Anion‐Cation Synergistic Regulation of Low‐Dimensional Perovskite Passivation Layer for Perovskite Solar Cells

Abstract

Mixing 2D and 3D perovskite together is an effective strategy to enhance the stability of perovskite solar cells (PSCs). This strategy has been widely used in many recent works. Typically, 2D layer is formed by introducing 2D spacer onto 3D surfaces through in situ intercalation reaction. However, this intercalation may not stop after the 2D layer is formed. Progressive migration of 2D spacer into 3D bulk leads to increased n‐values of 2D phases and deviation from optimized structural design. The high n‐value 2D perovskite is less stable than the low n‐value 2D perovskite and may be prone to degradation under external stresses. Here, a heteroatom ammonium ligand, thiomorpholine (SMOR) is found, which can effectively passivate the perovskite surface, and form a 1D phase or 2D phase depending on cation to anion ratio and the type of anions. Due to lower formation energy at 1:1 cation to anion ratio, 1D phase can prevent the formation of high‐n‐value 2D phase and show excellent thermal stability. The passivation of SMOR‐based 1D perovskite boosts the device efficiency to 25.6% (certified 24.7%). More importantly, the unpackaged device can maintain >80% of its initial efficiency after stable operation at 85 °C for 1000 h.

The sulfur‐hybridized aniline‐like molecule thiomorpholine (SMOR) forms n = 1 2D perovskite and is preferred to transfer into 1D perovskite instead of high n‐valued 2D perovskite after aging. Due to larger formation energy, chlorine anion is better than iodine anion to accelerate the formation of 1D phase. The PCE of the SMOR‐based 1D/3D perovskite passivated device reaches 25.5% and shows T₉₀ lifetime over 1000 h at 85°C without encapsulation.

1. Introduction

In recent years, the efficiency of single‐cell perovskite solar cells (PSCs) has exceeded 26% due to the high‐quality perovskite films and efficient interfacial passivation.^{[ 1} , ² , ^{3 ]} Coupled with its low‐cost solution preparation method, PSCs has become an emerging and alternative choice for solar energy harvesting.^{[ 4 ]} To achieve commercialization, the stability of PSCs still needs to be improved to pass of industrial lifetime‐assessment tests defined by the International Electrotechnical Commission (IEC 61 215:2016).^{[ 5 ]}

Although three‐dimensional (3D) perovskite is an ideal absorption layer, its stability is plagued by grain boundary defects and ion migration, especially under illumination and elevated temperature, which will aggravate its degradation.^{[ 6} , ^{7 ]} The 2D perovskite, on the opposite, may be inefficient as an active layer but is significantly more stable due to its long and stable cations. To achieve the coexistence of high efficiency and ultra‐high stability in perovskite solar cells, researchers have begun to construct 2D/3D heterostructures through in‐situ surface reactions with ammonium cations,^{[ 8} , ^{9 ]} such as phenylethylamine (PEA),^{[ 10} , ^{11 ]} butylamine (BA) and octyl ammonium (OA),^{[ 4 ]} which can effectively passivate defects, isolate air and water, and suppression migration.^{[ 12 ]}

However, the facile solution‐based in situ growth of 2D perovskites onto 3D perovskites lacks precise control over the phase purity, film thickness, orientation, and structural phase of 2D overlayer, limiting their use as interfacial passivation layers and also leads to unsatisfactory stability and efficiency. Jang et al. achieved solvent‐free growth of 2D BA₂PbI₄ perovskite on 3D films by controlling pressure, temperature, and time, demonstrating the importance of high‐quality 3D/2D interfaces.^{[ 13 ]} Similarly, Choi et al. used vacuum deposition to grow a highly ordered 2D RP (BA)₂PbI₄ perovskite layer in the out‐of‐plane direction on top of the 3D film, achieving controlled crystallization and preventing the formation of unwanted intermediate phases.^{[ 14 ]} However, this solid phase reaction has high requirements for equipment and preparation conditions and is complicated.

Therefore, the controllable preparation of high‐quality 2D/3D interfaces by solution method, limiting the penetration of 2D cations into the perovskite lattice, and inhibiting the gradual migration of 2D cations into 3D phase are the key problems in perovskite research.^{[ 15 ]} Gu et al. controlled the in situ surface passivation reaction by regulating the carbon chain length of aniline and obtained 2D perovskite passivation layers with optimal n = 1 and n = 2, respectively.^{[ 16 ]} By using the dielectric constant and Gutmann donor number of the solvent, Sidhik et al. selected acetonitrile as the 2D phase precursor solvent to prepare a pure phase, controllable n value, adjustable thickness, and energy level matching 2D perovskite passivation layer.^{[ 17 ]} Park et al. systematically studied the reaction depth of different ammonium ligands with perovskite surfaces and showed that the smallest aromatic ligand, anilinium, had the lowest ligand reactivity with 3D perovskites due to steric hindrance near the ammonium group, while the fluorinated derivatives of this ligand formed a strong interface structure.^{[ 18 ]} However, the 2D ammonium cations commonly used for passivation, especially under operating conditions, can disrupt the fragile corner‐sharing octahedral connections within the perovskite structure and migrate into 3D perovskite layers.^{[ 6} , ^{19 ]} This migration increases the n‐value of the 2D phase and reduces the thickness of the 2D layer, resulting in accelerated film degradation and device efficiency and stability drops.^{[ 20} , ²¹ , ^{22 ]}

Here, we constructed the smallest aromatic‐like ligand, thiomorpholine (SMOR), which had the lowest ligand reactivity with 3D perovskites and the sulfurized derivative of this ligand enhances intermolecular hydrogen bonding. Theoretical calculations found that the introduction of sulfur atoms enhances the adsorption energy with defects and can also provide band edges for the formed low‐dimensional perovskite, thereby improving charge transport.^{[ 23 ]} Moreover, this ligand can form both 1D and 2D perovskites with lower formation energies than high‐n‐value 2D phases. Since the migration of 2D cations in 3D phase is mainly driven by the concentration gradient, the lower stoichiometric ratio of SMOR to PbI₂ (1:1) in 1D phase than in the 2D phase (n = 1, 2:1) makes the 1D phase is free of gradient driven phase change. This unique ability ensures that the ligand does not produce a 2D phase with a high n value like other traditional amine salts (such as PEA) during in situ surface passivation,^{[ 22 ]} but instead ultimately exists stably in a 1D phase. At the same time, we have obtained accurate 1D perovskite phase passivation by properly adjusting the anions from iodine to chlorine, due to the low 1D phase formation energy in the chlorine environment. Finally, we report a power conversion efficiency of 25.6% (certified 24.7%) for an inverted PSC. Unpackaged devices maintain T80 (lifetime >80% of their initial efficiency) for >1000 h in thermal stability testing at 85 °C.

2. Design and Synthesis of Low Dimensional Perovskites

Previous studies have shown that larger ammonium ligands, such as A6P,^{[ 24 ]} can inhibit ligand insertion with robust organic‐inorganic networks, while certain small‐sized ammonium ligands exhibit lower reactivity to the 3D phase, such as anilinium.^{[ 18 ]} Therefore, the piperazine‐like salt, SMOR, with a small‐sized amine salt configuration and enhanced intermolecular interaction between its spacer ligand by hybridizing the sulfur atoms, could also have low reactivity to the 3D phase. Then, we synthesized the corresponding 1D and 2D perovskite crystals by reacting to the organohalide salts of SMORI with lead iodide in an acidic mixture of HI and H₃PO₂.

Through calculations, the structural parameters of 1D perovskite SMORPbI₃ and 2D perovskite (SMOR)₂PbI₄ were obtained. In Table S1 (Supporting Information), both SMORPbI₃ and (SMOR)₂PbI₄ belong to the orthorhombic system with space group P2₁/c.^{[ 25 ]} As shown in Figure  1a, the 1D SMORPbI₃ consists of an organic‐inorganic 1D chain perovskite structure composed of face‐sharing 1D [PbX₃]⁻ octahedral chains interlaced with thiomorpholine cations. As in its 2D form, it is the same as other RP perovskites. From Figure 1b, it can be seen that the experimental data and the simulated X‐ray diffraction peaks are in good agreement. The peaks at 10.5° and 11.3° for SMORPbI₃ perovskite belong to facets (002) and (011), respectively (Figure S1a, Supporting Information). The peaks at 6.8° and 14.1° for (SMOR)₂PbI₄ perovskite belong to facet (002) and (111), respectively (Figure S1b, Supporting Information). It is worth mentioning that in the XRD of the 2D phase, there are still a few 1D diffraction peaks. This is because, during the synthesis process, part of SMOR volatilizes, resulting in SMOR being less than twice the amount of lead iodide.

Figure 1.

a) Side view of constructs, b) experimental and simulated XRD, and c) energy bands and density of states of SMORPbI₃ and (SMOR)₂PbI₄ (d–f).

The formation of 1D and 2D phases depends on the ratio of SMOR to PbI₂. When SMOR is in excess and its ratio to PbI₂ is greater than or equal to 2, the n = 1 2D phase is formed. When PbI₂ is in excess, it transforms into n = 2 2D phase or 1D phase (the ratio of SMOR to PbI₂ for both phases is 1), but since the formation energy of 1D phase is lower, the formation of n = 2 2D phase is not energetically preferred. To prove it, we removed or added SMORI molecule from its 2D n = 1 and 1D films by annealing or adding SMORI containing IPA solvent, respectively. As shown in Figure S19 (Supporting Information), we prepared the white 1D and vermilion 2D films on a quartz plate by adjusting the SMOR to PbI₂ ratios at 1:1 and 2:1, respectively. When adding SMORI by its IPA solution, the film changed from colorless to pink (Figure S19b, Supporting Information), indicating a 1D to 2D n = 1 phase transition due to the presence of excess SMORI molecules. To remove the SMORI molecules, we annealed two films at 85 °C in the air. The 2D film gradually transferred into a mixed phase of 1D and 2D at one hour, and the complete transition to 1D phase took one day (Figure S19d, Supporting Information). The slow transition of 2D n = 1 phase to 1D phase suggests the driving force is due to evaporation of SMORI component from 2D phase. On the contrary, the 1D phase did not show any changes, confirming its good thermal stability.

To further confirm our assertion, the film transferred into 1D phase was treated with SMOR(IPA) solution. As expected, the film turned back to 2D n = 1 phase after dipped by the solution twice (Figure S19f, Supporting Information). With these observations, we can undoubtedly confirm the phase transition between 2D n = 1 phase to 1D phase is only determined by its SMOR to PbI₂ ratio. The transition formulae were summarized in Figure S19c (Supporting Information).

At the same time, we also calculated and analyzed the electronic structures of these two phases. As shown in Figure 1c,f, the band gaps of the 1D and 2D phases are 3.25 and 2.27 eV, respectively. Moreover, their conduction band is mainly composed of Pb‐p and I‐p, while the valence band is composed of S‐p, Pb‐p, and I‐p hybridization. Therefore, due to the presence of heteroatoms, sulfur, in the cations and their participation in the band edge construction, the band gap of the 2D perovskite composed of this cation is slightly smaller than that of the conventional 2D perovskite. This shows that not only organic ammonium salts with large π‐conjugated systems^{[ 26 ]} but also special heteroatoms can affect the frontier orbits of perovskites, thereby affecting the band edge structure of perovskites and even the carrier dynamics of perovskites.

3. Low Dimension/3D Heterostructures

We then attempted to construct low‐dimension/3D heterostructures by casting ammonium cation solutions onto 3D perovskite films. The resulting films were studied by XRD (Figure S2, Supporting Information) and found that PEA generated 2D perovskites with mixed n values (Figure 2b; Figure S2a, Supporting Information). It can also be seen from UV absorption (Figure 2d) that the PEAI sample initially showed two absorption peaks of n = 1 and n = 2. After 90 days of storage, the results showed that as PEA diffused into the 3D phase, the absorption peak of the n = 1 phase decreased and the absorption peak of the n = 2 phase increased, which indicated the trend of transformation from the conventional n = 1 low‐dimensional phase to the high n value phase (shown in Figure 2e). In contrast, although SMOR will generate a 2D phase with n = 1 at the beginning (Figure 2c; Figure S2b, Supporting Information), it will not generate other 2D phases with high n values and will only transform into a 1D phase after 90 days of storage (Figure 2d,f).

Figure 2.

a–c) Top and side view of SEM images of perovskite films with and without PEAI and SMORI treated. d) UV–vis absorption spectrum of perovskite films with and without PEAI and SMORI. e,f) Schematic diagram of phase evolution of PEAI and SMORI, respectively.

To clarify the phase evolution of SMOR, we conducted separate experiments and computational simulations on the reaction of SMORI and SMORCl molecules with lead iodide. From the experiment, it can be seen that SMORI molecules form both 1D and 2D phases on PbI₂ by spin coating (Figure  3a). The 1D film prepared by solution spin coating has a strong peak at 10.5°, while the peak at 11.3° is very weak. At the same time, the SMORCl molecule reacting with lead iodide only generates two weak 1D XRD peaks, without the formation of a 2D phase. The differences in phases of the two molecules can be also confirmed by the absorption spectra in Figure 3b. Theoretical calculations show that the formation energies of 1D and 2D perovskite with its n = 1 (SMOR)₂PbI₄ phases are much lower than that of 2D perovskite with n = 2 (SMOR)₂MAPb₂I₇ phases, which means when SMOR cation diffuses due to concentration gradient, it is thermodynamically preferred to form 1D phase rather than 2D phases with high n values (Figure 3c). This is why only the transition from n = 1 (SMOR)₂PbI₄ to 1D was observed in aged samples, but not the transition to high n values. Meanwhile, the calculation shows that the energy barriers between the n = 2 2D phase and the 1D phase are 0.78, 1.34, and 1.88 eV for iodine, bromine, and chlorine environments, respectively. This shows that the tendency to form a 1D phase is more obvious in the chloride environment.^{[ 25 ]} From Figure S2a (Supporting Information), it found that the intensity change of the low‐dimensional phase XRD peak of the film after passivation with bromide salt or chloride salt, which means that anions can indeed affect the low‐dimensional phase formed. As shown in Figure 3d, when in an iodine‐rich environment, SMOR and lead iodide tend to form a 2D phase with n = 1, while in a chlorine‐rich environment, SMOR is more likely to react with lead iodide to form a 1D perovskite.

Figure 3.

a) XRD and b) UV–vis adsorption spectrum of PbI₂ with SMORCl and SMORI. c) The formation energies of 1D and 2D phase. d) Schematic diagram of the reaction mechanism of SMOR and lead iodide.

4. Low Dimensional Perovskite Passivation Effect

Combined with UPS data (Figure S5, Supporting Information), the energy band alignment at the interface is shown in Figure 4a. It shows that the film with SMORCl (0.05 eV) has the smallest mismatch band alignment than SMORI (0.48 eV) which facilitates electron transfer. It is worth mentioning that when a layer of 2D perovskite is formed on the surface of the perovskite treated with a higher concentration of PEAI, its energy band is more suitable for hole transport rather than electron transport. This is also one of the reasons why most organic ammonium ion salts are very suitable for n‐i‐p device passivation but not for p‐i‐n device passivation. Therefore, for p‐i‐n devices, it is often better to limit the degree of passivation, which may sometimes lead to a less‐than‐ideal passivation effect and no obvious improvement in device performance (Table S2, Supporting Information). X‐ray photoelectron spectroscopy (XPS) spectra (Figure S6, Supporting Information) show the appearance of characteristic element peaks of passivation molecules.

Figure 4.

a) Interface band alignment, b) UV–vis absorption and photoluminescence (PL) intensity, c) time‐resolved PL (TRPL), d) electrochemical impedance spectroscopy (EIS), and e) Mott–Schottky plots of devices with and without passivation at 1 kHz. f) Binding energies of ammonium ligands with the typical V_I and Pb_I defects. And g–i) partial density of states (PDOS) of Pb_I defects with and without ammonium ligands, and the insets are the differential charge density. Yellow represents positive charge and green represents negative charge.

As shown in Figure 4b, the photoluminescence (PL) intensities of perovskite films treated with different salts are stronger than the control. Among different anions, the PL intensity of the SMORCl treatment is the highest, suggesting the best passivation to surface nonradiative defects. Time‐resolved PL (TRPL) decay measurements of perovskite films after passivation show enhanced average carrier lifetimes for PEAI (t = 3.40 µs), SMORI (t = 3.62 µs) and SMORCl (t = 3.78 µs) compared to the control (t = 2.19 µs). The reduced fast decay lifetime confirms that the nonradiative recombination of photogenerated carriers is suppressed after passivation treatment. Again, TRPL measurement also suggests that SMORCl has the best passivation to nonradiative defects. Electrochemical impedance spectroscopy (EIS) measurements show that the SMORCl‐treated devices have larger recombination resistance (R_rec) compared to the others (Figure 4d).^{[ 27 ]} The Mott‐Schottky analysis (Figure 4e) shows that the built‐in potential V_bi is larger after SMORCl (1.15 V) treatment, resulting in better carrier extraction and transport properties.^{[ 28 ]} This is consistent with the calculation results (Figure S13, Supporting Information), which show that in the 1D/3D heterojunction, more electrons are transferred from the 3D phase to the 1D phase, indicating that the 1D interface is more suitable for electron transport than the 2D interface.

It is noteworthy that various surface defects can change the potential distribution on the perovskite surface.^{[ 29 ]} The inset compares the topography and line scans of the surface potential variations for a typical perovskite region, as shown in the inset of Figure S8 (Supporting Information). These potential fluctuations can be up to ≈50–100 mV at the grain boundary. It can be seen that the potential energy at the grain boundary shows an overall downward trend after SMOR‐treated. To further measure the changes in perovskite surface defects, the space charge‐limited current (SCLC) studies with the structure ITO/SnO₂/perovskite/PCBM/Ag. The results show that the perovskite has a lower trap density after SMORCl treatment with a lower trap‐filed limit voltage (0.18 V) than the control (0.28 V). All these confirm that the introduction of SMORCl greatly suppresses the carrier recombination caused by defects. This is also reflected in the lower dark current of the SMOR‐treated device compared to the control (Figure S7, Supporting Information).

The type of defects that were passivated by SMOR cations are studied by calculation. It is found that the two most common defects: I vacancy (V_I) and Pb‐I antisite (Pb_I), with the lowest formation energy on the perovskite surface, can be strongly bounded by SMOR cations.^{[ 30} , ^{31 ]} Figure 4f shows that the binding energy of SMOR on both defects (V_I, −5.95 eV; Pb_I, −5.34 eV) is stronger than that of PEA (V_I, −4.81 eV; Pb_I, −2.47 eV), which may be due to the lesser steric hindrance from its smaller molecular size and stronger polarity caused by S atoms. At the same time, through the analysis of electronic states, it is shown that SMOR's stronger bonding ability can better combine with Pb_I defects than PEA, thereby reducing the defect state density.

5. Device Performance and Stability

The cross‐sectional SEM image of the inverted PSCs is shown in Figure  5a with a structure of ITO/SAM/perovskite/C60/BCP/Ag. And the current‐voltage (J‐V) scan of a champion small area device (0.0465 cm²) based on SMORCl passivation, which has a PCE of 25.6%, an open circuit voltage (V_OC ) of 1.18 V, a short circuit current density (J_SC) of 25.3 mA cm⁻², and a fill factor of 85.8%. The device was certified externally reaching 24.7% in PCE (Figure S11, Supporting Information). The device performances (Figure 5b) show that the PV performance of the SMOR‐treated devices is improved compared to the PEA device (24.7%), with an average PCE of 25.2% of SMORI and 25.5% of SMORCl (Figure 5b; Figure S12, Supporting Information), primarily from increased V_OC and fill factor (FF). It is worth mentioning that the average PCE of the PEAI‐treated devices is not much improved compared to the control devices (Figure S14, Supporting Information). As we mentioned earlier, although the PEA molecule has a certain passivation effect, it is not very suitable for p‐i‐n devices. The surface passivation effect of SMOR molecules is obvious, so the V_OC of SMOR‐treated devices is significantly improved. The efficiency of SMORCl is slightly better than that of SMORI devices, which is attributed to the better band matching of SMORCl. The band gap of the device is determined to be ≈1.53 eV based on the external quantum efficiency (EQE) spectrum (Figure 5d) and UV–vis (Figure 4b).

Figure 5.

Device performance. a) Cross‐sectional SEM image of the device structure and the scale bar is 500 nm. b) PCE statistics and c) current density–voltage (J‐V) curves for PEAI, SMORI, and SMORCl‐passivated devices. d) The IPCE of the SMOR‐passivated device. e) The stabilized power output of the different passivated devices. f) The degradation (the XRD peak ratio of PbI₂/(001)) of perovskite films stored at 85 °C under 60% RH in air. g) Thermal stability tracking of unencapsulated devices stored in N₂ glovebox at 85 °C. h) Storage stability of the devices stored in the N₂ glovebox.

To elucidate the effect of 1D and 2D perovskite passivation on device stability, we evaluated the thermal stability of unencapsulated small cells at 85 °C (ISOS‐D‐2 protocol, where ISOS is the International Summit on Organic Photovoltaics Stability). Due to the better suppression of ion migration by 1D perovskites and the low reactivity of small molecule benzene ring compounds, the 1D‐SMOR‐based device retained 80% of its initial PCE after 1000 h (Figure 5g). In contrast, the 2D‐PEAI‐based device only maintained 400 h under the same test conditions before dropping to 80% of its initial PCE. At the same time, in the steady‐state output test, the SMOR device is better than the PEAI (Figure 5e). In the nitrogen storage stability test, the SMOR device maintained 95% of its initial efficiency after 2500 h, while the PEA device based on 2D passivation had <90% of its initial efficiency (Figure 5h). To test the damp‐heat stability of different passivation strategies, we placed different perovskite films in the air (around RH 60%) and heated them to 85 °C (Figure S15, Supporting Information). By observing the degradation of the perovskite films, it was found that the SMOR device lasted up to seven days, while the control and PEAI‐treated samples began to turn yellow on the 4th and 5th days, respectively. Through the XRD tracking test (Figure 5f), the ratio of PbI₂ to (001) peak was used as a measure of the decomposition of perovskite film. The results showed that after 120 h, the lead PbI₂ of the SMOR device was less than 0.5, while that of PEA and control samples exceeded 1.0 after 50 h. This shows that passivation by the 1D SMOR perovskite layer can better inhibit ion migration and enhance the water resistance of the surface perovskite layer compared to 2D PEA passivation.^{[ 32} , ^{33 ]}

6. Conclusion

Organic amine salts are often used for perovskite surface passivation, which forms a low‐dimension/3D heterojunction structure on the 3D perovskite surface through an in situ intercalation reaction on the surface, thereby greatly improving the efficiency and stability of the device. Here, we selected aniline‐like organic amines carrying heteroatoms as cations which can form a more stable 1D perovskite with a self‐limited nature to transfer into 2D perovskite, such as PEA. In the meantime, the smaller Cl⁻ anions help to stabilize the 1D phase by enhancing its formation energy further. As a result, we successfully built a self‐limiting 1D perovskite layer on 3D perovskite and suppressed the generation of high n‐valued 2D phases. On this basis, we further obtained 1D/3D heterojunction passivated devices through anion regulation and obtained a certified efficiency of 24.7%. This device shows better stability than conventional 2D/3D. This inspired us to explore passivation schemes other than conventional 2D/3D heterojunctions, as well as to make surface passivation more controllable and stable through coordinated regulation of anions and cations. In addition, we also found through theoretical calculations that heteroatoms can introduce orbital electrons of different energy depths to ammonium cations, which can provide additional means for ammonium cations to affect the perovskite band edge. In general, by rationally designing the cations and anions of the passivating molecules, the in situ passivation reaction can be controlled to achieve the desired passivation effect.

Conflict of Interest

The authors declare no conflict of interest.

Author Contributions

S.L. and H.G. contributed equally to this work. S.L. performed conceptualization, methodology, investigation, and DFT calculation. H.G. performed methodology and validation. A.Z., X.Q., and Y.C. performed methodology and characterization on morphology measurement. J.G., C.X. performed characterization on the optical properties and performance measurement. Prof. H.P. performed validation and resources. Prof. G.X. performed supervision of experiments and funding acquisition. Prof. S.C. performed conceptualization, wrote the review, and performed editing, supervision, and funding acquisition.

Supporting information

Supporting Information

Li S., Gu H., Zhu A., Guo J., Xi C., Qiu X., Chen Y., Pan H., Chen J., Xing G., Chen S., Anion‐Cation Synergistic Regulation of Low‐Dimensional Perovskite Passivation Layer for Perovskite Solar Cells. Adv. Mater. 2025, 37, 2500988. 10.1002/adma.202500988

Data Availability Statement

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