Interfacial Dipole poly(2-ethyl-2-oxazoline) Modification Triggers Simultaneous Band Alignment and Passivation for Air-Stable Perovskite Solar Cells

To promote the performance of perovskite solar cells (PSCs), its theoretical power conversion efficiency (PCE) and high stability, elaborative defect passivation, and interfacial engineering at the molecular level are required to regulate the optoelectric properties and charge transporting process at the perovskite/hole transport layer (HTL) interfaces. Herein, we introduce for the first time a multifunctional dipole polymer poly(2-ethyl-2-oxazoline) (PEOz) between the perovskite and Spiro-OMeTAD HTL in planar n-i-p PSCs, which advances the PSCs toward both high efficiency and excellent stability by stimulating three beneficial effects. First, the ether–oxygen unshared electron pairs in PEOz chemically react with unsaturated Pb2+ on the perovskite surfaces by forming a strong Pb–O bond, which effectively reduces the uncoordinated defects on the perovskite surfaces and enhances the absorption ability of the resulting PSCs. Second, the dipole induced by PEOz at the perovskite/HTL interface can decrease the HOMO and LUMO level of Spiro-OMeTAD and optimize the band alignment between these layers, thereby suppressing the interfacial recombination and accelerating the hole transport/extraction ability in the cell. Third, the hygroscopic PEOz thin film can protect perovskite film from water erosion by absorbing the water molecules before perovskite does. Finally, the PEOz-modified PSC exhibits an optimized PCE of 21.86%, with a high short-circuit current density (Jsc) of 24.88 mA/cm2, a fill factor (FF) of 0.79, and an open-circuit voltage (Voc) of 1.11 V. The unencapsulated devices also deliver excellent operation stability over 300 h in an ambient atmosphere with a humidity of 30~40% and more than 10 h under thermal stress.


Introduction
Since Miyasaka and co-workers first introduced the organometallic lead perovskite materials as light absorbers in dye-sensitized solar cells [1], hybrid lead halide perovskite solar cells (PSCs) have aroused substantial attention as an emerging next-generation thinfilm photovoltaic technology due to their feasibility of low-temperature processability, remarkable PCE and inexpensive fabrication [2][3][4]. Over the past decade, the PCE of PSCs has dramatically advanced from 3.8% in 2009 to the currently certified value of 25.5%, which is steadily approaching the record efficiency (26.7%) of leader single-junction crystalline silicon solar cells, demonstrating great potential for the delivery of cost-effective solar energy [1,[5][6][7][8]. The ever-increasing efficiency of PSCs predominantly stems from the unique optoelectronic properties of perovskite materials, including excellent hole/electron HTL to tune its energy level and the quality of perovskite film. The results exhibited that the hole trap-state density and noncapacitive current could be effectively reduced and the champion PCE of 17.39% was achieved with the device structure of ITO/PEOz-PEDOT:PSS/CH 3 NH 3 PbI 3 /PCBM/Bphen/Ag [50]. Despite these achievements, we found that PEOz has rarely been employed in the perovskite/HTL interface in n-i-p planar PSCs, which is currently the most popular device model for the exploration of PSCs due to its excellent performance.
In this work, we illustrate for the first time that the dipole polymer PEOz can be employed to passivate the defects on perovskite surfaces and regulate the energy level of spiro-OMeTAD HTL, with a resulting optimization of band alignment for the planar n-i-p structured PSCs. It is found that the oxygen atom in PEOz chemically reacts with coordinatively unsaturated Pb2+ on perovskite films, and thus passivates those uncoordinated Pb2+ defects. Further, the dipole nature of PEOz reduces the HOMO of the Spiro-OMeTAD layer from −5.09 eV to −5.25 eV, resulting in a lower energy barrier between the perovskite and Spiro-OMeTAD layer. Based on these desirable features, further measurements reveal that the dipole PEOz modifier triggers enhanced absorption and photoluminescence lifetime, reduced nonradiative charge carrier recombination, improved charge transport ability as well as enlarged photo-voltage and fill factor (FF) in resulting PSCs. Consequently, the optimized PSC based on PEOz demonstrates a remarkable PCE of 21.86% with a high open-circuit voltage (Voc) of 1.11 V, a short-circuit current density (Jsc) of 24.88 mA/cm 2 , and an FF of 0.79. In addition, the hygroscopic property of PEOz ensures that the perovskite film is highly stable during operation and storage in the ambient air atmosphere via screening water molecules. Therefore, the unencapsulated devices maintain 92.29% of initial efficiency after 300 h of storage in ambient air with relative 30~40% humidity (RH) air. While heated at 85 • C for 10 h in a glovebox, the devices exhibit ca. 17% reduction in efficiency. All these results highlight the multi-functional role of the dipole and hygroscopic PEOz thin film in improving the performance of PSCs.

Fabrication and Characterization of Perovskite Solar Cells
ITO glass substrates were cleaned with detergent, deionized water, acetone, and absolute ethanol in the ultrasonic bath for 20 min, respectively. Then, the ITO substrates were dried with nitrogen and treated in a UV-ozone oven for 15 min. A SnO 2 with a concentration of 3.46 M was spin-coated on ITO substrates at 3000 rpm for 30 s and subsequently annealed at 150 • C for 30 min in air. The substrates with SnO 2 were then transferred into a glove box filled with highly pure N 2 . After that, the spin-coating procedure of the perovskite layer was performed as follows: First, 1.36 M PbI 2 and 0.24 M PbCl 2 were dissolved in the DMF solvent and stirred for 2 h at 75 • C; 70 mg MAI and 30 mg FAI were dissolved in 1 mL IPA with an additional 10 µL DMF. Second, the PbX 2 precursor solution was spin coated on the top of the SnO 2 -covered substrates at 3000 rpm for 45 s, and then the solution of MAI and FAI was spin coated on top of the PbX 2 substrates at 3000 rpm for 45 s. Finally, the samples were annealed at 100 • C for 10 min on a hotplate. After cooling down to room temperature, the PEOz in trifluoroethanol solutions (1 mg/mL) were spin coated onto the surface of the perovskite layer at 5000 rpm for 45 s and annealed at 80 • C for 10 min. Next, the hole transport material (HTM) was spin coated on ITO/SnO 2 /MA 0.7 FA 0.3 PbI 3−x Cl x /PEOz at 1000 rpm for 5 s and 4000 rpm for 45 s using the prepared HTM solution. The HTM solution contained 90 mg of Spiro-OMeTAD in 1 mL of chlorobenzene with 45 µL of LiTFSI/acetonitrile solution (170 mg/mL), 10 µL of tBP, and 75 µL of Co(III) complex FK209/acetonitrile solution (100 mg/mL). Finally, 100-nm-thick Ag layers were thermally evaporated as the top electrodes. The area of the device was 0.07 cm 2 defined by a shadow mask.

Device and Material Characterization
All current density-voltage (J-V) curves were recorded using a Keithley 2400 source meter unit under simulated AM 1.5 G illumination at an intensity of 100 mW/cm 2 with an XES-70S1 solar simulator. The system was calibrated using a NREL certified monocrystal Si photodiode detector before device testing. Scanning electron microscopy (SEM) images were obtained on a JSM-7800F SEM. The X-ray diffraction (XRD) test was conducted on a Bruker D8 Advance XRD. The UV-visible absorption spectra were measured on a Perkin-Elmer Lambda 950 spectrophotometer. The samples were prepared in the same process of device fabrication. The incident photon-to-current conversion efficiency (IPCE) spectra were recorded using a solar cell quantum efficiency measurement system (SCS10-X150, Zolix instrument. Co., Ltd., Beijing, China). Steady photo-luminescence (PL) and time-resolved photo-luminescence (TR-PL) were measured by using the Pico Quant Fluotime 300 with a 510 nm picosecond pulsed laser. Transient photocurrent (TPC) measurement was performed with a system excited by a 532 nm (1000 Hz, 3.2 ns) pulse laser. Transient photovoltage (TPV) measurement was performed with the same system excited by a 405 nm (50 Hz, 20 ms) pulse laser. A digital oscilloscope (Tektronix, D4105) was used to record the photocurrent or photovoltage decay process with a sampling resistor of 50 Ω or 1 MΩ, respectively. All the measurements of the solar cells were performed under an ambient atmosphere at room temperature without encapsulation.

Results and Discussion
3.1. Interactions of PEOz with MA 0.7 FA 0.3 PbI 3−x Cl x As presented in Figure 1a,b, the PEOz in trifluoroethanol solution was deposited on perovskite film by the spin-coating method, and after being annealed at 80 • C for 10 min, the PEOz modified samples were subjected to the follow-up measurements. In order to find the optimal passivation condition, three different concentrations of PEOz solution (0.5, 1, 1.5 mg/mL) were employed in this study. Noting that during the photocurrent density-voltage (J-V) measurements for PSCs based on different concentrations of PEOz with the device structure of ITO/SnO 2 /MA 0.7 FA 0.3 PbI 3−x Clx/PEOz/Spiro-OMeTAD/Ag, the device with 1 mg/mL PEOz yields the highest PCE of 21.86% because of its outstanding Jsc and FF ( Figure S1 and Table S1). So, in this work, we use the optimal PEOz concentration of 1 mg/mL in trifluoroethanol to fabricate the film and device samples for the following characterization and discussion.
To obtain an in-depth understanding of the interaction between the PEOz and perovskite film, X-ray photoelectron spectroscopy (XPS) measurements for pristine perovskite and the perovskite treated by 1 mg/mL PEOz were carried out. For simplicity, the pristine perovskite film and the control cell are named "W/O", and the ones treated with PEOz are labeled as "W/". The chemical state changes of the perovskite film induced by PEOz are probed, and the Pb 4f and O 1s spectra are shown in Figure 1c,d, respectively. For pristine perovskite film, two clear core-level Pb 4f characteristic peaks centered at 138.38 eV and 143.30 eV are observed, corresponding to Pb 4f 7/2 and Pb 4f 5/2 , respectively. While the O 1s peak located at 531.24 can also be detected from the pristine perovskite, which primarily originated from the absorbed oxygen on the perovskite surfaces. After PEOz treatment the two Pb 4f peaks shift toward higher binding energy centered at 138.55 eV and 143.41 eV, and the O 1s peak shifts to lower binding energy by 210 meV, manifesting the increase in the positive charge on the Pb element and negative charge on the O element [51]. We ascribe this phenomenon to the strong coordinate bonding between the O from PEOz chains and Pb 2+ ions from the perovskite surfaces. It is noted that the O in PEOz contains a lone pair of electrons, which makes it easy to form a chemical bond with Pb 2+ ions. Moreover, the high electronegativity of O leads to a partial charge transfer from Pb to the PEOz; the proposed working mechanism is schematically presented in Figure 1a [52]. The formation of Pb-O bonds certainly will passivate the uncoordinated surface Pb 2+ centers on perovskite film and repress the nonradiative recombination of charge carriers [53]. To obtain an in-depth understanding of the interaction between the PEOz and perovskite film, X-ray photoelectron spectroscopy (XPS) measurements for pristine perovskite and the perovskite treated by 1 mg/mL PEOz were carried out. For simplicity, the pristine perovskite film and the control cell are named "W/O", and the ones treated with PEOz are labeled as "W/". The chemical state changes of the perovskite film induced by PEOz are probed, and the Pb 4f and O 1s spectra are shown in Figure 1c,d, respectively. For pristine perovskite film, two clear core-level Pb 4f characteristic peaks centered at 138.38 eV and 143.30 eV are observed, corresponding to Pb 4f7/2 and Pb 4f5/2, respectively. While the O 1s peak located at 531.24 can also be detected from the pristine perovskite, which primarily originated from the absorbed oxygen on the perovskite surfaces. After PEOz treatment the two Pb 4f peaks shift toward higher binding energy centered at 138.55 eV and 143.41 eV, and the O 1s peak shifts to lower binding energy by 210 meV, manifesting the increase in the positive charge on the Pb element and negative charge on the O element [51]. We ascribe this phenomenon to the strong coordinate bonding between the O from PEOz chains and Pb 2+ ions from the perovskite surfaces. It is noted that the O in PEOz contains a lone pair of electrons, which makes it easy to form a chemical bond with Pb 2+ ions. Moreover, the high electronegativity of O leads to a partial charge transfer from Pb to the PEOz; the proposed working mechanism is schematically presented in Figure 1a [52]. The formation of Pb-O bonds certainly will passivate the uncoordinated surface Pb 2+

Characterization of MA 0.7 FA 0.3 PbI 3−x Cl x Films with PEOz
The effect of PEOz on the morphological properties of the MA 0.7 FA 0.3 PbI 3−x Cl x perovskite film is investigated by scanning electron microscopy (SEM) and atomic force microscopy (AFM), respectively. From the top surface SEM images in Figure 2a,b, we can find that both perovskite films present a closely packed surface morphology with triple junction grain boundaries. There is no great difference in grain size or any discernible pinholes observed for the two perovskite films. While after the PEOz treatment, some white extraneous substances with visible contrasts appear at the MA 0.7 FA 0.3 PbI 3−x Cl x grain boundaries, as indicated in Figure 2b. These substances are most likely assigned to the introduced PEOz since no PbI 2 diffraction signals are detected in the X-ray diffraction (XRD) patterns for the PEOz-treated perovskite film (see Figure 3b). It should be noted that deep-level defects usually exist in grain boundaries that cause fast non-radiation carrier recombination, so the PEOz located at grain boundaries should contribute to bringing down the boundary defects and related energy losses of the perovskite film and help to realize more compact  Figure 2c,d confirm the morphology evolution of the perovskite film with PEOz. The root-mean-square (RMS) roughness values of the MA 0.7 FA 0.3 PbI 3−x Cl x films decrease from 12.39 nm to 9.99 nm after 1 mg/mL PEOz treatment as a result of the passivation of boundary defects. The smoother perovskite surface with PEOz could enable better interface contacts between perovskite and upper HTLs, providing a more effective channel for carrier separation and extraction.
neous substances with visible contrasts appear at the MA0.7FA0.3PbI3−xClx grain boundaries, as indicated in Figure 2b. These substances are most likely assigned to the introduced PEOz since no PbI2 diffraction signals are detected in the X-ray diffraction (XRD) patterns for the PEOz-treated perovskite film (see Figure 3b). It should be noted that deep-level defects usually exist in grain boundaries that cause fast non-radiation carrier recombination, so the PEOz located at grain boundaries should contribute to bringing down the boundary defects and related energy losses of the perovskite film and help to realize more compact crystalline grains with smoother surfaces. The AFM results in Figure 2c,d confirm the morphology evolution of the perovskite film with PEOz. The root-mean-square (RMS) roughness values of the MA0.7FA0.3PbI3−xClx films decrease from 12.39 nm to 9.99 nm after 1 mg/mL PEOz treatment as a result of the passivation of boundary defects. The smoother perovskite surface with PEOz could enable better interface contacts between perovskite and upper HTLs, providing a more effective channel for carrier separation and extraction. Next, the influence of PEOz introduction on the optical properties and crystallographic structure of the MA0.7FA0.3PbI3−xClx films are studied. Figure 3a shows the UV-vis spectra of the perovskite film with and without PEOz. The PEOz-treated perovskite film obtains higher absorption intensities over the whole wavelength range, especially from 400 to 500 nm, indicating its better light absorption capacity induced by the improved film quality, which may make a great contribution to Jsc of PEOz-passivated PSC. In addition, the two films exhibit a similar typical absorption edge of ~785 nm that corresponds to the bandgap of MA0.7FA0.3PbI3−xClx (1.58 eV) [45]. The results demonstrate that the introduction of PEOz does not affect the band gap of MA0.7FA0.3PbI3−xClx perovskite, which is further verified by the XRD measurements. As shown in Figure 3b, both films exhibited diffraction peaks at 14.18°, 28.50°, and 31.98°, indexed to the (110), (220), and (310) planes of the MA0.7FA0.3PbI3−xClx crystals, respectively [44,45]. In addition, the intensities of the dominant diffraction peaks of MA0.7FA0.3PbI3−xClx film are clearly increased with the treatment of PEOz, and neither diffraction signals for PbI2 nor PEOz are detected, which indicates the better crystallinity of MA0.7FA0.3PbI3−xClx perovskite with fewer grain boundaries and intergranular defects. From these results, we can conclude that the PEOz treatment can beneficially passivate the surface and boundary defects, and hence boost the film quality of MA0.7FA0.3PbI3−xClx without affecting its crystal structure. To further verify the above inference, we measured the steady-state photoluminescence (PL) spectra and time-resolved photoluminescence (TRPL) spectra for perovskite films with and without PEOz treatment. The perovskite films were grown directly on insulated glass substrates. As shown in Figure 3c, the PEOz-treated perovskite film exhibits stronger PL intensity around 790 nm compared to the pristine one, indicating the en-  Next, the influence of PEOz introduction on the optical properties and crystallographic structure of the MA 0.7 FA 0.3 PbI 3−x Cl x films are studied. Figure 3a shows the UV-vis spectra of the perovskite film with and without PEOz. The PEOz-treated perovskite film obtains higher absorption intensities over the whole wavelength range, especially from 400 to 500 nm, indicating its better light absorption capacity induced by the improved film quality, which may make a great contribution to J sc of PEOz-passivated PSC. In addition, the two films exhibit a similar typical absorption edge of~785 nm that corresponds to the bandgap of MA 0.7 FA 0.3 PbI 3−x Cl x (1.58 eV) [45]. The results demonstrate that the introduction of PEOz does not affect the band gap of MA 0.7 FA 0.3 PbI 3−x Cl x perovskite, which is further verified by the XRD measurements. As shown in Figure 3b, both films exhibited diffraction peaks at 14.18 • , 28.50 • , and 31.98 • , indexed to the (110), (220), and (310) planes of the MA 0.7 FA 0.3 PbI 3−x Cl x crystals, respectively [44,45]. In addition, the intensities of the dominant diffraction peaks of MA 0.7 FA 0.3 PbI 3−x Cl x film are clearly increased with the treatment of PEOz, and neither diffraction signals for PbI 2 nor PEOz are detected, which indicates the better crystallinity of MA 0.7 FA 0.3 PbI 3−x Cl x perovskite with fewer grain boundaries and intergranular defects. From these results, we can conclude that the PEOz treatment can beneficially passivate the surface and boundary defects, and hence boost the film quality of MA 0.7 FA 0.3 PbI 3−x Cl x without affecting its crystal structure.
To further verify the above inference, we measured the steady-state photoluminescence (PL) spectra and time-resolved photoluminescence (TRPL) spectra for perovskite films with and without PEOz treatment. The perovskite films were grown directly on insulated glass substrates. As shown in Figure 3c, the PEOz-treated perovskite film exhibits stronger PL intensity around 790 nm compared to the pristine one, indicating the enhanced radiative recombination and weaker nonradiative recombination in it. This can be further evidenced by the TRPL results in Figure 3d, in which the semi-quantitative information about the nonradiative recombination of charge carriers is calculated with a bi-exponential decay function. Obviously, the PEOz-treated film yields the average carrier lifetime (τ ave ) of 205.93 ns, which is higher than that of the pristine perovskite film (178.43 ns). The elongated PL decay time of the former means a lower defect density and hence superior photoelectric properties for the PEOz-treated perovskite films. This is primarily ascribed to the strong coordinate bonding between PEOz and the perovskite, which beneficially passivates the defects on the perovskite surfaces and boundaries by effective repression of nonradiative recombination paths.

Effect of PEOz on the PSCs Photovoltaic Performance
The device structure used in this experiment is shown in Figure 4a and the polymer PEOz is introduced between the perovskite light-absorbing layer and the hole transport layer. The dipole effect of PEOz makes the HOMO energy level difference between the perovskite layer and the HTL 1.0 eV is reduced to 0.7 eV (Figure 4b) [54,55]. The lower the energy level difference, the more beneficial it is to inhibit electron-hole recombination and accelerate hole extraction, which is of great significance to the collection of photogenerated carriers and the improvement of battery performance.
To further explore the influence of PEOz on the HOMO energy of the perovskite layer and the HTL, we conducted ultraviolet photoelectron spectroscopy (UPS) tests. As indicated in Figure 4c,d, the HOMO energy level of the pristine perovskite film and PEOztreated perovskite film is 6.09 eV and 5.95 eV, respectively, according to the formula of HOMO = 21.22 − (Ecutoff − Eonset) [13]. The HOMO energy levels of the Spiro-OMeTAD film without and with a PEOz under layer are 5.09 eV and 5.25 eV, respectively (Figure 4e,f). The simultaneous change of HOMO energy levels of perovskite and Spiro-OMeTAD could promote the favorable band regulation between the perovskite and Spiro-OMeTAD layer with the corresponding energy barrier reduced by 0.3 eV, while a better hole transport with restrained interfacial non-radiative recombination can be expected, which further facilitates the increase in FF and Voc in the cell [13,56].
PEOz is introduced between the perovskite light-absorbing layer and the hole transpor layer. The dipole effect of PEOz makes the HOMO energy level difference between th perovskite layer and the HTL 1.0 eV is reduced to 0.7 eV (Figure 4b) [54,55]. The lower th energy level difference, the more beneficial it is to inhibit electron-hole recombination and accelerate hole extraction, which is of great significance to the collection of photogener ated carriers and the improvement of battery performance. To further explore the influence of PEOz on the HOMO energy of the perovskite laye and the HTL, we conducted ultraviolet photoelectron spectroscopy (UPS) tests. As indi cated in Figure 4c,d, the HOMO energy level of the pristine perovskite film and PEOz treated perovskite film is 6.09 eV and 5.95 eV, respectively, according to the formula o HOMO = 21.22 − (Ecutoff − Eonset) [13]. The HOMO energy levels of the Spiro-OMeTAD film without and with a PEOz under layer are 5.09 eV and 5.25 eV, respectively (Figur 4e,f). The simultaneous change of HOMO energy levels of perovskite and Spiro-OMeTAD could promote the favorable band regulation between the perovskite and Spiro-OMeTAD layer with the corresponding energy barrier reduced by 0.3 eV, while a better hol transport with restrained interfacial non-radiative recombination can be expected, whic further facilitates the increase in FF and Voc in the cell [13,56].
Taking into account the defect passivation effect of PEOz on the uncoordinated Pb2 of the perovskite film, the perovskite/HTL HOMO energy level is more matched, whic has a beneficial effect on the extraction and transport of holes; we use PEOz solution t process perovskite films, which has been used to prepare a device with a structure o ITO/SnO2/MA0.7FA0.3PbI3−xClx/PEOz/Spiro-OMeTAD/Ag. Figure 5a shows that the devic obtained by modifying the perovskite with 1 mg/mL PEOz solution exhibits an optimall high efficiency of 21.86%, which is about 21% higher than the 18.14% of the comparativ sample. Moreover, the detailed performance parameters are shown in Table 1  Taking into account the defect passivation effect of PEOz on the uncoordinated Pb2+ of the perovskite film, the perovskite/HTL HOMO energy level is more matched, which has a beneficial effect on the extraction and transport of holes; we use PEOz solution to process perovskite films, which has been used to prepare a device with a structure of ITO/SnO 2 /MA 0.7 FA 0.3 PbI 3−x Cl x /PEOz/Spiro-OMeTAD/Ag. Figure 5a shows that the device obtained by modifying the perovskite with 1 mg/mL PEOz solution exhibits an optimally high efficiency of 21.86%, which is about 21% higher than the 18.14% of the comparative sample. Moreover, the detailed performance parameters are shown in Table 1, a Jsc of 24.88 mA/cm 2 , a Voc of 1.11V, an FF as high as 79.32% far above the control devices, a Jsc of 22.22 mA/cm 2 , a Voc of 1.09 V, and an FF of 74.56%. These results verify the abovementioned dual effect of PEOz. In Figures 5b and S2, the steady output current densities were examined for the PSC treated by PEOz and the pristine PSC under the condition of maximum power points (MPPs, 0.93 and 0.92 V) and are 22.17 and 19.27 mA/cm 2 , with the corresponding PCE of 20.62% and 17.72%, respectively. The two stable PCE values are close to those calculated from the J-V curve, which not only effectively confirms the reliability of the J-V test, but also convincingly demonstrates that the introduction of PEOz has significantly improved the performance of PSC. In addition, the incident photo-to-current conversion efficiency (IPCE) measurements were performed. The results in Figure 5c manifest that the PSC treated by PEOz exhibits higher photon-electron transformation efficiencies over the whole response wavelength region, revealing a better perovskite/HTL interface for the extraction and transmission of charge carriers. What is more noteworthy is that, especially from 350 nm to 450 nm, PSC treated with PEOz has more obvious photo-tocurrent conversion efficiency, which is almost the same as the UV-vis spectra of Figure 3a, which more effectively proves that PEOz can improve the film quality and enhance its light absorption capacity. manifest that the PSC treated by PEOz exhibits higher photon-electron transformat ficiencies over the whole response wavelength region, revealing a better perovskite interface for the extraction and transmission of charge carriers. What is more notew is that, especially from 350 nm to 450 nm, PSC treated with PEOz has more obvious p to-current conversion efficiency, which is almost the same as the UV-vis spectra of F 3a, which more effectively proves that PEOz can improve the film quality and enha light absorption capacity. During the preparation of the device, it is affected by the gas atmosphere, tem ture, and heating time of the cross-solvent solution in the glove box, resulting in the lem of poor repeatability of the device. To evaluate the reproducibility of PSCs, we ure the PCE of 15 PSCs from three batches fabricated with pristine and treated by PSCs. The statistical results in Figure 5d and Table 1 show that the average PCE va increased from 17.34 ± 0.80 to 21.17 ± 0.69, after processing with PEOz. Meanwhi distribution deviation of PSCs treated by PEOz is reduced by 16% compared wi pristine PSCs. In addition, as shown in Figure S3 and Table 1, the average Jsc, Voc and the PSCs treated by PEOz are about 24.70 ± 0.18 mA/cm 2 , 1.10 ± 0.01 V and 77.35 higher than that of the pristine PSCs, which have an average Jsc, Voc and FF of abou ± 1.45 mA/cm 2 , 1.04 ± 0.05 V and 71.44 ± 3.12, respectively. These data are in accor with the J-V characteristics shown in Figure 5a, indicating a good reproducibility better performance of PEOz-treated devices. The hysteresis behavior of the PSCs wa  During the preparation of the device, it is affected by the gas atmosphere, temperature, and heating time of the cross-solvent solution in the glove box, resulting in the problem of poor repeatability of the device. To evaluate the reproducibility of PSCs, we measure the PCE of 15 PSCs from three batches fabricated with pristine and treated by PEOz PSCs. The statistical results in Figure 5d and Table 1 show that the average PCE value is increased from 17.34 ± 0.80 to 21.17 ± 0.69, after processing with PEOz. Meanwhile, the distribution deviation of PSCs treated by PEOz is reduced by 16% compared with the pristine PSCs. In addition, as shown in Figure S3 and Table 1, the average J sc , V oc and FF of the PSCs treated by PEOz are about 24.70 ± 0.18 mA/cm 2 , 1.10 ± 0.01 V and 77.35 ± 1.97 higher than that of the pristine PSCs, which have an average J sc , V oc and FF of about 20.77 ± 1.45 mA/cm 2 , 1.04 ± 0.05 V and 71.44 ± 3.12, respectively. These data are in accordance with the J-V characteristics shown in Figure 5a, indicating a good reproducibility and a better performance of PEOz-treated devices. The hysteresis behavior of the PSCs was also improved after surface passivation by PEOz ( Figure S4). Moreover, the hysteresis index (HI) is usually used to measure the hysteresis effect in PSCs. The specific formula is shown in Formula (1) [57].
where J scan + (0.8V oc ) and J scan − (0.8V oc ) represent the photocurrent density of 0.8 V in forward scanning and reverse scanning, respectively. The results show that the device with the PEOz layer achieved the lowest hysteresis index (2.6%), compared to the control device (4.9%). The improvement of hysteresis may be ascribed to the passivation layer of PEOz which decreased the surface traps of MA 0.7 FA 0.3 PbI 3−x Cl x , thereby reducing the nonradiative recombination. This also proves that the PEOz at the grain boundary shown by the SEM in Figure 2 helps to reduce the boundary defects and related energy loss of the perovskite film. The photocurrent (TPC) and photovoltage (TPV) tests are used to analyze the carrier dynamics of PSCs with PEOz and W/O PEOz [15,44]. The TPC results tested under shortcircuit conditions are shown in Figure 6a, the decay time of the PSCs introduced with PEOz is 1.58 µs, which is 0.64 µs lower than the 2.22 µs of the control devices. A smaller decay time indicates more efficient carrier extraction and transmission, which may be the reason for the obvious increase in J sc in Figure 5a. Meanwhile, Figure 6b shows the TPV results measured in the open-circuit state, compared with the basic devices, the photovoltage decay value of the PEOz-treated devices is increased from 0.22 ms to 0.37 ms. This result supports the weaker recombination of charge carriers in it, indicating the better perovskite crystallization quality by the perovskite film modification of the PEOz layer. We further tested the Mott-Schottky (M-S) and the electrochemical impedance sp tra (EIS) of the PSCs with W/O PEOz in the dark state, so as to explore the carrier reco bination behavior on a deeper level. In Figure 6c, C is the dark state capacitance, and is the built-in potential. Since the introduction of PEOz can make the energy levels of perovskite/HTL more matched, the PSCs treated with PEOz have a higher Vbi, which provide a greater separation force for photogenerated carriers, thereby inhibiting th recombination. It can be seen from the figure that the Vbi of the PSCs treated by PEOz We further tested the Mott-Schottky (M-S) and the electrochemical impedance spectra (EIS) of the PSCs with W/O PEOz in the dark state, so as to explore the carrier recombination behavior on a deeper level. In Figure 6c, C is the dark state capacitance, and V bi is the built-in potential. Since the introduction of PEOz can make the energy levels of the perovskite/HTL more matched, the PSCs treated with PEOz have a higher V bi , which can provide a greater separation force for photogenerated carriers, thereby inhibiting their recombination. It can be seen from the figure that the V bi of the PSCs treated by PEOz has increased from 0.98 V to 1.04 V. A higher V bi is conducive to the improvement of the device V oc , which is also consistent with the V oc result on the J-V curve of Figure 5a. The EIS in Figure 6d is mainly divided into two parts, the first part is that the large semicircle of the low-frequency part is related to the recombination resistance (Rrec) of the PSCs, and a large Rrec means a low carrier composite probability [15,45,58]. It can be seen that the Rrec of PSCs with PEOz is increased from 263 Ω to 538 Ω, indicating that PEOz can make the separation of photogenerated carriers of PSCs easier, resulting in the weaker internal recombination of PSCs. This is mainly because PEOz can effectively passivate defects in MA 0.7 FA 0.3 PbI 3−x Cl x films, which is also consistent with the TRPL analysis results in Figure 3d. The second part is the series resistance (Rs) of the high frequency part, which is also an important parameter. Compared with the basic device (307 Ω), the PEOz-treated device has lower Rs (50 Ω), which shows that PEOz can effectively reduce the series resistance of PSCs, to improve the FF value and improve the performance of the battery.

Effect of PEOz on the PSCs Stability
Finally, in order to study the effect of PEOz modification on the stability of the perovskite device, We observed the optical image of the degradation of the perovskite film at 30~40% relative humidity (RH) within 33 days (Figure 7a). It could be observed that small white dots started to appear in the pristine perovskite film the next day, these dots continued to expand until the black film was completely hydrolyzed into white at about 33 days. By contrast, it was until 33 days that the white dots begun to exist on the PEOz passivated perovskite film, showing higher water stability. Meanwhile, as shown in Figure S6, the contact angles of water on the perovskite film W/O PEOz and with PEOz are measured to be 60.92 • , and 39.62 • , respectively, suggesting that PEOz is hygroscopic, which may absorb water molecules in ambient atmosphere preferentially before the perovskite does, slowing down the decomposition of the perovskite layer.  In addition, we compared the thermal stability of an unpackaged battery in a glo box with water content of less than 1ppm and a temperature of 85 °C. As shown in Fig  7b, the device with PEOz can still maintain an initial efficiency of 83.13% after heating 10 h, while the device W/O PEOz shows only 79.05% of its initial efficiency under the sa In addition, we compared the thermal stability of an unpackaged battery in a glovebox with water content of less than 1ppm and a temperature of 85 • C. As shown in Figure 7b, the device with PEOz can still maintain an initial efficiency of 83.13% after heating for 10 h, while the device W/O PEOz shows only 79.05% of its initial efficiency under the same storing condition. This may be due to poor crystal quality instability of MA 0.7 FA 0.3 PbI 3−x Cl x exposed at high temperatures [59]. We also placed the unpackaged battery in a drying cabinet with a humidity of 30~40% and a temperature of 25 • C. In Figure 7c, the device modified by PEOz can still maintain an initial efficiency of 92.29% after storage for 300 h. By contrast, under the same conditions, the initial efficiency of the device W/O PEOz dropped to 83.46%. These results indicate that PEOz can ensure the stability of PSCs in a humid and hot environment, and effectively increase the service life of the device. This is primarily due to the PEOz wrapped around the perovskite grains since the decomposition of PSCs generally starts from the grain boundaries. With the preservation of grain boundaries to not being exposed to ambient atmosphere, the decomposition of PSCs is effectively suppressed [60].

Conclusions
In summary, we demonstrate for the first time, the successful application of multifunctional dipolar polymer PEOz as a simultaneous interfacial modifier and passivation agent at the perovskite/HTL interface in n-i-p planar PSCs. It is found that the lone pair electrons of the oxygen atoms in PEOz can bond with the uncoordinated lead ions on the perovskite surfaces and GBs, which effectively passivates the defect sites, reduces the trap density in perovskite film, and thereby facilitates the absorption ability and photoluminescence lifetime of the resulting PSCs. We also illustrate that the dipole of PEOz provides an optimized interface band alignment and boosts the hole transport/extraction couple with low non-radiation recombination at the perovskite/HTL interface. Moreover, the hygroscopic property of PEOz on top of the perovskite film contributes to the high stability of the PSCs during operation and under ambient air atmosphere. Consequently, with the above three benefits, we realized an optimized PSC efficiency of 21.83%, with a Voc of 1.11 V, Jsc of 24.88 mA/cm 2 , and FF of 0.79, much higher than the control device. Remarkably, the non-encapsulated PEOz-based PSCs show outstanding long-term stability against moisture and thermal stresses, retaining 92.29% or 83.13% of its initial PCE value after 300 h of storage in ambient air with relative 30~40% humidity (RH) air, or being heated at 85 • C for 10 h in a glovebox, respectively. As a result, these findings contribute to providing an important guide for effectively improving the performance and environmental stability of the perovskite materials and the resulting optoelectronic devices.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/polym14132748/s1, Figure S1: J-V characteristics of best-performing control device and devices fabricated with different concentrations of PEOz under AM 1.5G illumination (100 mW cm −2 ); Figure S2: Steady-state photocurrent and PCE output characteristics of the control device; Figure S3: Statistical results of (a) Jsc, (b) Voc and (c) FF for the device without and with PEOz, respectively; Figure S4: J-V curves with different scanning directions of the PSCs (a) without and (b) with PEOz layer, respectively; Figure S5: Water contact angles of the MA 0.7 FA 0.3 PbI 3−x Cl x films (a) without and (b) with PEOz deposition, respectively; Table S1: Photovoltaic parameters of best-performing control device and device prepared with different concentrations of PEOz.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no competing financial interest.