Solution-processed CuBO₂ hole transport layers for stable p–i–n perovskite solar cells

Abstract

Cu-based delafossite oxides have shown great potential as hole transport materials (HTMs) for perovskite solar cells (PSCs). As a member of the delafossite family, CuBO₂ has not received much attention despite its advantages of wide bandgap and high carrier mobility. Herein, a facile one-step sol–gel method is developed to fabricate CuBO₂ hole transport layers directly on FTO substrates for PSCs. Through precise optimization of the precursor concentration and pH value, flat and dense CuBO₂ films are obtained. Moreover, the influence of annealing temperature on the crystalline phase and photoelectrical properties of the CuBO₂ films is systematically investigated. Ultraviolet photoelectron spectroscopy results show that the valence band edge of the CuBO₂ film matches well with that of the perovskite layer, effectively facilitating hole extraction. The as-prepared inverted PSCs based on the CuBO₂ HTM achieve a remarkable power conversion efficiency of 18.26%. This study not only validates the potential of CuBO₂ as a promising HTM but also demonstrates a new avenue for novel inorganic HTMs in photovoltaic applications.

---

1. Introduction

Perovskite solar cells (PSCs) have been regarded as a promising next-generation solar power technology.^1–4 As an important component of PSCs, hole transport materials (HTMs) play a vital role in extracting photogenerated holes from the perovskite light-absorbing layer and suppressing carrier recombination. Traditional HTMs used in PSCs are organic p-type semiconductors. In early 2012, the small organic molecule 2,2′,7,7′-tetra[N,N-bis(4-methoxyphenyl)amino]-9,9′-spirobifluorene (spiro-OMeTAD) laid the foundation for high-efficiency PSCs.^5,6 To date, state-of-the-art normal (n–i–p) structured PSCs still use spiro-OMeTAD as HTMs. Meanwhile, poly[bis(4-phenyl)(2,4,6-trimethylphenyl)amine] (PTAA) and poly(3,4-ethylenedioxythiophene)/poly(styrenesulfonate) (PEDOT:PSS) HTMs are used in inverted (p–i–n) structured PSCs and require a high temperature annealing step. However, the complicated synthetic routes of these organic HTMs lead to a high price and low yield. Spiro-OMeTAD is amorphous and not conducive to the rapid transfer of carriers. It is necessary to dope lithium salts (bistrifluoromethanesulfonimide) followed by slow oxidation of the amorphous film in air to improve its conductivity. Nevertheless, lithium salts are highly hygroscopic, which induces degradation risks of the perovskite layers in humid environments.⁷ As for the PEDOT:PSS HTM, the valence band maximum has a large mismatch with that of perovskite, resulting in a relatively low V_OC.^8–10 Moreover, the hygroscopic and acidic characteristics of PEDOT:PSS trigger the stability issue of perovskites.

Many inorganic p-type semiconductors have properties such as high stability, high carrier mobility, and low price. In recent years, transition metal oxides¹¹ such as NiO,^12–14 CuO_x,^15–17 MoO_x,¹⁸ and CoO_x^19–21 have exhibited excellent performances as HTMs in PSCs. Excluding binary oxides, copper-based ternary oxides have emerged as potential HTMs owing to their advantages of wide bandgaps and high carrier mobility. These ternary oxides typically adopt two configurations, i.e., CuMO₂ and CuM₂O₄, corresponding to the delafossite and spinel structures, respectively.²² Most delafossite materials are p-type semiconductors and exhibit optical transparency owing to their wide bandgaps, making them highly promising for optoelectronic applications. In CuMO₂, Cu is a monovalent cation, while M is a trivalent cation, typically including elements such as Ga, Cr, Al, and Fe.²³ Compared to the binary oxides, the Cu⁺ ions in the CuMO₂ lattice are more closely spaced, leading to a wider bandgap. Additionally, CuMO₂ films exhibit high transparency in the visible light region. The hybridization of the d orbital of Cu with the p orbital of oxygen enhances hole mobility.²⁴ In recent years, delafossite materials such as CuCrO₂,^25–28 CuAlO₂,^29–31 CuFeO₂,³² and CuGaO₂^33,34 have been successfully applied in PSCs, and their highest PCE even exceeds that of organic PEDOT:PSS-based devices.

Theoretical calculations show that the holes in delafossite-type CuMO₂ compounds are mainly transported through the Cu hexagonal closed plane.³⁵ The bond length of Cu–Cu directly affects the conductivity of the material. Additionally, the band gap increases with the decrease of the ionic radius of M. Compared with other delafossite-type oxides, CuBO₂ has the smallest Cu–Cu bond length and the smallest radius of B ions at M site, which may contribute to its higher conductivity and largest band gap.^35–37 CuBO₂ was first reported in 2007, with a band gap of 2.2–4.5 eV, and the film is highly transparent in the visible light range and has a conductivity of 1.65 S cm⁻¹ at room temperature, which is higher than that of CuCrO₂.³⁸ CuBO₂ nanocrystals have been used as photocatalysts^39,40 and counter electrodes in dye-sensitized solar cells.³⁸ Although CuBO₂ is a potential HTM, the preparation of the film is challenging. The thickness of HTLs in PSCs is only tens to one hundred nanometers, and, in order to ensure that holes are extracted quickly, the film should cover the substrate uniformly to prevent leakage.^41,42 However, it is difficult to obtain delafossite-type materials at low temperature, while the high-temperature synthetic process leads to the agglomeration issue of nanoparticles, which makes it difficult to form dense films. CuBO₂ can be prepared by the sol–gel method,²⁴ which involves calcining precursors at elevated temperatures to produce powders. However, these powders often face challenges in redispersion within solvents. Alternative approaches, such as the hydrothermal and molten salt methods, have also been employed to prepare CuBO₂ powders.^39,43–45 However, these methods typically yield particles ranging from hundreds of nanometers to several micrometers in size which fail to meet the essential requirements for high-quality HTLs.

In this study, we developed a solution-processed approach for fabricating CuBO₂ nanocrystalline films as hole transport layers (HTLs) in p–i–n type PSCs (Fig. 1). Through meticulous optimization of precursor concentration and pH value, we achieved highly continuous and compact CuBO₂ films. Systematic investigations were conducted to examine the influence of post-annealing temperature on the phase composition, crystallinity, and optoelectronic properties of the resulting films. Notably, the energy level alignment of the CuBO₂ film demonstrated excellent compatibility with the perovskite layer. The CuBO₂-based inverted PSCs achieved a remarkable power conversion efficiency (PCE) of 18.26%, significantly outperforming the PEDOT:PSS-based devices (14.84%).

Fig. 1 Schematic of the preparation process of the CuBO₂ HTLs for p–i–n structured PSCs.

2. Experimental section

2.1. Preparation of CuBO₂ films

The precursor is prepared using a sol–gel method. The recipe for 0.4 M CuBO₂ precursor is as follows: 8 mmol Cu(NO₃)₂·3H₂O was dissolved in 10 mL deionized water and 8 mmol H₃BO₃ was dissolved in 10 mL 3 M HNO₃, and then the two solutions were mixed. Subsequently, 16 mmol of anhydrous citric acid was added as a chelating agent to the above solution. The pH values of the precursor were adjusted with ethylenediamine to be 0.5, 1.75 and 2.75. The above precursor was refluxed in a water bath at 75 °C for 8 h to obtain a lake blue sol which was aged for one day. Precursors with concentrations of 0.1, 0.2, and 0.3 M were also synthesized by the same method. Before preparing the films, FTO substrates were ultrasonically cleaned for 15 min with glass detergent, deionized water, and ethanol, sequentially. The surface organics were removed by ultraviolet ozone treatment after drying. 30 μL of precursor was dropped on the FTO surfaces and spin-coated at a rate of 3000 rpm for 30 s, then dried on a hot plate at 200 °C. After 15 minutes, the films were transferred to a muffle furnace for annealing at temperatures of 350, 450, and 550 °C for 60 min, with a heating rate of 10 °C min⁻¹, to obtain the HTLs.

2.2. Device fabrication

The perovskite layers and electron transport layers (ETLs) were prepared in a glove box. 1.4 M perovskite precursor solution was prepared by mixing FAI, PbI₂, MABr, PbBr₂ and CsI in DMF : DMSO mixed solvent (4 : 1 v : v) with a chemical composition of Cs_0.05(FA_0.83MA_0.17)_0.95Pb(I_0.83Br_0.17)₃.^41,46 50 μL of the perovskite precursor was spin-coated on the surface of CuBO₂ films at 1000 rpm for 10 s and 6000 rpm for 20 s. 200 μL of chlorobenzene as an anti-solvent was dripped on the perovskite film at 10 s before the end of the last procedure and then annealed at 105 °C for 30 min. After cooling to room temperature, 40 μL of a PC₆₁BM chlorobenzene solution (15 mg mL⁻¹) was dropped onto the surface of the perovskite films and spin-coated at 1500 rpm for 30 s, followed by annealing at 100 °C for 10 min. Finally, the silver electrodes were deposited onto the surface of the PC₆₁BM films using a vacuum evaporator under a vacuum of 6 × 10⁻⁴ Pa and an applied current of 50 A.

2.3. Characterizations

The X-ray diffraction (XRD) pattern of the samples was tested by the D2 HASER X-ray diffractometer. The scanning electron microscopy (SEM) images of the films were observed on a Hitachi SU8010. The energy dispersive X-ray spectrometry (EDX) pattern of the sample was tested by a ZEISS Gemini 300 SEM to determine the elements and distributions. The transmission and absorption spectra of the films were measured by a UV-3600 ultraviolet-visible spectrophotometer. X-ray photoelectron spectroscopy (XPS) and ultraviolet photoelectron spectroscopy (UPS) tests were performed on a Thermo Scientific ESCALAB 250Xi X-ray photoelectron spectrometer. The contact angles were measured using an AST VCA Optima XE droplet analyzer. The photoluminescence (PL) spectra and time-resolved photoluminescence (TRPL) spectra were measured using an Edinburgh FLS 1000 photoluminescence spectrometer. The performance of the PSCs was tested using a Newport 96160 solar simulator and a digital source meter (Keithley 2400). The light intensity was calibrated by an NREL-certified standard silicon solar cell with an intensity of 100 (mW cm⁻²). The electrochemical impedance spectroscopy (EIS) and Mott–Schottky plots were obtained using a Zahner–Elektrik electrochemical workstation.

3. Results and discussion

We firstly investigated the influence of the precursor concentration and pH value on the surface morphologies of CuBO₂ films. Fig. S1a–d (ESI†) show the surface SEM images of CuBO₂ films prepared with varying precursor concentrations ranging from 0.1 to 0.4 M at pH of 0.5 and annealing temperature of 550 °C. As the precursor concentration increases, the coverage of the film gradually enhances. The film prepared with a precursor concentration of 0.4 M shows a dense and smooth structure composed of tightly packed small crystalline grains fully covering the FTO substrate. It is noteworthy that a homogeneous film can be easily achieved under ambient conditions with a temperature above 20 °C and humidity below 60%. To explore the influence of pH on the film, precursors with pH values of 0.5, 1.75 and 2.75 were investigated. Fig. S2 (ESI†) shows that the crystalline size obviously increases with the increase of pH, and the smallest particles are obtained at pH 0.5. Adjusting the precursor pH significantly alters the surface charge density in the precursor solution, which in turn affects the chemical composition and interfacial tension of the particles. At pH 0.5, the surface charge reaches a maximum while interfacial tension is minimized, achieving a thermodynamic colloidal steady state that reduces secondary aging and inhibits particle growth.⁴⁶ In general, dense and flat HTLs composed of tiny particles minimize interfacial contact resistance, whereas larger particles induce carrier leakage and degrade device performance. Thus, a precursor concentration of 0.4 M at pH 0.5 was selected for further studies.

The annealing temperature has a significant influence on the phase evolution and optoelectronic properties of the final CuBO₂ films. Fig. 2 shows the surface morphologies of films annealed at 350 °C, 450 °C and 550 °C, and all films exhibit a flat surface. From the cross-sections, it is found that the film thickness decreases from 55 nm to about 20 nm when the annealing temperature increases from 350 °C to 550 °C. This could be attributed to the incomplete decomposition of the precursor at a lower temperature, while a higher temperature promotes the transformation of the loosely amorphous phase to a densely crystalline phase. The thickness reduction is indeed beneficial to increase the light transmittance and reduce the series resistance. Since hindered charge transport can reduce the photocurrent density of PSCs, reducing the thickness of the HTL film can help optimize the device performance when the film fully covers the FTO substrate.⁴⁶ The surface wettability of the CuBO₂ films is also investigated. The contact angles of the films gradually decrease with the increase of the annealing temperature, and the film annealed at 550 °C shows the best spreadability. Due to its flat and smooth surface, the perovskite precursor can spread smoothly without ozone treatment, which helps the perovskite to form a good crystalline film.^47,48

Fig. 2 Top-view and cross-sectional SEM images of the CuBO₂ films annealed at (a) and (b) 350 °C, (c) and (d) 450 °C and (e) and (f) 550 °C. The insets are contact angles.

To study the phases formed at different temperatures, the corresponding powder samples were used for XRD characterization. Fig. S3 (ESI†) shows digital photos of the precursor and three powder samples. When the temperature increases from 350 °C to 550 °C, the powders change from black to dark green. Fig. 3a shows the XRD patterns, and the samples obtained at 350 °C and 450 °C have more obvious burrs, indicating their relatively poor crystallinity. The diffraction peaks at 36.4° and 42.3° of the sample at 350 °C correspond to the (111) and (200) crystal planes of Cu₂O, and the peaks at 43.32° and 50.45° correspond to the Cu. Therefore, a mixture of Cu₂O and Cu is formed at 350 °C. The sample annealed at 450 °C has two high-intensity diffraction peaks at 35.54° and 38.63° which are attributed to the (002) and (111) crystal planes of CuO. Therefore, when the temperature rises to 450 °C, Cu₂O is oxidized to CuO. The sample annealed at 550 °C has obvious diffraction peaks at 35.4°, 36.4°, 38.5° and 42.3° which correspond to the CuBO₂ standard card (PDF# 28-1256).³⁸ To corroborate the structural characteristics and particle size distribution of CuBO₂ nanocrystals, transmission electron microscopy (TEM) analysis was performed, as shown in Fig. S4 (ESI†). We observed an average particle size of CuBO₂ of 6–10 nm and a lattice fringe spacing of 2.53 Å corresponding to the (100) crystallographic plane, which is in agreement with the corresponding interplanar spacing derived from XRD analysis.

Fig. 3 (a) XRD patterns of CuBO₂ annealed at different temperatures. (b) UV-Vis transmission spectra. (c) UV-Vis absorption spectra and the corresponding Tauc plots. (d) Dark I–V curves of the devices (the inset is a schematic diagram of the structure). (e) Ultraviolet-photoelectron spectroscopy of the CuBO₂ film. (f) Secondary electron cut-off region and valence band region.

Scanlon et al.⁴⁹ have studied the geometric structure and electronic properties of CuBO₂ through density functional theory (DFT) and GGA+U and HSE06 functional correction of electronic correlation effect. The valence band maximum (VBM) is mainly dominated by the Cu 3d state which hybridizes with the O 2p state, which is beneficial to hole conduction. According to the conductivity and Hall measurements, the estimated Hall mobility is 100 cm² V⁻¹ s⁻¹.⁵⁰ The calculated hole effective mass is low (1.71 m_e), which indicates that it has excellent p-type conductivity potential. The conduction band minimum (CBM) is composed of Cu d/s and O s states, and B has a small contribution. Furthermore, the high conductivity of CuBO₂ is attributed to the short Cu–Cu distance (2.49 Å), which promotes the polaron jump conduction of holes.⁵¹ Compared with other CuMO₂ (M = Al, Cr) delafossite oxides, the introduction of B minimizes the interference to the valence band characteristics of Cu₂O and retains good p-type characteristics.

EDX was employed to verify the elemental composition. The CuBO₂ powders annealed at 550 °C were ultrasonically dispersed in ethanol and subsequently deposited on silicon wafers for characterization. Fig. S5a (ESI†) shows the spectrogram obtained from the point scan, while the inset is the corresponding SEM image. The peaks clearly indicate the presence of three elements: Cu, B, and O. Fig. S5b (ESI†) shows the EDX mapping, which exhibits a uniform distribution of Cu and O elements with well-defined contours. The B element shows a relatively lower distribution, which can be attributed to its low atomic weight affecting the percentage of B atoms excited by electron bombardment during EDX measurements. In addition, XPS was performed to further analyze the chemical states of the film. Fig. S6a (ESI†) shows the full XPS spectrum, which reveals the presence of copper, boron and oxygen elements. Fig. S6b (ESI†) shows the high-resolution XPS spectrum of Cu, and the peaks are assigned to the Cu 2p_3/2 and Cu 2p_1/2 orbitals, accompanied by satellite peaks between them. The Cu 2p_3/2 core energy level spectrum is deconvoluted into two distinct peaks at 932.7 eV and 935.2 eV, corresponding to Cu⁺ and Cu²⁺, respectively (Fig. S6c, ESI†). Similarly, the Cu 2p_1/2 core energy level spectrum in Fig. S6d (ESI†) is also deconvoluted into two peaks at 953.1 eV and 954.9 eV which are attributed to Cu⁺ and Cu²⁺ in the surface oxides. Therefore, copper in CuBO₂ primarily exists in the +1 oxidation state (d¹⁰ electronic configuration). A small amount of Cu²⁺ signal observed at binding energies of 935.2 eV and 954.9 eV indicates that part of the surface of CuBO₂ film is oxidized under environmental conditions. The presence of Cu²⁺ in the split peaks, along with the satellite peaks (Cu²⁺), is a common phenomenon in delafossite oxides.^39,52 The peak located at 191.65 eV corresponds to B³⁺ in boron oxide (Fig. S6e, ESI†). Boron in CuBO₂ is usually in the +3 oxidation state (B³⁺), forming a [BO₃]³⁻ triangular planar unit (similar to other borates).⁵³ The oxidation state of B³⁺ in the film is stable, and the depth profile changes little.⁵⁴ The core energy level of O 1s is deconvoluted into two peaks at 530.7 eV and 531.8 eV (Fig. S6f, ESI†). The dominant peak at 530.7 eV is associated with lattice oxygen in CuBO₂, while the peak at 531.8 eV corresponds to surface-adsorbed oxygen. The subsurface is a Cu⁺/Cu²⁺ mixture and stable B³⁺. The bulk phase is pure Cu⁺ (CuBO₂ phase), and the stoichiometric ratio of B³⁺ and O²⁻ is maintained.^31,38

Fig. 3b shows the UV-Vis transmission spectra of films deposited on FTO substrates. The films annealed at three different temperatures exhibit excellent transmittance, and the transmittance increases gradually with the increase of temperature. The CuBO₂ film annealed at 550 °C has the highest transmittance of 84%. The absorption spectra of films annealed at different temperatures are shown in Fig. 3c. The bandgap of the samples is fitted according to the Tauc plots converted from the absorption curves. The bandgap (E_g) of CuBO₂ film is 3.57 eV, which verifies the characteristic of wide bandgap. To compare the electrical conductivities of the films at different temperatures, the dark state I–V curves were obtained using a device structure of FTO/HTL/Ag. The magnitude of the film conductivity is directly reflected by the slope in Fig. 3d. When the temperature increases to 550 °C, the conductivity of the CuBO₂ film is significantly improved. Therefore, the conductivity of the CuBO₂ film is higher than that of the CuO_x film. In addition, the low crystallinity of the film may be disadvantageous for charge transport at low temperatures. At high temperatures, the crystallinity increases and the dislocation density and strain in the film decrease, thus improving the crystallization quality of the film.^55,56 Therefore, the following investigations mainly focus on the CuBO₂ film annealed at 550 °C.

To determine the flat band potential (E_fb) and carrier concentration (N_D) of CuBO₂ film, Mott–Schottky tests were performed in 0.1 M Na₂SO₄ solution using a three-electrode system with FTO/CuBO₂/Ag as the working electrode, Pt as the counter electrode, and Ag/AgCl as the reference electrode. As shown in Fig. S7 (ESI†), the negative slope of the fit indicates the p-type semiconductor property of the CuBO₂, and the E_fb is determined to be 0.28 V from the linearly extrapolated intercept of the curve, which is consistent with previous research.³⁸ Since there is no previous report on the dielectric constant of CuBO₂, we used the similar delafossite-type material (CuFeO₂, ε = 20)³² to calculate N_D according to the formula ,⁵⁷ which results in 7.35 × 10¹⁶ cm⁻³ as calculated from the fitting slope of the Mott–Schottky plot.

Since PSCs are composed of multi-layer functional films, these films form interfaces. Whether electrons and holes can be effectively extracted at these interfaces mainly depends on whether the energy level at the interface is matched, which directly affects the collection efficiency of carriers.^58,59 Therefore, UPS was performed for the CuBO₂ film to study its energy level positions. Fig. 3e shows the UPS full spectrum of CuBO₂ film, and Fig. 3f corresponds to the secondary electron cut-off region and valence band region. The secondary electron cut-off edge (E_cut-off) is determined to be 18.29 eV. By subtracting the E_cut-off from the excitation energy (21.1 eV) of the He–I spectrum, the Fermi level (E_F) is determined to be 2.8 eV, which is close to the value obtained from density functional theory.²⁴ The difference between the valence band maximum (VBM) and E_F is 2.57 eV, so the final VBM is −5.39 eV relative to the vacuum level.

Fig. 4a shows the cross-sectional SEM image of a typical PSC based on the CuBO₂ THL. The thickness of the CuBO₂ HTL is about 20 nm, while the perovskite layer is about 400 nm, consisting of large and perpendicular crystal grains. Fig. 4b shows the schematic diagram of the energy level structure according to the UPS results. The VBM (−5.39 eV) of the CuBO₂ HTL matches well with the VBM of mixed perovskite Cs_0.05(FA_0.83MA_0.17)_0.95Pb(I_0.83Br_0.17)₃ (−5.44 eV),⁶⁰ which greatly minimizes the open-circuit voltage loss caused by energy level mismatch. The conduction band maximum (CBM) of CuBO₂ HTL is higher than that of perovskite, which forms an electron barrier and hinders the diffusion of electrons from the perovskite to the HTL. This type-II band alignment leads to spatial separation of electrons and holes across the interface. It helps to reduce electron–hole recombination, allow holes to be extracted by the HTL more efficiently, and improve hole mobility, thus improving device performance.⁶¹

Fig. 4 (a) Cross-sectional SEM image of PSC with CuBO₂ as the HTL. (b) Schematic of the energy level structure. (c) J–V curves of PSCs based on CuBO₂ and PEDOT:PSS HTLs. (d) Dark state J–V curves. (e) Steady-state power output curves. (f) Mott–Schottky curves.

Fig. 4c shows the J–V curves of the PSCs based on CuBO₂ and PEDOT:PSS HTLs. The champion PCE of CuBO₂ PSCs reaches 18.26%, with a V_OC of 1.05 V, J_SC of 22.51 mA cm⁻², and FF of 77.40%. The PCE of PSCs based on organic PEDOT:PSS HTL is 14.84%, with a V_OC of 0.99 V, J_SC of 21.68 mA cm⁻², and FF of 69.33%. The statistical distribution of performance parameters including V_OC, J_SC, FF and PCE is shown in Fig. S8 (ESI†). It can be seen that the average value of each performance parameter based on the CuBO₂ device has been improved and shows good repeatability. Fig. S9 (ESI†) shows the J–V curves of PSCs with an area of 1.08 cm² based on CuBO₂ and PEDOT:PSS HTLs. The champion PCE of CuBO₂ PSCs reaches 14.72% and that of the PEDOT:PSS-based PSCs is 10.21% (Fig. S9, ESI†). The statistical distribution of performance parameters is shown in Fig. S10 (ESI†). CuBO₂-based PSCs show better performance parameters and good repeatability compared to PEDOT:PSS-based PSCs. The higher FF of CuBO₂ PSCs could be attributed to the good spreadability of the perovskite precursor solution, which makes perovskite grains larger, reduces bulk defects and inhibits non-radiative recombination, which has consistently posed a challenge for devices utilizing PEDOT:PSS as the HTL.^62,63 Moreover, the higher V_OC of CuBO₂ PSCs is attributed to the matched energy level between CuBO₂ and perovskite films. To verify the light absorption abilities of perovskite films based on different HTLs, we analyzed their UV-vis spectra. Compared with that of the PEDOT:PSS device, the absorbance of the CuBO₂-based device is slightly increased, as shown in Fig. S11 (ESI†). Additionally, in order to verify the optical conversion abilities of different devices, we measured the incident photon-to-current efficiency (IPCE) spectra. As shown in Fig. S12 (ESI†), the integrated J_SC values of CuBO₂-based and PEDOT:PSS-based devices are 21.70 and 20.84 mA cm⁻², respectively, which are basically consistent with the J_SC values measured by the J–V curves.

The dark J–V curves of the PSCs were also measured to explore the current leakage.⁶⁴ As shown in Fig. 4d, both PSCs show quite low dark current densities, which indicates that the current leakage of both devices is very low. Furthermore, to evaluate the defect density of perovskite films on different HTLs, we fabricated hole-only devices with the structure FTO/CuBO₂ (or PEDOT:PSS)/perovskite/spiro-OMeTAD/Ag and performed space-charge-limited current (SCLC) measurements. Fig. S13 (ESI†) shows the current variation of the devices under bias voltage in the dark state. The defect state density N_t can be calculated by the formula V_TFL = eN_td²/2ε₀ε.⁶⁵ From the SCLC curves, it can be concluded that the V_TFL of the device is reduced from 0.75 to 0.59 V when CuBO₂ is used as the HTL, and the calculated N_t of the device based on CuBO₂ is 1.90 × 10¹⁶ cm⁻³, which is obviously lower than that of the device based on PEDOT:PSS (3.08 × 10¹⁶ cm⁻³). This is because the good interfacial contact between CuBO₂ and perovskite reduces the defect density. Additionally, the high hole mobility of CuBO₂ shortens carrier dwell time, thereby mitigating recombination losses caused by carrier localization.⁶⁶

Fig. 4e is the steady-state output of J_SC for the two devices obtained by maintaining the bias voltage at the maximum power point (MPP) under illumination conditions.^67,68 When the CuBO₂ based PSC is operated under illumination at MPP (V_max = 0.912 V), it maintains a J_SC of 20.13 mA cm⁻². In contrast, the PEDOT:PSS-based PSC exhibits a J_SC of 20.05 mA cm⁻² at a significantly lower voltage of 0.738 V. To explore the photo-generated charge recombination process in the device,⁶⁹ the electrochemical impedance spectra (EIS) were measured by applying −0.9 V bias in the frequency range of 0.1 Hz to 100 kHz under dark condition, and the obtained Nyquist plots are shown in Fig. S14 (ESI†). According to the equivalent circuit fittings, the CuBO₂ and PEDOT:PSS PSCs exhibit charge recombination resistances (R_rec) of 6174 Ω and 4135 Ω, respectively, indicating that the charge recombination in CuBO₂-based devices is lower, which can effectively promote charge transfer and reduce non-radiative recombination.⁷⁰ To study the mechanism of accelerating hole transport, the built-in potential (V_bi) of the device was characterized through Mott–Schottky tests. Fig. 4f shows the Mott–Schottky curves of devices based on different HTLs (tested at 5 kHz). The results show that the built-in voltage (V_bi) of CuBO₂-based PSCs is 1.01 V, which is higher than that of the PEDOT:PSS device (0.94 V). The increase of built-in voltage is related to the increase of the driving ability of charge transfer and extraction.^71,72 The V_OC variation trends are consistent with the J–V measurements under light. This indicates that the higher-quality perovskite thin film provides a stronger charge driving force, leading to enhanced carrier transport efficiency and reduced carrier recombination.⁴⁶V_bi not only promotes the separation and collection of photogenerated carriers, but also determines the level of V_OC. Therefore, the improvement of V_OC is also attributed to the enhancement of V_bi, resulting in less energy loss.

The dependence of V_OC and J_SC on light intensity (P_light) was examined to detect the charge recombination mechanism in these devices and is shown in Fig. S15(a) and (b) (ESI†). Firstly, the J_SC as a function of light intensity was measured and a linear correlation can be found, following the proportional relationship between ln J_SC and ln P_light: J_SC ∝ P^α_light.⁷³ The slope of the CuBO₂-based PSCs is 0.9842 and that of the PEDOT:PSS-based PSCs is 0.9723. Both are close to 1, indicating that the carrier extraction efficiency is high, and there is only a slight monomolecular (Shockley–Read–Hall⁷⁴) process. This result is consistent with the previously reported result.⁷⁵ Then, we studied the relationship between V_OC and light intensity, following the voltage-light intensity dependence relationship V_OC = nk_BT ln(P_light)/q. As shown in Fig. S15(b) (ESI†), the slopes of V_OC are close to kT/q, indicating that the Shockley–Read–Hall recombination is dominated by the bulk effect.⁷⁶ The slope decreases from 1.20kT/q for the PEDOT:PSS-based device to 1.08kT/q for the CuBO₂-based device, which means that the interface is improved due to the CuBO₂ gradient energy level arrangement and that trap-assisted recombination is significantly reduced.

To gain insight into the carrier dynamics at the CuBO₂/perovskite interface, photoluminescence (PL) and time-resolved photoluminescence (TRPL) were performed on the perovskite layers deposited on the FTO and CuBO₂ films. As shown in Fig. 5a, the PL intensity is significantly reduced when there is a CuBO₂ HTL on the FTO substrate. The significant PL quenching indicates that CuBO₂ has good hole extraction ability. Fig. 5b shows the TRPL spectra; the curves are fitted by selecting the double exponential equation , and the average carrier lifetimes are calculated according to the equation .¹⁹ The average carrier lifetimes (τ_avg) are calculated from the fast decay time τ₁ and the slow decay time τ₂, and the fitting parameters are summarized in Table 1. The τ_avg is reduced from 363.9 to 75.9 ns with the depositing of CuBO₂ HTL, which realizes the effective extraction of holes.

Fig. 5 (a) PL spectra and (b) TRPL spectra of perovskite films deposited on CuBO₂ and FTO.

Table 1 Fitting parameters and the average carrier lifetime

Sample	τ 1 (ns)	A 1	τ 2 (ns)	A 2	τ avg. (ns)
Perovskite	363.78	48.85	364.10	52.79	363.9
CuBO2/perovskite	4.41	3.72	76.15	78.40	75.9

---

Finally, stability tests were performed on the devices based on the CuBO₂ and PEDOT:PSS HTLs. The unencapsulated devices were stored in a desiccator (∼40% relative humidity), and the PCEs of the devices were measured periodically. Fig. 6 shows the downward trend of the normalized PCEs. There is no obvious degradation of the device based on inorganic CuBO₂ HTL. After nearly 200 hours, the PCE still maintains more than 85% of the initial PCE, while the device based on organic PEDOT:PSS has only about 70% of the original PCE. Additionally, we tested the light stability of the PSCs based on CuBO₂ and PEDOT:PSS, as shown in Fig. S16 (ESI†). All devices were stored unpackaged in the glove box, and their reverse scanning efficiencies were measured regularly. Obviously, the degradations of the two devices are significantly different. The photoelectric conversion efficiency of the device based on PEDOT:PSS showed a rapid degradation after 120 h, and only 67% of the original efficiency remained after 200 h. However, the CuBO₂-based PSCs retained 82% of its original efficiency after 200 h degradation. To further evaluate the stability of the device, the unpackaged devices were subjected to a damp-heat conditions (85 °C, 85% relative humidity). As shown in Fig. S17 (ESI†), the CuBO₂-based PSC retained 77% of its initial PCE under moist heat test condition (ISOS-D-3), while the PEDOT:PSS-based PSC exhibited a lower retention rate of 63% over the same aging period, thus proving the advantage of inorganic HTL in promoting long-term device stability.

Fig. 6 Air stability of unencapsulated PSCs based on CuBO₂ and PEDOT:PSS HTLs. The insets are photographs of the aged perovskites.

4. Conclusion

In summary, we developed a facile one-step sol–gel method to prepare CuBO₂ HTLs for PSCs. By optimizing the precursor concentration, pH value and annealing temperature, dense and homogeneous films were yielded. The as-obtained CuBO₂ film exhibited optimal characteristics, with a maximum optical transmittance of 84%, a band gap of 3.57 eV and a carrier concentration of 7.35 × 10 cm⁻³. Moreover, the valence band edge of CuBO₂ (−5.39 eV) showed excellent alignment with that of perovskite (−5.44 eV). The inverted p–i–n PSCs based on CuBO₂ HTLs achieved a remarkable PCE of 18.26%, surpassing that of PEDOT:PSS-based devices. Meanwhile, the inorganic HTL based device demonstrated superior open-circuit voltage and environmental stability. The devices maintained 85% of their initial efficiency after 200 hours in ambient air (40% humidity). This work establishes CuBO₂ as a promising inorganic HTM, which is foreseen to pave a new way for inorganic HTMs in the fields of PSCs and other photovoltaic technologies.

Author contributions

Shichao Wang: conceptualization, data curation, methodology, investigation, formal analysis, and writing – original draft. Jiangshan Shi: data curation and writing – original draft. Jianhui Li: investigation and data curation. Yuanqiang Wang: validation and writing – review & editing. Jingxia Yang: supervision and visualization. Yichuan Rui: conceptualization, funding acquisition, resources, supervision, project administration, and writing – review & editing.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its ESI.†

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (52202178), Natural Science Foundation of Shanghai (22ZR1426300) and the Shanghai Sailing Program (19YF1417600).

References

1. M. Cheng, Y. Duan, D. Zhang, Z. Xie, H. Li, Q. Cao, Z. Qiu, Y. Chen and Q. Peng, Adv. Mater., 2025, 24, 113.
2. S. Liu, J. Li, W. Xiao, R. Chen, Z. Sun, Y. Zhang, X. Lei, S. Hu, M. Kober-Czerny, J. Wang, F. Ren, Q. Zhou, H. Raza, Y. Gao, Y. Ji, S. Li, H. Li, L. Qiu, W. Huang, Y. Zhao, B. Xu, Z. Liu, H. J. Snaith, N.-G. Park and W. Chen, Nature, 2024, 632, 536–542.
3. P. D. Zhu, D. Wang, Y. Zhang, Z. Liang, J. B. Li, J. Zeng, J. Y. Zhang, Y. T. Xu, S. Y. Wu, Z. X. Liu, X. Y. Zhou, B. H. Hu, F. He, L. Zhang, X. Pan, X. Z. Wang, N. G. Park and B. M. Xu, Science, 2024, 383, 524–531.
4. Z. Liu, R. Lin, M. Wei, M. Yin, P. Wu, M. Li, L. Li, Y. Wang, G. Chen, V. Carnevali, L. Agosta, V. Slama, N. Lempesis, Z. Wang, M. Wang, Y. Deng, H. Luo, H. Gao, U. Rothlisberger, S. M. Zakeeruddin, X. Luo, Y. Liu, M. Grätzel and H. Tan, Nat. Mater., 2025, 24, 252–259.
5. H.-S. Kim, C.-R. Lee, J.-H. Im, K.-B. Lee, T. Moehl, A. Marchioro, S.-J. Moon, R. Humphry-Baker, J.-H. Yum, J. E. Moser, M. Grätzel and N.-G. Park, Sci. Rep., 2012, 2, 591.
6. Y. Xu, X. Wang, Z. Jin, B. Li, W. An, Q. Zhang and Y. Rui, Solar RRL, 2023, 7, 2300302.
7. L. Ye, J. Wu, S. Catalán-Gómez, L. Yuan, R. Sun, R. Chen, Z. Liu, J. M. Ulloa, A. Hierro, P. Guo, Y. Zhou and H. Wang, Nat. Commun., 2024, 15, 7889.
8. M. Yahiro, S. Sugawara, S. Maeda, Y. Shimoi, P. Wang, S.-I. Kobayashi, K. Takekuma, G. Tumen-Ulzii, C. Qin, T. Matsushima, T. Isaji, Y. Kasai, T. Fujihara and C. Adachi, ACS Appl. Energy Mater., 2021, 4, 14590–14598.
9. T. Li, Y. Rui, X. Wang, J. Shi, Y. Wang, J. Yang and Q. Zhang, ACS Appl. Energy Mater., 2021, 4, 7002–7011.
10. Y. Rui, T. Li, B. Li, Y. Wang and P. Müller-Buschbaum, J. Mater. Chem. C, 2022, 10, 12392–12401.
11. Z. Yu, A. Hagfeldt and L. Sun, Coord. Chem. Rev., 2020, 406, 213143.
12. D. Di Girolamo, F. Di Giacomo, F. Matteocci, A. G. Marrani, D. Dini and A. Abate, Chem. Sci., 2020, 11, 7746–7759.
13. X. Yin, J. Zhai, P. Du, N. Li, L. Song, J. Xiong and F. Ko, ChemSusChem, 2020, 13, 1006–1012.
14. B. Zhang, J. Su, X. Guo, L. Zhou, Z. Lin, L. Feng, J. Zhang, J. Chang and Y. Hao, Adv. Sci., 2020, 7, 1903044.
15. C.-C. Tseng, L.-C. Chen, L.-B. Chang, G.-M. Wu, W.-S. Feng, M.-J. Jeng, D. W. Chen and K.-L. Lee, Sol. Energy, 2020, 204, 270–279.
16. Z. S. Zhuang, L. L. Qiu, L. K. Dong, Y. Chen, Z. D. Chu, X. Y. Ma, P. F. Du and J. Xiong, Polym. Compos., 2020, 41, 2145–2153.
17. C. Zuo and L. Ding, Small, 2015, 11, 5528–5532.
18. J. S. Hong, Z. J. Xu, D. Lungwitz, J. Scott, H. M. Johnson, Y. H. Kim, A. Kahn and B. P. Rand, ACS Energy Lett., 2023, 8, 4984–4992.
19. Y. Dou, D. Wang, G. Li, Y. Liao, W. Sun, J. Wu and Z. Lan, ACS Appl. Mater. Interfaces, 2019, 11, 32159–32168.
20. B. Li, Y. Rui, J. Xu, Y. Wang, J. Yang, Q. Zhang and P. Müller-Buschbaum, J. Colloid Interface Sci., 2020, 573, 78–86.
21. Y. Zhang, J. Ge, B. Mahmoudi, S. Förster, F. Syrowatka, A. W. Maijenburg and R. Scheer, ACS Appl. Energy Mater., 2020, 3, 3755–3769.
22. A. N. Banerjee and K. K. Chattopadhyay, in Reactive sputtered wide-bandgap p-type semiconducting spinel AB₂O₄ and delafossite ABO₂ thin films for “transparent electronics”, Reactive Sputter Deposition, ed. D. Depla and S. Mahieu, 2008, vol. 109, pp. 413–484.
23. K. Rajeshwar, M. K. Hossain, R. T. Macaluso, C. Janáky, A. Varga and P. J. Kulesza, J. Electrochem. Soc., 2018, 165, H3192.
24. S. Santra, N. S. Das, D. Sen and K. K. Chattopadhyay, J. Phys. D: Appl. Phys., 2014, 47, 505301.
25. W. A. Dunlap-Shohl, T. B. Daunis, X. Wang, J. Wang, B. Zhang, D. Barrera, Y. Yan, Julia W. P. Hsu and D. B. Mitzi, J. Mater. Chem. A, 2018, 6, 469–477.
26. B. Yang, D. Ouyang, Z. Huang, X. Ren, H. Zhang and W. C. H. Choy, Adv. Funct. Mater., 2019, 29, 1902600.
27. B. Zhang, S. Thampy, W. A. Dunlap-Shohl, W. Xu, Y. Zheng, F.-Y. Cao, Y.-J. Cheng, A. V. Malko, D. B. Mitzi and J. W. P. Hsu, Mg Doped CuCrO₂ as Efficient Hole Transport Layers for Organic and Perovskite Solar Cells, Nanomaterials, 2019, 1311.
28. H. Zhang, H. Wang, H. Zhu, C.-C. Chueh, W. Chen, S. Yang and A. K. Y. Jen, Adv. Energy Mater., 2018, 8, 1702762.
29. I. Y.-Y. Bu, Y.-C. Lu, Y.-S. Fu and C.-T. Hung, Optik, 2020, 210, 164505.
30. F. Igbari, M. Li, Y. Hu, Z.-K. Wang and L.-S. Liao, J. Mater. Chem. A, 2016, 4, 1326–1335.
31. A. Savva, I. T. Papadas, D. Tsikritzis, A. Ioakeimidis, F. Galatopoulos, K. Kapnisis, R. Fuhrer, B. Hartmeier, M. F. Oszajca, N. A. Luechinger, S. Kennou, G. S. Armatas and S. A. Choulis, ACS Appl. Energy Mater., 2019, 2, 2276–2287.
32. S. Akin, F. Sadegh, S. Turan and S. Sonmezoglu, ACS Appl. Mater. Interfaces, 2019, 11, 45142–45149.
33. B. Lee, A. J. Yun, J. Kim, B. Gil, B. Shin and B. Park, Adv. Mater. Interfaces, 2019, 6, 1901372.
34. R. Li, Y. Dou, Y. Liao, D. Wang, G. Li, J. Wu and Z. Lan, J. Colloid Interface Sci., 2022, 607, 1280–1286.
35. T. Zhao, Q.-L. Liu and Z.-Y. Zhao, J. Phys. Chem. C, 2019, 123, 14292–14302.
36. K. J. Zhu, G. Mul and A. Huijser, ChemSusChem, 2024, 17, 10.
37. T. Li, Y. Rui, X. Zhang, J. Shi, X. Wang, Y. Wang, J. Yang and Q. Zhang, J. Alloys Compd., 2020, 849, 156629.
38. T. Jiang, M. Bujoli-Doeuff, Y. Farré, E. Blart, Y. Pellegrin, E. Gautron, M. Boujtita, L. Cario, F. Odobel and S. Jobic, RSC Adv., 2016, 6, 1549–1553.
39. S. Santra, D. Das, N. S. Das and K. K. Nanda, ACS Appl. Energy Mater., 2019, 2, 260–268.
40. S. Santra, N. S. Das and K. K. Chattopadhyay, Mater. Res. Bull., 2013, 48, 2669–2677.
41. B. Zhu, B. Li, G. Ding, Z. Jin, Y. Xu, J. Yang, Y. Wang, Q. Zhang and Y. Rui, ACS Appl. Mater. Interfaces, 2024, 16, 28560–28569.
42. Y. Xu, Y. Rui, X. Wang, B. Li, Z. Jin, Y. Wang, W. An and Q. Zhang, Sol. Energy Mater. Sol. Cells, 2023, 250, 112088.
43. P. Das, N. S. Das, K. Sardar, B. Das, A. Adalder and K. K. Chattopadhyay, Mater. Chem. Phys., 2023, 301, 10.
44. S. Santra, N. S. Das, N. Besra, D. Banerjee and K. K. Chattopadhyay, Langmuir, 2017, 33, 9961–9971.
45. X. Fan, Y. Rui, X. Han, J. Yang, Y. Wang and Q. Zhang, J. Power Sources, 2020, 448, 227405.
46. Z. Jin, B. Li, Y. Xu, B. Zhu, G. Ding, Y. Wang, J. Yang, Q. Zhang and Y. Rui, J. Mater. Chem. C, 2023, 11, 6730–6740.
47. Z. H. Zhao, W. L. Liu, T. F. Kong, Y. J. Liu, W. T. Chen, P. Gao and D. Q. Bi, Adv. Funct. Mater., 2024, 2419393.
48. Q. Niu, Y. Deng, D. Cui, H. Lv, X. Duan, Z. Li, Z. Liu, W. Zeng, R. Xia, W. Tan and Y. Min, J. Mater. Sci., 2019, 54, 14134–14142.
49. D. O. Scanlon, A. Walsh and G. W. Watson, Chem. Mater., 2009, 21, 4568–4576.
50. M. Snure and A. Tiwari, Appl. Phys. Lett., 2007, 91, 092123.
51. D. O. Scanlon, A. Walsh, B. J. Morgan, G. W. Watson, D. J. Payne and R. G. Egdell, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79, 035101.
52. A. Jedidi, S. Rasul, D. Masih, L. Cavallo and K. Takanabe, J. Mater. Chem. A, 2015, 3, 19085–19092.
53. C. Ruttanapun, J. Appl. Phys., 2013, 114, 113108.
54. M. Snure and A. Tiwari, Appl. Phys. Lett., 2007, 91, 092123.
55. K. Ravindra, M. Hari Prasad Reddy and S. Uthanna, Mater. Today: Proc., 2017, 4, 12505–12511.
56. S. Dugu, R. W. Smaha, S. Quadir, A. Treglia, J. Martin, S. O'Donnell, S. Mahatara, G. Teeter, S. Lany, J. R. Neilson and S. R. Bauers, Phys. Rev. Mater., 2024, 8, 104411.
57. D. Di Girolamo, N. Phung, M. Jošt, A. Al-Ashouri, G. Chistiakova, J. Li, J. A. Márquez, T. Unold, L. Korte, S. Albrecht, A. Di Carlo, D. Dini and A. Abate, Adv. Mater. Interfaces, 2019, 6, 1900789.
58. W. B. Peng, Y. Zhang, X. Y. Zhou, J. W. Wu, D. Wang, G. P. Qu, J. Zeng, Y. T. Xu, B. Jiang, P. D. Zhu, Y. F. Du, Z. T. Li, X. Lei, Z. X. Liu, L. Yan, X. Z. Wang and B. M. Xu, Energy Environ. Sci., 2025, 18, 874–883.
59. S. You, F. T. Eickemeyer, J. Gao, J.-H. Yum, X. Zheng, D. Ren, M. Xia, R. Guo, Y. Rong, S. M. Zakeeruddin, K. Sivula, J. Tang, Z. Shen, X. Li and M. Grätzel, Nat. Energy, 2023, 8, 515–525.
60. J. Shi, B. Li, Q. Zhang and Y. Rui, J. Alloys Compd., 2022, 890, 161879.
61. W. Peng, Y. Zhang, X. Zhou, J. Wu, D. Wang, G. Qu, J. Zeng, Y. Xu, B. Jiang, P. Zhu, Y. Du, Z. Li, X. Lei, Z. Liu, L. Yan, X. Wang and B. Xu, Energy Environ. Sci., 2025, 18, 874–883.
62. Y. Jiang, X. Dong, L. Sun, T. Liu, F. Qin, C. Xie, P. Jiang, L. Hu, X. Lu, X. Zhou, W. Meng, N. Li, C. J. Brabec and Y. Zhou, Nat. Energy, 2022, 7, 352–359.
63. Z. Ren, S. Luo, X. Shi, Y. Dou, T. Liu, L. Wang, K. K. Tsang, F. Wang, Y. Zhao, Y. Liu, X. Hu, X. Peng, W. Liu, H. Yan and S. Chen, Sci. China: Chem., 2024, 67, 1941–1945.
64. B. Yang, S. Peng and W. C. H. Choy, EcoMat, 2021, 3, e12127.
65. Y. Lv, B. Cai, Q. Ma, Z. Wang, J. Liu and W.-H. Zhang, RSC Adv., 2018, 8, 20982–20989.
66. J. Tao, C. Zhao, Z. Wang, Y. Chen, L. Zang, G. Yang, Y. Bai and J. Chu, Energy Environ. Sci., 2025, 18, 509–544.
67. L. Zhang, C. Liu, X. Wang, Y. Tian, A. K. Y. Jen and B. Xu, Adv. Funct. Mater., 2019, 29, 1904856.
68. Y. Xu, Y. Rui, X. Wang, B. Li, Z. Jin, Y. Wang and Q. Zhang, Appl. Surf. Sci., 2023, 607, 154986.
69. C. Zhao, Q. Zhang, Y. Lyu, X. Lian, K. Wang, F. Shen, M. Luo, H. Liu, H. Han, F. Xie, X. Mo, Y. Yang, J. Xu, A. K. Y. Jen, H. Zhang and J. Yao, ACS Energy Lett., 2024, 9, 1405–1414.
70. R. Liu, Y. Yu, C. Liu, H. Yang, X.-L. Shi, H. Yu and Z.-G. Chen, Sci. China: Chem., 2022, 65, 2468–2475.
71. T. Bu, J. Li, F. Zheng, W. Chen, X. Wen, Z. Ku, Y. Peng, J. Zhong, Y.-B. Cheng and F. Huang, Nat. Commun., 2018, 9, 4609.
72. X. Wang, Y. Zhao, B. Li, X. Han, Z. Jin, Y. Wang, Q. Zhang and Y. Rui, ACS Appl. Mater. Interfaces, 2022, 14, 22879–22888.
73. C. Chen, Y. Zhai, F. Li, F. Tan, G. Yue, W. Zhang and M. Wang, J. Power Sources, 2017, 341, 396–403.
74. W. Shockley and W. T. Read, Phys. Rev., 1952, 87, 835–842.
75. C. L. Davies, M. R. Filip, J. B. Patel, T. W. Crothers, C. Verdi, A. D. Wright, R. L. Milot, F. Giustino, M. B. Johnston and L. M. Herz, Nat. Commun., 2018, 9, 293.
76. R. Das Adhikari, M. J. Patel, H. Baishya, D. Yadav, M. Kalita, M. Alam and P. K. Iyer, Chem. Soc. Rev., 2025, 54, 3962–4034.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5tc01169h

---