Pure Chloride 2D/3D Heterostructure Passivation for Efficient and Stable Perovskite Solar Cells

To date, organic–inorganic hybrid perovskite solar cells (PSCs) have reached a certified efficiency of 25.7%, showing their great potential in industrial commercialization. However, defects at the surface and grain boundaries hinder their device performance and long‐term stability. Herein, long‐chain dodecylammonium halides (DACl, DABr, and DAI) to treat the perovskite surface and improve the device performance are introduced. It is found that the three passivators can all form 2D perovskites but with different halide compositions. The DACl‐treated perovskite forms a pure chloride DA2PbCl4 2D perovskite, while the DABr and DAI‐treated surfaces form a pure iodide DA2PbI4 2D perovskite. Compared with the DA2PbI4 layer, it is found that the DA2PbCl4 passivation layer can more effectively passivate defects, improve carrier separation at the perovskite surface, and optimize the energy alignment between the perovskite film and hole transport layer. As a result, a champion power conversion efficiency of 23.91% is achieved for the DACl‐treated PSCs. Moreover, the device maintains around 95% of its initial efficiency after 1000 h storage under relative humidity of 10% at 25 °C.

studied the passivation effect of 2D perovskites from three bromide-based alkylammonium salts with different alkyl chain lengths (butylammonium bromide, octylammonium bromide, and dodecylammonium bromide) for wide-bandgap PSCs. The results also showed that long-alkyl-chain organic cations possessed a greater passivation effect than short-chain organic cations. Mahmud et al. [23] demonstrated that the incorporation of octylammonium chloride (OACl) provided both bulk and surface passivation. The Cl À ions could penetrate the bulk 3D perovskite to passivate lattice vacancy defects, while the 2D/3D heterostructure provided localized surface passivation. Benefiting from these, a high PCE of 23.62% was achieved for the OACl passivated device. Recently, He and coworkers [24] applied dodecylammonium bromide (DABr) to react with excessive PbI 2 to form 2D/3D heterojunction, which passivated surface defects and increased the built-in electric field of the device to improve carrier transport capability. As a result, an improved PCE of 21.81% was achieved, and the unencapsulated device maintained 64% of its initial efficiency after 1000 h storage in ambient air.
Previous studies have shown that alkylammonium halide with a longer chain length exhibited better passivation effect, while halide ions can also influence the passivation process. Herein, we introduced long-chain dodecylammonium chloride (DACl), dodecylammonium bromide (DABr) and dodecylammonium iodide (DAI) to enhance the performance of devices via forming 2D/3D heterojunctions. Our results showed that the DACltreated perovskite formed DA 2 PbCl 4 2D perovskite, while DABr and DAI-treated surfaces formed DA 2 PbI 4 2D perovskite.
Compared with the DA 2 PbI 4 2D perovskite, the DA 2 PbCl 4 2D perovskite exhibited more efficient passivation capability and could improve carrier lifetime and optimize the energy alignment between the perovskite film and hole transport layer (HTL) more effectively. These contributed to significantly improved opencircuit voltage (V oc ), enhanced short-current density ( J sc ), and fill factor (FF) of the devices. As a result, a champion PCE of 23.91% was achieved for the DACl-treated PSCs, and the device maintained around 94% of its initial efficiency after 1000 h of storage under relative humidity (RH) of 10% at 25°C.

Results and Discussion
To evaluate the effect of dodecylammonium halide precursor solution, the PSCs with DACl, DABr, and DAI as passivation layers on the perovskite surface were fabricated, with the structure of glass/tin-doped indium oxide (ITO)/SnO 2 /perovskite/ passivation layer/spiro-OMeTAD/Ag. To optimize the concentration of the passivation solution, three different concentrations of each dodecylammonium halide precursor were applied to fabricate PSCs ( Figure S1-S3, Supporting Information). It can be seen that for all the three passivators, the V oc increases at higher concentrations. However, the J sc decreases with increased precursor concentration due to the formation of a thicker passivation layer, increasing the carrier transport barrier at the surface of the 3D perovskite. [25,26] The optimum concentration for DACl, DABr, and DAI passivators to achieve high-performance PSCs is 1 mg mL À1 . Figure 1 shows the statistical distribution of the  photovoltaic parameters of the PSCs with passivator treatments, along with the control devices. All the three types of passivated devices exhibit higher PCE than the control devices, mainly attributed to the significantly enhanced V oc and slightly increased J sc and FF. Among the three passivated device variations, the DACl-treated PSCs exhibit the best performance, with the largest improvement in V oc and the highest average and champion PCE.
To study the interaction mechanism between the dodecylammonium halide and the perovskite film, X-ray photoelectron spectroscopy (XPS) was carried out on the control and passivated perovskite films. As illustrated in Figure 2a, the peaks of Pb 2þ for control film located at 138.2 and 143.1 eV correspond to Pb 4f 7/2 and Pb 4f 5/2, respectively. [27] Compared with the control film, Pb 4f 7/2 and Pb 4f 5/2 peaks of the three passivated films all shift Figure 2. XPS spectra of a) Pb 4f and b) N 1s orbital for control and passivated films. c) XRD spectra of the control and passivated films. d) Enlarged XRD spectra at 3°-10°. e) XRD spectra of 2D DA 2 PbCl 4 , DA 2 PbBr 4 , and DA 2 PbI 4 perovskites (n = 1). f ) Schematic illustration of the formation of DA 2 PbCl 4 2D perovskite after incorporation of DACl.
to higher binding energies, indicating that all the three passivators form interactions with Pb 2þ ions. [28][29][30] For the N 1s spectra (Figure 2b), two individual peaks at 400.1 and 402.0 eV for the control film are assigned to C-N and C-N þ , respectively. [31] With the addition of the dodecylammonium halides passivators, the peak intensity at 402.0 eV is enhanced, which proves the interactions between the passivators and the perovskite film. The spectra of Cl and Br are shown in Figure S4 and S5, Supporting Information, respectively. The significantly enhanced intensities of Cl 2p peaks at 198.2 and 199.8 eV and Br 3d peaks at 65.0, 68.7 and 71.7 eV after incorporation of DACl and DABr, respectively, further prove the interactions between dodecylammonium halide and the perovskite films.
To investigate the morphology change of the perovskite surface, top-view scanning electron microscope (SEM) images were performed for the control and passivated perovskite surfaces. As shown in Figure S6, Supporting Information, no significant change in surface morphology was observed. Energy-dispersive X-ray spectroscopy (EDX) was carried out to further investigate the distribution of Cl element, as depicted in Figure S7, Supporting Information. It can be seen that Cl element was detected and was well distributed in the control perovskite film, indicating that plenty of residual Cl À remained in the bulk perovskite after annealing to improve the perovskite quality. Meanwhile, for the passivator-treated films, no significant increase of Cl signal was observed compared with the control film. Thus, we concluded that the extra Cl À addition from DACl was less dominant in improving the bulk perovskite quality, as the Cl À addition was small compared with the residual Cl À in the bulk perovskite.
The X-ray diffraction (XRD) patterns for the control and dodecylammonium halide passivated perovskite films are presented in Figure 2c. It can be seen that the peak intensities at 14.04°o f the α-perovskite (001) plane are significantly enhanced after the incorporation of the passivators. This indicates that dodecylammonium halide addition can effectively improve the crystallinity of the perovskite films. The enlarged pattern depicts the additional low-angle peaks of the films (Figure 2d). Small peaks at 7.24°and 7.26°were observed for the films treated with DAI and DABr, respectively. For the DACl-treated film, peaks at 6.28°a nd 9.42°were observed, while no peaks at low angles were observed for the control film. It has been widely reported that the incorporation of long-chain alkylammonium halide could form 2D or quasi-2D perovskite on the surface of the bulk perovskite. [22][23][24] Thus, to specify the new phase, XRD measurements for pure 2D perovskite (n = 1) with different halide formations were carried out. The DA 2 PbX 4 (X for Cl, Br, and I) precursor solution was prepared by mixing DAX and PbX 2 in the DMF:DMSO (1:4) solution with the mole ratio of 2:1. The precursor solution was subsequently spin coated to the ITO/glass substrate to prepare the 2D perovskite film. As depicted in Figure 2e, the 2D perovskite DA 2 PbI 4 exhibits a diffraction peak at 7.24°, the 2D perovskite DA 2 PbBr 4 exhibits diffraction peaks at 6.34°and 9.62°, and the 2D perovskite DA 2 PbCl 4 exhibits diffraction peaks at 6.28°and 9.44°. After analyzing these XRD patterns, we concluded that DA 2 PbCl 4 2D perovskite was formed on the 3D bulk perovskite after incorporating with DACl precursor, as illustrated in Figure 2f. In contrast, DA 2 PbI 4 2D perovskite was formed after DABr or DAI treatments. Thus, combined with the XPS results and the device performance, we concluded that the significantly enhanced device performance after incorporating dodecylammonium halide is attributed to the formation of 2D/3D heterostructures, which could possibly improve the perovskite crystallinity and reduce defects on the surface of the perovskite films. In addition, the best performance of DACl-treated devices is possibly attributed to more effective passivation of 2D DA 2 PbCl 4 perovskite than DA 2 PbI 4 .
To investigate the passivation effect of the 2D/3D heterostructures induced by the dodecylammonium halide precursors, steady-state photoluminescence (PL) measurements were performed for the perovskite films on glass substrates. As shown in Figure 3a, a significant enhancement in PL intensity of the three passivated perovskite films is observed compared with the control film. The DACl-treated perovskite film shows the highest PL intensity, followed by the DABr and DAI-treated films. The improved PL intensity indicates that nonradiative recombination is greatly suppressed at the surface of the perovskite film, and DACl exhibits the best passivation capability, which is consistent with the V oc enhancements of the passivated devices. A slight blueshift is also observed for the peak position of the DACl-treated perovskite film. The UV-vis absorption spectra of perovskite film without or with the passivation treatments are depicted in Figure S8, Supporting Information. A slight blueshift is also observed for the DACl-treated film, which is in accordance with the PL intensity results. The E g values of the perovskite films were derived to be 1.56 eV for the DACl-treated perovskite film, and 1.55 eV for others from the Tauc plots, as illustrated in Figure S9, Supporting Information. Time-resolved photoluminescence (TRPL) spectra were conducted to examine the charge recombination behavior of the control and 2D/3D perovskite films, as illustrated in Figure 3b. The decay spectra were fit with a biexponential function, and the fitting parameters are listed in Table S1, Supporting Information. The fast decay component (τ 1 ) is related to the nonradiative recombination derived by the defects on the surface and grain boundaries of the perovskite film, and the slow decay process (τ 2 ) is related to the radiative recombination of photogenerated free carriers in the bulk of the perovskite. [32,33] The 2D/3D perovskite films from DACl, DABr, and DAI treatments all exhibit prolonged τ 1 and τ 2 values, and the 2D/3D heterostructure formed by DACl possesses the longest τ 1 (171 ns) and τ 2 (1454 ns) compared with τ 1 (64 ns) and τ 2 (240 ns) of the control perovskite film. This indicates that the DA 2 PbCl 4 2D perovskite can greatly suppress nonradiative recombination and improve charge carrier lifetime at the surface of the perovskite film. Notably, the best passivation effect of DACl among dodecylammonium halides is in agreement with the previous findings, which show a better passivation effect of OACl (OA: octylammonium) than OAI and OABr. [23] To investigate the trap state density of the perovskite films, space-charge-limited current (SCLC) measurements were performed through simplified devices with the configuration of ITO/SnO 2 /perovskite/passivation layer/PCBM/Ag. As shown in the double-logarithmic scale I-V plots in Figure 3c, the trap-filled limit voltage V TFL is significantly reduced for the three passivated electron-only devices, which can be used to evaluate the trap density of the perovskite by Equation (1).
where ε 0 , ε, e, and L are the vacuum permittivity, relative permittivity, elementary charge, and the thickness of the perovskite film, respectively. The thickness of our perovskite films is around 600 nm ( Figure S10, Supporting Information), and the relative permittivity is around 35. [34]  To investigate the effect of passivators on the energy level of the perovskite films, ultraviolet photoelectron spectroscopy (UPS) measurements were performed, as shown in Figure S11, Supporting Information. According to the UPS results, the valence band maximum (VBM) was calculated to be À5.59, À5.36, À5.44, and À5.43 eV for control, DACl, DABr, and DAI-treated films. [35] Combined with the E g values of the perovskite films, the energy-level alignment of the perovskite films and SnO 2 and spiro-OMeTAD is shown in Figure 3d. [36] All the three passivated films show an upshift of the VBM compared with the control film. The VBM values of DABr and DAI-treated films are similar, while less energy difference between the VBM of DACl-treated film and spiro-OMeTAD is achieved. This indicates that the formation of DA 2 PbCl 4 2D/3D heterostructure can optimize the band alignment between the perovskite film and spiro-OMeTAD, which is favorable for hole transport and thus reduced V oc loss. [37,38] To better understand the mechanism of the improved device performance by the 2D/3D heterostructures, a series of characterizations at the device level were performed. The built-in potential (V bi ) of the control and passivated PSCs were characterized by the Mott-Schottky plots, [39,40] as illustrated in Figure 4a. From the plots, the V bi values of the control and DACl, DABr, DAI-passivated devices are 0.82, 0.98, 0.94, and 0.90 V, respectively. The higher built-in electric field is recognized to facilitate the transport of photogenerated carriers, mitigate the accumulation of charges, and widen the depletion region www.advancedsciencenews.com www.advenergysustres.com in perovskites for suppressing the carrier recombination, [33,41,42] which is beneficial for the V oc improvement and in agreement with the enhanced V oc results of our passivated PSCs. To further investigate the internal charge transfer behaviors of the devices, electrochemical impedance spectroscopy (EIS) measurements were performed in the dark, [43,44] and the Nyquist plots are shown in Figure 4b. The results were fit using the equivalent circuit from the inset of the figure, and the fit R s and R rec values are listed in Table S2, Supporting Information. All the devices show similar R s values, owing to the same device structure and hole transport material (spiro-OMeTAD). [45] However, the three dodecylammonium halide passivated devices all exhibit significantly enhanced R rec values compared with the control device, indicating that carrier separation is dramatically improved in the passivated devices. Moreover, the DACl-treated devices possess the highest R rec value, which proves that the DA 2 PbCl 4 2D/3D perovskite formed by the DACl addition can significantly reduce the trap state density of the perovskite film at the device  level. [24,46] Figure 4c shows the dark current density of the control and passivated devices. Significant decreases in dark current density of the DACl, DABr, and DAI-passivated devices are observed compared with the control device. This indicates that the 2D/3D heterostructure improves the ohmic contact of the passivated devices, which is beneficial for the charge transfer within the PSC under illumination. [36,47] Meanwhile, the lower dark current density implies that the 2D/3D perovskite could suppress the leakage current, thus leading to less charge recombination and improved V oc and J sc of the devices. [48,49] Figure 4d shows the reverse scan J-V curves of the champion devices based on pristine and 2D/3D perovskites, and the photovoltaic parameters of the champion devices are listed in Table 1. The control devices show the best PCE of 22.66%, with a J sc of 24.69 mA cm À2 , V oc of 1.132 V, and FF of 81.01%. In contrast, the DACl-passivated devices exhibit the best PCE of 23.91%, with significantly enhanced V oc (1.174 V) and slightly improved J sc (24.80 mA cm À2 ) and FF (82.11%). Meanwhile, the DABr and DAI-passivated devices also show improved champion device PCEs up to 23.78% and 23.45%, respectively, owing to dramatically improved V oc (1.165 and 1.163 V for DABr and DAI-treated devices, respectively) and slightly improved J sc (24.97 and 24.79 mA cm À2 for DABr and DAI-treated devices, respectively) and FF (81.76% and 81.29% for DABr and DAItreated devices, respectively). The incident photon-to-current conversion efficiency (IPCE) measurements for the corresponding devices are depicted in Figure 4e. The integrated current densities for the control, DACl, DABr, and DAI-treated devices were calculated to be 24.21, 24.52, 24.47, and 24.46 mA cm À2 , respectively, which are in agreement with the J sc results obtained from the J-V characterization. The improved current density of DAClpassivated devices is attributed to the DA 2 PbCl 4 2D/3D heterostructure leading to reduced trap state density, suppressed nonradiative recombination at the perovskite surface, along with reduced leakage current of the devices. To evaluate the hysteresis behavior of the PSCs, J-V characterizations from both forward and reverse scans were performed, with the curves shown in Figure S12, Supporting Information, and photovoltaic parameters listed in Table S3, Supporting Information, respectively. The control device shows significant hysteresis, while the three passivated devices exhibit reduced hysteresis, which further proves that defects of the perovskite films have been effectively suppressed by the 2D/3D heterostructures. [50] To evaluate the operational stability of the devices, steady-state PCEs under 1 sun illumination at maximum power point (V max , 0.99 V for all the devices) for 300 s were performed for the control and passivated devices, as illustrated in Figure 4f and S13, Supporting Information. Stabilized PCEs of 22.29%, 23.33%, 23.23%, and 22.86% were achieved for control, DACl, DABr, and DAIpassivated devices, respectively. These results indicate that the incorporated 2D/3D heterostructures could significantly enhance the operational stability of the devices and further confirm the reliability of the positive effect of DA 2 PbCl 4 formed by DACl addition on the device performance.
To investigate the humidity stability of the 2D/3D heterostructure, UV-vis absorption spectra were carried out for the control and passivated perovskite films stored under ambient air condition with RH of 40% at 25°C. As illustrated in Figure 5a-d, the absorbance of the control perovskite film exhibits a larger drop from day 1 to day 12 compared with the passivated perovskite films, and the band edge also shows a larger redshift. The reduced absorbance and redshift of the band edge are possibly attributed to the decomposition of the MA-based perovskite under humid conditions, which possesses a larger bandgap than the FA-based perovskite. [51] Thus, the absorbance results suggest that the 2D/3D heterostructure could improve the moisture resistance of the perovskite film and suppress the decomposition of the bulk perovskite. The insets show images of contact angle measurements of water droplets on the control and passivated perovskite films. It can be seen that the contact angles on the passivated films are all increased compared with the pristine film, further proving that the long alkyl chain from the 2D perovskite increases the hydrophobicity of the bulk perovskite film. To evaluate the long-term stability of the devices, J-V curves of the unencapsulated control and passivated devices stored under RH of 10% at 25°C were performed, as shown in Figure 5e-g. The photovoltaic parameters of the devices are listed in Table S4, Supporting Information. The control device maintained around 82% of its initial PCE (from 22.42% to 18.34%) after 44 days of storage, mainly due to significantly reduced V oc (1.125 to 1.075 V) and FF (82.17% to 70.96%). In contrast, all the three passivated devices exhibit improved humidity stability, with over 90% of their initial efficiency after 44 days of storage. Moreover, the DACl-passivated device shows the best humidity stability, with about 95% of its initial efficiency (from 22.88% to 21.70%) after 44 days of storage, with slightly reduced J sc (from 24.28 to 23.93 mA cm À2 ) and V oc (from 1.157 to 1.136 V) and high FF (from 81.42% to 80.20%). These results further prove that the DA 2 PbCl 4 2D/3D heterostructure by DACl incorporation can improve the photovoltaic properties and long-term stability of the devices.

Conclusion
In summary, we introduced dodecylammonium halides to treat the 3D perovskite film to form a 2D/3D heterostructure and passivate defects at the perovskite surface. The DACl-treated perovskite formed DA 2 PbCl 4 2D perovskite, while DABr and DAI-treated surfaces formed DA 2 PbI 4 2D perovskite. All the three passivated PSCs show improved performance due to reduced defect traps and improved carrier lifetime through the formation of the 2D/3D heterostructure, while the DA 2 PbCl 4 2D perovskite can passivate perovskite surface defects, enhance the built-in electric field across the film, and optimize the energy alignment between the perovskite film and HTL more effectively, thus contributing to significantly improved open-circuit voltage (V oc ) and enhanced short-current density ( J sc ) and FF of the devices. As a result, a champion PCE of 23.91% was achieved for the DACl-treated PSCs. Meanwhile, the long alkyl chain from the 2D perovskite can increase the hydrophobicity of the bulk perovskite film and form a waterproof layer to suppress the decomposition of the perovskite and
Solution Preparation: The SnO 2 electron transport material solution was prepared by diluting the SnO 2 colloid precursor with deionized water (1:2, volume ratio). The PbI 2 precursor solution was prepared by dissolving 691.5 mg PbI 2 and 19.5 mg (5%) CsI in 1 mL mixture of DMF and DMSO solution (9:1, volume ratio). The organic salt solution was prepared by dissolving GAI (5 mg Device Fabrication: ITO conductive glasses were sequentially cleaned by ultrasonic cleaning for 20 min in glass cleaner, deionized water, ethyl alcohol, and isopropyl alcohol, respectively. Before deposition of SnO 2 electron transport layer (ETL), a 10 min plasma treatment was applied to the ITO substrates to increase the hydrophilicity of the surface. Afterward, SnO 2 was spin coated onto the ITO substrate at 4000 rpm for 20 s, and the SnO 2 film was then annealed in air at 150°C for 15 min. For the PbI 2 layer, the PbI 2 precursor solution was spin coated onto the ETL at 2000 rpm for 30 s and then annealed in glovebox at 70°C for 1 min. The organic salt solution was spin coated onto the PbI 2 film at 1700 rpm for 30 s, and the film was annealed for 15 min at 150°C in ambient air. For the interface passivation of PSCs, 1 mg mL À1 DAX solution (in isopropyl alcohol) was spin coated on the perovskite film at 5000 rpm for 30 s and then annealed in a glovebox at 100°C for 5 min. For HTL, the spiro-OMeTAD solution was spin coated onto the perovskite film at 3000 rpm for 30 s. Finally, a 90 nm Ag electrode was thermally evaporated onto the HTL. For the pure 2D perovskite film, the 2D precursor solution was spin coated onto the ITO substrate at 1000 rpm for 10 s and 4000 rpm for 30 s. At 32 s, 120 μL of anisole as the antisolvent was deposited onto the film. The film was then annealed in a glovebox at 100°C for 30 min.
Characterization: The surface and cross-sectional structures of the samples were obtained via field-emission SEM (JEOL JSM-7610 F plus), operated at an electron beam voltage of 5 kV. The EDX was performed by JEOL JSM-7610F plus. XRD measurements were carried out on a SmartLab X-ray diffractometer (Rigaku, Japan). XPS was performed by the Thermo Fisher Scientific K-alphaþ. UPS results were obtained by Thermo Scientific ESCALAB 250Xi. Steady-state PL spectra were acquired on the XPQY-EQE (CL-01A) equipment (Xipu Optoelectronics Corporation) in the N 2 -filled glovebox. UV-vis absorption spectra were collected on an instrument supplied by Xipu electronics in the glovebox.TRPL was measured by FLS920 (Edinburgh Instruments, Ltd.) with a pulsed excitation at 375 nm. J-V curves were collected by a Keithley 2400 source meter and an AAAgrade solar simulator (Enli. Tec.). The light intensity was calibrated by a NREL-calibrated silicon solar cell equipped with an infrared cutoff filter (KG-5). The tested devices were measured with a black metal mask (0.119 cm 2 ). The test range of reverse scanning was 1.25 to 0 V with a step rate of 0.02 V s À1 . Dark current analysis and SCLC data were collected by the Keithley 2400 in the dark. The EIS were measured on the CHI660E potentiostat system with bias voltage of 1 V. The Mott-Schottky analysis was acquired on the Zahner electrochemical workstation. IPCE was obtained from the Enli QE-R equipment.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.