Skip to main content
NCBI home page
Chemical Science logoLink to Chemical Science
. 2026 Jun 23;17(32):15338–15347. doi: 10.1039/d6sc03304k

Polyvinyl-based hole-transporting materials processed with non-destructive and green solvents for tin–lead perovskite solar cells and all-perovskite tandems

Jiayi Liu a,†, Yun Zhao a,†, Liqing Zhan a,†, Shuo Zhang a,d,✉, Haorui Tang a, Rujun Ma a, Xinwen Long a, Lihui Zhou a, Qifeng Zhu c, Shuang Yang c, Weihong Zhu a,b,d, Yongzhen Wu a,b,d,✉
PMCID: PMC13326693  PMID: 42394798

Abstract

The widely used poly(3,4-ethylenedioxythiophene):polystyrenesulfonate (PEDOT:PSS) in tin–lead (Sn–Pb) mixed perovskite solar cells limits scalable fabrication and operational stability of all-perovskite tandem solar cells. The aqueous-based and acidic PEDOT:PSS can penetrate the interconnection layer and degrade the underlying wide bandgap subcell. Here, we report a series of polyvinyl-based hole-transporting materials (PE-MPs) that are processable in non-aqueous green solvents. The deposition of PE-MPs is non-destructive to underlying perovskite stacks, and the resulting thin layer further protects the bottom subcell during the deposition of the Sn–Pb perovskite, providing a wide processing window for tandem fabrication. Through optimization of the monomer ratio, the best-performing PE-MP2 enables both efficient hole extraction and robust interfacial adhesion, greatly improving performance and stability of Sn–Pb devices and all-perovskite tandems. The Sn–Pb perovskite solar cell achieves a power conversion efficiency of 22.72%, and the all-perovskite tandem cell reaches 29.08% (certified 28.69%). More importantly, the all-perovskite tandems retain 90% of their initial efficiency after 570 hours of maximum power point tracking, ranking among the most efficient and stable all-perovskite tandems without PEDOT:PSS.

The widely used poly(3,4-ethylenedioxythiophene):polystyrenesulfonate (PEDOT:PSS) in tin–lead (Sn–Pb) mixed perovskite solar cells limits scalable fabrication and operational stability of all-perovskite tandem solar cells.graphic file with name d6sc03304k-ga.jpg

Introduction

Perovskite solar cells (PSCs) have emerged as promising candidates for next-generation photovoltaics owing to their exceptional optoelectronic properties and cost-effectiveness.1 To surpass the theoretical Shockley–Queisser limit of single-junction PSCs, all-perovskite tandem solar cells (TSCs) have been developed by integrating a wide-bandgap (WBG, 1.70–1.90 eV) top subcell with a narrow-bandgap (NBG, 1.20–1.30 eV) bottom subcell.2 Although the power conversation efficiency (PCE) of TSCs has reached 30.1%, the inferior stability of NBG tin–lead (Sn–Pb) NBG perovskite subcells based on poly(3,4-ethylenedioxythiophene):polystyrenesulfonate (PEDOT:PSS) is still limiting the operational stability of all-perovskite tandem solar cells.3

PEDOT:PSS, an aqueous-based hole-transporting material (HTM), is widely employed in tin-based devices due to its excellent conductivity, high transparency and suitable energy level alignment.4 However, the hygroscopicity and acidity of aqueous-based PEDOT:PSS accelerate degradation of overlying Sn–Pb perovskite5,6 and potentially penetrate the underlying interconnection layer (ICL), triggering decomposition of the WBG subcell and irreversibly compromising device efficiency and stability.7 To address these issues, interface engineering and structural modification have been explored.8 Alternatively, emerging HTMs including inorganic oxides, self-assembled monolayers (SAMs) and polymers have been developed.9–13 NiOx usually suffers from detrimental interfacial reactions and requires additional surface modification.9 While novel SAMs have shown significant progress in lead-based PSCs, their mismatched highest occupied molecular orbital (HOMO) levels remain a critical problem in Sn–Pb systems, impeding their application in all-perovskite TSCs.11 Polymeric HTMs like poly[bis(4-phenyl)(2,4,6-trimethylphenyl)amine] (PTAA) exhibit good environmental stability, but their hydrophobic nature complicates the deposition of uniform perovskite layers and exacerbates interfacial non-radiative recombination.13,14 Therefore, it remains a significant challenge to develop a new multifunctional HTM to replace PEDOT:PSS in Sn–Pb PSCs for efficient, stable and scalable all-perovskite TSCs.

Vinyl polymerization, an important method based on chain-growth polymerization of monomers containing carbon–carbon double bonds (C Created by potrace 1.16, written by Peter Selinger 2001-2019 C), serves as a powerful and controllable strategy for constructing polymeric materials with well-defined structures.15 Meanwhile, the polyvinyl backbone with suitable side chains can impart excellent solubility in green solvents, thereby facilitating the formation of high-quality thin films.16 Furthermore, through the optimization of monomer types and ratios, copolymerization can integrate the functions of different side chains to fine-tune key physicochemical properties, such as energy levels, hole mobility, and solvent processability, thus contributing to favourable energy level alignment and interfacial adhesion with the photoactive layer.17

In this work, we designed and synthesized a series of polyvinyl-based and green-solvent-processable HTMs (PE-MPs) by copolymerizing triphenylamine units and phosphonic acid units. Through optimization of the monomer ratio, the optimal PE-MP2 integrates multiple functionalities. First, the soft, non-conjugated polyvinyl backbone endows PE-MPs with excellent non-aqueous solvent processability, enabling device fabrication with green and non-destructive solvents such as ethyl acetate, which broadens the processing window of the hole-transporting layer of NBG subcells (NBG-HTL) in TSCs. Secondly, the deliberately selected units collectively ensure efficient hole extraction and transport and robust interfacial anchoring to both the indium tin oxide (ITO) substrate and the perovskite layer. Moreover, the excellent solubility facilitates the formation of a homogeneous and compact film of PE-MP2, which provides robust protection for the underlying layers during the deposition of the upper perovskite layer. The Sn–Pb perovskite solar cell and all-perovskite tandem device employing the optimized PE-MP2 achieved champion PCEs of 22.72% and 29.08% (certified 28.69%), respectively, which are comparable to those of PEDOT:PSS-based counterparts. Furthermore, the resulting tandem device showed markedly enhanced operational stability, retaining 90% of its initial PCE after 570 hours of continuous maximum power point tracking, which is one of the most efficient and stable all-perovskite tandems without PEDOT:PSS.

Results and discussion

Molecule design and fundamental properties of PE-MPs

We designed and synthesized a series of polymeric HTMs, named PE-MPs (Fig. 1a and S1), which are based on polyvinyl backbones chosen for their superior solubility and film-forming ability. The functional side chains in PE-MPs were selected to combine a suitable energy level with enhanced anchoring ability at the ITO/perovskite interface. First, the triphenylamine units exhibit a shallower HOMO level compared to widely used carbazole,18 and the introduction of methoxy groups further stabilizes the HOMO level due to their electron-donating effect,19 promoting favourable energy level alignment with the valence band maximum (VBM) of the Sn–Pb mixed perovskite. Additionally, the flexible polyvinyl main chain enables the phosphonic acid groups to play a dual role at the buried interface: anchoring to ITO and interacting with the perovskite, thereby strengthening the chemical interactions and enhancing the interfacial adhesion. Three specific polymers, namely PE-MP1, PE-MP2, and PE-MP3, were synthesized by systematically varying the molar ratio of 4-methoxy-N-(4-methoxyphenyl)-N-(4-vinylphenyl)aniline to vinylphosphonic acid at 6 : 1, 3 : 1, and 1 : 1, respectively. The detailed synthesis procedure is shown in Fig. S1. The monomer was characterized by proton nuclear magnetic resonance (1H NMR, Fig. S2), 13C NMR (Fig. S3), and mass spectrometry (Fig. S4). The molecular weight of the polymers was determined by gel permeation chromatography (GPC), including the number-average (Mn) and weight-average (Mw), which are provided in Table S1. In addition, thermogravimetric analysis (TGA) confirmed the excellent thermal stability of the PE-MPs, with all polymers exhibiting high decomposition temperatures (Td, corresponding to 5% mass loss) above 350 °C (Fig. S5).

Fig. 1. (a) Molecular structure of PE-MPs. Products PE-MP1, PE-MP2, and PE-MP3 were obtained at the 4-methoxy-N-(4-methoxyphenyl)-N-(4-vinylphenyl)aniline (x) and vinylphosphonic acid (y) molar feed ratios of 6 : 1, 3 : 1, and 1 : 1, respectively. (b) UV-vis spectra of PE-MPs. (c) CV curves of the ferrocene (Fc/Fc+) reference and PE-MPs measured in CH2Cl2. (d) Energy level diagrams of PE-MPs determined from CV and UV-vis results; the data of PEDOT:PSS and Sn–Pb perovskite are taken from the literature.20 (e) Photographs of PE-MP1, PE-MP2, and PE-MP3 solutions in ethyl acetate (EA) and the corresponding films deposited on ITO. (f) UPS spectra of ITO/PE-MPs and ITO/PEDOT:PSS films. (g) In 3d and (h) Sn 3d XPS spectra of pristine ITO and ITO/PE-MPs samples. (i) Schematic diagram of the deposition of polymers using non-aqueous solvent.

Fig. 1

Open in a new tab

To explore the relationship between the structure and photoelectric properties, we conducted ultraviolet-visible spectroscopy (UV-vis) on PE-MPs. As shown in Fig. 1b, PE-MPs showed a strong absorption peak between 250 and 400 nm, which corresponds to the characteristic peak of PE-MTPA, a homopolymer of the 4-methoxy-N-(4-methoxyphenyl)-N-(4-vinylphenyl)aniline unit (Fig. S1 and S6). The intensity of this peak decreased progressively from PE-MP1 to PE-MP3, demonstrating that the content of triphenylamine in polymers could be effectively tuned by varying the molar feed ratios. Additionally, cyclic voltammetry (CV) measurements were conducted to investigate the redox behaviour of PE-MPs (Fig. 1c), and the resulting data were used to calculate the HOMO energy levels. As shown in Fig. 1d, the HOMO energy levels of PE-MP1, PE-MP2 and PE-MP3 were all around −5.30 eV. Furthermore, the hole mobility of PE-MPs was characterized using space-charge-limited current (SCLC) measurements. The hole mobility slightly increased when the proportion of triphenylamine units in copolymers was increased (1.03 × 10−6, 9.63 × 10−7 and 7.53 × 10−7 cm2 V−1 s−1 for PE-MP1, PE-MP2 and PE-MP3, respectively, Fig. S7). This is rationalized by the fact that charge transport is mainly contributed by triphenylamine units.

The solubility of PE-MPs was then evaluated in four solvents: chlorobenzene, anisole, 2-methyltetrahydrofuran (2Me-THF), and ethyl acetate. In all cases, the solubility was determined to be higher than 20 mg mL−1 (Fig. S8). Ethyl acetate was selected as the optimal solvent as it is a green and non-destructive solvent for perovskites. Uniform and transparent hole-transporting layers (HTLs) can be successfully fabricated on a glass/ITO substrate from solutions of the PE-MPs in ethyl acetate (Fig. 1e), confirming their excellent processability. The resulting films exhibited slightly higher transmittance than PEDOT:PSS in the 380–500 nm range (Fig. S9). Moreover, all PE-MPs solutions displayed excellent photostability as negligible variation was observed in absorption spectra in the 250–350 nm range after aging for 3 days (Fig. S10) under a xenon lamp solar simulator, which is conducive to enhancing operational stability of Sn–Pb PSCs.

To confirm energy level alignment between the PE-MPs and the Sn–Pb perovskite, the energy levels were measured using ultraviolet photoelectron spectroscopy (UPS) on films of PE-MPs and PEDOT:PSS fabricated on ITO substrates (Fig. 1f). The UPS results showed that the HOMO of both PE-MP2 and PE-MP3 thin films is located at −5.10 eV, closer to the VBM of Sn–Pb perovskite than that of PEDOT:PSS (Fig. S11). Both the CV and UPS results suggest that the HOMOs of PE-MPs are more favourable than that of PEDOT:PSS for aligning with Sn–Pb perovskite.20 The interfacial interaction between PE-MPs and ITO was further investigated by X-ray photoelectron spectroscopy (XPS). Notable shifts in the In 3d and Sn 3d peaks of the ITO substrate were observed after the deposition of PE-MPs (Fig. 1g and h), which are ascribed to chemical interactions between In/Sn atoms and the phosphoryl oxygen of PE-MPs. Owing to this robust interfacial anchoring, we propose that the PE-MPs can achieve uniform and dense thin-film deposition on relatively rough ITO substrates using ethyl acetate as the processing solvent (Fig. 1i).17

Characterization of Sn–Pb perovskite films on different HTLs

We measured the contact angles of water and Sn–Pb perovskite precursor solution on different HTLs (Fig. S12a and b). Compared to hydrophilic PEDOT:PSS, the copolymer-based PE-MPs exhibited larger contact angles for both liquids. This slightly enhanced hydrophobic character did not hinder the formation of fully covered perovskite films (Fig. S12c) and might facilitate growth of larger grains.21 As the proportion of phosphonic acid groups in PE-MPs increased, the contact angles decreased moderately, indicating that some of the hydrophilic phosphonic acid groups in polymer were exposed and oriented upward, allowing interaction with the Sn–Pb perovskite. The corresponding NBG Sn–Pb perovskite films deposited on PEDOT:PSS, PE-MP1, PE-MP2, and PE-MP3 are hereafter denoted as PEDOT:PSS/NBG, PE-MP1/NBG, PE-MP2/NBG, and PE-MP3/NBG, respectively.

The top-view scanning electron microscopy (SEM) images showed that all Sn–Pb perovskites exhibited compact morphology with grain sizes approaching one micrometre (Fig. 2a). The perovskite on PE-MPs displayed a flatter surface and larger grains than that on PEDOT:PSS. We noted that morphology of the buried surface demonstrated a significant difference: the PEDOT:PSS/NBG and PE-MP1/NBG films showed uneven, loosely packed grains with visible voids at grain boundaries, whereas PE-MP2/NBG and PE-MP3/NBG films exhibited more uniform bottom surfaces (Fig. 2b). The improved morphology reveals that the introduction of phosphonic acid groups is conducive to regulating crystallization of Sn–Pb perovskite films. We further analysed the crystal structure of Sn–Pb perovskite films on different HTLs using X-ray diffraction (XRD) measurements. While all Sn–Pb perovskite films showed a relatively pure perovskite phase, the film based on PE-MP2 exhibited the highest diffraction intensities and minimal PbI2 (2θ = 12.6°), indicating that PE-MP2 improves the crystallinity of the perovskite (Fig. 2c). To investigate the effect of different HTLs on the stability of Sn–Pb perovskite, we further tracked the thermal stability of Sn–Pb perovskite films deposited on different substrates during heating at 65 °C in a nitrogen atmosphere. After aging for 3 days, the intensity of XRD peaks for the Sn–Pb perovskite film deposited on ITO/PEDOT:PSS decreased. In contrast, all films deposited on ITO/PE-MPs displayed enhanced thermal stability, and PE-MP2/NBG exhibited the best stability under the same conditions (Fig. S13). We speculated that a moderate proportion of phosphonic acid units contributes to improving the crystallization and enhancing stability of Sn–Pb perovskite, but excessive phosphonic acid groups might interact with perovskite, compromising thermal stability of the Sn–Pb perovskite film.22 Collectively, the SEM and XRD results suggest that the Sn–Pb perovskite films deposited on PE-MP2 possess a higher crystallization quality than those on the other HTLs.

Fig. 2. (a and b) Top-view SEM images of the surface (a) and bottom surface (b) of Sn–Pb perovskite films deposited on different HTLs, scale bars: 1 µm. Comparison of (c) XRD patterns, (d) PL spectra, and (e) PLQY of Sn–Pb perovskite films fabricated on different HTLs.

Fig. 2

Open in a new tab

We further evaluated the photoelectric properties of the Sn–Pb perovskite based on different HTLs by absorbance and photoluminescence (PL) characterization. The absorption spectra of these Sn–Pb perovskite films deposited on different substrates are quite similar (Fig. S14). However, their PL intensity and PL quantum yields (PLQYs) are different. PE-MP2/NBG exhibited the highest PL intensity and PLQY (1.22%), even higher than those of PEDOT:PSS/NBG (Fig. 2d and e). The enhanced PL and PLQY of PE-MP2/NBG with respect to PE-MP1/NBG suggest that the phosphonic acid groups in PE-MPs contribute to defect passivation of the bottom surface of the Sn–Pb perovskite. However, an excessive proportion of phosphonic acid groups in PE-MPs may be detrimental to the perovskite crystallization, as evidenced by the decreased crystallinity in the XRD results (Fig. 2c). The highest PL and PLQY values achieved by PE-MP2 indicated that PE-MP2 was the optimal copolymer for application in Sn–Pb perovskite solar cells, probably owing to the appropriate proportion of the two functional units.

To investigate the effect of these copolymers on interfacial mechanical strength, we conducted double cantilever beam (DCB) delamination measurement. As depicted in Fig. S15 and Table S2, all perovskite films fabricated on copolymers withstood higher traction stress under mechanical pulling and exhibited stronger interfacial adhesion than PEDOT:PSS/NBG. Notably, the interfacial adhesion strength correlated positively with the proportion of anchoring phosphonic acid units in the copolymers. Specifically, increasing the phosphonic acid ratio (from 14% to 50%) in the copolymers enhanced the traction stress at the ITO/HTL/perovskite interface by a factor of approximately 1.5. The enhanced interfacial mechanical integrity demonstrated that phosphonic acid groups in the copolymer can simultaneously interact with both the underlying ITO and the overlying perovskite layer, improving the interfacial adhesion and hence enhancing the operational stability under external stress.14

Performance of single-junction Sn–Pb perovskite solar cells

We first employed an intensity-dependent PLQY measurement to quantify the quasi-Fermi level splitting (QFLS) and to derive pseudo J–V curves (detailed in the SI) of Sn–Pb perovskite films, allowing an evaluation of the potential photovoltaic performance of HTL/perovskite half stacks.23 At intermediate to high excitation intensities, PE-MP2/NBG always showed a higher QFLS than Sn–Pb perovskites on other HTLs (Fig. 3a). Moreover, the pseudo J–V curves extracted from the QFLS results revealed that PE-MP2/NBG achieved the highest pseudo-PCE of 24.08% and VOC of 0.875 V, even better than those of PEDOT:PSS/NBG (Fig. 3b).

Fig. 3. (a) QFLS as a function of excitation light intensity for samples with the structure of ITO/HTM/Sn–Pb perovskite/passivation, derived from PLQY measurements. (b) Calculated pseudo J–V curves of samples based on different HTLs. (c) The PCE statistics of Sn–Pb PSCs based on different HTLs. (d) J–V curves under forward scan (F) and reverse scan (R) of champion devices based on PEDOT:PSS and PE-MP2. (e) EQE spectra and integrated JSC of champion devices. (f) Normalized PCE evolution of unencapsulated devices stored in a nitrogen-filled glovebox in the dark at 65 °C. Devices based on PE-MP2 and PEDOT:PSS exhibited initial PCEs of 22.63% and 21.45%, respectively.

Fig. 3

Open in a new tab

We further fabricated inverted PSCs with the architecture of ITO/PEDOT:PSS or PE-MPs/FA0.8MA0.1Cs0.1Sn0.5Pb0.5I3/ethane-1,2-diaminium iodide (EDAI2)/fullerene (C60)/2,9-dimethyl-4,7-diphenyl-1,10-phenanthroline (BCP)/Ag. The statistical photovoltaic parameters of Sn–Pb perovskite solar cells are displayed in Fig. 3c and S16. The PE-MP2 based devices exhibited significant improvements in VOC and fill factor (FF) compared to PEDOT:PSS, where the average VOC was prompted from 0.859 V to 0.869 V, achieving a considerably enhanced average PCE of 21.86% compared to 20.64% for PEDOT:PSS-based devices. As shown in Fig. 3d and S17, the PE-MP2 device exhibited negligible hysteresis, with a champion PCE of 22.72% (a VOC of 0.878 V, current density (JSC) of 32.84 mA cm−2, and FF of 78.79%). It is worth noting that this is one of the highest efficiencies for non-PEDOT:PSS-based Sn–Pb PSCs (Table S3). As depicted in Fig. 2e, the external quantum efficiency (EQE)-integrated JSC values for the PE-MP2 and PEDOT:PSS devices were 32.59 and 32.05 mA cm−2 respectively, closely matching the JSC values of J–V measurement. Transient photocurrent (TPC) and transient photovoltage (TPV) measurements further showed that the PE-MP2 device exhibited faster photocurrent decay (1.72 vs. 2.69 µs) and slower photovoltage decay (17.23 vs. 8.82 µs), indicating improved charge extraction and suppressed nonradiative charge recombination when compared to PEDOT:PSS devices (Fig. S18).

Beyond photovoltaic performance, we further examined the storage stability of unencapsulated devices by keeping them at room temperature in a N2 atmosphere. As shown in Fig. S19, the PE-MP2 devices retained 90% of their initial efficiencies after 1100 hours, whereas the PCEs of the PEDOT:PSS devices rapidly dropped to 90% of their initial PCEs at around 370 hours. In addition, under continuous heating at 65 °C, the PE-MP2 devices also demonstrated enhanced stability by retaining 90% of their initial PCEs after 300 hours, outperforming the PEDOT:PSS devices, which showed a T90 of less than 130 hours (Fig. 3f). It is known that weak interfacial adhesion between perovskite and its substrate tends to trigger delamination, nanovoids, or cracks at the buried interface during aging at an elevated temperature.17 The enhanced storage and thermal stability can be attributed to superior intrinsic stability of PE-MP2 and the robust buried interface realized by its dual-sided anchoring properties.

Improved fabrication of all-perovskite tandems

To examine the impact of solution-based deposition of Sn–Pb perovskite and NBG-HTLs on tandem device fabrication, we first constructed a WBG perovskite based semi-tandem stack (Fig. 4a) with the architecture of glass/ITO/HTL/FA0.8Cs0.2Pb(I0.6Br0.4)3/EDAI2/PCBM/C60/SnOx/indium tin oxide nanoparticles (ITO NCs). The ITO NC layer provides a favourable substrate with a fully hydroxyl-covered surface for phosphonic acid anchoring.24 To quantify the potential damage to WBG perovskites during the fabrication of the Sn–Pb subcell, we performed QFLS analysis on the semi-stack of the tandem device in different stages (Fig. 4b and S20). The fabrication of the interconnection layer (ICL, including electron-transporting PCBM/C60, atomic layer deposition based SnOx and spin-coated ITO NCs) resulted in a 25–30 mV drop in QFLS (Fig. S20).25,26 The subsequent deposition of the HTL and Sn–Pb perovskite further decreased the QFLS (Fig. 4b), while the degradation was far more severe when using aqueous PEDOT:PSS as the NBG-HTL. We attributed this difference to the aqueous-based and acidic PEDOT:PSS, which likely penetrates the porous ITO NC layer, thereby compromising the interconnection layer (ICL) and damaging the underlying WBG perovskite. When using PE-MP2 (processed from ethyl acetate) as the NBG-HTL, the decrease in QFLS was largely mitigated.

Fig. 4. (a) Schematic of the semi-tandem stack and the experimental procedure for fabricating the NBG subcells, including simulated deposition with a cosolvent (DMF : DMSO = 3 : 1, v/v) and actual deposition of the Sn–Pb perovskite precursor on PE-MP2 or PEDOT:PSS HTLs. (b) QFLS for different layer stacks. The terms HTL and NBG denote PEDOT:PSS or PE-MP2 and Sn–Pb perovskite, respectively. (c) XRD patterns and (d) UV-vis absorption spectra of WBG perovskite after deposition of PE-MP2 or PEDOT:PSS under standard and prolonged processing. Cross-sectional SEM images: (e) device with PEDOT:PSS prepared by standard spin-coating, (f) device with PEDOT:PSS after prolonged processing, (g) device with PE-MP2 prepared by standard spin-coating, and (h) device with PE-MP2 after prolonged processing.

Fig. 4

Open in a new tab

Based on this finding, we further investigated the fabrication time window of different NBG-HTLs on a semi-tandem half stack. Specifically, we prolonged the dwell time of HTL solutions during the spin-coating process to simulate a situation in scalable fabrication like blade-coating. When the dwell time reached 20 seconds, the PEDOT:PSS sample exhibited visible holes on the film surface along with yellowing at the edges, indicating perovskite degradation and formation of lead iodide (Fig. S21); conversely, the sample with PE-MP2 as the NBG-HTL showed negligible change on the surface. Corresponding SEM surface images revealed that the distribution of ITO NCs became more non-uniform, with more and enlarged pinholes after prolonging deposition of PEDOT:PSS, verifying that aqueous-based PEDOT:PSS penetrates the porous ITO NC layer and damages the ICL (Fig. S22). Benefiting from non-destructive solvent processing, the deposition of PE-MP2 did not cause noticeable morphological differences in the semi-tandem stack regardless of dwell time. Meanwhile, the XRD pattern of the PEDOT:PSS-coated semi-tandem stack showed an obvious decrease in peaks of the perovskite phase along with the emergence of new peaks attributed to lead iodide and hydrated perovskite,27 indicating degradation of the WBG perovskite (Fig. 4c). By contrast, the semi-tandem stack with PE-MP2 retained its original structure and crystallinity after prolonged deposition. Similarly, the UV-vis absorption spectrum of the PEDOT:PSS-based stack exhibited a shift in the absorption edge and a significant reduction in the 400–500 nm range after prolonged deposition, which were attributed to perovskite decomposition (Fig. 4d). We further fabricated complete tandem devices. When extending the processing time, the PEDOT:PSS-based tandems displayed substantial degradation in the underlying layer, especially the WBG perovskite layer (Fig. 4e and f). In sharp contrast, the tandems based on PE-MP2 demonstrated well-preserved morphology with intact configurations (Fig. 4g and h). These results collectively confirm that non-destructive-solvent-processable PE-MP2 significantly improves the deposition compatibility and extends the processing window of NBG-HTLs during the fabrication of TSCs.

Following the deposition of Sn–Pb perovskite, QFLS measurements showed a further reduction in QFLS for underlying WBG perovskite, which presumably originated from damage to the ICL (including the NBG-HTL) caused by strong polarity of the NBG perovskite precursor (Fig. 4b). A more pronounced decrease in QFLS was observed in the PEDOT:PSS-based samples than in the PE-MP2 based samples. To further investigate the influence of the NBG perovskite deposition on the ICL and WBG subcell, we simulated the deposition process of the NBG perovskite layer by treating the stacks with the cosolvent using the same procedure (Fig. 4a). From the pictures of samples before and after cosolvent processing, the PEDOT:PSS-based sample exhibited many yellow holes (Fig. S23) along with decreased XRD diffraction intensity and UV-vis absorption, indicating degradation of the WBG perovskite layer (Fig. S24). In contrast, the PE-MP2 based film showed no noticeable changes under the same conditions. We attribute the marked difference of WBG perovskite to the different protection of NBG-HTLs. The PEDOT:PSS film would swell in strongly polar solvents, such as N,N-dimethylformamide (DMF),28 thus becoming more heterogeneous and less compact, which impaired its barrier function against the polar NBG precursor; however, the orthogonal-solvent-processable PE-MP2 provides effective protection for the underlying layers from NBG perovskite deposition.

We further deposited the Sn–Pb perovskite to fabricate complete tandem devices and changed the dwell time of the perovskite precursor. Cross-sectional SEM images of devices based on PEDOT:PSS showed that prolonging the dwell time (from 0 to 5 seconds) damaged the underlying ICL and adversely affected the upper NBG perovskite growth (Fig. S25). In contrast, devices based on PE-MP2 showed no obvious morphological changes after extending the deposition. The durable and enhanced protection through the PE-MP2 layer mainly stems from two factors. On one hand, the non-destructive-solvent-processable PE-MP2 does not damage the underlying ICL and causes negligible damage to the WBG subcell. On the other hand, the excellent solubility and orthogonal solubleness of PE-MP2 ensure a homogeneous and stable NBG-HTL during the deposition of NBG perovskite, providing robust protection for the underlying layers. Accordingly, PE-MP2 broadens the processing window of the NBG-HTL and greatly improves the fabrication of all-perovskite tandem solar cells, offering great potential for large-area fabrication of stable TSCs.

Performance of all-perovskite tandem solar cells

With the improved NBG perovskite subcells, we further fabricated monolithic all-perovskite tandems with a device configuration of ITO/HTL/WBG perovskite/EDAI2/PCBM/C60/SnOx/ITO NCs/PEDOT:PSS or PE-MP2/NBG perovskite/EDAI2/PCBM/C60/BCP/Ag (Fig. 5a). Similarly, the PE-MP2 based tandem devices exhibited significant improvement in photovoltaic performance compared to the PEDOT:PSS based tandems (Fig. 5b and S26). The improvement in VOC and FF is attributed to suppressed nonradiative recombination loss at the PE-MP2/NBG interface and enhanced protection through uniform and compact PE-MP2 films. We prolonged the dwell time during deposition of PEDOT:PSS or PE-MP2 to simulate a broad processing window in scalable fabrication. As shown in Fig. 5b and S26, the PEDOT:PSS-based tandem devices displayed severe PCE drop after prolonging the dwell time, which was in accordance with destroyed morphology and structure of WBG subcells mentioned in the previous discussion. In stark contrast, the PE-MP2 tandems nearly retained the original photovoltaic performance under the same conditions.

Fig. 5. (a) The schematic structure of all-perovskite tandem solar cells. Comparative performance of all-perovskite tandem solar cells fabricated with PEDOT:PSS or PE-MP2 using standard spin-coating and prolonged processing with a 20 s dwell time: (b) PCE statistics, (c) J–V characteristics, and (d) EQE curves of champion devices. (e) Summary table of PCE and VOC for all-perovskite solar cells using non-PEDOT:PSS HTMs.5,12,24,25 (f) Continuous MPP tracking of encapsulated tandem solar cells under simulated AM 1.5G illumination in ambient air with a relative humidity (RH) of 30–50%. Devices based on PE-MP2 and PEDOT:PSS exhibited initial PCEs of 28.69% and 27.33%, respectively.

Fig. 5

Open in a new tab

Fig. 5c shows the J–V curves of champion tandem devices based on different HTLs under various conditions. The best-performing PE-MP2 tandem achieved a PCE of 29.08%, with a VOC of 2.183 V, a JSC of 16.05 mA cm−2, an FF of 82.98%, and a steady-state PCE of 28.61% during constant operation of 400 seconds (Fig. S27). The JSC values obtained by integrating the EQE spectra of WBG and NBG subcells were 16.17 and 15.97 mA cm−2 respectively, which was almost consistent with the JSC obtained from J–V measurement (Fig. 5d). We sent one of the PE-MP2 based TSCs to an accredited photovoltaic calibration laboratory for third-party independent measurements. The tandem device delivered a certified PCE of 28.69%, with a VOC of 2.192 V, FF of 82.50%, and JSC of 15.86 mA cm−2 (Fig. S28). Encouragingly, this is one of the highest performances reported for the new HTM-based all-perovskite tandems besides PEDOT:PSS (Fig. 5e and Table S3).

We further investigated the operational stability of encapsulated tandems under simulated one-sun illumination with maximum power point (MPP) tracking in an air atmosphere (Fig. 5f). The PEDOT:PSS tandem retained 90% of the initial efficiency after only 160 hours, while the PE-MP2 tandem device retained 90% of its initial efficiency after operating for 570 hours. Moreover, the thermal stability of unencapsulated tandem devices was monitored (Fig. S29). The PEDOT:PSS-based TSCs only retained 90% of their initial PCEs for approximately 100 hours under thermal aging of 65 °C, while TSCs with PE-MP2 can still retain 90% of their initial PCEs for over 340 hours. The improved operational and thermal stability confirm the enhanced stability of TSCs through non-destructive-solvent-processable and robust PE-MP2 and rank among the most efficient and stable tandems without PEDOT:PSS (Table S3).

Conclusions

In summary, we have demonstrated a new type of polyvinyl-based HTM that contains triarylamine and phosphonic acid as side chains, as an alternative to PEDOT:PSS for application in Sn–Pb perovskite solar cells and all-perovskite tandem solar cells. The polyvinyl backbone provides excellent green-solvent-processability, which enables non-destructive deposition onto the underlying ICL and WBG subcell. The PL characterization of semi-tandem stacks revealed that these new HTMs minimize potential degradation and maintain device integrity while broadening the processing window for tandem fabrication. By tuning the ratio of triphenylamine units to phosphonic acid units, PE-MP2 realizes efficient hole extraction and robust interfacial adhesion, thereby enhancing both device efficiency and operational stability. PE-MP2-based devices achieve champion power conversion efficiencies of 22.72% for Sn–Pb perovskite solar cells and 29.08% (certified 28.69%) for all-perovskite tandem solar cells, representing one of the highest performances among PEDOT:PSS-free TSCs. Furthermore, all perovskite tandem solar cells employing PE-MP2 as the HTL for Sn–Pb subcells exhibit significantly enhanced operational stability, with a T90 of 570 hours under MPPT, ranking among the most stable all-perovskite tandems without PEDOT:PSS. Overall, our work highlights the significance of non-destructive and durable NBG-HTLs during tandem fabrication and offers a new multifunctional polymer-based HTM to replace PEDOT:PSS through rational design, which paves the way for efficient and stable all-perovskite TSCs.

Author contributions

Y. Z., L. Z. and J. L. conceived the idea, Y. Z. synthesized PE-MP molecules with the assistance of L. Z. and performed related characterization, J. L. conducted device characterization and wrote the manuscript with Y. Z., R. M. assisted in some experiments and contributed to the fabrication of all-perovskite solar cells together with H. T. and X. L., Y. W. and W. Z. supervised the project, participated in the investigation, and, together with S. Z. and L. Z., revised the manuscript, L. Z. performed the SEM characterization of samples, and Q. Z. and S. Y. conducted TPC and TPV characterization of the tin–lead perovskite PSCs. All authors contributed to the discussion and review of this manuscript.

Conflicts of interest

There are no conflicts to declare.

Supplementary Material

SC-017-D6SC03304K-s001

Acknowledgments

The work was supported by the National Natural Science Foundation of China (22425502, T2488302, W2412114, 22179037, and 22509061), the Guangxi Science and Technology Innovation Platform Program (“Leitai” Action Plan – Guangxi Laboratory Capacity Building) (LT2504240035), the Shanghai Municipal Science and Technology Major Project (24DX1400200 and 24DZ3001003), the Shanghai Pilot Program for Basic Research (22TQ14001001), the Program of Introducing Talents of Discipline to Universities (B16017), the Shanghai Sailing Program (24YF2708200), Fundamental Research Funds for the Central Universities, the China Postdoctoral Science Foundation (BX20240117 and 2025M780061), and the Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China (JYB2025XDXM404).

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6sc03304k.

Notes and references

  1. (a) Liu S. Li J. Xiao W. Chen R. Sun Z. Zhang Y. Lei X. Hu S. Kober-Czerny M. Wang J. Ren F. Zhou Q. Raza H. Gao Y. Ji Y. Li S. Li H. Qiu L. Huang W. Zhao Y. Xu B. Liu Z. Snaith H. J. Park N. G. Chen W. Nature. 2024;632:536. doi: 10.1038/s41586-024-07723-3. [DOI] [PubMed] [Google Scholar]; (b) Liu G. Yang G. Feng W. Wu W.-Q. Chem. Sci. 2026;17:4846. doi: 10.1039/D5SC09981A. [DOI] [PMC free article] [PubMed] [Google Scholar]; (c) Xiong Z. Zhang Q. Cai K. Zhou H. Song Q. Han Z. Kang S. Li Y. Jiang Q. Zhang X. You J. Science. 2025;390:638. doi: 10.1126/science.adw8780. [DOI] [PubMed] [Google Scholar]; (d) Lv J. Le L. Yao S. Huang Z. Chen Y. Chem. Sci. 2025;16:18559. doi: 10.1039/D5SC05266A. [DOI] [PMC free article] [PubMed] [Google Scholar]; (e) Liu L. Huang X. Jen A. K. Chem. Sci. 2026;17:5351. doi: 10.1039/D6SC00127K. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. (a) Lin R. Xiao K. Qin Z. Han Q. Zhang C. Wei M. Saidaminov M. I. Gao Y. Xu J. Xiao M. Li A. Zhu J. Sargent E. H. Tan H. Nat. Energy. 2019;4:864. doi: 10.1038/s41560-019-0466-3. [DOI] [Google Scholar]; (b) Wang C. Song Z. Li C. Zhao D. Yan Y. Adv. Funct. Mater. 2019;29:1808801. doi: 10.1002/adfm.201808801. [DOI] [Google Scholar]; (c) Zhao D. Ding L. Sci. Bull. 2020;65:1144. doi: 10.1016/j.scib.2020.04.013. [DOI] [PubMed] [Google Scholar]
  3. Chen B. Yu Z. Onno A. Yu Z. Chen S. Wang J. Holman Z. C. Huang J. Sci. Adv. 2022;8:eadd0377. doi: 10.1126/sciadv.add0377. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Chen H.-W. Li C. Chin. J. Polym. Sci. 2019;38:435. doi: 10.1007/s10118-020-2373-2. [DOI] [Google Scholar]
  5. Fu S. Sun N. Xian Y. Chen L. Li Y. Li C. Abudulimu A. Kaluarachchi P. N. Huang Z. Wang X. Dolia K. Ginger D. S. Heben M. J. Ellingson R. J. Chen B. Sargent E. H. Song Z. Yan Y. Joule. 2024;8:2220. doi: 10.1016/j.joule.2024.05.007. [DOI] [Google Scholar]
  6. Liu T. Jiang Y. Qin M. Liu J. Sun L. Qin F. Hu L. Xiong S. Jiang X. Jiang F. Peng P. Jin S. Lu X. Zhou Y. Nat. Commun. 2019;10:878. doi: 10.1038/s41467-019-08843-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. (a) He J. Wang J. Zhu H. Wu R. Chen X. Li D. Jiang Y. Wang K. Zhan C. Gao P. Ma Y. Zhu R. Li J. Tian C. Wei Z. Meng L. Zhou Y. Chen W. Zhao Y. Liu Z. Zhao Q. Energy Environ. Sci. 2026;19:1192. doi: 10.1039/D5EE06282A. [DOI] [Google Scholar]; (b) Wang G. Cui Y. Ou M. Yu Y. Xiao Y. Yuan H. Zhang T. Chen Z. Yang N. Ma L. Yu R. Zhang S. Hou J. Joule. 2026;10:102263. doi: 10.1016/j.joule.2025.102263. [DOI] [Google Scholar]
  8. (a) Jin X. Cao J. Sui S. Xie J. Yan W. Han L. Duan J. Hu M. Zhang H. Wang F. Dong J. Wang A. Yuan W. Huang W. Qin T. Angew. Chem., Int. Ed. 2026;65:e18800. doi: 10.1002/anie.202518800. [DOI] [PubMed] [Google Scholar]; (b) Lai H. Zhu J. Kuo R. T. Aeberhard U. Lin Z. H. Mrkyvkova N. Luo J. Siegrist S. Wechsler P. Shih C. J. Siffalovic P. Chen C. Tiwari A. N. Lin T. Y. Zhao D. Fu F. Nat. Commun. 2025;16:11468. doi: 10.1038/s41467-025-66479-0. [DOI] [PMC free article] [PubMed] [Google Scholar]; (c) Pan H. Bai Y. Sun K. Yang M. Tian R. Meng Y. Gao J. Wang Y. Wang J. Zhou S. Song Z. Xiaoyi L. Liu C. Ge Z. Nat. Commun. 2025;16:9221. doi: 10.1038/s41467-025-64274-5. [DOI] [PMC free article] [PubMed] [Google Scholar]; (d) Sun X. Wu H. Li Z. Zhu R. Li G. Su Z. Zhang J. Gao X. Pascual J. Abate A. Li M. Angew. Chem., Int. Ed. 2024;63:e202409330. doi: 10.1002/anie.202409330. [DOI] [PubMed] [Google Scholar]; (e) Zhu J. Xu Y. Luo Y. Luo J. He R. Wang C. Wang Y. Wei K. Yi Z. Gao Z. Wang J. You J. Zhang Z. Lai H. Ren S. Liu X. Xiao C. Chen C. Zhang J. Fu F. Zhao D. Sci. Adv. 2024;10:eadl2063. doi: 10.1126/sciadv.adl2063. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Han Q. Wei Y. Lin R. Fang Z. Xiao K. Luo X. Gu S. Zhu J. Ding L. Tan H. Sci. Bull. 2019;64:1399. doi: 10.1016/j.scib.2019.08.002. [DOI] [PubMed] [Google Scholar]
  10. (a) Kapil G. Bessho T. Sanehira Y. Sahamir S. R. Chen M. Baranwal A. K. Liu D. Sono Y. Hirotani D. Nomura D. Nishimura K. Kamarudin M. A. Shen Q. Segawa H. Hayase S. ACS Energy Lett. 2022;7:966. doi: 10.1021/acsenergylett.1c02718. [DOI] [Google Scholar]; (b) Zhang Z. Zhu R. Tang Y. Su Z. Hu S. Zhang X. Zhang J. Zhao J. Xue Y. Gao X. Li G. Pascual J. Abate A. Li M. Adv. Mater. 2024;36:e2312264. doi: 10.1002/adma.202312264. [DOI] [PubMed] [Google Scholar]; (c) Wang J. Yu Z. Astridge D. D. Ni Z. Zhao L. Chen B. Wang M. Zhou Y. Yang G. Dai X. Sellinger A. Huang J. ACS Energy Lett. 2022;7:3353. doi: 10.1021/acsenergylett.2c01578. [DOI] [Google Scholar]; (d) Yuan H. Li X. Zhang W. Hu Y. Xu J. You T. Weng Q. Mao Y. Shu T. Fang J. Adv. Mater. 2025;37:e2502680. doi: 10.1002/adma.202502680. [DOI] [PubMed] [Google Scholar]
  11. (a) Zhang X. Pascual J. Li Z. Zhang X. Su Z. Zhang J. Gao X. Hou B. Li G. Abate A. Li M. Sci. Bull. 2025;70:556. doi: 10.1016/j.scib.2024.12.004. [DOI] [PubMed] [Google Scholar]; (b) Zhu J. Huang X. Luo Y. Jiao W. Xu Y. Wang J. Gao Z. Wei K. Ma T. You J. Jin J. Wu S. Zhang Z. Liang W. Wang Y. Ren S. Wang C. Chen C. Zhang J. Zhao D. Nat. Commun. 2025;16:240. doi: 10.1038/s41467-024-55492-4. [DOI] [PMC free article] [PubMed] [Google Scholar]; (c) Zhumagali S. Li C. Marcinskas M. Dally P. Liu Y. Ugur E. Petoukhoff C. E. Ghadiyali M. Prasetio A. Marengo M. Pininti A. R. Azmi R. Schwingenschlögl U. Laquai F. Getautis V. Malinauskas T. Aydin E. Sargent E. H. De Wolf S. Adv. Energy Mater. 2025;15:2404617. doi: 10.1002/aenm.202404617. [DOI] [Google Scholar]
  12. Wu J. Hu M. Dai Q. Alkan E. A. Barabash A. Zhang J. Liu C. Hauch J. A. Han G. F. Jiang Q. Wang T. Seok S. I. Brabec C. J. Angew. Chem., Int. Ed. 2025;64:e202424515. doi: 10.1002/anie.202424515. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Chen J. Du J. Cai J. Ouyang B. Li Z. Wu X. Tian C. Sun A. Zhuang R. Wu X. Chen C. Cen T. Li R. Xue T. Zhao Y. Zhao K. Chen Q. Chen C.-C. ACS Energy Lett. 2025;10:1117. doi: 10.1021/acsenergylett.4c03370. [DOI] [Google Scholar]
  14. Zhan L. Zhang S. Gao W. Zheng W. Li Z. Jiang X. Ning Z. Han L. Li Z. Stolterfoht M. Zhu W. H. Wu Y. Adv. Mater. 2025;37:e2506048. doi: 10.1002/adma.202506048. [DOI] [PubMed] [Google Scholar]
  15. (a) Sturzel M. Mihan S. Mulhaupt R. Chem. Rev. 2016;116:1398. doi: 10.1021/acs.chemrev.5b00310. [DOI] [PubMed] [Google Scholar]; (b) Xu Y. Bu T. Li M. Qin T. Yin C. Wang N. Li R. Zhong J. Li H. Peng Y. Wang J. Xie L. Huang W. ChemSusChem. 2017;10:2578. doi: 10.1002/cssc.201700584. [DOI] [PubMed] [Google Scholar]
  16. Pan Z. Peng D. Zhao X. Xu W. Bao Y. Feng Z. Zou Q. Xu B. Wang Y. Gao H. Yin C. Li R. Wang J. Huang W. Adv. Funct. Mater. 2023;33:2304881. doi: 10.1002/adfm.202304881. [DOI] [Google Scholar]
  17. Zhan L. Zhang S. Li Z. Li W. Zhang H. He J. Ji X. Liu S. Yu F. Wang S. Ning Z. Li Z. Stolterfoht M. Han L. Zhu W. H. Xu Y. Wu Y. Angew. Chem., Int. Ed. 2025;64:e202422571. doi: 10.1002/anie.202422571. [DOI] [PubMed] [Google Scholar]
  18. Zhan L. Zhang L. Li Y. Cai H. Wu Y. ACS Appl. Energy Mater. 2025;8:3985. doi: 10.1021/acsaem.5c00155. [DOI] [Google Scholar]
  19. Ji X. Zhao Y. Chen X. Zhang S. Zhan L. Zhang H. Zheng W. Zhu W. H. Wu Y. Chem. Sci. 2025;16:8569. doi: 10.1039/D4SC08810G. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Sun Q. Li Z. Cheng T. Ma G. He D. Su G. Chen H. Yin L. He Z. Phys. Chem. Lett. 2025;16:2348. doi: 10.1021/acs.jpclett.4c03648. [DOI] [PubMed] [Google Scholar]
  21. Bi C. Wang Q. Shao Y. Yuan Y. Xiao Z. Huang J. Nat. Commun. 2015;6:7747. doi: 10.1038/ncomms8747. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Fei C. Zhang Y. Wang M. Yang Y. Shi X. Wang S. Guo H. Guo C. Liu H. Li M. Gu H. Barlow S. You W. Wang M. Marder S. R. Huang J. Science. 2026;391:eadz7969. doi: 10.1126/science.adz7969. [DOI] [PubMed] [Google Scholar]
  23. (a) Caprioglio P. Smith J. A. Oliver R. D. J. Dasgupta A. Choudhary S. Farrar M. D. Ramadan A. J. Lin Y. H. Christoforo M. G. Ball J. M. Diekmann J. Thiesbrummel J. Zaininger K. A. Shen X. Johnston M. B. Neher D. Stolterfoht M. Snaith H. J. Nat. Commun. 2023;14:932. doi: 10.1038/s41467-023-36141-8. [DOI] [PMC free article] [PubMed] [Google Scholar]; (b) Caprioglio P. Stolterfoht M. Wolff C. M. Unold T. Rech B. Albrecht S. Neher D. Nat. Commun. 2019;9:1901631. [Google Scholar]; (c) Ye F. Zhang S. Warby J. Wu J. Gutierrez-Partida E. Lang F. Shah S. Saglamkaya E. Sun B. Zu F. Shoaee S. Wang H. Stiller B. Neher D. Zhu W. H. Stolterfoht M. Wu Y. Nat. Commun. 2022;13:7454. doi: 10.1038/s41467-022-34203-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. (a) Liu C. Lin R. Wang Y. Gao H. Wu P. Luo H. Zheng X. Tang B. Huang Z. Sun H. Zhao S. Guo Y. Wen J. Fan F. Tan H. Angew. Chem., Int. Ed. 2023;62:e202313374. doi: 10.1002/anie.202313374. [DOI] [PubMed] [Google Scholar]; (b) Wu P. Wen J. Wang Y. Liu Z. Lin R. Li H. Luo H. Tan H. Adv. Energy Mater. 2022;12:202202948. [Google Scholar]
  25. Thiesbrummel J. Peña-Camargo F. Brinkmann K. O. Gutierrez-Partida E. Yang F. Warby J. Albrecht S. Neher D. Riedl T. Snaith H. J. Stolterfoht M. Lang F. Adv. Energy Mater. 2022;13:2202674. doi: 10.1002/aenm.202202674. [DOI] [Google Scholar]
  26. Yang F. Tockhorn P. Musiienko A. Lang F. Menzel D. Macqueen R. Köhnen E. Xu K. Mariotti S. Mantione D. Merten L. Hinderhofer A. Li B. Wargulski D. R. Harvey S. P. Zhang J. Scheler F. Berwig S. Roß M. Thiesbrummel J. Al-Ashouri A. Brinkmann K. O. Riedl T. Schreiber F. Abou-Ras D. Snaith H. Neher D. Korte L. Stolterfoht M. Albrecht S. Adv. Mater. 2023;36:2307743. doi: 10.1002/adma.202307743. [DOI] [PubMed] [Google Scholar]
  27. Ma C. Eickemeyer F. T. Lee S.-H. Kang D.-H. Kwon S. J. Grätzel M. Park N.-G. Science. 2023;379:173. doi: 10.1126/science.adf3349. [DOI] [PubMed] [Google Scholar]
  28. Li H. Zhang C. Ma Y. Li Z. Xu Y. Mai Y. J. Energy Chem. 2019;32:71. doi: 10.1016/j.jechem.2018.06.013. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

SC-017-D6SC03304K-s001
SC-017-D6SC03304K-s001.pdf (2.8MB, pdf)

Data Availability Statement

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6sc03304k.

Articles from Chemical Science are provided here courtesy of Royal Society of Chemistry

RESOURCES