Competitive-binding buried interlayer for year-round reproducible perovskite solar cells

Abstract

Homogenizing the upper surface through posttreatment has made great progress in perovskite solar cells. In contrast to the exposed surface, there are no practical remedies if imperfections form randomly at the hidden buried interface after perovskite film generation. Here, we reveal a severe distribution of residual lead iodide, voids, and grain-surface concavities at the buried interface, which severely trap carriers in inactive regions. To address these challenges, we introduce a potassium dihydrogen phosphate competitive-binding interlayer that systematically reduces residual solvents at the buried interface through strong chemical interactions. Homogenized buried interface along with facilitated perovskite film quality and charge extraction have been achieved, enabling year-round improvements in photovoltaic performance and reproducibility. The resultant devices achieve a champion power conversion efficiency (PCE) of 26.3% (certified at 25.8%) for a 0.07–square centimeter device and 25.17% for a 1.028–square centimeter device. The device also demonstrates exceptional stability, maintaining 97% of its initial PCE after 1000 hours of continuous maximum power point tracking.

A competitive-binding interlayer was introduced to homogenize the buried interface of perovskite film.

INTRODUCTION

Metal halide perovskite solar cells (PSCs) have attracted widespread interest in both academia and industry as a promising photovoltaic technology to help relieve the urgent global energy demand (1–6). High-quality perovskite films and favorable charge extraction are essential prerequisites for achieving high performance (7–14). However, imperfections at the heterointerfaces between perovskite films and charge-selective layers remain a great challenge, undermining these critical factors and limiting device reproducibility, particularly under year-round variations in temperature and humidity (15, 16). Surface posttreatment strategies such as two-dimensional (2D)/3D heterojunction engineering and chemical polishing strategies have greatly contributed to homogenizing the upper surface, improving the charge extraction and suppressing the nonradiative recombination (17–20). In contrast, the buried interface at the bottom of perovskite films is difficult to investigate and regulate (21). Once imperfections such as voids, grain-surface concavities, and impurities form at this buried interface (22–24), no feasible postfabrication remedies exist.

To homogenize the buried interface, it is crucial to identify the root causes of these imperfections. Perovskite films typically formed by the reaction between PbI₂ with organic ammonium salts on a bottom charge transport layer, using cosolvent systems comprising polar aprotic solvents [e.g., N,N-dimethylformamide (DMF), acetonitrile (ACN), and high-boiling-point solvents such as dimethyl sulfoxide (DMSO), N-methylpyrrolidone, and γ-valerolactone] to chelate with PbI₂ and form [PbI₆]⁴⁻-solvent intermediates that facilitate crystal growth (25–27). However, the interactions between these coordinated solvents and the bottom transport layer are often overlooked, leading to the retention of solvents at the bottom of perovskite films. These residual solvents promote the formation of amorphous regions and voids at the buried interface (28, 29), and their hygroscopic nature absorbs moisture, hindering the conversion from intermediate complexes to perovskites and inducing the generation of harmful phases (30, 31). These issues are further exacerbated when annealing in ambient air to enlarge grain size during perovskite films fabrication (32). While pretreating the buried interface to regulate coordinated solvents could mitigate these issues, the high solubility of most treatment molecules in perovskite precursor solutions greatly reduces their effectiveness during film formation (23, 33). Nevertheless, it is very urgent to develop a method to regulate coordinated solvents, homogenize the buried interface and illustrate the key factors that fundamentally determine the crystal quality.

In this work, we introduced a potassium dihydrogen phosphate (PDP) competitive-binding interlayer to mitigate these issues. Systematical density functional theory (DFT) calculations revealed that PDP had a remarkably stronger binding ability with SnO₂ than PbI₂, making it hard to be dissolved in PbI₂ precursor solution during the perovskite films generating process. Further DFT calculations revealed that coordinative solvents even presented much stronger binding abilities with bottom SnO₂ electron transport layer than perovskite, which exacerbated the residual solvents remaining at the buried interface. PDP competed with the coordinative solvents to interact with PbI₂ through the strong chelation and lone-pair electrons in PDP molecule structure avoided the binding between PDP and coordinative solvents, which enabled lessening the residual solvents at the buried interface. As a result, homogenized buried interface along with improved perovskite film-quality and charge extraction have been realized. The resultant devices achieved a champion power conversion efficiency (PCE) of 26.3% (certified at 25.8%) for a 0.07-cm² device and 25.17% for a 1.028-cm² device, demonstrating year-round improvement in photovoltaic performance. Besides, this competitive-binding interlayer strongly stabilized the perovskite structure, hampering the perovskite degradation and irreversible ion migration, enabling 97% of its initial PCE maintaining after 1000 hours of operation.

RESULTS

Intermolecular interactions at the buried interface

We initially studied the intermolecular interactions at the buried interface via systematical DFT calculations. DMF and DMSO cosolvent system were used as the coordinated solvents due to their chemical binding ability with PbI₂ by forming Pb-O═C and Pb-O═S bonds, showing an adsorption energy (E_ad) of 0.04 and 0.05 J/m² with PbI₂ (fig. S1), respectively. However, DMF and DMSO also had strong binding ability with SnO₂ through the chemical bonds of Sn-OC and Sn-O═S, presenting an even higher E_ad of 0.15 and 0.17 J/m², respectively. To avoid the chemical binding of solvents with SnO₂ thus to reduce the residual solvents at the buried interface, PDP interlayer containing lone-pair electrons was introduced between SnO₂ layer and perovskite film. PDP can simultaneously bind with SnO₂ transporter and perovskite at the buried interface via the chemical bonds of Sn-O═P, Sn-O-P, and Pb-O-P (Fig. 1A). To tell the preferential binding, we calculated the E_ad values of PDP with different molecules including PbI₂, formamidinium iodide (FAI), and SnO₂ (Fig. 1B and fig. S2). PDP exhibited an E_ad of 0.29 J/m² with SnO₂, which pronouncedly exceeded those of PbI₂ and FAI, implying that PDP would preferentially anchor with SnO₂ than PbI₂ and hard to be dissolved in PbI₂ precursor solution during the perovskite film generation.

Fig. 1. Intermolecular interactions and crystallization at the buried interface.

(A) DFT modeling of PDP interacting with SnO₂ transporter and perovskite through chemical bonds formation. (B) The calculated E_ad between different molecules. (C) ¹H-nuclear magnetic resonance of the coordinative solvents residual in the control and target wet perovskite films. (D and E) In situ PL spectra of the control (D) and target (E) perovskite films during spin-coating the organic ammonium salts. The overlapped curves are PL spectra at 2 s. (F) Evolution of PL intensity at 775-nm wavelength. (G) XRD patterns of the control and target wet perovskite films before annealing process. (H) XRD patterns of the control and target perovskite films after annealing process. a.u., arbitrary unit.

X-ray photoelectron spectroscopy (XPS) measurements were then conducted to analyze the chemical states of SnO₂. After introducing the PDP interlayer, there appeared characteristic peaks corresponding to P and K atom (fig. S3), and the characteristic peaks corresponding to Sn atom slightly shifted toward lower binding energy (fig. S4), indicating that the lone-pair electrons in PDP molecule structure were successfully transferred to Sn⁴⁺ (34). Further analyses about the O 1s spectra of SnO₂ indicated that the content of O²⁻ improved after PDP modification (fig. S5), implying the passivated oxygen vacancy defects for electron transfer (35). Moreover, the contact angle of SnO₂ film only increased from 0° to 9° after introducing the PDP interlayer, which demonstrates the well-preserved wetting property (fig. S6). Owing to the electronegativity difference between P atom and O atoms, PDP was also chemically bonded to PbI₂. To explore the binding ability, we added PDP to PbI₂ precursor solution with a molar ratio of 10%. The color of PbI₂ solution changed from yellow to brown (fig. S7), suggesting that PbI₂ has interacted with PDP to form PDP-[PbI₆]⁴⁻ compound (36). Fourier transform infrared spectroscopy measurements were further performed to explore the interaction. After mixing PDP and PbI₂, the P═O stretching peak at 1251 cm⁻¹ and P─O stretching peak at 1038 cm⁻¹ shifted to lower wave numbers (fig. S8), agreeing with the transferring of lone-pair electron in PDP toward Pb²⁺ (37). These results revealed that PDP tended to compete with the coordinative solvents to interact with PbI₂ through the strong chelation, benefiting for lessening the residual solvents at the buried surface.

We then investigated the contents of residual DMF and DMSO using 1H nuclear magnetic resonance spectroscopy. Wet perovskite films before annealing process were scraped off from the substrates and then dissolved in deuterium oxide. As shown in Fig. 1C, peaks at 7.83, 2.75, and 2.61 parts per million (ppm) were assigned to the -CH in FA⁺, -CH in DMF, and -CH₃ in DMSO, respectively (23, 38). For the PDP modified sample, the atomic ratio of DMF/FA⁺ and DMSO/FA⁺ decreased, well demonstrating that the residual solvents have been successfully lessened.

Homogenizing the buried interface

We then investigated the perovskite crystallization and film morphologies after lessening the residual solvents at the buried interface in detail. Scanning electron microscopy (SEM) measurements were conducted to detect the morphologies of PbI₂ films. There appeared a more pronounced porous structure in the PbI₂ film after introducing the PDP interlayer (fig. S9). We attributed this phenomenon to the evaporation of solvents. PDP competed with solvents to coordinate with [PbI₆]⁴⁻ octahedrons and led to more solvents evaporating, creating the porous structure. Porous PbI₂ film benefits for the reaction between PbI₂ and FAI to form FAPbI₃ perovskite (39).

In situ photoluminescence (PL) characterizations were performed to explore the perovskite-generating process during spin-coating. The peak of PL spectra located at about 775 nm during the spin-coating process (Fig. 1, D and E) represents the generation of α-phase FAPbI₃. As shown in Fig. 1F, both samples showed a rapid increase followed by a sharp quenching, closely relating to the reaction process of PbI₂ and FAI and the initial evaporation of solvents as the beginning of spin-coating (40). After introducing the PDP interlayer, there showed a quicker increase in PL intensity and the peak time was advanced by 2.7 s. The PL spectrum at 2 s further evidenced the stronger PL intensity and lower full width at half maximum (FWHM) for the target sample. For this process, the formed FAPbI₃ crystals mainly exist as the initial nucleus that are momentous for the following crystals growth. On the basis of these results, the introduced PDP interlayer was demonstrated to be able to redound the reaction between PbI₂ and FAI, thereby accelerating the nucleation.

X-ray diffraction (XRD) measurements were used to analyze the phases in the wet perovskite films before annealing process. Both samples showed a distinct diffraction peak at 14°, assigning to the (001) plane of α-phase FAPbI₃ (Fig. 1G). However, some impurity-related diffraction peaks such as 2H-phase at 11.8° were detected in the control sample (41). In contrast, impurity-related peaks were notably reduced along with a more pronounced peak of α-phase FAPbI₃ (the FWHM decreased from 0.122 to 0.110) after introducing the PDP interlayer, which tightly corresponds to the accelerated nucleation and the reduction of residual solvents at the buried interface. After further annealing process at 145°C in the air, a pronounced diffraction peak associating with PbI₂ was observed at 12.8° in control sample (Fig. 1H). Notably, there did not appear any diffraction peak of PbI₂ before annealing, indicating that PbI₂ phase was generated during the annealing process through phase degradation. After introducing the PDP interlayer, the PbI₂ phase was eliminated and impurity phases were suppressed. Meanwhile, the crystallization of α-phase FAPbI₃ (the FWHM of the (001) diffraction peak decreased from 0.156 to 0.152) was facilitated.

To study the buried interface more clearly, we peeled off the perovskite films carefully to the bare glass sheets using epoxy (figs. S10 and S11). An abundant distribution of imperfections including PbI₂ and voids along with grain surface concavities were detected at the buried interface of control sample (Fig. 2A). For the control sample annealing in the air, residual solvents at the buried interface easily absorb the moisture due to the hygroscopic nature, producing hydrated compounds that highly affect the FAPbI₃ crystal facets and accelerate the degradation of FAPbI₃ to PbI₂ (31). Evaporation of these residual solvents during the annealing process likely generated the void, grain surface concavities or space for the distribution of produced PbI₂. After lessening the residual solvents by the PDP interlayer, homogenized buried interface along with enlarged grain size of perovskites was achieved (Fig. 2B). Lessening the residual solvents at the buried interface is favorable for suppressing the generation of amorphous regions and voids as well as the harmful phase thus homogenizing the buried interface as the schematic illustration shown in fig. S12, which benefits for the nonradiative recombination suppression.

Fig. 2. Film characterizations.

(A and B) SEM images of the buried interfaces obtained by peeling off the control (A) and target (B) perovskite films. (C and D) The temperature-dependent PL spectra of control (C) and target (D) perovskite films. (E and F) The FWHMs of the control (E) and target (F) sample as a function of temperature. (G and H) 2D TA spectra of control (G) and target (H) perovskite films. (I) TA decay curves of control and target perovskite films. (J and K) Schematic diagram of electron extracting behavior in the control (J) and target (K) buried interface.

Nonradiative recombination suppression at the buried interface

Nonradiative recombination suppression achieved by film quality improvement and charge extraction optimization was then explored. To study the effect of film quality improvement, we then conducted steady-state PL and time-resolved PL (TRPL) on the buried interfaces obtained by peeling off the perovskite films (fig. S13). PDP enhanced the PL intensity by three times and increased the lifetime from 468 to 1155 ns, demonstrating the improved perovskite film quality at the buried interface. We also carried out extremely low temperature–dependent PL measurements (80 to 300 K) to calculate the electron-phonon coupling coefficients (Г_op) so as to assess the effect of homogenized buried interface on carrier relaxion of perovskite films (Fig. 2, C to F, and note S1). As the temperature increased, both samples showed the peak shapes broadening due to the electron-phonon coupling effect and there appeared an evident bend at 190 K owing to the phase transition from α phase to β phase of formamidinium (FA)-based perovskite (42). After homogenizing the buried interface, the calculated Г_op decreased from 45.04 to 41.68 eV. This implied that the coupling between charge carriers and longitudinal optical phonons has been weakened and less energy dissipated into the lattice, prolonging the carrier lifetime and diffusion length.

To evaluate the defects in perovskite films, we performed the thermal admittance spectroscopy (TAS) measurements as shown in fig. S14. After introducing the PDP interlayer, the energy difference between the conduction band minimum and the defect state level (E_a) reduced and the integrated defect density decreased visibly, indicating that the defects were greatly passivated. Moreover, after introducing the PDP interlayer, the stress in the perovskite film was well released (fig. S15). Meanwhile, PL mapping image of PDP-modified buried interface showed a more homogeneous and stronger PL intensity (fig. S16), highlighting the potential of this buried interface–homogenizing technique on large-scale PSCs fabrication.

To study the effect of this buried interface–homogenizing strategy on charge extraction, ultraviolet photoelectron spectroscopy (UPS) measurements were then used to calculate the energy levels at the buried interface as shown in figs. S17 and S18. Notably, PbI₂ presented a p-type semiconductor characteristic, seriously blocking the electron extraction if distributing at the bottom of perovskite grains (fig. S19). Eliminating the PbI₂ in buried interface via PDP interlayer is ‌fundamental‌ for boosting the charge extraction. In addition to homogenizing the buried interface, SnO₂ film also showed a upshift Fermi level value of −4.08 eV after introducing the PDP interlayer, indicating the lower energy barrier between SnO₂ film and perovskite layer, which further facilitated the electron extraction. We then conducted the transient absorption (TA) spectroscopy to evaluate the electron extraction at the buried interface. As shown in Fig. 2 (G to I), more rapid decay kinetics was detected in the target sample, indicating that the accelerated electron extraction has been achieved at the perovskite/SnO₂ heterointerface. For the control buried interface, p-type semiconductor impurities, voids, and grain surface concavities accumulated carriers in the invalid region and created barrier for the electron extraction (Fig. 2J). By contrast, a homogenous and smooth electron extracting structure has been achieved after introducing the buried interface–homogenizing technique (Fig. 2K). The improved film-quality coupling with the boosted charge extraction contributes to the nonradiative recombination suppression at the buried interface, which is conducive to elevating the open-circuit voltage (V_oc) and fill factor (FF) of PSCs.

Photovoltaic performance of PSCs

We fabricated the n-i-p PSCs with a structure of fluorine-doped tin oxide (FTO)/SnO₂/perovskite/2,2′,7,7′-tetrakis N, N-di-p-methoxy-phenylamine)-9,9′-spirobifluorene (spiro-OMeTAD)/Au to evaluate the photovoltaic performance improvement. The statistical photovoltaic performances of devices with different concentrations of PDP were shown in fig. S20. The control PSC showed a PCE of 25.32 and 24.30% under reverse and forward scans, respectively (Fig. 3A). The target PSC showed a pronounced increase in PCE values along with a suppressed hysteresis, reaching a champion PCE of 26.32% for reverse scanning and 26.04% for forward scanning (Fig. 3B). Notably, the champion PCE had a high V_oc of 1.204 V along with a high FF of 84.86%, rendering a record V_oc × FF, which reached 88.6% of the Shockley-Queisser limit. The target PSCs also achieved a certified PCE of 25.8% (fig. S21). External quantum efficiency (EQE) characterizations provided the consistent J_sc results with the J-V curves (fig. S22). The target device showed a slightly increased integrated J_sc from 25.49 to 25.74 mA/cm² due to the improved EQE values at the short wavelength range originated from the effects of optimized buried interface including the improved charge extraction and perovskite crystallization. The target PSC also showed a remarkable improvement in stabilized power out efficiency, increasing from 24.92 to 26.22% (Fig. 3C). After enlarging the aperture area of PSC to 1.028 cm², target PSC showed a more pronounced enhancement in PCE value from 22.53 to 25.17% (Fig. 3D), indicating the potential of this technique on large-scale PSCs application.

Fig. 3. Photovoltaic performance of devices.

(A and B) J-V curves of the control (A) and target (B) devices under forward and reverse scans. (C) Stabilized power out of the control and target devices. (D) J-V curves of the control and target large-area devices with aperture area of 1.028 cm². (E) EQE_EL–current density curves of control and target devices. (F) Detailed FF loss analysis of control and target devices. (G) Statistical performances of the control and target devices fabricated in different seasons (20 devices for each type in one season). The PCE optimized from 24.63 ± 0.40% to 25.71 ± 0.20%, 23.57 ± 0.86% to 25.53 ± 0.16%, 24.53 ± 0.48% to 25.71 ± 0.16%, and 24.53 ± 0.49% to 26.04 ± 0.14% for spring, summer, autumn and winter, respectively.

Electroluminescence quantum yield (EQE_EL) performances obtained by treating the devices as light-emitting diodes (LEDs) under forward-bias voltages were used to assess the V_oc losses caused by the nonradiative recombination in devices (∆V_oc^nonrad). The target PSC showed an improved EQE_EL value from 3.16 to 7.88% at the corresponding current density (Fig. 3E). We calculated the ∆V_oc^nonrad, where the target PSC showed a lower ∆V_oc^nonrad of 0.066 V than the control PSC of 0.089 V. Meanwhile, the target device showed an enhanced built-in potential, which is beneficial for exciton separation (fig. S23), coinciding with the improvement in the V_oc. Moreover, the target device presented a smaller charge transport resistance and a larger recombination resistance than the control one (fig. S24), implying that the charge recombination at the buried interface was effectively suppressed. More specifically, we calculated the nonradiative recombination loss and transport loss via the FF loss analysis (note S2). The nonradiative recombination was suppressed from 5.73 to 4.23% and transport loss from 5 to 3.4% (Fig. 3F), suggesting the effect of this buried interface–homogenizing technique on nonradiative recombination suppression and charge transport improvement to improve the V_oc × FF.

To broaden the potential of this buried interface–homogenizing technique, we fabricated the PSCs in four seasonal characteristic conditions all year round (the average temperature/humidity were ⁓20°C/⁓40% relative humidity (RH), ⁓35°C/⁓60% RH, ⁓25°C/⁓35% RH, and ⁓10°C/⁓20% RH for spring, summer, autumn, and winter, respectively). The corresponding results were presented in Fig. 3G and figs. S25 to S28. For the control PSCs, there showed a poor repeatability especially in summer with high temperature and humidity, wherein the highest PCE is 24.85% while the lowest PCE is 22.03%. In contrast, the target PSCs showed slight difference in the four seasons and presented a pronounced improvement in year-round photovoltaic performance, highlighting the potential of this buried interface–homogenizing technique on year-round PSCs fabrication.

Durability of perovskite films and PSCs

In addition to homogenizing the buried interface, PDP interlayer can also strongly stabilize the perovskite structure via chemical bonds, showing potential on mitigating the perovskite degradation and ion migration in PSCs. We then placed both films under harsh condition with 85°C heating but not isolating light for 100 hours to evaluate the stability of perovskite films. First, we conducted SEM measurements on the buried interface of aged perovskite films as shown in Fig. 4 (A and B). For the control buried interface, we found that there appeared a severer distribution of PbI₂ and lots of white spots at the bottom of perovskite grains after aging, indicating the terrible degradation in perovskites. In contrast, homogenized buried interface still well presented in target sample after aging. The white spots could be metallic lead (Pb⁰) generated by the degradation of PbI₂ during the aging process, as evidenced by the XPS results shown in Fig. 4C. XRD results further confirmed the perovskites degradation in control perovskite film and suppressed by this buried interface–homogenizing technique (Fig. 4D). To evaluate the ion migration in the device, we conducted the temperature-dependent conductivity measurements (note S3) (43). We noted that the ion-migration activation energy increased from 0.146 to 0.428 eV after homogenizing the buried interface (Fig. 4E), implying the potential on obstructing the ion migration in device. Time-of-flight secondary-ion mass spectrometry (TOF-SIMS) measurements further confirmed the suppressed iodine ion migration from perovskite layer to charge-selective layers (Fig. 4F). The PDP interlayer effectively homogenized the buried interface and strongly anchored the perovskite structure, suppressing the iodine vacancy formation as well as the ion migration.

Fig. 4. Stabilities of perovskite films and devices.

(A and B) SEM images of the buried interfaces of the control (A) and target (B) perovskite films after aging. (C) XPS results of the control and target perovskite films after aging. (D) XRD patterns of the control and target perovskite films after aging. (E) Temperature-dependent conductivity measurements of control and target perovskite films. (F) TOF-SIMS measurements of control and target devices after aging. (G) Long-term MPPT stability measurements of the control and target devices under continuous white light illumination with an intensity of 100 mW cm⁻² at ~55°C in N₂. (H) Long-term heating stability measurements of the control and target devices under 85°C. h, hours.

We further investigated the operational stability of PSCs using maximum power point tracking (MPPT) under continuous white light illumination with an intensity of 100 mW cm⁻² at ~55°C in N₂. The control device retained 79% of its initial PCE after 1000 hours of MPPT. After introducing the PDP interlayer, the device showed an increased operational stability, retaining 97% of its initial PCE after 1000 hours (Fig. 4G). We attribute the enhanced operational stability to the effects of PDP on homogenizing the buried interface, passivating the defects, suppressing the ion migration, and stabilizing the perovskite structure. We further studied the thermal stability of PSCs under 85°C heating as shown in Fig. 4H. The control device only retained 46% of its initial PCE after 1000 hours of heating. After introducing the PDP interlayer, the heating stability largely improved and the target device retained more than 90% of its initial PCE after 1000 hours of heating.

DISCUSSION

Ensuring imperfections at the buried interface remains a tremendous challenge, which can be further exacerbated when annealing in ambient air to enlarge grain size during perovskite films fabrication, severely limiting device performance and reproducibility, particularly under year-round variations in temperature and humidity. We have systematically studied the origin that generates the imperfections at the buried interface and introduced a PDP competitive-binding interlayer to homogenize the buried interface. PDP can preferentially bind with bottom SnO₂ electron transporters and also coordinate with PbI₂. PDP competes with the solvents to coordinate with [PbI₆]⁴⁻ octahedrons, lessening the residual coordinative solvents at the buried interface. As a result, homogenized buried interface along with facilitated perovskite film quality and charge extraction have been achieved, enabling year-round improvements in photovoltaic performance and reproducibility. The target devices achieve a champion PCE of 26.3% for a 0.07-cm² device and a champion PCE of 25.17% for a 1.028-cm² device. Besides, both the perovskite film and device present an increased durability even if storing at heating conditions without isolating light. The device can retain 97% of its initial PCE after 1000 hours of operation under MPPT with 1-sun illumination at ~55°C. We hope that this work could attract more attention on the hidden buried interface homogenization and provide guidance for the fabrication of efficient, stable, and reproducible PSCs.

MATERIALS AND METHODS

Materials

DMF, DMSO, isopropanol (IPA), bis(trifluoromethane)sulfonimide lithium salt (Li-TFSI), ACN, chlorobenzene (CB), 4-tert-butylpyridine (tBP), and potassium iodide (KI) were purchased from Sigma-Aldrich. SnO₂ colloid precursor [tin (IV) oxide, 12% in H2O colloidal dispersion], octylammonium iodide (OAmI), poly[bis(4-phenyl)(2,4,6-trimethylphenyl)amine] (PTAA), and methylammonium chloride (MACl) were purchased from Xi’an Polymer Light Technology Corp. TPFB (4-isopropyl-4′-methyldiphenyliodonium Tetrakis (pentafluorophenyl) borate, >98%) was purchased from TCI. Lead iodide (PbI₂, 99.99%), FAI, and Spiro-OMeTAD (99%) were purchased from Advanced Election Technology CO. Ltd. PDP was obtained from Aladdin. Hydrogen peroxide aqueous solution (H₂O₂, 30%) was purchased from Sinopharm Chemical Reagent Co. Ltd.

Device fabrication

FTO substrates were cleaned in the sequences of detergent solution, deionized (DI) water, acetone, and ethanol. Then, substrates were dried using a nitrogen stream followed by an ultraviolet ozone process for 15 min, and 4 nm of SnO₂ were deposited using atomic layer deposition (ALD). Next, SnO₂ colloid solutions (SnO2 colloid precursor was diluted in H₂O₂ and DI water with a volume ratio of 1:1:4) were spin-cast onto the substrates at 4000 rpm for 30 s and annealed at 150°C for 30 min in ambient air (30 to 60% RH). Then, substrates were again processed an ultraviolet ozone process for 15 min and turned to N₂-filled glove box. PDP solutions prepared by diluting 1.5 mg of PDP in 1 ml of DI water were deposited onto the SnO₂ films at 3000 rpm for 30 s and annealed at 100°C for 10 min in ambient air (30 to 60% RH). For the perovskite films fabrication process, PbI₂ solutions (691.5 mg PbI₂ powder, 7.0 mg KI were dissolved in 1 ml of DMF/DMSO mix solution with a volume ratio of 9:1) were first spin-coated at 1500 rpm for 30 s and annealed at 70°C for 1 min. Next, organic amine salt solutions (90 mg FAI and 16 mg MACl were dissolved in 1 ml of IPA solution) were deposited on PbI₂ film at 2000 rpm for 30 s and annealed in the air (30 to 40% RH) at 147°C for 13 min. For the posttreatment, OAmI solutions with a concentration of 3 mg/ml were deposited on the perovskite film at 4000 rpm for 30 s. Spiro-OMeTAD solutions (prepared by dissolving 72.3 mg of Spiro-OMeTAD, 28.8 μl of tBP, and 17.5 μl of Li-TFSI solution in 1 ml of CB, the concentration of Li-TFSI solution was 520 mg ml⁻¹ in ACN) were deposited at 3000 rpm for 20 s. For the device thermal stability and MPPT test, PTAA solutions (prepared by dissolving 30 mg of PTAA with 3 mg of TPFB doping in 1 ml of CB) were deposited at 4000 rpm for 30 s to replace the Spiro-OMeTAD as hole-transport layer. Last, 80 nm of Au was thermally evaporated to serve as the mental electrode.

Characterizations

SEM images were recorded using a field-emission SEM (Zeiss GeminiSEM 500). XRD patterns of perovskite films were performed by using a Rigaku smartlab XRD instrument with Cu Kα radiation under operating conditions of 40 kV and 44 mA. XPS and UPS measurements were conducted using an XPS/UPS system (ESCLAB 250Xi, Thermo Fisher Scientific) with He Iα radiation. TOF-SIMS (Tescan SOLARIS) spectrometer was used for depth profile analysis of perovskites with 30 KV and 30 pA. TA spectra were obtained using an ultrafast TA spectrometer (ORPHEUS-HP, Light Conversion). PL and TRPL were performed using a Delta Flex fluorescence spectrum spectroscopy (HORIBA). Current density-voltage characteristics (J-V curves) of PSCs were recorded by using a standard AM 1.5 G solar simulator (class AAA, 450 W xenon lamp, Newport 94043A) with a Keithley 2400 Source Meter, and the light intensity was calibrated using a KG-5 filtered Si diode. EQEs were performed using a QE/IPCE system (Enli Technology Co. Ltd.). TAS spectra characterizations were performed with a CHI760E electrochemical workstation (Shanghai Chenhua Instruments Inc.) at different temperatures ranging from 250 to 320 K. Operational stabilities were conducted by using a white LED source (wavelength from 410 to 850 nm) with an AM 1.5 filter and an irradiation intensity of 100 mW cm⁻² in an N₂ atmosphere using a customized photovoltaic test system integrated into a glovebox (PLVT-G8001X-16B).

DFT simulations

The DFT calculations were carried out using the Vienna ab initio simulation package (44, 45). The core-valence interaction was described by the projector-augmented wave method (46, 47). The cutoff energy for basis functions was 520 eV. The generalized gradient approximation of the Perdew-Burke-Ernzerh functional was used for exchange correlation (48). The Grimme’s DFT-D3 scheme was used for the inclusion of van der Waals interactions (49). A rectangular box measuring 19.43 Å by 18.94 Å by 65 Å is selected to accommodate one DMF/DMSO/PDP molecule and PbI₂ or FAI terminated perovskite (001) surface, ensuring a large vacuum size (>20 Å) between adjacent components. The heterojunction model consisting of perovskite (001) surface, SnO₂ (001) surface, and PDP was chosen with size of 19.25 Å by 19.01 Å by 54.00 Å, yielding small lattice mismatch of 1.8% (along x direction) and 0.5% (along y direction) between perovskite and SnO₂. A large vacuum spacing exceeding 20 Å was used to avoid interaction between adjacent layers.

Supplementary Materials

This PDF file includes:

Supplementary Notes S1 to S3

Figs. S1 to S28

References