Perovskite solar cells (PSCs) exhibit great efficiency, low cost, and easy solution processing, demonstrating great potential as next-generation photovoltaics for future energy supply. The photoelectric conversion efficiency (PCE) of organic–inorganic hybrid PSCs has rapidly increased from 3.8 to 26.1% in just a few years due to optimization of the light absorption layer, continuous exploration, and innovation in materials for the hole transport and electron transport layers (ETL) in PSC devices. This achievement brings them closer to the theoretical value and places them on par with leading silicon solar cells. (1−5) Unfortunately, hybrid perovskites still suffer from inevitable degradation under air and light induced by the destruction of high polar covalent bonding between the monovalent organic cation and octahedral PbI₂. (6−8) Replacing organic cations in the organic–inorganic hybrid PSCs with inorganic Cs⁺ to generate all-inorganic CsPbX₃ perovskites may improve the stability under environmental pressure. (9−12) Among all-inorganic CsPbX₃ perovskites, CsPbI₃ possesses a proper bandgap of 1.73 eV, achieving a high efficiency of approximately 21.75%. (13) However, despite the optimum bandgap, black phase CsPbI₃ still suffers from an imperfect tolerant element triggering a metastable stage steadiness, leading to quick and easy degradation to a nonphotoactive yellow stage with an extensive bandgap of 2.82 eV under an ambient environment. (14)

By comparison, tribrominated cesium lead bromide (CsPbBr₃) perovskite shows great environmental tolerance under illumination, humidity, heat, and oxygen attack, making it potentially suitable for use in photoelectric devices, communication systems, electrical power plants, satellites, spacecrafts, space-based solar power plants, and so on. (15−17) Nevertheless, fewer CsPbBr₃ PSCs have been reported so far when compared to organic–inorganic hybrid PSCs due to their big optical bandgap that results in reduced spectral absorption, leading to the recombination loss at interfaces and grain boundaries owing to the energetic mismatch. (18) To solve these issues, tremendous efforts have been devoted to improving the intermediate energy levels for filling the bandgap and accelerating the charge-carrier extraction for better photoelectric performance, including optimization of the structure of the electron transport layer and the hole transport layer. (19,20) Spiro-OMeTAD is a common hole transport material used in perovskite solar cells. However, its small organic structure, complex synthesis process, poor long-term environmental stability, and relatively high cost seriously restrict their commercial use in the future. (21) In this view, nickel oxide (NiO) has attracted considerable attention for use in the hole transport layer owing to its several advantages: (22−25) First, the highest occupied orbital of NiO is higher than that of perovskite, which shows excellent energy-level alignment with hole transport materials and perovskite, enhancing the extraction ability of the holes in the perovskite. Second, NiO demonstrates excellent chemical stability in the air environment due to its inorganic oxide nature, ensuring the stable performance of the battery. Third, an ideal hole mobility and block electrons reduce the loss of holes and avoid carrier recombination. Four, the materials are easy to prepare and have a low cost. However, NiO_x-based PSCs still suffer from several problems. For example, the mismatched interfacial energy level at NiO_x and the perovskite results in a big energy offset. (26) Also, NiO_x films prepared at room temperature usually contain large amounts of impurities and surface defects, accumulating holes at the perovskite/NiO_x interface. (27) These challenges can be addressed by introducing organic molecules between the perovskite and NiO layers to significantly improve the hole extraction and transport. (28,29) However, little work has been carried out in this field of research, opening new avenues for investigating high-efficency PSCs.

An FTO/SnO₂/CsPbBr₃/NiO-l-Cys/carbon perovskite solar cell device constructed by a multistep spin-coating method is shown in Figure 1a, and the details can be found in the Experimental Section. The cross-sectional scanning electron microscopy (SEM) image (Figure 1i) shows a seamless connection between the perovskite and carbon electrode layers. Excitons are generated after sunlight absorption by the CsPbBr₃ perovskite layer and are divided into electrons and holes. The FTO glass collects electrons through the SnO₂ electron transport layer. At the same time, the holes reach the carbon electrode through the nickel oxide hole transport layer, generating a current between the two electrodes. As depicted in Figure 1b–d, the perovskite film fabricated by the multiprocedure spin-coating approach shows compact and pinhole-free morphology. After altering the CsPbBr₃ perovskite film with NiO and NiO-l-Cys, the mean crystal grain size remains almost unchanged, while the grain boundary gap narrows and the perovskite film is smooth and dense. AFM images of the perovskite films reveal root-mean-square (R_MS) values of 42.2, 35.9, and 39.6 nm for the pristine CsPbBr₃ perovskite film, modified film with NiO, and modified film with NiO-l-Cys, respectively. These results corroborate the improved flatness. Notably, the nickel oxide modified by l-Cys (Figure 1d) possesses better morphology than nickel oxide (Figure 1c), consistent with the transmission electron microscopy (TEM) image (Figure S1 of the Supporting Information). These micromorphological changes indicate reduced defects in the CsPbBr₃ perovskite film after NiO-l-Cys deposition, displaying an inhibited nonradiative recombination process and enhanced charge transport performance. During annealing, the C₃H₇NO₂S molecules (Figure S2) anchor to the CsPbBr₃ perovskite surface owing to the interaction between the S atom-containing lone pair of electrons and unharmonious positively charged ions (Pb²⁺ and Cs⁺). The role of the defect passivation effect was further studied by preparing a Cys-free hole transport layer.

Since the quality of absorbing layer films would directly affect the carrier transport performance and optical absorption quality, the CsPbBr₃ crystal phase was further characterized to clarify the role of NiO and NiO-l-Cys in the perovskite films. In Figures 1e and S3, three feature peaks are observed at 2θ = 15.2, 21.7, and 26.5, corresponding to (100), (110), and (111) facets of perovskite stage (JCPDS No: 54-0752), respectively. Thus, the perovskite structure does not change after depositing the hole transport layer on the perovskite surface. Nevertheless, the characteristic peaks shift to higher angles, meaning a decreased lattice constant, (30) caused by the lattice contraction induced by a smaller ion radius of Ni²⁺ (0.69 Å) compared to that of Pb²⁺ (1.190 Å), validating the partial substitution of Pb²⁺ by Ni²⁺ ions. The contracted lattice volume is favorable to stabilizing the perovskite (more proper τ value) and generating more energy. (31)

The surface chemical states and cooperation between l-Cys and perovskite films were further investigated by XPS characterization (Figure S4). Compared to CsPbBr₃, the binding energy of Pb 4f_7/2 in CsPbBr₃/NiO-Cys changed from 138.14 to 138.04 eV, while that of Pb 4f_5/2 decreased from 142.95 to 142.87 eV (high-resolution graphs in Figure 1f). (32,33) According to the lower binding energy of Pb 4f, the lone pair of electrons of the S or N elements of Cys interacts with the CsPbBr₃ lattice. The peak of S 2p at 158.7 eV in Figure 1g is attributed to the Pb–S bond. Since S is significantly smaller than Br, S is embedded in the CsPbBr₃ perovskite lattice gap instead of replacing Br, causing the perovskite lattice expansion. (34,35) As seen in Figure S5, the N 1s peak of the l-Cys molecule is located at 399.89 eV and attributed to deprotonated -NH₂, suggesting the adsorption of N atoms of l-Cys on the perovskite film surface. (36) In Figure S6, the EDS test mapping corroborates the XPS data, further confirming the existence of Ni, S, and O elements in the CsPbBr₃/NiO-Cys film.

The overall electron allocation on the l-Cys surface was obtained by calculating the electrostatic potential mapping (ESP, Figure 1h). The color change from red to blue indicates a reduction in the charge density from −6.513 × 10^–2 to 6.513 × 10². The higher electron density of the l-Cys molecule is induced by a more negative ESP of –COOH, determining whether –COOH can preferentially combine with –OH on NiO. (37,38) The –SH groups interact with Pb to form a strong coordination linkage between l-Cys and CsPbBr₃ because S atoms donate electrons to the oxygen vacancy. XRD and XPS show l-Cys acting as an interfacial chemical bridge between NiO and the CsPbBr₃ perovskite film for an improved interface association.

The integrated XPS, ESP, and XRD spectra reveal l-Cys’ association at the perovskite/NiO interface. In general, l-Cys can interact with the NiO side via –COOH and the perovskite side through -SH through chemical bridging at the perovskite/NiO interface. Such bilateral associations affect the interface properties, leading to enhanced contact performance and interface charge extraction/passivation. The roles of such a heterojunction interface in photovoltaic performance were further investigated by constructing all-inorganic PSCs with a configuration of FTO/SnO₂/CsPbBr₃ or (NiO or NiO-l-Cys)/carbon. Compared to the control CsPbBr₃ device, the PCEs of NiO-l-Cys-altered PSCs show an enhancement (Figure 2d). As summarized in Table 1, the tailored CsPbBr₃/NiO-l-Cys PSC realizes a maximum PCE of 9.61% with a V_OC of 1.614 V, a short-circuit current density (J_SC) of 7.23 mA·cm^–2, and a fill factor (FF) of 82.2% (J–V curves, Figure 2f) under sunlight irradiation. These values are much greater than those of the control device with a typical PCE of 7.35%, a short-circuit current (J_SC) of 6.65 mA·cm^–2, a V_OC of 1.504 V, and a fill factor of 73.50%, demonstrating the universality of NiO-l-Cys for improved device efficiency (Table S1). The photovoltaic parameters in Figures 2d and S7 revealed excellent repeatability and high performance of the NiO-l-Cys device with greatly suitable outcomes in IPCE spectra (Figure 2e) and stable power outputs (Figure 4d). In particular, significantly increased V_OC was obtained for the CsPbBr₃/NiO-l-Cys device. Since the splitting of quasi-Fermi degrees and reintegration inside the perovskite would control the V_OC, (39) the improvement can be caused by control of the interface exposed to the upper perovskite film when considering the nearly unvaried morphology and optical transmittance of the underlying ETL on the FTO substrate. This was confirmed by the optimized energy structure associated with an interfacial charge extraction–transfer action.

UPS testing of CsPbBr₃, CsPbBr₃/NiO, and CsPbBr₃/NiO-l-Cys films was used to identify the evolution of interfacial energetics. The binding energy cutoff (E_cutoff) and binding energy onset (E_onset) are illustrated in Figure 2b. The corresponding work function values (WF_s) and valence band maximum (E_VB) values of CsPbBr₃, CsPbBr₃/NiO, and CsPbBr₃/NiO-Cys films are 3.57, 3.59, and 3.63 eV and −5.84, −5.76, and −5.54 eV, respectively, calculated by the equation: WF_s = 21.2 eV – E_cutoff and E_VB = −WF_s – E_onset. (40,41) The energy level value of carbon is about −5.0 eV. (42) Thus, p-type doping of NiO greatly conforms to previous reports, (43) and the energy barrier for hole extraction from perovskite is found to be greatly decreased. Furthermore, the detailed energy-level diagrams are illustrated in Figure 2c, which show the conduction band (CB) and the valence band (VB) of CsPbBr₃, CsPbBr₃/NiO, and CsPbBr₃/NiO-Cys films based on the E_onset and bandgaps derived from Figure S8. From the optimized interfacial energy structure, such as the conversion from downward band-bending to upward band-bending owing to an up-shifted E_F of CsPbBr₃, CsPbBr₃/NiO, and CsPbBr₃/NiO-Cys films, the increased energy difference between the CB of CsPbBr₃, CsPbBr₃/NiO, and CsPbBr₃/NiO-Cys and the E_F of carbon (ΔE_e) (1.51, 1.59, vs 1.81 eV) and the reduced energy difference between the VB of CsPbBr₃, CsPbBr₃/NiO, and CsPbBr₃/NiO-Cys and the E_F of carbon (ΔE_h) (0.84, 0.76, vs 0.54 eV), the hole accumulation in the perovskite film is effectively suppressed, thus reducing the recombination loss and improving the voltage output. (37) The schematic diagram of carrier dynamics in the perovskite and hole transport layer interfaces is illustrated in Figure 2a. The excitons in the optimized perovskite film are separated and transferred smoothly, inhibiting the charge recombination.

Since the interfacial charge extraction–transfer behavior is highly dependent on the disadvantageous flaws generated during conversion and crystal development, the trap-state density (n_trap) of each perovskite film was quantized by space-charge-limited current (SCLC) measurements under dark conditions and electron-only structures (Figure 3a). n_trap can be calculated from the onset voltage (V_TFL) of the trap-filled limit region using eq 1: (44) where e represents the primary charge, L and ε represent the thickness and relative dielectric constant of the CsPbBr₃ film, respectively, and ε₀ indicates the vacuum permittivity. The calculations reveal remarkably decreased n_trap values from 2.65 × 10¹⁶ for the control CsPbBr₃ device to 8.75 × 10¹⁵ and 7.55 × 10¹⁵ for the CsPbBr₃ layer modified with NiO and NiO-l-Cys, respectively. Hence, the perovskite films possess reduced trap states. The influence of NiO-l-Cys alternation on carrier reintegration kinetics was clarified by exciting the films by using a laser at a wavelength of 385 nm. The PL and TRPL spectra in Figure 3c,d show opposite responses to traditional PL quenching in the CsPbBr₃/NiO-l-Cys-based perovskite film. The quenched PL intensity is related to the presence of Cys, which enhances the interface contact of perovskite/NiO by forming chemical bonding, further driving the interfacial electron extraction. (45) The mean carrier lifetime (τ_ave) can be calculated by eq 2: (46) where τ₁ represents the fast decay time caused by electron extraction by the hole transport layer and τ₂ reveals the slow decay time related to trap reintegration in the bulk perovskite film. A₁ and A₂ indicate their amplitudes. As shown in Table S2, the CsPbBr₃/NiO-l-Cys-based perovskite film displayed a shorter τ_ave (8.89 ns) than that of the CsPbBr₃-based perovskite film (11.13 ns). Therefore, the Cys alternation may drive the interface charge extraction according to the grown built-in electric field (V_bi, Figure 3b) originating from the capacitance–voltage (C–V) curves based on the Mott–Schottky equation, eq 3: (47) where A refers to the active surface, e represents the primary charge, ε denotes the related dielectric constant of the perovskite, ε₀ depicts the vacuum permittivity, and N represents the doping density of the perovskite. Compared to the control device, the CsPbBr₃/NiO-l-Cys-tailored device exhibits a higher V_bi, suitable for better diode characteristics of PSCs. Also, the leakage current of CsPbBr₃/NiO-l-Cys is significantly suppressed when compared to the control device (Figure 4a). The light-intensity dependence curves of V_OC in Figure 3b are evaluated by eq 4: (48)

The V_OC is directly proportional to the logarithm of I, and n refers to an ideal element related to monomolecular reintegration. T, k, and q represent the absolute temperature, Boltzmann constant, and elemental charge, respectively. The Shockley–Read–Hall (SRH) level and bimolecular reintegration are lower for smaller slope values. (49,50) According to Figure 4b, the slope of the control device is estimated to be (1.87) kT/e. By comparison, lower slopes of (1.56) and (1.42) kT/e are recorded for the devices modified with CsPbBr₃/NiO and CsPbBr₃/NiO-l-Cys, respectively. Hence, the lower disadvantageous flaw leads to recombination restraining due to decreased unharmonious Pb²⁺ in the CsPbBr₃ film. Meanwhile, the NiO-l-Cys-altered device displays a greater R_rec value (400 Ω) than the control CsPbBr₃ PSC device (R_rec of 340 Ω), indicating a greatly controlled charge reintegration in PSCs (Figure 4c). The higher R_rec values originating from the smaller trap-state density can stop the charge recombination and drive carrier extraction in the PSC device. Thus far, it is concluded that the CsPbBr₃/NiO-l-Cys structure induced energy degree control and defect reduction synergistically decide the enhanced V_OC.

The PCE longevity tests of PSC devices with and without NiO-l-Cys are compared in Figure 5a,b. The optimal PSC without encapsulation maintains 83% of the initial PCE after 30 days of storage in air with 0% relative humidity (RH) and a temperature of 80 °C. By comparison, the control PSC shows obvious degradation with a loss of 48% of its initial PCE. Meanwhile, the humidity stability of unencapsulated devices is evaluated at a high RH of 80% and 25 °C and displays a stronger humidity resistance of CsPbBr₃ /NiO-l-Cys devices after 30 days of aging, equivalent to a conservation rate of 98%. This value is higher than that of the control device (87%). The expanded contact angle from 48.9 to 60.0 indicates a higher degree of hydrophobicity for the CsPbBr₃ /NiO-l-Cys film (Figure 5c). By fabricating this all-in-one film to heal the CsPbBr₃/NiO-l-Cys interface effectively, the interface defects are suppressed to trigger phase degradation, thereby improving the long-term stability of the self-encapsulation devices.