Dual-Resistance of Ion Migration and Moisture Erosion via Hydrolytic Crosslinking of Siloxane Functionalized Poly(Ionic Liquids) for Efficient and Stable Perovskite Solar Cells

Open AccessCCS ChemistryRESEARCH ARTICLE16 Jun 2022Dual-Resistance of Ion Migration and Moisture Erosion via Hydrolytic Crosslinking of Siloxane Functionalized Poly(Ionic Liquids) for Efficient and Stable Perovskite Solar Cells Lingyun Gong†, Jia Yang†, Wangping Sheng, Yang Zhong, Yang Su, Licheng Tan and Yiwang Chen Lingyun Gong† College of Chemistry and Chemical Engineering, Institute of Polymers and Energy Chemistry (IPEC), Nanchang University, Nanchang 330031 , Jia Yang† College of Chemistry and Chemical Engineering, Institute of Polymers and Energy Chemistry (IPEC), Nanchang University, Nanchang 330031 , Wangping Sheng College of Chemistry and Chemical Engineering, Institute of Polymers and Energy Chemistry (IPEC), Nanchang University, Nanchang 330031 , Yang Zhong College of Chemistry and Chemical Engineering, Institute of Polymers and Energy Chemistry (IPEC), Nanchang University, Nanchang 330031 , Yang Su College of Chemistry and Chemical Engineering, Institute of Polymers and Energy Chemistry (IPEC), Nanchang University, Nanchang 330031 , Licheng Tan *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] College of Chemistry and Chemical Engineering, Institute of Polymers and Energy Chemistry (IPEC), Nanchang University, Nanchang 330031 and Yiwang Chen *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] College of Chemistry and Chemical Engineering, Institute of Polymers and Energy Chemistry (IPEC), Nanchang University, Nanchang 330031 National Engineering Research Center for Carbohydrate Synthesis/Key Lab of Fluorine and Silicon for Energy Materials and Chemistry of Ministry of Education, Jiangxi Normal University, Nanchang 330022 https://doi.org/10.31635/ccschem.022.202201871 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail The inevitable ion migration that occurs within ionic polycrystalline perovskite film results in inferior long-term stability of perovskite solar cells (PVSCs) that cannot meet the commercial requirements. Here, a novel poly(ionic liquid) named poly-1-vinyl-3-propyltrimethoxysilane imidazolium chloride (PImIL-SiO) is first introduced into perovskite to strengthen grain boundaries (GBs) and construct dual-functional barriers against internal ion migration and external moisture erosion for fabricating highly efficient and stable PVSCs. PImIL-SiO-containing imidazolium cations and pendant siloxane groups contribute to passivation of bulk defects and anchoring of GBs, which effectively hinders ion migration channels, thus reducing perovskite film phase separation and device hysteresis. Furthermore, the intrinsically hydrophobic PImIL-SiO automatically forms a secondary protective barrier to endow the perovskite film with ultrahigh moisture corrosion resistance through the hydrolyzation reaction of siloxane with the permeated moisture. Consequently, the PImIL-SiO-modified PVSCs achieve a champion power conversion efficiency (PCE) of 22.46%, accompanied by excellent thermal and humidity stabilities where the non-encapsulated devices retain 87% of the initial PCE after aging at 85 °C for 250 h and >85% of the initial PCE over 1100 h in air with a relative humidity of 50–70%. Download figure Download PowerPoint Introduction Over the past several years, perovskite solar cells (PVSCs) have made astounding advancements with the power conversion efficiency (PCE) continuously improving from 3.8% to 25.5%.1–5 However, long-term stability has emerged as a critical obstruction to the wholesale application of perovskite photovoltaics. In addition to being easily degraded by extraneous factors, for instance, illumination, moisture, oxygen, and heat, ion migration in perovskite is also a crucial reason for perovskite degeneration.6–10 Significantly, ion migration is inevitable as a result of the ionic polycrystalline nature of perovskite film, which undoubtedly causes severe phase segregation, structural destruction, and associated device hysteresis.11–13 Because of the quick crystallization rate during the process of perovskite formation, a large density of point defects (Schottky defects) can generate and provide migration paths for the movable ions.14 Additionally, grain boundaries (GBs) in polycrystalline film are demonstrated to offer ample space for ion movement and dominate the migration process.15,16 Therefore, understanding the ionic migration characteristics and achieving the GBs passivation together with ion immobilization are of great importance for highly efficient and stable PVSCs. To date, various strategies have been developed to passivate defects and inhibit ion migration, such as perovskite composition engineering, anti-solvent engineering, and additive engineering.17–19 Among them, ionic liquids (ILs) are widely applied as additives due to their excellent properties of wide electrochemical windows and good conductivities, which have made significant contributions to improving the quality and stability of perovskite film.20–25 Kim et al.26 has applied imidazolium (IM)-based IL IMBF4 to obtain excellent device efficiency and stability, where IM cations can efficiently passivate defects at GBs and BF4 anions can relax lattice strain simultaneously. Lin et al.27 has incorporated piperidinium-based IL [BMP]+[BF4]− to prevent phase segregation and decrease pinhole formation in the perovskite absorber layer, enhancing the open-circuit voltage and device efficiency. However, the stability of perovskite with such small molecular ILs is insufficient to satisfy the demands of practical applications because these ILs are inadequate to anchor GBs and are supposed to migrate with ions in perovskite film.28,29 Furthermore, perovskites modified by ILs with hydrophilic groups are more susceptible to moisture erosion caused by the IL water absorption, which is not conducive to the durable stability of perovskite devices. To overcome above issues, we employed poly-1-vinyl-3-propyltrimethoxysilane imidazolium chloride (PImIL-SiO) as a novel poly(ionic liquid) (PIL) additive to reduce defect state density and restrain ion migration in perovskite for preparing highly efficient and stable PVSCs. The supportive interaction between PImIL-SiO and uncoordinated Pb2+ can provide the synergic actions of defects passivation and lattice stress release, which significantly improves the quality of perovskite film.30 Meanwhile, the long-chain PImIL-SiO can effectively anchor GBs and hinder ion migration channels to reduce phase separation of perovskite film and device hysteresis, which ultimately enhances the light and thermal stabilities of PVSCs. Furthermore, the introduction of siloxane-based pendant groups into PILs can improve their intrinsic hydrophobicity, and further construct a secondary protective barrier against moisture erosion by the hydrolyzation reaction of PImIL-SiO with permeated moisture at the GBs,31–33 which improves the humidity stability of PVSCs. Consequently, the unencapsulated devices based on PImIL-SiO possess excellent thermal and moisture stabilities for maintaining nearly 87% of the original efficiency after aging in N2 at 85 °C for 250 h and >85% of the initial efficiency after aging in relative humidity (RH) of 50–70% at 25 °C for 1100 h. Experimental Methods Materials N,N-dimethylformamide (DMF, 99.8% purity), dimethyl sulfoxide (DMSO, 99.9% purity), acetonitrile (99.8% purity), chlorobenzene (CB, 99.8% purity), and 4-tert-butyl pyridine were purchased from Sigma-Aldrich (United States) and used as received without further purification. Lead iodide (PbI2, 99.9985% purity), cesium iodide (CsI, 99.99% purity), tin(IV) oxide (15% in H2O colloidal dispersion liquid), and lithium bis(trifluoromethylsulfonyl) imide (LiTFSI, >98% purity) were purchased from Alfa Aesar (United Kingdom). Formamidinium iodide (FAI, >99.5% purity), methylammonium chloride (MACl, >99.5% purity), and methylammonium bromide (MABr, >99.5% purity) were purchased from Xi'an Polymer Light Technology Corp. (Xi'an, Shaanxi, China). 2,2′,7,7′-Tetrakis [N,N-di(4-methoxyphenyl) amino]-9,9′-spirobifluorene (spiro-OMeTAD, 99% purity) was purchased from Luminescence Technology Corp. (Taiwan, China). 2,2′-Azobis (2-methylpropionitrile) (AIBN), (3-chloropropyl)triethoxysilane, and 1-vinylimidazole were purchased from Innochem (Beijing, China). 1-Vinyl-3-methylimidazolium chloride was purchased from Energy Chemical (Shanghai, China). Indium tin oxide (ITO) (transmission > 95%) substrates were purchased from South China Science & Technology Company Ltd. (Shenzhen, China). Synthesis of PImIL-SiO (3-Chloropropyl)triethoxysilane (0.1 mol) and 0.1 mol 1-vinylimidazole were put into the reactor and reacted at 120 °C for 24 h. After the product was washed with ethyl acetate three times, the residual solvent was removed in a vacuum atmosphere, and the liquid product ImIL-SiO was quantitatively obtained. The ImIL-SiO and AIBN (as initiator, the mass ratio of initiator to ImIL-SiO is 1%) were dissolved in ethanol in a three-necked flask, then stirred and kept at 120 °C for 12 h in a nitrogen atmosphere. The reaction solution was washed three times with ethyl acetate until the washing solution was not turbid and then rotary evaporated. The resulting thick brownish-black liquid was PImIL-SiO. Preparation of precursor solutions A mixture of 1.6 M PbI2, 1.3 M FAI, 0.14 M MABr, 0.3 M MACl, and 0.076 M CsI in 1 mL of a mixed solvent of DMF/DMSO (volume ratio of 4:1) was used to prepare the Cs0.05FA0.85MA0.10Pb(I0.97Br0.03)3 perovskite precursor solution. The perovskite precursor solutions were stirred overnight in a nitrogen-filled glove box and used without any further treatment. PVSCs fabrication First, the ITO glass substrates were cleaned by sequential ultrasonic treatment with detergent solution, deionized water, and isopropyl alcohol for 20 min. The ITO glass substrates were dried by nitrogen (N2) flow and treated under UV-ozone plasma for 10 min. SnO2 solution was prepared by dispersing in deionized water (volume ratio of 1∶3), which was spin-coated on ITO at 3000 rpm for 30 s, then annealed at 150 °C for 30 min in air. Then the fully dissolved perovskite precursor solution (50 μL, without or with 0.5, 0.8, 1.0 wt % PImIL or PimIL-SiO) was spin-coated onto the SnO2 layer with the following spin process: first, 1000 rpm for 5 s and then 5000 rpm for 30 s. The CB solvent was dropped onto the substrate 25 s after the second spin-coating step. The substrate was then heated at 150 °C for 10 min to form perovskite films. After the substrate cooled to room temperature, the spiro-OMeTAD solution was spin-coated on perovskite films at 3000 rpm for 30 s. Finally, 90 nm Ag anode was deposited by thermal evaporation (rate of 1.0 Å s−1). The device area was 0.04 cm2. Device characterizations Current density–voltage (J–V) characteristics were measured using a source meter (Keithley 2400), equipped with a light source (100 mW/cm2) under an AM 1.5 G filter. The standard silicon solar cell was corrected from NREL (United States), and the currents were detected under the solar simulator (Enli Tech, Taiwan, China, 100 mW cm−2, AM 1.5 G irradiation). The reverse scan range was from 1.2 V to −0.2 V, with 20 mV for each step. The ultraviolet–visible (UV–vis) spectra were conducted using a Shimadzu UV-2600 spectrophotometer (Japan). The steady-state photoluminescence (PL) and time-resolved photoluminescence (TRPL) spectra were recorded by an FLS920 spectrometer (Edinburgh Instruments Ltd., United Kingdom). Fourier transform infrared spectroscopy (FTIR) spectra were recorded on a Shimadzu IRPrestige-21 spectrometer (Japan). X-ray photoelectron spectroscopy (XPS) and ultraviolet photoelectron spectroscopy (UPS) measurements were performed in an ESCALAB 250Xi (Thermo Fisher, United States) by using Al Kα X-ray source under high vacuum (10−9 mbar). Scanning electron microscopy (SEM) images were conducted on an SU8020 scanning electron microscope operated at an acceleration voltage of 8 kV. X-ray diffraction (XRD) measurements were recorded on a Rigaku D/Max-B X-ray diffractometer (Japan) with Bragg-Brentano part focusing geometry. Electrochemical impedance spectroscopy tests were conducted on the electrochemical workstation with a bias voltage of 0.8 V and an applied frequency range of 10 MHz–100 kHz. External quantum efficiency (EQE) values were measured under monochromatic illumination (Oriel Cornerstone 260 1/4 m monochromator equipped with an Oriel 70613NS QTH lamp), and the calibration of the incident light was performed using a monocrystalline silicon diode. In an interlaced mode, the dual-beam time-of-flight secondary ion mass spectrometry (ToF-SIMS) depth profiling characterizations were obtained on a ToF-SIMS 5 instrument (IONTOF GmbH, Japan). A pulsed 30 keV Bi+ ion beam was used as the analysis beam with a beam current of 1.08 pA. The analysis area was 100 × 100 μm2 that was at the center of the sputter crater of 300 × 300 μm2. A 2 keV Cs+ ion beam was used as the sputtering beam with a beam current of 120 nA. Results and Discussion The interaction between perovskite and poly(ionic liquids) A novel PIL PImIL-SiO has been successfully designed to establish dual-resistance of ion migration and moisture erosion (Figure 1a). The chemical structure of PImIL-SiO is shown in Figure 1b. We speculate that the imidazolium (IM) cations in PImIL-SiO help provide chemical anchoring sites for the passivation of uncoordinated Pb ions in the perovskite film. Simultaneously, the siloxane groups can hydrolytically condense with the permeated moisture, effectively hindering ion migration channels and forming a secondary barrier to endow the perovskite film with ultrahigh moisture corrosion resistance. The molecular structure of the employed poly-1-vinyl-3-methylimidazolium (PImIL) and synthetic route to PImIL-SiO are displayed in Supporting Information Figure S1. Proton nuclear magnetic resonance (1H NMR) characterization ( Supporting Information Figure S2) was implemented to confirm the molecular structure of PImIL-SiO. XPS provides ample evidence for the molecular interactions of perovskite with PILs additives. The main peaks of Pb 4f7/2 and 4f5/2 with saturated coordination are demonstrated in Figure 1c, where the PImIL-SiO-modified perovskite film shows a large shift of 0.45 eV toward lower binding energy but a relatively small change for the PImIL-modified film (0.30 eV) compared with the control sample. We ascribe the noteworthy shift toward lower binding energy of PImIL and PImIL-SiO-modified perovskite films to the positively charged IM cations passivating the negatively charged Pb–I anti-sites in the perovskite. Additionally, the Si-O groups in PImIL-SiO can passivate numerous electronic defects to improve efficiency and stability by strong Pb–O coordination bonds. Related interactions between PImIL-SiO and perovskite were also observed in the FTIR spectra ( Supporting Information Figure S3), where the stretching vibration peak of Si–O–CH3 at 2843 cm−1 shifts to 2837 cm−1 and the IM skeleton stretching vibration peaks blue shift to larger wavenumbers from 1546 cm−1 and 1454 cm−1 to 1550 cm−1 and 1462 cm−1, respectively. These results explain that the undercoordinated Pb2+ ions are well passivated by IM and Si–O groups. Additionally, the similar shift of the I 3d peaks toward higher binding energy in the XPS spectra (Figure 1d) suggests that the existing IM cations in PImIL and PImIL-SiO form electrostatic interactions with I− to inhibit migration. The appearance of the Si 2p peak in the XPS spectra indicates that PImIL-SiO partially exists on the surface of the perovskite film ( Supporting Information Figure S4). The ratio of O–C and O–Si changed along with the PImIL-SiO-modified perovskite film stored in high humidity. As presented in Figure 1e, the area ratio of O–C and O–Si was less than 1 after siloxane hydrolytic condensation, whereas the original area ratio was close to 1, proving that PImIL-SiO was partially crosslinked in the perovskite film. Furthermore, FTIR spectra have verified the occurrence of the hydrolytic crosslinking reaction in PImIL-SiO at high humidity, as shown in Figure 1f,g. After PImIL-SiO crosslinking, the stretching vibration peak of Si–O–Si was detected at 1051 cm−1, whereas the disappearance of the peak at 2840 cm−1 was assigned to the stretching vibration of Si–O–CH3 in pristine PImIL-SiO. The above results indicate that PImIL-SiO can passivate defects and anchor GBs for improving crystalline quality and stability of perovskites. Figure 1 | Schematic illustration of (a) the PVSCs based on PImIL-SiO-passivated perovskite film, and (b) the formed dual-resistance of ion migration and moisture erosion with hydrolytic crosslinking of siloxane. XPS spectra of (c) Pb 4f and (d) I 3d core level of the control and PIL-modified perovskite films with PImIL and PImIL-SiO, respectively. XPS spectra of (e) O 1s core level for perovskite film with original PImIL-SiO and crosslinked PImIL-SiO. The corresponding area radio of O–C and O–Si is 0.93 and 0.23, respectively. (f) FTIR spectra of original PImIL-SiO and crosslinked PImIL-SiO and (g) corresponding enlarged view of Si–O–CH3 and Si–O–Si stretching vibrations. Download figure Download PowerPoint Effect of poly(ionic liquids) on perovskite crystallization The morphology of the perovskite film was investigated through SEM. Top-view images and crystal size distributions of the perovskite grains are displayed in Figure 2a–c and Supporting Information Figure S5, respectively. The control film demonstrates small grains (clustered at 500–700 nm) and non-uniform morphology with pinholes and squamiform PbI2 (Figure 2a). In contrast, the incorporation of PImIL promotes a uniform morphology with a larger grain size of 700–900 nm (Figure 2b), while the flaky structure PbI2 still exists. Upon the PImIL-SiO incorporation, PbI2 disappears and the perovskites become more homogeneous and denser with the grain size further increasing to 900–1100 nm (Figure 2c). These results illustrate that the perovskite film quality can be better regulated by introducing PImIL-SiO on account of its low volatile property and robust interaction with perovskite components, which can slow the crystallization rate to provide larger perovskite grains. The XRD patterns also indicate that the incorporation of PImIL and PImIL-SiO poses a significant effect on crystallization control and PbI2 inhibition in perovskite films (Figure 2d). Although the UV–vis absorption spectra do not increase obviously, the significant enhancement of steady-state PL intensity for PImIL-SiO-modified perovskite film provides convincing evidence, indicating that PImIL-SiO plays a vital role in high-quality perovskite (Figure 2e). As is well documented in recent studies, residual tensile strain (σ) caused by defects in perovskite film is a crucial origin of inherent PVSCs instability.34–36 The strain in the control and PIL-modified perovskite films was investigated using the grazing-incidence XRD (GIXRD) technique with the 2θ-sin2ψ method (Figure 2f–h). The GIXRD peaks for all perovskite films constantly shift to the smaller 2θ as the value of ψ increases, indicating that the crystal plane distance d (012) has increased, and all films have been subjected to tensile strain. Generally, perovskite film stress (σ) can be calculated by fitting 2θ as a function of sin2ψ, and the magnitude of residual strain can be reflected by the slope of the fitted line. According to the slope of the linearly fitted 2θ-sin2ψ (Figure 2i), the PImIL-SiO-modified perovskite film exhibits a significantly reduced negative slope compared with the control film, meaning that the incorporation of PImIL-SiO can efficiently release the residual tensile strain to some extent due to the stronger anchoring effect in eliminating the bulk defects. The release of residual stress can contribute to the formation of uniform perovskite film with fewer defects density, which improves the carrier transport at the interface and inhibits ion migration, thus resulting in enhancing device efficiency and stability. Figure 2 | SEM images of (a) control and PIL-modified perovskite films with (b) PImIL and (c) PImIL-SiO. (d) XRD patterns and (e) UV–vis absorption integrated with PL spectra of control and PIL-modified perovskite films with PImIL and PImIL-SiO. GIXRD patterns with different instrumental ψ values (10–50°) for (f) control and PIL-modified perovskite films with (g) PImIL and (h) PImIL-SiO. (i) The corresponding linear fit of 2θ-sin2ψ for control and PIL-modified perovskite films. Download figure Download PowerPoint Characterization of defect and photovoltaic performance To comprehensively evaluate the passivation characteristics of PILs in perovskite films, the carrier transport dynamics and nonradiative recombination properties were investigated. The TRPL decays of different perovskite films are shown in Figure 3a, and the corresponding parameters are fitted with the bi-exponential decay function as in eq 1 I ( t ) = A + B 1 * exp ( − i / τ 1 ) + B 2 * exp ( − i / τ 2 ) (1)to calculate the intensity average lifetime. The TRPL measurement exhibiting the perovskite film based on PImIL-SiO has a much longer carrier lifetime (τave = 221 ns) than those of control (τave = 109 ns) and PImIL-modified film (τave = 169 ns). We attribute the long carrier lifetime to the larger grain size and fewer bulk defects in PImIL-SiO-modified perovskite, which restrains the non-radiative recombination. In addition, the Nyquist diagrams (Figure 3b) of the control and PIL-modified devices show only a single semicircle, which indicates the charge recombination resistance (Rrec) value. The extracted Rrec value of the PImIL-SiO-modified device is 34 kΩ, which is more significant than the control device (12 kΩ) and PImIL modified device (24 kΩ). The larger Rrec value indicates that there is superior charge transfer performance and suppressed carrier recombination in the device based on PImIL-SiO. The dark-current measurements ( Supporting Information Figure S6) show a reduced reverse current density of the PImIL-SiO-modified device compared with the control and PImIL modified device, which reflects an enhanced shunt resistance as well as a restrained charge carrier recombination and leakage current. The trap densities of different films have been further estimated by space-charge limited current measurement with a structure of ITO/SnO2/perovskite (control, with PImIL or PImIL-SiO)/PCBM/Ag. As evidenced in Figure 3c, the trap density decreases from 8.26 × 1015 cm−3 for the control film to 6.25 × 1015 and 5.47 × 1015 cm−3 for PImIL and PImIL-SiO-modified films, respectively. These results further prove that PImIL-SiO suppresses the occurrence of shallow trap states concentrated at GBs, which effectively improves the robustness of the perovskite crystal and the optoelectronic performance. Figure 3 | (a) TRPL spectra and (b) Nyquist plots of control and PIL-modified perovskite films with PImIL and PImIL-SiO. (c) I–V curves of the electron-only devices based on ITO/SnOx/perovskite (control, with PImIL and PImIL-SiO)/PCBM/BCP/Ag. (d) Energy level diagram of the devices with control and PIL-modified perovskite layers. (e) J–V curves under AM 1.5 G illumination for the best-performing PVSCs and (f) PCE statistics of 20 devices based on control and PIL-modified perovskite layers. The insert table is statistical photovoltaic parameters of control and PIL-modified PVSCs with PImIL and PImIL-SiO under standard AM 1.5 G illumination (100 mW cm−2). (g) EQE of the control and PImIL-SiO-modified PVSCs. J–V curves under reverse and forward scan directions of the control and PImIL-SiO-modified PVSCs (without any encapsulation) before (h) and after (i) aging under 25 ± 5 °C, 20% ± 5% RH for 25 days. Download figure Download PowerPoint To intuitively evaluate the impact of PILs on the photoelectric performance of the device, PVSCs with a formal structure of ITO/SnO2/perovskite (with or without PImIL or PImIL-SiO)/spiro-OMeTAD/Ag have been fabricated. The J–V characteristics curves of PVSCs based on perovskite with different PImIL-SiO suggest the optimum concentration of 0.8 wt % ( Supporting Information Figure S7), which is adopted in the following discussion. The corresponding optimum photovoltaic parameters are given in Supporting Information Table S1. The corresponding schematic of the energy band alignment is described in Figure 3d, and the energy bands of these perovskite films are computed from UPS ( Supporting Information Figure S8). The J–V curves of the best-performing PVSCs are shown in Figure 3e. Notably, the control device yields a PCE of 20.68%, with a short circuit current density (Jsc) of 24.21 mA·cm−2, an open-circuit voltage (Voc) of 1.12 V, and a fill factor (FF) of 75.85%. In contrast, the PCE of the best-performance device based on PImIL-SiO increases to 22.46%, with a Voc of 1.13 V, a Jsc of 24.82 mA·cm−2, and an FF of 79.61%. The enhancement of the device performance, especially the FF, is ascribed to the improved film morphology and the reduced defect density. The PCE statistics of 20 devices based on the control and PIL-modified perovskite films further certify the excellent reproducibility of the PVSCs (Figure 3f), and the corresponding distribution of Voc, Jsc, and FF are shown in Supporting Information Figure S9. The integrated Jsc values from the EQE spectra agree well with the J–V measured values (Figure 3g). Moreover, the higher and much swifter photocurrents were measured by conducting the steady-state photocurrent and stabilized PCE output at the maximum power point ( Supporting Information Figure S10). Inhibition of ion migration In the test process of the J–V curve of PVSCs, researchers often observe that the J–V curves obtained by multiple measurements are inconsistent when changing the direction and speed of the voltage scan, which is called "current hysteresis".37,38 The control and PImIL-SiO-modified devices show different J–V characteristics in Figure 3h, with less hysteresis in the PImIL-SiO-modified device. The difference in the current hysteresis between the control and PImIL-SiO-modified devices is more extensive after aging in air with 25% RH for 25 days (Figure 3i). Considering that the current hysteresis is related to the applied voltage, scanning direction, scanning speed, and device design factors, the PCE of the aged PVSCs (in air with 25% RH for 25 days) under forward and reverse scans with different scan rates were compared (Figure 4a). Hysteresis factor (HF) is defined according to the following expression: Hysteresis factor = ( PCE forward − PCE reverse ) / PCE forward (2) Figure 4 | (a) PCE measured at forward and reverse scan directions under different scan rates and (b) corresponding hysteresis factor (HF) of control and PIL-modified PVSCs with PImIL and PImIL-SiO after aging in air with 25% RH for 25 days. (c–e) PL spectra and (f) the shift spectral centroids for control and PIL-modified perovskite films upon illumination for 0–40 min (The composition of perovskite for PL spectra here is Cs0.05(FA0.87MA0.13)0.95Pb(I0.83Br0.17)3.). ToF-SIMS depth profiles for (g) control and PIL-modified PVSCs with (h) PImIL and (i) PImIL-SiO after aging at 85 °C for 3 days. Download figure Download PowerPoint The hysteresis effects in PVSCs are usually reduced with a slower scan speed, where a similar trend can be observed in Figure 4b.39,40 Starting with a large HF of 0.36 at 100 mV/s, the HF of the control device falls to 0.25 at a slow scan rate of 0.04 mV/s. In contrast, devices based on PIL-modified perovskite layers always keep their small HF values at about 0.1 within the whole scan range. Theoretical calculations and experimental results demonstrate that ion migration and subsequent accumulation at the interface are main intrinsic origins of the current hysteresis in PVSCs. The consistently small HF value suggests that macromolecular PILs have outstanding ion-immobilization ability for robust PVS