In addition, the O‐1s spectra (Figure S8, Supporting Information) displayed distinct peaks at 532.0 and 533.0 eV, corresponding to –OH group and adsorbed H₂O, respectively.^{[ 52 ]} While the level of adsorbed H₂O remained relatively consistent due to the perovskite's exposure to ambient conditions, the –OH group signal was notably higher for cis‐ACHACl/FA compared to trans‐ACHACl/FA and significantly higher than that of FA, indicating the stronger formation affinity of cis‐ACHACl. Figure 2b shows [I]/[Pb] atomic ratio calculated from the XPS measurements. While the FA film shows the [I]/[Pb] ratio of 2.73, this value increases to 2.94 for cis‐ACHACl/FA and 2.87 for trans‐ACHACl/FA, indicating iodine‐rich surfaces following the RPP formation.^{[ 53} , ^{54 ]}

Figure 2.

XPS characterization of the Pb‐4f chemical state in a), and surface [I]/[Pb] ratios calculated from XPS measurements in b) for films with FA, cis‐ACHACl/FA, and trans‐ACHACl/FA. c) SCLC method was used to evaluate the trap density in hole‐only devices based on FA, cis‐ACHACl/FA, and trans‐ACHACl/FA films. d) UV–vis absorbance spectra were obtained for perovskite films, including FA, cis‐ACHACl/FA, and trans‐ACHACl/FA. The bandgap of each perovskite film was calculated using a Tauc plot. e) Steady‐state PL quenching spectra was acquired for FA, cis‐ACHACl/FA, and trans‐ACHACl/FA films. f) TRPL spectra for Glass/Perovskite/Spiro‐OMeTAD structures with perovskite films of FA, cis‐ACHACl/FA, and trans‐ACHACl/FA.

The impact of 2D passivation on interfacial defect passivation was evaluated by measuring the density of trap states (N _trap) in perovskite thin‐films using space‐charge‐limited current (SCLC) measurements, as shown in Figure 2c. The trap‐filling‐limited voltage (V _TFL) for hole‐only devices was extracted from the dark current density–voltage (J–V) curves, and N _trap was calculated using the following Equation (1):^{[ 55 ]}

$$N_{trap}=2·{\epsilon }_r·{\epsilon }_0·V_{TFL}/\left(e·L^2\right)$$ (1)

where ɛ _r is the dielectric constant of the perovskite (46.9),^{[ 56 ]} ɛ ₀ is the vacuum permittivity (8.854 × 10⁻¹⁴ F cm⁻¹), e is the electric charge (1.602 × 10⁻¹⁹ C), and L is the thickness of the perovskite layer (600 nm). The V _TFL values were determined to be 0.52, 0.55, and 0.59 V with corresponding N _trap values of 8.49 × 10¹⁵, 7.48 × 10¹⁵, and 7.92 × 10¹⁵ cm⁻³ for devices with FA, cis‐ACHACl/FA, and trans‐ACHACl/FA, respectively. The cis‐ACHACl/FA device exhibits the most effective suppression of hole trap states at the interface, which can be attributed to the defect passivation effect of cis‐ACHACl on the perovskite, as evidenced by the FTIR and XPS analysis. This trap density reduction is crucial for enhancing charge transport and minimizing recombination losses, ultimately contributing to improved photovoltaic performance, as will be discussed in the PSC device performance section.

The optical absorption properties of FA, cis‐ACHACl/FA, and trans‐ACHACl/FA films were evaluated using a UV‐Vis spectrophotometer. As shown in Figure 2d, the absorbance spectra for all samples exhibit generally similar features, with a slight increase in absorbance below 600 nm observed in the 2D/3D films. Additionally, a Tauc plot derived from the absorbance spectra indicates E _g of 1.55 eV for FA, cis‐ACHACl/FA, and trans‐ACHACl/FA films, respectively. Overall, the 2D/3D perovskite samples exhibit the same E _g as the 3D perovskite sample, even after the RPP formation. In contrast, as shown in Figure 2e, the steady‐state PL spectra of the samples exhibit slight blue‐shifted PL peaks. The 3D perovskite exhibits a PL peak at 793 nm, which blue‐shifts slightly to 791 nm upon forming the 2D/3D perovskite heterostructure. Additionally, the FWHM of the PL peaks narrows to 47 and 48 nm for cis‐ACHACl/FA, and trans‐ACHACl/FA films, respectively, indicating a lower density of sub‐bandgap states.^{[ 57 ]} Furthermore, steady‐state PL measurements without HTL were conducted to isolate the intrinsic optoelectronic properties of the perovskite layer and examine the direct impact of RPP formation. The PL spectra without the HTL reveal a notable increase in PL intensity after RPP formation, indicating enhanced radiative recombination due to the suppression of non‐radiative recombination pathways as shown in Figure S9 (Supporting Information).

The carrier dynamics of the perovskite films were investigated via time‐resolved PL (TRPL) measurements, as shown in Figure 2f. For these measurements, spiro‐OMeTAD thin‐film was as the HTL atop the perovskite film, and back‐side excitation was applied with a 405 nm laser through the glass substrate. As shown in Table S1 (Supporting Information), the TRPL decay curves were fitted to a biexponential model to evaluate fast (τ ₁) and slow (τ ₂) decay components, which are attributed to trap‐assisted charge recombination and carrier transport to HTL, as well as radiative recombination in the bulk, respectively.^{[ 58} , ⁵⁹ , ^{60 ]} For the untreated FA sample, τ ₁ and τ ₂ are 397.4 and 2014 ns. Following post‐treatment with cis‐ACHACl/FA, these values are reduced to 318.3 and 1119 ns, respectively, compared to trans‐ACHACl/FA, which exhibits τ ₁ = 357.0 ns and τ ₂ = 1598 ns. The reduction in τ ₁ indicates a substantial decrease in trap density at the interface, effectively suppressing non‐radiative recombination and facilitating more efficient charge transfer to the HTL.^{[ 61} , ^{62 ]} Similarly, the decrease in τ ₂ suggests that radiative recombination in the bulk occurs more rapidly, likely as a secondary effect of enhanced charge extraction at the interface induced by the surface treatment. To further investigate the intrinsic recombination dynamics of the perovskite and the 2D passivation layer, we conducted TRPL measurements on perovskite/glass structures without HTL. The untreated FA sample exhibited of τ ₁ of 158.5 ns and τ ₂ of 1392.9 ns, which increased to 301.5 ns and 2017.1 ns, respectively, after cis‐ACHACl/FA passivation, confirming effective defect passivation and reduced non‐radiative recombination.^{[ 63} , ^{64 ]} In contrast, trans‐ACHACl/FA showed a more moderate increase, with τ ₁ at 175.6 ns and τ ₂ at 1450.9 ns, as shown in Figure S10 and Table S2 (Supporting Information). These findings demonstrate that cis‐ACHACl/FA not only minimizes non‐radiative interfacial recombination but also facilitates faster photogenerated charge transport to the HTL, resulting in enhanced carrier extraction efficiency.

To investigate the origin of the enhanced carrier transport properties, ultraviolet photoelectron spectroscopy (UPS) analysis was carried out. As shown in Figure 3a, the valence band offset (VB _offset) and the cut‐off binding energy of the secondary electron edge (E _cutoff) were measured to construct the electronic band alignment diagram within the PSCs. In the 2D/3D heterostructure, trans‐ACHACl/FA exhibits a valence band maximum (E _VBM) of −6.09 ± 0.10 eV for 2D perovskite slightly higher than that of FA (−6.14 ± 0.10 eV). Meanwhile, the 2D perovskite in cis‐ACHACl/FA shows an upshifted E _VBM at −5.99 ± 0.10 eV. The E _g of cis‐ACHA₂Cl₂PbI₂ and trans‐ACHA₂Cl₂PbI₂ thin‐films were determined to be 3.05 ± 0.10 and 2.90 ± 0.10 eV, respectively, as shown in Figure S11 (Supporting Information). Figure 3b shows a schematic diagram of band alignment among FA, cis‐ACHACl, trans‐ACHACl, Spiro‐OMeTAD, and Au, indicating a more favorable energy alignment that enhances hole transport to the HTL.^{[ 65} , ^{66 ]}

Figure 3.

a) Secondary electron edge and valence band offset for FA, cis‐ACHACl/FA, and trans‐ACHACl/FA films was achieved through UPS spectra. b) Schematic illustration of the energy level distribution for FA, cis‐ACHACl/FA, and trans‐ACHACl/FA films in PSCs. Note that the energy levels of Spiro‐OMeTAD and Au are referenced from the literature.^{[ 77 ]} KPFM images of perovskite films are presented, depicting the average surface potential as follows. c) FA: −10.36 mV, d) cis‐ACHACl/FA: 122.58 mV, e) trans‐ACHACl/FA: 102.17 mV.

In addition, Kelvin probe force microscopy (KPFM) further examined energy levels and their spatial uniformity, as shown in Figure 3c–e. The averaged surface potentials were measured to be −10.3, 122.5, and 102.1 mV for FA, cis‐ACHACl/FA, and trans‐ACHACl/FA, respectively, consistent with the reduced work functions and upward in Fermi level shifts observed in the UPS data, as summarized in Table S3 (Supporting Information). Furthermore, a comparison of the potential variation along the dotted line scans shows that cis‐ACHACl/FA exhibits a more homogeneous potential profile, with a standard deviation of potential (σ _V = 32.4 mV), compared to FA (σ _V = 39.0 mV) and trans‐ACHACl/FA (σ _V = 41.8 mV).

By utilizing the 2D/3D perovskite heterostructure, we fabricated planar n‐i‐p type PSCs. Figure 4a shows a cross‐sectional scanning electron microscopy (SEM) image of the PSC structure, comprising Au / spiro‐OMeTAD / RPP / FAPbI₃ / SnO₂ / fluorine‐doped tin oxide (FTO). As depicted in Figure 4b and detailed in Table S4 (Supporting Information), J–V measurements under 1‐Sun illumination (AM1.5G, 100 mW cm⁻²) showed that the best‐performing PSCs with an antireflection coating achieved PCEs of 22.16%, 23.46%, and 22.88% for FA, cis‐ACHACl/FA, and trans‐ACHACl/FA, respectively. One of the best‐performing PSC with cis‐ACHACl/FA was independently certified by the Korea Institute of Energy Research (KIER), achieving an average forward and reverse J–V scan efficiency of 22.77% under steady‐state 1‐Sun illumination condition, as shown in Figure S12 (Supporting Information). Additionally, the external quantum efficiency (EQE) spectrum for the certified PSC was measured, as shown in Figure S13 (Supporting Information).

Figure 4.

a) Cross‐sectional SEM image of n‐i‐p structured PSC treated with cis‐ACHACl. b) J–V curves of the best‐performing PSCs utilizing FA, cis‐ACHACl/FA, and trans‐ACHACl/FA films with anti‐reflection coating under 1 sun illumination. c–e) V _OC, FF, and PCE distribution of FTO‐based PSCs incorporating FA, cis‐ACHACl/FA, and trans‐ACHACl/FA films. Error bars represents the standard deviation calculated from 24 devices. f) Nyquist plots illustrating the EIS characteristics of PSCs incorporating FA, cis‐ACHACl/FA, and trans‐ACHACl/FA films. g) Ideality factor was determined by analyzing the relationship between V _OC and kT/q ln(J _SC), with light intensity ranging from 10 to 100 mW cm⁻², for PSCs incorporating FA, cis‐ACHACl/FA, and trans‐ACHACl/FA films. h) Shelf stability of FA‐, cis‐ACHACl/FA‐, and trans‐ACHACl/FA‐based PSCs was monitored over 1056 h under storage conditions of 25% relative humidity and 25 °C. i) MPPT was measured for unencapsulated PSCs fabricated with FA, cis‐ACHACl/FA, and trans‐ACHACl/FA over 150 h in a dry room (3–5% relative humidity, 25 °C).

Statistical data of 24 devices per condition is presented in Figure 4c–e and Figure S14 (Supporting Information). In addition, averaged results from 15 PSCs fabricated on indium‐tin‐oxide (ITO) coated glass substrates, further confirm improved PCEs in PSCs treated with cis‐ACHACl‐driven by increases in open‐circuit voltage (V _OC) and fill factor (FF), as shown in Figure S15 (Supporting Information) and Table S5 (Supporting Information). These improvements in and V _OC and FF are indicative of a reduced density of non‐radiative recombination centers and more efficient charge transport, respectively. The EQE spectra in the 300–900 nm wavelength range for PSCs fabricated on ITO substrates, are presented in Figure S16 (Supporting Information). The integrated short‐circuit current density (J _SC) calculated from the EQE spectra were 23.04, 23.13, and 23.18 mA cm⁻² for FA, cis‐ACHACl/FA, and trans‐ACHACl/FA, respectively–deviating by 1.1%, 1.3%, and 1.3% from the measured J _SC. Additionally, we measured the performances of PSCs with varying ratios of cis‐ACHACl and trans‐ACHACl for 2D/3D perovskite heterostructure formation, as summarized in Table S6 (Supporting Information). An increase in PCE was observed as the proportion of cis‐ACHACl increases, driven by improvements in V _OC and FF values. These results suggest that cis‐ACHACl enables more efficient defect passivation and charge extraction by facilitating the RPP layer formation compared to trans‐ACHACl.^{[ 67} , ⁶⁸ , ^{69 ]} Moreover, to verify the universality of our passivation strategy, we applied it to p‐i‐n structured PSCs with a Cu/BCP/C₆₀/2D perovskite/Perovskite/PTAA/ITO architecture and compared their performance, as shown in Figure S17 (Supporting Information). The best‐performing devices achieved PCEs of 18.71% (cis‐ACHACl/FA) and 18.18% (trans‐ACHACl/FA), both surpassing the FA‐only device (17.91%). Statistical analysis confirmed that cis‐ACHACl/FA enhances V _OC and FF, indicating effective defect passivation and improved charge extraction. In contrast, trans‐ACHACl/FA increases V _OC but slightly reduces FF, likely due to incomplete 2D interfacial layer formation. These results demonstrate that our passivation strategy is effective in both n‐i‐p and p‐i‐n architectures, confirming its broad applicability.

Electrochemical impedance spectroscopy (EIS) was conducted to investigate the carrier dynamics of PSCs, by fitting parameters derived from the equivalent circuit model, including series resistance (R _S), charge transfer resistance (R _CT), and recombination resistance (R _Rec). As shown in Figure 4f and summarized in Table S7 (Supporting Information), PSCs with cis‐ACHACl/FA and trans‐ACHACl/FA exhibit lower R _CT of 5.16 × 10³ and 6.14 × 10³ Ω cm², respectively, relative to 6.68 × 10³ Ω cm² for FA. The R _Rec for PSCs with cis‐ACHACl/FA and trans‐ACHACl/FA increase to 6.43 × 10⁴ and 5.60 × 10⁴ Ω cm², respectively, compared to 4.91 × 10⁴ Ω cm² for FA. The reduced R _CT and increased R _Rec indicate that cis‐ACHACl/FA not only facilitate enhanced interfacial carrier transfer, but also significant suppression of non‐radiative recombination, demonstrating the effectiveness of interface defect passivation.^{[ 70} , ^{71 ]} Furthermore, Mott–Schottky analysis, by examining the capacitance‐voltage (C⁻²–V) characteristics, enables the determination of the built‐in potential (V _bi), a key parameter for understanding interfacial energetics at the transport layer/perovskite junction.^{[ 72} , ^{73 ]} The results indicate that cis‐ACHACl passivation increases V _bi from 0.57 (FA) to 0.62 V, thereby strengthening the built‐in electric field and enhancing charge transport, whereas trans‐ACHACl‐treated devices exhibit a V _bi of 0.60 V, as shown in Figure S18 (Supporting Information).^{[ 72} , ^{73 ]} The trap‐assisted recombination in PSCs was further analyzed via illumination‐intensity‐dependent V _OC measurements, as shown in Figure 4g. The diode ideality factor (n) was determined using the following Equation (2):

$$V_{\mathrm{oc}}=\left(n·k·T/q\right)·\mathrm{In}\left(J_{\mathrm{sc}}/J_{\mathit{0}}+\mathit{1}\right)$$ (2)

where k is the Boltzmann constant, T is the absolute temperature, q is the elementary charge, and J ₀ is the saturated current density in the dark.^{[ 74 ]} The n for FA PSC was measured to be 1.92, whereas the cis‐ACHACl/FA and trans‐ACHACl/FA PSCs exhibited lower n values of 1.51 and 1.62, respectively, indicating reduced trap‐assisted recombination near junction.^{[ 75 ]}

Furthermore, the stability of the PSCs was examined to investigate the effect of the 2D perovskite layer. As shown in Figure 4h, PSCs stored over 1000 h under ambient air conditions (30‐40% relative humidity, 25 °C) were monitored for performance degradation, retaining 83.1%, 96.8%, and 90.8% of their initial PCEs–18.10%, 20.72%, and 20.23%, respectively–for FA, cis‐ACHACl/FA, and trans‐ACHACl/FA. To evaluate thermal stability, devices with Poly[bis(4‐phenyl)(2,4,6‐trimethylphenyl)amine] (PTAA) as the HTL were aged at 60 °C for 336 h under ambient conditions. The cis‐ACHACl/FA, trans‐ACHACl/FA, and FA‐treated devices retained 64%, 52%, and 34% of their initial PCE, respectively, highlighting the superior thermal stability of cis‐ACHACl/FA‐based devices, as shown in Figure S19 (Supporting Information). In addition, continuous maximum power point tracking (MPPT) measurements were carried out on PSCs without encapsulation under 3–5% relative humidity at 25 °C for 150 h. As shown in Figure 4i, the PSCs yielded PCE retention of rates of 44%, 70%, and 61% for FA, cis‐ACHACl/FA, and trans‐ACHACl/FA, respectively. This enhanced stability is attributed to the suppression of metallic Pb⁰ defects and the stronger binding of halide species facilitated by the –OH and –NH₃ ⁺ groups of cis‐ACHACl.^{[ 31} , ^{76 ]}

3. Conclusion

In summary, we have successfully demonstrated that controlling the spatial atomic arrangement of RPP ligand molecules can enhance both the performance and stability of PSCs. By utilizing cyclohexyl‐based ligands in cis‐and trans‐configurations, we compared their unique capabilities to passivate surface defects within 2D/3D heterojunction perovskites. The cis configuration, with its stronger molecular polarity and localized electron density, more effectively suppresses surface defects compared to the trans configuration. This advantage is primarily attributed to reducing metallic Pb⁰ defects, thereby lowering the density of deep‐trap states. Furthermore, the RPP layer promotes hole extraction to the HTL by inducing an upward shift in the valence band edge. Consequently, the cis‐ACHACl‐treated PSCs demonstrate a PCE of 23% under AM 1.5G illumination as well as improved operational stability. This work highlights the crucial role of molecular configuration in RPP ligands for achieving efficient and stable 2D/3D PSCs.

4. Experimental Section

Materials

ITO (8 Ω sq.⁻¹) and FTO (7 Ω sq.⁻¹) substrates were purchased from Asahi. Formamidinium iodide (>99.99%), methylammonium bromide (>99.99%), and methylammonium chloride (>99.99%) were acquired from Greatcell Solar Materials. Lead iodide (99.999%), DI water, and a 15% colloidal SnO₂ solution in H₂O were obtained from Alfa Aesar. Lead bromide (>98.0%) and 4‐Isopropyl‐4′‐methyldiphenyliodonium tetrakis(pentafluorophenyl)borate (>98.0%) were sourced from Tokyo Chemical Industry Co. Spiro‐OMeTAD and Co(III)‐TFSI were purchased from Luminescence Technology. Poly[(9,9‐bis(30‐((N,N‐dimethyl)‐N‐ethylammonium)‐propyl)‐2,7‐fluorene)‐alt‐2,7‐(9,9‐dioctylfluorene)] dibromide (PFN‐P2) was purchased from 1‐Material. PTAA was obtained from (10000 Da, MS Solution). cis‐4‐Aminocyclohexanol hydrochloride (97%), trans‐4‐Aminocyclohexanol hydrochloride (97%), N,N‐Dimethylformamide (99.99%), dimethyl sulfoxide (99.9%), chlorobenzene (99.8%), acetonitrile (99.8%), 2‐propanol (99.5%), N‐Methyl‐2‐Pyrrolidinone (99.5%), 4‐tert‐butylpyridine (99.9%), and Li‐TFSI (99.99%) were obtained from Sigma‐Aldrich. Diethyl ether (99.5%) was purchased from Samchun.

Device Fabrication

Pre‐patterned FTO‐coated glass substrate as cleaned by ultrasonication using in 70 °C DI water with 2% (v/v) Hellmanex, acetone, isopropanol alcohol, and de‐ionized water for 15 min, sequentially. Pre‐patterned ITO‐coated glass substrate underwent a cleaning process through ultrasonication in acetone, isopropanol alcohol, and de‐ionized water for 15 min each, sequentially. Subsequently, the substrates were dried using a nitrogen gun and an oven at 110 °C for 1 h.

n‐i‐p structured PSC

A 20‐nm‐thick SnO₂ layer was prepared by spin‐coating a SnO₂ colloid solution mixed with deionized water at a 1:4 (v/v) ratio onto the substrate. The spin‐coating was performed at 4000 rpm for 30 s, followed by annealing at 170 °C for 1 h. The perovskite precursor solution was meticulously prepared by mixing the following concentrations: 1.46 m PbI₂, 1.4 m FAI, 0.5 m MACl, 0.012 m MABr, and 0.012 m PbBr₂. The solvents used were dimethylformamide and dimethyl sulfoxide at a volumetric ratio of 8:1. The solution was stirred continuously for 12 h within a controlled environment under a nitrogen atmosphere inside a glove box. After filtering the solution through a 0.2 µm PTFE filter, 25 µL of the solution was dispensed onto the substrate and spin‐coated at 4000 rpm for 20 s with a ramp rate of 2000 rpm s⁻¹. During the spin‐coating, 500 µL of diethyl ether was dropped at the center of the substrate 8 s after the initiation of the spinning program. The substrates were then transferred to a hot plate and annealed at 130 °C for a duration of 60 min. For the formation of the 2D perovskite layer on 3D perovskite, a solution containing 5 mm of cis‐4‐Aminocyclohexanol hydrochloride or trans‐4‐Aminocyclohexanol hydrochloride in isopropyl alcohol was spin‐coated on the substrate at 4000 rpm for 20 s, followed by annealing at 85 °C for 3 min. Next, 20 µL of spiro‐OMeTAD solution, containing 36 mg of spiro‐OMeTAD, 15.8 µL of 4‐tert‐butylpyridine, 9.2 µL of Li‐TFSI solution (520 mg Li‐TFSI in 1 mL of acetonitrile), and 5 µL of FK 209 Co(III)‐TFSI salt solution (375 mg FK 209 in 1 mL of acetonitrile), all dissolved in 0.4 mL of chlorobenzene, was spin‐coated at 2500 rpm for 30 s. Finally, 100‐nm‐thick Au electrode was thermally‐evaporated. For the champion device, an MgF₂ antireflection layer coating was thermally‐evaporated.

p‐i‐n structured PSC

p‐i‐n structured PSCs were fabricated on ITO‐coated glass substrates. A PTAA solution (6 mg mL⁻¹ in chlorobenzene) was prepared with the addition of 4‐isopropyl‐4′‐methyldiphenyliodonium tetrakis(pentafluorophenyl)borate to enhance hole transport properties. This solution was spin‐coated at 5000 rpm for 30 s, followed by annealing at 100 °C for 10 min. Subsequently, a PFN‐P2 solution (0.4 mg mL⁻¹ in methanol) was spin‐coated at 3000 rpm for 30 s to modify the interface. The perovskite precursor solution was meticulously prepared by dissolving Cs_0.05FA_0.85MA_0.1PbI₃ with 0.2 m MACl in a solvent mixture of dimethylformamide and N‐Methyl‐2‐Pyrrolidinone at a volumetric ratio of 4:1, achieving a final perovskite concentration of 1.65 m. A total of 40 µL of the perovskite solution was spin‐coated at 5000 rpm for 8 s with an acceleration rate of 2000 rpm s⁻¹. Immediately after deposition, the samples underwent vacuum‐flash assisted solution processing, where they were exposed to a pressure of less than 60 mTorr for 30 s. The films were then annealed at 130 °C for 30 min to complete crystallization. Following perovskite deposition, C₆₀ (20 nm) and BCP (6 nm) layers were sequentially deposited via thermal evaporation. Finally, a 100 nm Cu electrode was deposited using thermal evaporation.

To fabricate hole‐only‐type samples for SCLC measurements, a PTAA solution (1.5 mg ml⁻¹ in toluene) was spin‐coated on the FTO substrate at 4000 rpm for 30 s, followed by annealing at 100 °C for 10 min. Subsequently, perovskite, Spiro‐OMeTAD, Au layers were deposited.

Device Characterization

A solar simulator (Newport, Oriel Class AAA) equipped with an AM 1.5G filter and a sourcemeter (Keithley 2450) were used for the J–V measurements. The illumination intensity was calibrated to 1‐Sun condition (100 mW cm⁻²) using an NREL‐traceable Si reference cell (91150‐KG5, Newport). The J– V curves were obtained by measuring at a scan rate of 10 mV s⁻¹ in both the reverse direction (from 1.2 to −0.1 V) and the forward direction (from −0.1 to 1.2 V). The active area was regulated by employing a metal shadow mask with a size of 0.0619 cm². Long‐term stability testing involved monitoring the maximum power point under continuous illumination (Newport, Oriel Class AAA, AM 1.5G, 100 mW cm⁻²) within a low humidity environment (relative humidity 3–5%, temperature 25 °C). EIS for Nyquist plot were measured under dark condition covering a frequency range of 10¹–10⁶ Hz. The SCLC measurements were conducted in the dark box, using a sourcemeter (Keithley 2602B).

Thin‐Film Characterization

GI‐WAXS (Xenocs, Xeuss 2.0 HR SAXS/WAXS) data were acquired with MetalJet‐D2 (Excillum) as the X‐ray source and Pilatus 3R 1 M (Dectris) as the detector. XRD (Bruker, D8‐Advance) was performed in a θ–2θ configuration with a scanning interval of 2θ between 3° and 50° using Cu‐Kα X‐ray source (λ = 1.5406 Å). The PL spectra were measured using a fluorescence lifetime spectrometer equipped with a 405‐nm‐wavelength picosecond laser (PicoQuant, FlouTime 300). The cross‐sectional micrograph of PSCs was examined using a field emission SEM (Zeiss, Sigma 300). The AFM (Park Systems, NX‐10) was utilized to capture images depicting the contact potential difference (V _CPD) between the conducting tip and the sample surface. A silicon cantilever with a 30 nm radius of curvature and a platinum tip was employed at a scan rate of 0.6 Hz to acquire dual images, capturing both AFM topography and KPFM bias information simultaneously. Optical transmittance and absorption spectra of the films were acquired through the use of an UV–vis spectrophotometer (Agilent, Cary 5000). XPS and UPS (Kratos, AXIS SUPRA⁺) were used to determine the elemental composition, WF, and valence band characteristics. The C‐1s peak at 284.8 eV served as the calibration standard when analyzing the data. XPS measurements were performed using a monochromatic Al‐Kα X‐ray source (1486.6 eV) operated at 15 keV and 20 mA (300 W), under ultrahigh vacuum conditions (<5x10⁻¹⁰ Torr). For UPS, a He‐I gas discharge lamp emitting at 21.22 eV was employed. FTIR analysis was performed in attenuated total reflection (ATR) mode using an infrared spectrometer (Bruker, TENSOR27).

Simulation

The molecular ESP was computed using density functional models with the ωB97X‐D functional and 6–31G basis set in a vacuum. The calculations were performed with the Spartan (v.9.0.2, Wavefunction Inc.). ESP maps were generated to visualize the electrostatic charge distribution, with electron‐rich regions represented in red and electron‐deficient regions shown in blue.

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supporting Information

Jung Y., Cho S. H., Kim S., Lee J., Park K. T., Lee Y. S., Spatial Atomic Arrangement of Cyclohexyl‐Based Ligands for Enhanced Interface Passivation in 2D/3D Perovskite Solar Cells. Small 2025, 21, 2501564. 10.1002/smll.202501564

Data Availability Statement

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