Chemically suppressing redox reaction at the NiO_x/perovskite interface in narrow bandgap perovskite solar cells to exceed a power conversion efficiency of 20%

Abstract

NiO_x as a type of inorganic hole-transporting layer (HTL) material in narrow bandgap perovskite solar cells (NBG PSCs) showed exceptional stability but suffered a considerably poorer performance compared with NBG PSCs with commonly used PEDOT:PSS as the HTL. Herein, we found that redox reactions would occur at the interface between Ni³⁺ on the NiO_x surface and the easily oxidized Sn²⁺ in the perovskite, causing considerable non-radiative recombination centers. On this basis, we proposed a bifacial reduction strategy at the interface to boost the performance of NBG PSCs. By using a reductive reagent ascorbic acid to reduce the Ni³⁺/Ni²⁺ ratio on the surface of NiO_x beforehand, the possibility of contact between Ni³⁺ on the surface of NiO_x and perovskite is chemically reduced substantially, suppressing the redox reaction between them as well as the non-radiative recombination at the interface. By applying this strategy, the device's power conversion efficiency is elevated from 17.81% to 20.48%, with 91% remaining after 1128 hours of storage in a nitrogen-filled glovebox.

---

Introduction

With a reported power conversion efficiency (PCE) of 3.8%¹ in 2009 and to a certified 25.7%² in merely 12 years, the perovskite solar cell has attained much attention as one of the next-generation photovoltaic devices due to its advantages, such as high absorption coefficient, long carrier diffusion length, high carrier mobility, and adjustable bandgap.^3–6 The active layer in PSCs has a structure of ABX₃ (B = Pb, Sn). Partially substituting Pb with Sn can absorb long-wavelength photons, broadening the absorption spectral range and forming a narrow bandgap perovskite (NBG). Thus, a possibility to achieve a higher theoretical PCE is provided for a single junction solar cell⁷ and realize all-perovskite tandem solar cells after coupling with a wide bandgap perovskite solar cell to surpass the Shockley–Queisser limit of single junction solar cells.^8–11

The inferior performance of NBG PSCs can be ascribed to both perovskite intrinsic properties and matching of energy levels among functional layers. On the one hand, the existence of Sn²⁺ in perovskite impairs the device stability due to the inherent strong reducing property. To solve this problem, some reducing additives (such as ascorbic acid, tea polyphenol, caffeic acid, and 4-hydrazinobenzoic acid) have been reported to prolong the lifetime.^12–15 On the other hand, the mismatch of the energy band between the perovskite film and carrier-transporting layers (CTLs) leads to nonradiative recombination and energy losses.¹⁶ Interface modification by inserting a thin layer of other materials between the perovskite layer and CTLs has also been proven to be an effective way for performance enhancement.^17–19

The hole-transporting layer (HTL) plays a pivotal role in PSCs, especially for the inverted device structure, in which some stringent criteria must be complied such as good physical and chemical stability, band alignment, and high transmittance. NBG inverted PSCs with PEDOT:PSS as HTL have received predominant PCE after years of efforts.^20,21 However, PEDOT:PSS is characterized by acidity and hydrophilicity that would accelerate the degradation of the upper perovskite layer and thus damage the device stability.^22,23 Additionally, the severe nonradiative recombination caused by the unaligned energy level at the interface of PEDOT:PSS and perovskite layer results in a large energy loss for carrier extraction and consequent V_OC loss. Inorganic p-type materials, such as NiO_x, CuSCN, and CuI, are considered to be more predominant HTLs than PEDOT:PSS because of the naturally robust chemical stability and relatively deeper valence band maximum (VBM).^24,25 Particularly, NiO_x is most notable for its various film preparation methods. These benefits might present a better performance and stability of PSCs with NiO_x as HTL than PEDOT:PSS. However, few works have been reported for efficient NiO_x-based NBG PSCs, resulting in the slow development of this field. Chi et al. compared PEDOT:PSS and NiO_x as HTL in NBG PSCs in 2018, and found that NiO_x would lead to a high-quality perovskite film with higher device performance.²⁶ Han et al. applied NiO_x prepared at low temperature into NBG PSCs, and received a PCE value of 17.6%. Moreover, 95% of the initial PCE still remained after storage in a glove box for 102 days.²⁷ Although it is stable, the performance of the NiO_x-based NBG PSCs is still relatively low compared with its counterpart. To comprehend the reasons why the NiO_x-based NBG PSCs suffered such a poor performance, we have to turn our focus back on Pb-based PSCs since studies on the NBG PSCs based on NiO_x are very limited. Boyd et al. believed that in inverted PSCs, Ni^≥3+ on the surface of NiO_x, acting as Brønsted proton acceptors and Lewis electron acceptors, would deprotonate and oxidize methylammonium iodide (MAI) and formamidinium iodide (FAI) (both provide protons and electrons) in the precursor solution during the fabrication process. Thus, I₂, methylamine, and formamidine are produced and evaporated to form segregated PbI₂ at the interface and in perovskite films due to the absence of A-site cations. To solve this issue, they proposed two strategies: one is to add excess MAI or FAI to compensate for the A-site ions deficiency; the other is to use A-site component post-treat NiO_x films.²⁸ Some inert polymers and small molecules were deposited on NiO_x to separate both and to avoid their direct contact physically.^29–34 Moreover, Wang et al. came up with an in situ bifacial passivation strategy using HI to simultaneously reduce the Ni³⁺/Ni²⁺ ratio and increase the crystallization of perovskite.³⁵ In NBG PSCs, however, Sn²⁺ is easier to be oxidized and would similarly show deteriorated performance.

In this work, we demonstrated the reason why NBG PSCs with NiO_x as HTL suffered from poor performance. It is attributed to the redox reaction that happens between Ni³⁺ on the NiO_x surface and easily oxidizable Sn²⁺ in NBG. Accordingly, we propose a bifacial reduction strategy (BRS) to chemically suppress such redox reaction. The ratio of Ni³⁺/Ni²⁺ was decreased after a reductive reagent, ascorbic acid (AA, the structure is drawn in Fig. S1†) was modified on the NiO_x film. In addition, excess AA can prevent the oxidation of Sn²⁺ in NBG perovskite. The device performance can be improved greatly with the aid of BRS. Finally, NBG PSCs with a champion PCE of 20.48% was achieved with a structure of indium tin oxide (ITO)/NiO_x/AA/(FASnI₃)_0.6(MAPbI₃)_0.4/fullerene (C₆₀)/bathocuproine (BCP)/Ag, which is a high PCE of inverted NiO_x-based NBG PSCs. Furthermore, 91% of its initial PCE was retained after storage in a glove box for 1128 hours.

Results and discussion

Based on the Pb-based PSCs reported by McGehee et al.,²⁸ we speculate that a similar redox reaction would happen in NBG PSCs at the interface of NiO_x/perovskite between Sn²⁺ and Ni³⁺O_x, and all of the redox reactions can be described as follows:

HAPbI₃ ↔ HA⁺ + PbI₂ + I⁻ (1)

HASnI₃ ↔ HA⁺ + SnI₂ + I⁻ (2)

HA⁺ ↔ H⁺ + A(g) (3)

SnI₂ + 2I⁻ ↔ SnI₄ + 2e⁻ (5)

Ni³⁺O_x + e⁻ + H⁺ ↔ Ni²⁺O_xH (6)

As is illustrated, Ni³⁺ can act both as Lewis electron acceptors and Brønsted proton acceptors, oxidizing the iodide species and deprotonating the cationic amines (HA⁺, represents MA⁺ or FA⁺). In inverted NBG PSCs with NiO_x, obviously, eqn (5) is easier to proceed forward compared with eqn (4), as the standard redox potential of Sn⁴⁺/Sn²⁺ at 0.15 V is much lower than that of I₂/I⁻ at 0.54 V.³⁶ In this case, the Sn⁴⁺ formed by the redox reaction at the interfaces will cause recombination centers.

In order to corroborate our hypothesis that Ni³⁺ would oxidize Sn²⁺ and rule out the reducing effect from I⁻, SnCl₂ as the Sn²⁺ source and MACl as the H⁺ source were dissolved into DMF (Sn²⁺–H⁺ solution) and dropped onto a NiO_x film for the reaction process. Afterward, DMF was used for rinsing, and then the substrate was annealed on a hotplate. The X-ray photoelectron spectrum (XPS) was recorded to quantitatively measure the Ni³⁺/Ni²⁺ ratio on the surface of the NiO_x film. The result of Ni 2p_3/2 presented in Fig. 1a displays three primary peaks: 853.89 eV corresponding to Ni²⁺, 855.69 eV for Ni³⁺, as well as the satellite peak at 860.87 eV.^27,37 In comparison to bare NiO_x, the result obtained from the NiO_x film after the reaction shows a reduced Ni³⁺/Ni²⁺ ratio from 2.56 to 1.95 (Fig. 1a and b), demonstrating the redox reaction between Sn²⁺ and Ni³⁺. Such reaction would result in Sn²⁺ and HA⁺ vacancies at the interface, inducing p-doping, and thus leading to the formation of nonradiative recombination centers.^38–40 We chose a widely used AA as a reducing reagent to provide both electrons and protons for the redox reaction needed on the NiO_x surface prior to the deposition of the NBG perovskite layer. Dark black (characteristic color of Ni³⁺) NiO_x NPs visibly turn into a green color (characteristic color of Ni²⁺) after AA is added (Fig. S2†), manifesting their redox reaction. We then prepared the AA modified NiO_x film (rinsed) and measured them (the result is shown in Fig. 1c). A reduced Ni³⁺/Ni²⁺ ratio of 2.08 was also observed, further confirming the reducing effect from AA on NiO_x. Furthermore, in order to testify whether the redox reaction between Ni³⁺ and Sn²⁺ happened after AA treatment, we used AA to modify NiO_x (rinsed after), and then made the film react with the Sn²⁺–H⁺ solution. Fig. 1d displays a further decease of the Ni³⁺/Ni²⁺ ratio from 2.08 down to 1.96, which indicates that the redox reaction between bare NiO_x and Sn²⁺–H⁺ solution is largely suppressed. In addition, the XPS results (Fig. S3a–d†) originating from the O 1s results confirm the decrease of the Ni³⁺/Ni²⁺ ratio. The reason of the decrease will be discussed in a later section.

Fig. 1 High resolution XPS of (a) bare NiO_x film reacting with (b) Sn²⁺ and MA⁺, (c) AA, (d) AA, and then reacting with Sn²⁺ and MA⁺.

We modified NiO_x films with various concentrations of AA. To identify whether the modification would affect the optical properties of NiO_x films, we used UV-vis for characterization and the structure of glass/ITO/NiO_x/AA was made. No obvious difference is observed in the transmittance of NiO_x (Fig. S4a†), which indicates that the NiO_x film treated with AA does not influence its optical properties, and thus hardly affects the light incident into the upper perovskite layer. Furthermore, atomic force microscopy (AFM) was applied to test whether the surface properties were altered after AA treatment on the NiO_x film. The morphologies of the NiO_x film displays no significant change after AA treatment, as shown in Fig. S4b and c.† Furthermore, the root mean square (RMS) for NiO_x and NiO_x/AA are 9.38 and 9.10 nm, respectively. We also continued to test the absorbance of perovskites (Fig. S5a†) on NiO_x substrates treated with different concentrations of AA, and found no distinct change. Moreover, the Tauc plot (Fig. S5b†) displays the same bandgap value, which further indicates that a minor mass of AA treatment would not affect the perovskite layer's optical properties.

Based on the unenhanced absorbance of the perovskite, we speculated that the crystallinity of the perovskite would not be changed after AA modification. To test our hypothesis, we examined the morphology of NBG grown on NiO_x treated with the same AA concentrations as above via field emission scanning electron microscope (FESEM). As shown in Fig. S6a–e,† the top-view FESEM images of the NBG films show almost no difference regardless of the selected concentration. The cross-sectional FESEM images of the 0 and 0.8 mg per mL AA modification at the interface of the perovskite and NiO_x with a structure of ITO/NiO_x/AA/NBG are displayed in Fig. S6f and g.† Both exhibit vertical perovskite crystals without lateral grain boundaries. Moreover, all XRD patterns (Fig. S7a†) display similarly high diffraction peaks at 14.07° and 28.33°, corresponding to the preferred orientations at the (110) and (220) planes, respectively. These results are consistent with the negligible change on the grain sizes, and no peak shifts (Fig. S7b and c†) are observed for both (110) and (220) planes after AA treatments. These results are also the same as we presumed: the NiO_x film surfaces treated with AA have a negligible impact on the growth of the perovskite crystals. We then considered that the low performance of the narrow-bandgap PSCs with NiO_x as HTL could be attributed to the interface of NiO_x/NBG, rather than the crystallinity of the bulk perovskite.

Because the XRD results and FESEM images showed no discernible changes after AA modification, the optimal concentration must be figured out prior to further study. Fig. 2a displays the device fabrication process, and the device structure of ITO/NiO_x/(AA)/NBG/C₆₀/BCP/Ag is shown in Fig. 2b. As it can be seen clearly in Fig. S8,† the performance is enhanced firstly up to 0.8 mg per mL AA and then declines. To ensure the reproducibility, we fabricated 40 devices for each treatment condition, and made statistical graphs and performance chart (Fig. S9 and Table S1†). The statistical distribution on PCE with and without the optimal concentration of the AA-modified NiO_x is drawn in Fig. 2c. Fig. 2d shows the J–V curves of the champion devices: the control champion device displays a rather low performance with a short-circuit current density (J_SC) of 29.31 mA cm⁻², open-circuit voltage (V_OC) of 0.853 V, fill factor (FF) of 71.24% and a PCE of 17.81%, while the champion BRS device exhibits a J_SC of 30.62 mA cm⁻², V_OC of 0.877 V, and FF of 76.27%, resulting in an elevated PCE of 20.48%. To the best of our knowledge, this is a relatively high PCE with NiO_x as HTL in NBG PSCs, and we list the relevant information in Table S2.† Therefore, 0.8 mg mL⁻¹ was defined as the optimal concentration for further characterization hereinafter. Fig. 2e displays the external quantum efficiency (EQE) spectra and integrated J_SC of the champion devices based on bare NiO_x and BRS-NiO_x. The values of the integrated J_SC are 28.97 and 30.01 mA cm⁻², respectively, which is in good agreement with the J_SC value from the J–V curves. To evaluate the stability of the device under operation, maximum power point tracking was measured with their bias voltages under one sun AM 1.5 G illumination (Fig. 2f). The steady-state PCE of the BRS device at a bias voltage is 20.18% after 200 s illumination and the control device displays a PCE of 17.44%. Both are consistent with the efficiencies obtained from the J–V curves. The J–V curves of the champion control and BRS devices measured under the reverse and forward scans are shown in Fig. S10.† The hysteresis index (HI) was calculated from the formula: HI = PCE_reverse/PCE_forward. The specific parameters are listed in Table S3.† Both control and BRS devices show the same HI of 1.10. The unchanged HI may be ascribed to the unmatched HTL with ETL.⁴¹

Fig. 2 (a) The fabrication process of PSC devices. (b) Structure of devices. (c) Statistical PCE distribution of the control and BRS PSCs. (d) The J–V curves measured under 100 mW cm⁻² AM 1.5 G illumination, and (e) corresponding EQE spectrum and integrated J_SC of the champion devices. (f) Maximum power point (MPP) tracking under 100 mW cm⁻² AM 1.5 G illumination. (g) The J–V curves of the BRS devices with different rinsing times of DMF on the AA modified NiO_x.

As the AA is soluble in DMF and DMSO, a series of tiny amounts of AA was doped into the NBG perovskite precursor solutions to determine whether AA has a dominant effect in the bulk perovskite. Fig. S11a–c† show the XRD patterns of perovskite films doped with various concentrations of AA grown on NiO_x film. As is clearly displayed, the crystallization peak strength remains unchanged. The peak centers of the (110) and (220) plane-preferred orientation do not shift, suggesting that AA did not enter the perovskite lattice. Then, Fourier transform infrared (FTIR) spectrometry was applied to find out whether PbI₂ (representative for SnI₂ here as it is extremely easy to oxidize with O₂), MAI, FAI had interactions with AA. The results (Fig. S11d–f†) show that all of the main components in the NBG perovskite did not have interactions with AA. We then doped the same series of concentrations of AA into the NBG perovskite solution and fabricated devices. The performance is drawn in Fig. S12.† No changes are found between the samples with only minor amounts of doping and the control devices. This indicated that it would hardly affect the device performance although the trace amount of AA can be dissolved into the perovskite precursor solution. To further demonstrate that AA has a greater effect at the interface rather than in the perovskite layer, we then rinsed the AA–NiO_x films for two, four and six times with DMF. The XRD patterns (Fig. S13a–c†) still show only a negligible variation for the peak position and the intensities. Furthermore, the device performance displayed in Fig. 2g suggests that the AA produces the main effect at the interface.

Based on the above results, we suppose that the decrease of the Ni³⁺/Ni²⁺ ratio is because AA provides both protons and electrons for the redox reaction needed, as mentioned in the chemical eqn (6) prior to the perovskite film deposition on NiO_x. This prevents their reaction with each other, thus decreasing the HA⁺ and Sn²⁺ vacancies at the interface. In addition, AA as a reducing reagent with a tiny amount at the interface can prevent Sn²⁺ from oxidation.¹² Thus, the scheme of the redox reaction at the interface and the BRS working mechanism presented in Fig. 3 can be deduced. FTIR measurements were conducted to detect the oxidation form of AA after it reacted with NiO_x. In Fig. S14,† two peaks at 1797 and 1630 cm⁻¹ corresponding to ν(C=O) in dehydroascorbic acid (oxidized form of AA) appeared,^42,43 which further confirm the redox reaction between AA and NiO_x and provide evidence for our speculated mechanism.

Fig. 3 Scheme of the redox reaction at the NiO_x/NBG interface in the control and BRS device. Ni³⁺ on the NiO_x surface would oxidize Sn²⁺ and deprotonate HA⁺ in the NBG perovskite precursor, resulting in both Sn²⁺ and HA⁺ vacancies at the interface after perovskite deposition. When AA with a good reducing effect is modified on the NiO_x film prior to NBG deposition, both protons and electrons are provided for the redox reaction needed at the NiO_x surface, preventing the redox reaction between Ni³⁺ among Sn²⁺, HA⁺. Thus, Sn²⁺ and HA⁺ vacancies at the interface are decreased.

To quantitatively assess the defect densities of the perovskite films grown on bare NiO_x and NiO_x/AA, the hole-only devices were made (the structures are shown in the graph inseted in Fig. 4a and b). The space charge limited current (SCLC) measurement was carried out, and the trap density (n_t) is calculated by the following equation:

where V_TFL indicates the trap-filled limit voltage, L is the thickness of the perovskite film, e is the elementary charge, ε₀ and ε are the vacuum dielectric constant and relative dielectric constant of perovskite (here, 35 is selected⁴⁴), respectively. Based on the equation above, Fig. 4a and b show that the V_TFL decreases from 0.545 to 0.177 V, corresponding to n_t dropping from 4.99 × 10¹⁵ to 1.62 × 10¹⁵ cm⁻³ after AA modification. This indicates that BRS can reduce the trap-assisted recombination centers induced by HA⁺ deprotonation and Sn²⁺ oxidation⁴⁵ caused by Ni³⁺ at the interface.

Fig. 4 The space-charge-limited-current (SCLC) module for (a) bare NiO_x hole-only device and for (b) BRS hole-only device. (The inset is the device structure) (c) Mott–Schottky measurements of PSCs. (d) UPS measurements of the bare NiO_x film and rinsed AA treated NiO_x film. (e) Energy diagram of the layers in the PSC device. (f) V_OC dependence on the light intensity of the control and BRS devices.

To understand the origin of the upgraded V_OC of the BRS devices, Mott–Schottky study was performed. By this manner, the built-in potential (V_bi) of the devices could be determined by fitting the linear part of the C⁻²–V curves according to eqn (II):

where C is the measured capacitance, and A, e, and N are the active area, elementary charge, and carrier concentration, respectively. ε₀, ε, and e have the same meanings as those in eqn (I), and V is the applied bias voltage. As shown in Fig. 4c, the control device produces a built-in voltage of 0.41 V, which is less than that of the BRS device by 0.04 V. The enlarged V_bi would accelerate dissociation of the photogenerated carriers and inhibit dark recombination, thus increasing the dynamics in the separation rate of the carriers, which is beneficial for higher V_OC. In addition, this would result in an extended depletion region.⁴⁶ We then investigated the band structure of the NiO_x films through ultraviolet photoelectron spectroscopy (UPS). As displayed in Fig. 4d, the bare NiO_x film shows an unmatched valence band maximum (VBM) with a value of −5.20 eV. Meanwhile, the VBM dropped to −5.40 eV after AA treatment and rinsing, demonstrating a better matched alignment with the NBG perovskite and a lower hole extracting energy loss. The Fermi level increased relative to the VBM by around 0.1 eV after AA modification, indicating less p-type doping on the surface of the NiO_x film. In other words, there was less Ni³⁺ on it. Fig. 4e delineates the energy diagram of PSCs. To investigate whether the redox reaction-induced non-radiative recombination was suppressed after BRS, we further measured the dependence of V_OC on light intensity. The values of the ideality factor (n) were determined as 1.51 and 1.22 for the control and BRS devices (Fig. 4f), respectively, showing a decreased n value. This indicates that the trap-assisted charge recombination caused by the redox reaction at the interface between NiO_x and the perovskite film are effectively suppressed.

Fig. 5a displays the Nyquist plot of electrochemical impedance spectroscopy, and the inset graph is its equivalent circuit utilized for analyzing the data. The R_s and R_rec shown in Table S4† are the series resistance and recombination resistance, respectively. The R_rec of the BRS device with 1865 Ω is larger than that of the control device with 332 Ω, implying that the suppressed non-radiative recombination resulted from the pre-reduced Ni³⁺/Ni²⁺ ratio and less Sn²⁺ oxidation and HA⁺ deprotonation induced Sn²⁺ and HA⁺ vacancies at interface. Subsequently, steady-state photoluminescence (PL) and time-resolved PL (TRPL) spectra were performed to better understand the dynamics of the carrier at the interface. We deposited NBG perovskite films on bare NiO_x and BRS NiO_x. The results shown in Fig. 5b suggest that a larger PL quenching appeared after BRS, which could be caused by: (1) trap-assisted recombination states resulting from the intrinsic poor perovskite film quality and (2) the improvement of the hole extraction from the perovskite to NiO_x. The first possibility can be excluded as the trap-states decrease greatly, as discussed above through the SCLC measurement. For further quantifying the carrier quenching lifetime and to better understand the extraction process of the carrier from the perovskite at the interface, the TRPL spectra of the NBG perovskite film grown on NiO_x and AA–NiO_x were carried out (Fig. 5c). The corresponding curves were fitted through a bi-exponent decay function. τ₁ and τ₂ correspond to the fast and slow decay lifetimes, respectively, and the average carrier decay lifetimes (τ_ave) were calculated (Table S5†). A lower τ_ave of BRS (26.259 ns) than that of the control (41.169 ns) shows a faster carrier decay, which is anastomotic with the PL results. The transient photocurrent (TPC) was measured. Fig. 5d displays a faster TPC attenuation (1.13 μs) after AA treatment. These results all suggest a better carrier extraction at the NiO_x/NBG interface.

Fig. 5 (a) Nyquist plots of the corresponding PSCs under dark condition (embedded with the equivalent circuit). (b) Steady-state photoluminescence spectra and (c) time-resolved photoluminescence spectra of the perovskite film deposited on bare and BRS NiO_x (the laser was incident from the glass side). (d) Transient photocurrent of control and BRS devices.

Finally, the long-term stability of the unencapsulated NBG PSCs in a N₂-filled glove box with an oxygen concentration of ≤10 ppm and with a H₂O concentration of ≤0.1 ppm was measured. Fig. 6 shows that the PCE of the control device degrades to 83% of its initial value after 1128 h. Meanwhile, the BRS device retains 91% of its initial PCE value within the same storage time, suggesting that the BRS we proposed would not impair the stability of the NBG PSCs, while improving its performance.

Fig. 6 Stability of V_OC, J_SC, FF and PCE of the unencapsulated control and BRS PSCs in a nitrogen-filled glovebox with a H₂O concentration of ≤0.1 and oxygen concentration of ≤10 ppm.

Conclusions

We have demonstrated that the Ni³⁺ on the NiO_x surface would react with Sn²⁺ in NBG perovskite. Such a reaction would introduce a mass of recombination centers at the interface of NiO_x/NBG, leading to a poor performance of the NBG PSCs. After confirming this hypothesis, we proposed a bifacial reduction strategy to reduce Ni³⁺ on the NiO_x surface to Ni²⁺ using AA to provide the protons and electrons needed for the redox reaction, thus chemically avoiding the contact between Ni³⁺ and Sn²⁺. Through this strategy, the redox reaction-induced recombination centers decreased, which was mainly attributed to the decreased Ni³⁺/Ni²⁺ ratio and Sn²⁺, HA⁺ vacancies at interface. This was better for the band alignment and carrier transportation. Thus, a lower energy loss in the extraction process of the interface and a high PCE with 20.48% were finally achieved. Our bifacial reduction strategy from the aspect of chemical separation provides a new train of thought to overcome the redox reaction at the interface of NiO_x/NBG, contributing to the improvement of the NBG PSCs performance toward the development of future stable photoelectric devices.

Conflicts of interest

The authors declare no competing interests.

Acknowledgements

The authors acknowledge financial support from the National Natural Foundation of China (grant no. 11774293 and 12074321), and the Chongqing Science and Technology Commission (cstc2020jcyj-msxm1120). This work was also supported by the Fundamental Research Funds for the Central Universities (SWU118105) and Academic Budding Project (grant number ZS-XM[2021]1-4). The authors would like to thank Prof. Xianju Zhou, from the College of Science, Chongqing University of Posts and Telecommunications, for help with the PL measurement. The authors also would like to thank Mr Chuanyao Luo for guidance in the synthesis of the NiO_x nanoparticles.

Notes and references

1. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc., 2009, 131, 6050–6051.
2. https://www.nrel.gov/pv/assets/pdfs/cell-pv-eff-emergingpv-rev220630.pdf .
3. P. P. Boix, K. Nonomura, N. Mathews and S. G. Mhaisalkar, Mater. Today, 2014, 17, 16–23.
4. M. A. Green, A. Ho-Baillie and H. J. Snaith, Nat. Photonics, 2014, 8, 506–514.
5. J. P. Correa-Baena, M. Saliba, T. Buonassisi, M. Gratzel, A. Abate, W. Tress and A. Hagfeldt, Science, 2017, 358, 739–744.
6. R. Prasanna, A. Gold-Parker, T. Leijtens, B. Conings, A. Babayigit, H. G. Boyen, M. F. Toney and M. D. McGehee, J. Am. Chem. Soc., 2017, 139, 11117–11124.
7. X. Y. Jiang, Z. H. Zang, Y. Y. Zhou, H. S. Li, Q. Wei and Z. J. Ning, Acc. Mater. Res., 2021, 2, 210–219.
8. P. Umari, E. Mosconi and F. De Angelis, Sci. Rep., 2014, 4, 4467.
9. Y. Ogomi, A. Morita, S. Tsukamoto, T. Saitho, N. Fujikawa, Q. Shen, T. Toyoda, K. Yoshino, S. S. Pandey, T. L. Ma and S. Hayase, J. Phys. Chem. Lett., 2014, 5, 1004–1011.
10. G. E. Eperon, T. Leijtens, K. A. Bush, R. Prasanna, T. Green, J. T. W. Wang, D. P. McMeekin, G. Volonakis, R. L. Milot, R. May, A. Palmstrom, D. J. Slotcavage, R. A. Belisle, J. B. Patel, E. S. Parrott, R. J. Sutton, W. Ma, F. Moghadam, B. Conings, A. Babayigit, H. G. Boyen, S. Bent, F. Giustino, L. M. Herz, M. B. Johnston, M. D. McGehee and H. J. Snaith, Science, 2016, 354, 861–865.
11. Y. Q. Yao, F. Lv, L. Luo, L. P. Liao, G. Wang, D. B. Liu, C. Y. Xu, G. D. Zhou, X. S. Zhao and Q. L. Song, Sol. RRL, 2020, 4, 1900396.
12. X. B. Xu, C. C. Chueh, Z. B. Yang, A. Rajagopal, J. Q. Xu, S. B. Jo and A. K. Y. Jen, Nano Energy, 2017, 34, 392–398.
13. H. X. Ban, Q. Sun, T. Zhang, H. Li, Y. Shen and M. K. Wang, Sol. RRL, 2020, 4, 1900457.
14. H. Liu, L. Wang, R. Li, B. Shi, P. Wang, Y. Zhao and X. Zhang, ACS Energy Lett., 2021, 6, 2907–2916.
15. J. Cao, H. L. Loi, Y. Xu, X. Guo, N. Wang, C. k. Liu, T. Wang, H. Cheng, Y. Zhu, M. G. Li, W. Y. Wong and F. Yan, Adv. Mater., 2021, 34, 2107729.
16. M. Wang, H. X. Wang, W. Li, X. F. Hu, K. Sun and Z. G. Zang, J. Mater. Chem. A, 2019, 7, 26421–26428.
17. Y. Zhong, C. Li, G. Xu, C. Xu, J. Dong, D. Liu, D. Lu, J. You, C. Gao and Q. Song, Chem. Eng. J., 2022, 436, 135134.
18. X. Jiang, W. S. Subhani, K. Wang, H. Wang, L. J. Duan, M. Y. Du, S. P. Pang and S. Z. Liu, Adv. Mater. Interfaces, 2021, 8, 2001994.
19. B. B. Liu, H. Bi, D. M. He, L. Bai, W. Q. Wang, H. K. Yuan, Q. L. Song, P. Y. Su, Z. G. Zang, T. W. Zhou and J. Z. Chen, ACS Energy Lett., 2021, 6, 2526–2538.
20. R. X. Lin, J. Xu, M. Y. Wei, Y. R. Wang, Z. Y. Qin, Z. Liu, J. L. Wu, K. Xiao, B. Chen, S. M. Park, G. Chen, H. R. Atapattu, K. R. Graham, J. Xu, J. Zhu, L. D. Li, C. F. Zhang, E. H. Sargent and H. R. Tan, Nature, 2022, 603, 73–78.
21. G. Kapil, T. Bessho, Y. Sanehira, S. R. Sahamir, M. Chen, A. K. Baranwal, D. Liu, Y. Sono, D. Hirotani, D. Nomura, K. Nishimura, M. A. Kamarudin, Q. Shen, H. Segawa and S. Hayase, ACS Energy Lett., 2022, 7, 966–974.
22. M. Kim, M. Yi, W. Jang, J. K. Kim and D. H. Wang, Polymers, 2020, 12, 129.
23. S. Ma, W. Y. Qiao, T. Cheng, B. Zhang, J. X. Yao, A. Alsaedi, T. Hayat, Y. Ding, Z. A. Tan and S. Y. Dai, ACS Appl. Mater. Interfaces, 2018, 10, 3902–3911.
24. A. C. Nkele, A. C. Nwanya, N. M. Shinde, S. Ezugwu, M. Maaza, J. S. Shaikh and F. I. Ezema, Int. J. Energy Res., 2020, 44, 9839–9863.
25. X. C. Yang, H. X. Wang, B. Cai, Z. Yu and L. C. Sun, J. Energy Chem., 2018, 27, 650–672.
26. D. Chi, S. H. Huang, M. Y. Zhang, S. Q. Mu, Y. Zhao, Y. Chen and J. B. You, Adv. Funct. Mater., 2018, 28, 1804603.
27. Q. L. Han, Y. Wei, R. X. Lin, Z. M. Fang, K. Xiao, X. Luo, S. Gu, J. Zhu, L. M. Ding and H. R. Tan, Sci. Bull., 2019, 64, 1399–1401.
28. C. C. Boyd, R. C. Shallcross, T. Moot, R. Kerner, L. Bertoluzzi, A. Onno, S. Kavadiya, C. Chosy, E. J. Wolf, J. Werner, J. A. Raiford, C. de Paula, A. F. Palmstrom, Z. S. J. Yu, J. J. Berry, S. F. Bent, Z. C. Holman, J. M. Luther, E. L. Ratcliff, N. R. Armstrong and M. D. McGehee, Joule, 2020, 4, 1759–1775.
29. X. M. Lian, J. H. Chen, S. Q. Shan, G. Wu and H. Z. Chen, ACS Appl. Mater. Interfaces, 2020, 12, 46340–46347.
30. X. Y. Hou, F. J. Li, X. Zhang, Y. F. Shi, Y. X. Du, J. B. Gong, X. D. Xiao, S. Q. Ren, X. Z. Zhao and Q. D. Tai, Sol. RRL, 2021, 5, 2100287.
31. J. Zhang, J. Long, Z. Huang, J. Yang, X. Li, R. Dai, W. Sheng, L. Tan and Y. Chen, Chem. Eng. J., 2021, 426, 131357.
32. Y. X. Guo, J. Ma, H. B. Wang, F. H. Ye, L. B. Xiong, H. W. Lei and Z. J. Tan, Adv. Mater. Interfaces, 2021, 8, 2100920.
33. W. B. Han, G. H. Ren, Z. Q. Li, M. N. Dong, C. Y. Liu and W. B. Guo, J. Energy Chem., 2020, 46, 202–207.
34. L. Zheng, Y. Xuan, J. Wang, S. Bao, X. Liu and K. Zhang, J. Mater. Chem. A, 2022, 10, 7251–7262.
35. H. Y. Wang, Z. Q. Huang, S. Q. Xiao, X. C. Meng, Z. Xing, L. Rao, C. X. Gong, R. S. Wu, T. Hu, L. C. Tan, X. T. Hu, S. H. Zhang and Y. W. Chen, J. Mater. Chem. A, 2021, 9, 5759–5768.
36. W. M. Latimer, Soil Sci., 1939, 48(4), 349.
37. S. Wang, Y. Li, J. Yang, T. Wang, B. Yang, Q. Cao, X. Pu, L. Etgar, J. Han, J. Zhao, X. Li and A. Hagfeldt, Angew. Chem., Int. Ed., 2022, 61, e202116534.
38. R. X. Lin, K. Xiao, Z. Y. Qin, Q. L. Han, C. F. Zhang, M. Y. Wei, M. I. Saidaminov, Y. Gao, J. Xu, M. Xiao, A. D. Li, J. Zhu, E. H. Sargent and H. R. Tan, Nat. Energy, 2019, 4, 864–873.
39. J. Xu, A. Maxwell, M. Y. Wei, Z. W. Wang, B. Chen, T. Zhu and E. H. Sargent, ACS Energy Lett., 2021, 6, 4220–4227.
40. Q. Y. Chen, J. C. Luo, R. He, H. G. Lai, S. Q. Ren, Y. T. Jiang, Z. X. Wan, W. W. Wang, X. Hao, Y. Wang, J. Q. Zhang, I. Constantinou, C. L. Wang, L. L. Wu, F. Fu and D. W. Zhao, Adv. Energy Mater., 2021, 11, 2101045.
41. D. H. Kong and N. G. Park, Adv. Mater., 2019, 31, 1805214.
42. E. S. Wagner, W. White, M. Jennings and K. Bennett, Biochim. Biophys. Acta, 1987, 902, 133–136.
43. K. B. Kokoh, F. Hahn, A. Metayer and C. Lamy, Electrochim. Acta, 2002, 47, 3965–3969.
44. H. Chen, Z. J. Peng, K. M. Xu, Q. Wei, D. N. Yu, C. C. Han, H. S. Li and Z. J. Ning, Sci. China Mater., 2021, 64, 537–546.
45. B. Li, B. H. Chang, L. Pan, Z. H. Li, L. Fu, Z. B. He and L. W. Yin, ACS Energy Lett., 2020, 5, 3752–3772.
46. M. Du, S. Zhao, L. Duan, Y. Cao, H. Wang, Y. Sun, L. Wang, X. Zhu, J. Feng, L. Liu, X. Jiang, Q. Dong, Y. Shi, K. Wang and S. Liu, Joule, 2022, 6, 1931–1943.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2ta06211a

‡ Equal contribution.

---