Improving the ecological index in the solution processing of hybrid perovskites†

Niusha Heshmati, Niklas Almandinger, Thomas Fischer and Sanjay Mathur*
Department of Chemistry, Institute of Inorganic and Materials Chemistry, University of Cologne, Cologne, 50939, Germany. E-mail: sanjay.mathur@uni-koeln.de

First published on 17th July 2025

---

---

Introduction

Lead halide perovskite solar cells were first discovered in 2009, with the initial devices achieving a modest power conversion efficiency (PCE) of 3.8%. Since then, extensive interdisciplinary research and consequent optimization of processing methods have propelled the PCE values to exceed 26% for single-junction configurations.^1,2 A key factor in the rise of perovskite research has been the easy accessibility to materials and their device integration, enabled by solution-based procedures, which are both cost-effective and highly reproducible.^3,4 Although solar energy is widely regarded as a promising clean energy source, a major barrier in the commercialization of perovskite solar cells relates to the toxicity of materials and solvents essential for their synthesis and processing.⁵

First, the use of lead (Pb) as an essential B site cation in the ABX₃ perovskite structure presents a significant environmental challenge. Efforts to substitute lead with elements from groups 14 and 15 of the periodic table have been explored; however, the PCE values in devices based on these alternative compositions fall short in competition with the lead-based PSCs.⁶ To this end, research also focuses on the encapsulation of perovskite solar cells (PSCs) to ensure long-term stability and prevent lead leakage.^7,8 A second major environmental concern is the extensive use of organic solvents, particularly volatile and toxic solvents like N,N-dimethylformamide (DMF) that are inhibitive for ambient processing due to their potential health risks.⁹ To emphasize the massive amount of required solvent, producing 1 GW of solar power with a blade-coated perovskite (15% module efficiency) demands about 3500 liters of solvent.¹⁰ Although vapor phase deposition methods offer a solvent-free alternative, they require high vacuum equipment which increases costs, processing complexity, and energy demand. Particularly, for mixed-cation and mixed-anion perovskite compositions, achieving precise stoichiometry control remains a challenge mostly due to the different vapor pressure of each component.^11,12

While significant efforts have been made to counteract lead toxicity in perovskite solar cells, comparatively less attention has been given to reducing the environmental impact by identifying green alternative solvents.^13,14 Recently, solvent engineering has received recognition as a potential strategy to develop solvent systems that are both high-performing and environmentally sustainable for PSC applications.¹⁵ However, a key challenge in identifying green solvents capable of dissolving Pb halides (PbI₂, PbBr₂, and PbCl₂) is producing perovskite inks based on a single, stable solution suitable for the one-step spin coating method. As a result, green solvent research has primarily focused on simpler perovskite compositions (e.g. FAPbI₃, MAPbI₃).¹⁶

Besides DMF and dimethyl sulfoxide (DMSO), perovskite film depositions have been successfully conducted using other polar aprotic solvents (non-proton donating) such as γ-butyrolactone (GBL), N-methyl-2-pyrrolidone (NMP) and dimethylacetamide (DMAc).¹⁷ The chemical appeal of polar aprotic solvents lies in their coordinative abilities to function as a Lewis-base and form adducts with metal species such as Pb²⁺ by donating electron pairs.^15,17,18 Recent reports have demonstrated that the polarity of these solvents significantly influences the coordination and stabilization of the lead-species in the solution.¹⁹

Oxygen (O) and nitrogen (N) are well-studied donor atoms, provided by solvents like γ-butyrolactone (C₄H₆O₂; GBL), N-methyl-2-pyrrolidone (C₅H₉NO; NMP), dimethylformamide (C₃H₇NO; DMF), dimethyl sulfoxide (C₂H₆OS; DMSO), acetonitrile (C₂H₃N; ACN), and water (H₂O). Solvent–solute analysis of perovskite solutions shows the predominance of [PbI_x]^2−x iodoplumbate clusters, whose nuclearity is primarily determined by the coordination ability of the solvents used.¹⁸ A common parameter to measure this coordination ability is Gutmann's donor number (DN), which correlates with the affinity of the solvent molecules to bind to Pb²⁺ species. To fabricate perovskite thin films with uniform grain sizes and defect-free structures needed for high-performance devices, it is crucial to achieve full dissolution of the ionic precursor salts. Also, simultaneously enabling the formation of high nuclearity clusters can reduce the perovskite's nucleation barrier by providing a high local concentration of precursor species. DMSO, with its strong oxygen donor capability, exhibits higher affinity for binding to Pb²⁺ compared to DMF, and that is evident in their respective donor numbers of 29.8 kcal mol⁻¹ and 26.6 kcal mol⁻¹.^20,21 Typically, solvents with high donor numbers like DMSO compete with halide anions (I⁻, Br⁻, Cl⁻) for coordination sites around Pb²⁺, thus inhibiting the formation of halidoplumbates, and consequently forming stable precursor solutions. Conversely, solvents with lower donor numbers coordinate weakly to Pb²⁺ and promote halidoplumbate formation which is detrimental to obtaining phase-pure perovskites.²² This highlights the importance of DMSO's contribution in forming higher nuclearity [PbI_x]^2−x complexes with relatively low toxicity,²³ and the potential to create stable and eco-friendly solvent systems, while ensuring high-quality thin film formation.^17,18 Therefore, mixed-solvent systems are chemically engineered to modify and optimize solution characteristics such as polarity, viscosity, and solvent–solute interactions.¹⁵ Volatile green solvents such as ethanol and acetonitrile represent a significant advancement in this research.²⁴

However, their rapid nucleation rates may limit grain growth, which is essential for optimal device performance and stability.^25,26 On the other hand, the slower evaporation of aprotic solvents with controlled evaporation rates can result in larger grains with higher crystallinity, further improving the device performance.²⁷

Dihydrolevoglucosenone (Cyrene™) derived from renewable and biodegradable cellulose is a green alternative for perovskite solar cell fabrication.^28,29 As an aprotic solvent, Cyrene minimizes the risk of proton-mediated degradation of perovskite films.¹³ Given the presence of the carbonyl group, it can interact with lead ions to stabilize the precursor solution by forming intermediate complexes, similar to protic solvents.³⁰ However, Cyrene has a relatively high viscosity (14.5 cP compared to 0.92 cP for DMF, 1.65 cP for NMP, and 2.20 cP for DMSO) that can hinder the formation of perovskite inks with optimal rheology for various deposition techniques.^26,29 To counteract the coating challenges posed by high-viscosity solvents like Cyrene and achieve smooth film formation with full coverage, in this work solvent-blending is considered by developing solvent systems that retain the benefits of ecological solar cell processing while optimizing film properties.^29,31 A solvent-blend consisting of Cyrene and 2-methyltetrahydrofuran (2-MeTHF) in a ratio of 80:20 vol% (denoted as Cy-THF) was developed that significantly lowered the viscosity (3.72 cP) to become a promising alternative for high-volume hybrid perovskite ink production.

In this work, the mixed-solvent system combines the low-toxicity and ecological considerations of perovskite inks with enhanced film properties. Additionally, the optimized solvent blend enables a one-step perovskite film deposition with anti-solvent treatment. For making the entire process eco-friendly (Fig. 1), ethyl acetate (CH₃CO₂CH₂CH₃; EA) was used as a green antisolvent with a high threshold limit value and a lower boiling point (77 °C), compared to commonly used antisolvents like toluene (C₆H₅CH₃; TL) and chlorobenzene (C₆H₅Cl; CB).^32,33 Ultimately, the solar cells made using a green solvent achieve 95% of the performance of the DMF reference solar cells with 13.3% efficiency and higher stability.

Results and discussion

Solvent and solute interactions

The reported methods for manufacturing lead halide perovskite thin films commonly involve a solvent mixture of DMF:DMSO with 80:20 vol%.^34,35 This research aims to obtain a high-quality thin film processed from a sustainable green solvent-based solution centered around a Cyrene: 2-methyltetrahydrofuran (80:20 vol%) blend denoted as (Cy-THF). Fig. 2a shows 40 mg mL⁻¹ perovskite precursor salts in Cy-THF. PbI₂ and PbBr₂ remain insoluble as yellow and white solids respectively resulting from the insufficient polarity and low hydrogen bonding affinity (δ_H) value of Cy-THF; therefore, co-solvent addition is necessary. DMSO was chosen as a co-solvent due to its low toxicity and the strong coordination effect on PbI₂ which was shown to be essential for full dissolution of the perovskite salts and formation of high-quality crystal phases.^34,36 FAI can be fully ionized in Cy-THF because of its low bonding energy and results in a clear solution without DMSO. Experiments optimized the Cy-THF ratio, annealing conditions, and antisolvent evaluation detailed in the ESI, Sections 2.1 and 2.2† for MAPbI₃ perovskite. Additionally, to gain a deeper understanding of Cy-THF:DMSO solution properties, coordinating additives and suitability for perovskite precursor solutions, ¹H-NMR data of fresh and aged 1 M MAPbI₃ solutions are provided in ESI Section 2.3.† These results represent the suitability of the solvent mixture because of the major role that DMSO plays in coordinating PbI₂ to make the MAI-PbI₂-DMSO intermediate phase.¹⁵ In view of the high stability reported for Cs_0.05(MA_0.17FA_0.83)_0.95Pb(I_0.83Br_0.17)₃ perovskite (referred to as a triple-cation perovskite), within this research attempts were made to develop green solvents for triple cation perovskites.³⁷ In addition, 0.8 M concentration was chosen for the triple cation perovskite to reduce the Pb amount and enhance solution stability.³⁸ The development of a green solvent-based triple cation perovskite with a single final ink formulation remains unexplored, and the complexity of the solution chemistry presents a significant challenge.³⁹

As a starting point, the Cy-THF:DMSO:DMF (60:30:10 vol%) and DMF-free (70:30:0 vol%) solvent blends were evaluated to determine the possibility of complete DMF removal for the triple cation solution. The green solvent perovskite thin films were prepared by one-step spin coating onto FTO/mesoporous TiO₂ (m-TiO₂) and treated with EA as an antisolvent, followed by annealing at 150 °C for 30 minutes. The degree of crystallinity was studied by X-ray diffraction (XRD) and Fig. 2b shows the XRD patterns. The measured signals show that the desired cubic lattice (100), (110), (111), (200), and (220) planes for the triple cation composition have been obtained for both thin films. Surprisingly, the full width at half maximum (FWHM) for all peaks especially (100) and (200) decreases for the DMF-free sample indicating higher crystallization.⁴⁰ Additionally, the presence of PbI₂ due to the incomplete formation of perovskite can significantly affect optoelectronic properties by increasing non-radiative charge carrier recombination and inducing surface defects, which is visible for both films' XRD patterns at 2θ = 12.6°.^41,42 To evaluate the photoabsorbance properties of perovskite thin films the UV-Vis spectra were measured and are shown in Fig. 2c. A clear absorption edge can be observed around a wavelength of 780 nm for both films that can be related to a bandgap energy of 1.59 eV, which agrees with the literature for the triple cation composition.² For the DMF-free thin film the absorbance spectrum is similar to that of the 10% DMF sample, and the absorption edge alone has a 5 nm red shift that can indicate the crystallinity enhancement for this sample, confirming the XRD results.⁴³ Although the removal of DMF has shown promising results and the addition of 30 vol% DMSO dissolves PbI₂ and forms a clear stable solution, the high viscosity of the Cy-THF:DMSO (70:30 vol%) solution remains a challenge that needs to be addressed.

Solvent engineering for perovskite inks

To adapt to the different coating methods, the precursor solution must confront issues such as viscosity, solution stability, and film wettability. To address the Cy-THF:DMSO (70:30 vol%) high viscosity challenge, the technique of “viscosity-blending” was employed which involves the incorporation of a new solvent with lower viscosity.²⁹ Acetonitrile (ACN) as a HSE recommended solvent,⁶ with a very low viscosity of 0.38 cP at 25 °C, has gained attention for perovskite solar cell fabrication as an antisolvent, and also as an additive in the established DMF:DMSO solvent system.^44,45 It has been demonstrated that DMSO with a high donor number of 29.8 kcal mol⁻¹ as a main solvent and ACN with a low DN of 14.1 kcal mol⁻¹ (ref. 21) as a cosolvent can significantly influence Pb²⁺ coordination and formation of high valence coordination spheres.¹⁸ Perovskite precursor solutions were prepared with a foundation of Cy-THF:DMSO (70:30 vol%) and ACN was added in amounts of 5, 7.5, 10, and 12.5 vol% to the precursor solution over multiple experiments to investigate its influence on the applied solvent system.¹⁷ The addition of ACN showed an immediate improvement for the solutions, and the viscosity significantly reduced and full dissolution of the perovskite salts were obtained after heating at 50 °C for only 30 minutes. In addition to high coordination by DMSO, the formation of Lewis acid-base pairs between ACN and PbI₂ enhanced the solubility of salts.⁴⁵ The precursor solutions showed rising film wettability, resulting in higher-quality thin films of cubic α perovskite.

Fig. 3a illustrates the XRD pattern, revealing the formation of a cubic α-phase with high crystallinity for films prepared with 7.5 and 10 vol% ACN addition. Peak intensities correlate with the microstructural change of the fabricated films, and intensity enhancement indicates preferred cubic lattice orientation improvement.⁴⁶ Addition of 7.5 vol% ACN leads to the reduction of PbI₂ hexagonal (marked peak at ∼12.6°) which corresponds to the presence of intermediate phases such as FAI-PbI₂-DMSO/Cyrene.⁴⁷ To study the film homogeneity, the samples were analyzed via scanning electron microscopy (SEM) presented in Fig. 3b. The SEM images for 7.5 vol% ACN show flake-shape grains resulting in incomplete surface coverage. Hence, utilization of acetonitrile in the perovskite precursor addresses several challenges, including viscosity, dissolution, and film wettability. Based on these results, the optimal amount of acetonitrile ranges between 7.5 and 10 vol% addition; however, full surface coverage and homogeneous film formation are still not achieved.

Stabilization of perovskite by volatile salt addition

With the appropriate solvent and solute interactions established in a previous section the focus was on exploring additional treatments that directly influence the nucleation of the deposited perovskites. An approach that has garnered attention over recent years is the addition of volatile salt additives like methylammonium chloride (MACl) as a transitional stabilizer that preserves the crystal structure of a given perovskite while promoting the formation of stable, photoactive phases.^48,49 CH₃NH₃Cl will react with the perovskite precursor and form complexes like PbI₂·CH₃NH₂ and PbI₂·HCl that lead to the gradual release of ions to control the nucleation and growth process, and CH₃NH₂ and HCl will evaporate during the annealing.²³ This approach has been geared towards the FAPbI₃ perovskite because of the common transition between the photoactive black perovskite α-phase and a photoinactive yellow δ-phase,^2,50,51 and addition of MACl resulted in pure α-phase crystallization. Given that the predominant component of the triple cation perovskite is FAI, this method could be an effective approach to enhance the film properties. The perovskite precursor solutions were prepared by adding MACl salt in amounts of 20% (wt% in relation to FAI) to the optimized solution of Cy-THF:DMSO (70:30) + 7.5–10 vol% of acetonitrile. The XRD measurements in Fig. 4a compared the effect of ACN amount and MACl addition on the crystallinity of thin films. Notably, XRD patterns of samples with MACl additive showed no detectable PbI₂, suggesting the absence of unreacted PbI₂ crystals, which can delay PSC degradation.⁵² Moreover, the FWHM and intensity for the cubic perovskite planes have drastically increased, and direct crystallization into the photoactive black perovskite α-phase films was achieved via MACl addition. The XRD peak at 2θ ≈ 36° corresponding to FTO substrate diffraction is considerably reduced in MACl treated films, which confirms the higher surface coverage. To shed light on this hypothesis the samples were analyzed via SEM, and Fig. 4b and c show considerable enhancement for grain shape and film coverage observed with 20 wt% MACl addition. However, the 10 vol% ACN-MACl film presented in Fig. 4b shows less homogeneity with more defects and pinholes in comparison to 7.5 vol% ACN (Fig. 4c). Therefore, this comparison clearly illustrates a notable enhancement in film crystallinity and growth associated with the addition of MACl. The grain size distribution for the optimized 7.5 vol% ACN and 20 wt% MACl film (Fig. 4c inset) ranges from 0.3 cm to 1.3 μm, with an average size of 0.8 μm. The grain size significantly influences the optoelectronic properties of polycrystalline lead halide perovskite films.⁵³ Accordingly, UV-Vis spectra and steady-state photoluminescence (PL) spectra are presented in Fig. 4d. The absorption spectra have a clear absorption edge at 767 nm corresponding to the calculated bandgap of ∼1.61 eV, emphasizing the formation of a photoactive perovskite phase. The photoluminescence spectra also show a single and well-defined emission peak, showing high quality thin films with increased photoelectric properties.⁵⁴ The presence of a single emission peak near the absorption edge suggests the exclusive presence of perovskite in a single crystallographic phase, consistently observed in our XRD results.^55,56

Control of the nucleation and growth of perovskite solar cells

Apart from altering the lead and organic precursors, the inclusion of certain ligand compounds, which act as electron donors (Lewis bases), can significantly influence the properties of perovskite films.⁵⁷ Thiourea (SC(NH₂)₂) as an organic compound has emerged as a promising candidate for enhancing grain size and improving film coverage when incorporated as an additive in a mixed halide perovskite solution. Incorporating thiourea (SC(NH₂)₂) as a Lewis base with a sulfur donor into the precursor solution is reported to control the perovskite phase evolution as well as nucleation and subsequent crystal growth processes.⁵⁸ The strong coordinating behavior of thiourea alters the solution and crystallization dynamics by increasing the Gibbs free energy, thereby reducing the number of nuclei formed.⁵⁹ Consequently, perovskite crystal growth initiates from fewer nucleation sites, leading to the development of larger grain sizes characterized by reduced grain boundary density.^60,61 Thiourea molecules containing CS and NH₂ functional groups exhibit dual roles: they coordinate with undercoordinated Pb²⁺ ions through Lewis acid–base interactions to mitigate defects and simultaneously establish hydrogen bonds with FA organic cations by forming the intermediate phase FAI·PbI₂·SC(NH₂)₂, thereby enhancing the stability of the perovskite framework.^62,63

The use of sequential ethyl acetate (EA) antisolvent processing eliminated the remaining thiourea and enlarged the grain size.⁶³ Therefore, 20 wt% thiourea was added to the perovskite precursor, referred to as the “final green solvent” and compared with a reference DMF:DMSO solution.

Final solution evaluation

Implementation of a single-step spin coating process requires precise control over physical properties such as density, viscosity, and film wettability of the precursor solution.⁶⁴ Given the colloidal nature of the perovskite ink, dynamic light scattering (DLS) measurement was performed to investigate the size distribution of the particles in precursor solutions.⁶⁵ The DLS results in Fig. 5a show the particle size distribution to be centered between 0.8 and 2 nm for the reference ink. In contrast, the optimized green solution blend displays a broader colloidal size distribution, with particles centered around 2 to 9 nm. The larger average particle size correlates well with the higher viscosity observed in green solvent systems. In addition, the donor strengths of the solvents can influence the critical particle size to stabilize larger nuclei. The perovskite droplet contact angle was measured on FTO/mesoporous-TiO₂ substrates, showing (Fig. 5b) improved wettability for the green solution blend droplet with a contact angle of 8.6°, while the DMF-reference solution droplet has a contact angle of 13.2°.⁶⁶

A comparative evaluation of the rheological characteristics of perovskite inks in DMF:DMSO, acting as the reference solution, and the green solvent-based perovskite precursor (both at 0.8 M concentration) was performed at 25 °C. In high-shear regimes of ∼1–1000 s⁻¹, the viscosity and flow behavior of perovskite precursor inks directly influence film formation, microstructural uniformity, and process scalability. It has been observed in previous reports that perovskite inks can exhibit either Newtonian or shear-thinning (pseudoplastic) behavior, depending on their composition, solvent system, and presence of additives.^64,67 Fig. 5c and d show that both solutions exhibit shear-thinning (pseudoplastic) behavior at low shear rates and transition to Newtonian behavior at higher shear rates.⁶⁸ Notably, the green ink exhibited a nearly constant viscosity (Newtonian behavior) around 10 s⁻¹, suggesting formation of stable perovskite precursor frameworks. This stable rheological profile across a wide shear range is advantageous for scalable, high-throughput coating methods, where predictable viscosity simplifies flow modeling and process control.^64,69 At 1 s⁻¹, the viscosities of the DMF:DMSO (34.0 cP) and green solvent (36.9 cP) inks are comparable, suggesting similar viscosity drag during early wetting stages. However, at 1000 s⁻¹, the DMF-reference ink shear thins to 1.06 cP, while the green solvent maintains a higher viscosity of 6.7 cP, indicating higher resistance to flow under dynamic coating conditions. This is manifested in more pronounced perovskite domains in green solvent processed films, which is expected to improve the charge transport properties due to the decreased grain boundary fraction, although enhancing the surface roughness.

Film coverage and grain size distribution

Considering the thiourea effect, the annealing temperature for thin films was chosen to start at 60 °C and ramped up to 150 °C over 30 minutes.⁷⁰ The low evaporation rate of the solvent helps to maintain the solution concentration close to the critical level for a long period. This results in a low concentration of nuclei, allowing the initially formed nuclei ample time and space to grow into large crystals.⁷¹

The surface of the final green optimized perovskite film was compared with that of the reference DMF:DMSO perovskite film using scanning electron microscopy (Fig. 6a and c). The surface morphologies demonstrate that the addition of 20 wt% thiourea to the green solvent resulted in a fully covered, defect-free perovskite surface. The grain size distribution (Fig. S6†) revealed a significant increase in crystal grain size for the green solvent system, with an average grain size exceeding 1.5 μm, which is more than six times higher than that of the DMF:DMSO film with a 225 nm average grain size. In addition, atomic force microscopy (AFM) analysis was employed to examine the surface morphology and roughness of the perovskite films. The results, presented in Fig. 6b and d, confirmed a progressive increase in grain size for the perovskite samples, consistent with the trends observed in SEM. Additionally, surface roughness increased in the same order, with root mean square (RMS) roughness values of 19.3 nm for the reference and 52.6 nm for the green solvent. This trend is further supported by 3D surface plot images, which clearly illustrate the gradual increase in surface roughness across the green solvent-based perovskite films. Increase of grain size has potential positive effects by reducing charge recombination due to fewer grain boundaries.⁷² However, higher roughness may negatively impact the interface between the perovskite layer and the charge transport layers, potentially reducing charge extraction of the solar cells.

To verify the optoelectronic properties of the optimized perovskite light absorber film, UV-Vis and PL spectrophotometry were conducted and the corresponding spectra are shown in Fig. 7a. The characteristic absorbance onset observed at 782 nm for the green solvent and 790 nm for the reference film. A gradual increase in the absorbance up to 500 nm is clear for both films, while absorbance for the green ink is considerably higher than that of the reference film.⁷³ Besides, normalized PL results in a high emission peak for both samples with a narrower FWHM for the green solvent sample. To disclose the molecular interaction of the triple cation perovskite with green solvents and additives, Fourier transform infrared spectroscopy (FTIR) was performed on thin films and results are shown in Fig. 7b. Both films exhibit identical FTIR spectra, with absorption peaks observed at wavenumbers of 1350 cm⁻¹ and 1477 cm⁻¹. These peaks are attributed to the vibrational modes of organic “–CH₃” groups present in the perovskite films. The corresponding absorption peaks of the amino groups in the perovskite films locate at the range of 1600–1750 cm⁻¹ and 3200–3500 cm⁻¹.⁷⁴ Specifically, the N–H bending mode at 1468 cm⁻¹ (symmetric) is observable, and selective monitoring of FA⁺ and MA⁺ is achieved through the measurement of the absorption associated with the CN stretching mode of FA⁺ (MA⁺) at 1720 cm⁻¹, which corresponds to the symmetric bending of NH₃. Vibrations of N–H stretching of FA/MA indicate a dipole interaction with the components in solutions.^75,76

The peak observed at 1617 cm⁻¹ is attributed to the stretching of the carbonyl (CO) bonds, which results from the effective coordination of DMF and Cy-THF solvent with the Lewis acid Pb²⁺.²³ Additionally, the peak at 1020 cm⁻¹ corresponds to the SO stretching associated with DMSO,⁷⁷ which contributes to the molecule's negative polarity.

The XRD results of the final green solvent thin film, compared with the reference films shown in Fig. 7c, demonstrate a significantly higher quality polycrystalline green perovskite thin film. The film exhibits a cubic phase with a (100) preferred orientation, consistent with previous studies highlighting the positive effects of thiourea addition on film crystallinity.⁷⁸

Solar cell performance

As the final aim of this study, perovskite solar cells were fabricated using the optimized precursor system developed in this work (referred to as the green solvent) and compared with reference cells prepared using a DMF:DMSO solvent (referred to as the reference). The cells were assembled in a mesoscopic n–i–p structure: glass/FTO/mesoporous TiO₂/perovskite/Spiro-OMeTAD/Au, as shown in Fig. 8a. The cross-sectional SEM image illustrates a compact and homogeneous perovskite layer processed with the green solvent, exhibiting an approximate thickness of 500 nm.

The photovoltaic properties are presented in Fig. 8b, where the average performance of 28 cells measured for each category shows that green solvent perovskite cells exhibit behaviour almost identical to the reference sample. On average, the green solvent cells achieve 95% of the performance of the reference cells. However, the green solvent cells exhibit slightly lower averages for both J_SC and the fill factor. In perovskite solar cells, the interfaces between functional layers are critical for efficient carrier extraction and charge transport, as they directly impact device performance. Unwanted losses, such as those caused by interfacial recombination, can often lead to a reduction in J_SC, as noted in previous studies.^79,80 Our AFM results reveal higher surface roughness for the green solvent perovskite films, which likely contributes to the slightly reduced average performance. Increased roughness can create non-ideal interfaces, enhance recombination losses and hinder charge transport. Further evidence of these interfacial effects is shown in Fig. 8c, which illustrates the J–V curve of the champion cells. The green solvent cells display a slightly higher degree of hysteresis, a phenomenon that can be directly linked to the increased interfacial recombination.

Further research on surface modifications strategies could reduce surface roughness while maintaining large grain size, ensuring a well-defined interface between the perovskite and charge transport layers.^81,82 This could significantly enhance both device performance and stability beyond those of DMF-based solar cells.

Green solvent thin film stability

The deposited films were kept under extreme conditions with humidity ranging between 60% and 80%, ambient light, and 85 °C for two weeks. Fig. 9a shows images from the reference and final green solvent samples during this period. The degradation of the reference perovskite film is evident by its color change from black to yellow (characteristic of PbI₂) after six days. In contrast, the green solvent thin film shows less color variation, indicating greater thermal and moisture stability.⁸³

This is further confirmed by XRD analysis on day seven, as shown in Fig. 9b. The reference sample almost completely lost its perovskite-related peaks, with intensified PbI₂ trigonal Pm1 peaks appearing.⁸⁴ Although the green solvent film also underwent some degradation to PbI₂, its perovskite peaks remain fully detectable. To explain this difference first the solution stability can be considered. DMF-based perovskites readily undergo hydrolysis at room temperature when coordinated with a metal center (like Pb²⁺) and break down into formic acid (HCOOH) and dimethylamine ((CH₃)₂NH) (DMA).⁸⁵ DMA is a stronger Lewis base compared to DMF or DMSO, and it forms stable complexes with Lewis acids such as Pb²⁺, Cs⁺, and PbX⁺. Formate ions have stronger bidentate ligand interactions with Cs⁺ and Pb²⁺ than halides and displace halide ions (X⁻) from PbX₂, forming Pb(HCOO)₂, which affects the overall stability of the solution. Thus, the equilibrium of precursor solution complexes is influenced by DMA⁺ and formate (HCOO⁻) and reduces solution stability.⁸⁶ Secondly, the degradation rate depends on grain size; smaller grains increase boundaries, accelerating perovskite degradation in humid environments.⁸⁷ Therefore, green solvent films, with larger grain sizes (hundreds of nm to micrometers), offer greater stability than DMF films. In addition, enhancing crystal quality improves longevity,⁸⁸ and our research confirms that improved crystallinity, especially in the (100) facet, enhances green solvent film stability.

To investigate the impact of the green solvent system on device stability, the solar cells presented in Fig. 8 were remeasured for PCE, after storing them for five months in a nitrogen-filled glovebox. As shown in Fig. 9c–e the highest and the average PCE of the green solvent-based devices remained marginally higher than those of the DMF-based reference, suggesting a modest stability benefit. However, a more pronounced hysteresis was observed in the green solvent devices after aging, indicating possible ion migration or interfacial instability. This may be attributed to the higher perovskite surface roughness and increased solution viscosity, which can lead to the formation of pinholes and voids in perovskite films.⁸⁹ In addition, both devices exhibited a reduction in PCE mostly due to the considerable loss of fill factor, possibly attributed to (i) the diffusion of gold from the top electrode that can penetrate the perovskite or transport layers to introduce recombination sites and increase series resistance (observed here),⁹⁰ and (ii) degradation of the hole transport layer, where Spiro-OMeTAD is known to suffer from dopant loss or oxidation state changes over time, even under inert conditions.⁵⁰

Conclusions

This work demonstrates the potential of green solvent systems in efficient processing of high-performance triple cation perovskite Cs_0.05(MA_0.17FA_0.83)_0.95Pb(I_0.83Br_0.17)₃. Our optimized 70:30 vol% Cy-THF:DMSO solvent mixture produces a fully dissolved precursor solution, while the addition of acetonitrile (7.5 vol%) with a donor number of 14.1 kcal mol⁻¹ and a low viscosity of 0.38 cP effectively reduces the overall viscosity of the solution. Addition of 20 wt% methylammonium chloride (MACl, CH₃NH₃Cl) and thiourea (SC(NH₂)₂) regulates ion release leading to the formation of a fully covered thin film with enhanced grain size. Notably, all solvents and additives were combined into a single solution, and the clear resulting ink is ready after less than one hour of stirring at 50 °C. This represents a significant achievement in simplifying the fabrication process for multi-cation perovskites, enabling the use of a single ink with a green solvent system. Additionally, the use of a lower solution molarity (0.8 M) reduces lead consumption. Finally, the ecological index of our approach was further increased by replacing commonly used anti-solvents with ethyl acetate (EA). Therefore, in this work, we successfully eliminated DMF from the solvent system by using 70% cellulose base solvent blend Cyrene_2-MeTHF and 30% DMSO to fully dissolve the perovskite precursor. These additives facilitated the direct formation of a highly crystalline, photoactive cubic-phase perovskite with excellent optoelectronic properties. Achieving 95% of the performance of the DMF-based reference perovskite with a fully green solvent and antisolvent triple-cation solar cell is a significant highlight of this work. This demonstrates the potential of green solvent approaches to deliver high-efficiency perovskite solar cells while promoting more sustainable and environmentally friendly fabrication processes.

Data availability

The data supporting this article have been included as part of the ESI.†

Author contributions

Niusha Heshmati: conceptualization, data curation, formal analysis, investigation, methodology, validation, visualization, writing – original draft, writing – review & editing. Niklas Almandinger: investigation, formal analysis, visualization. Thomas Fischer: methodology. Sanjay Mathur: funding acquisition, project administration, resources, supervision, writing – review & editing.

Conflicts of interest

The authors declare no conflicts of interest.

Acknowledgements

S. Mathur acknowledges the University of Cologne for financial and infrastructural support. The project funding received in the framework of DFG priority program “Perovskite Semiconductors: From Fundamental Properties to Devices” (SPP 2196)” is gratefully acknowledged. Special thanks are due to Dr Seeni Meera Kamal Mohamed (DLR, Cologne, Germany) for rheology measurement, Dr Ionela Lindfors-Vrejoiu (Physics department, University of Cologne) for conducting the AFM measurements, Tobias Lapić for thin-film photoluminescence measurements, and Ziyaad Aytuna for SEM measurements. We also sincerely appreciate Merck KGaA, Darmstadt, Germany for providing the green solvent samples.

Notes and references

1. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc., 2009, 131, 6050 CrossRef CAS PubMed .
2. F. Ünlü, E. Jung, J. Haddad, A. Kulkarni, S. Öz, H. Choi, T. Fischer, S. Chakraborty, T. Kirchartz and S. Mathur, APL Mater., 2020, 8(7), 070901 CrossRef .
3. D. Zhou, T. Zhou, Y. Tian, X. Zhu and Y. Tu, J. Nanomater., 2018, 2018, 1 CrossRef .
4. M. Cherasse, N. Heshmati, J. M. Urban, F. Ünlü, M. S. Spencer, M. Frenzel, L. Perfetti, S. Mathur and S. F. Maehrlein, Small, 2025, 21, 2500977 CrossRef CAS PubMed .
5. L. Schmidt-Mende, V. Dyakonov, S. Olthof, F. Ünlü, K. M. T. Lê, S. Mathur, A. D. Karabanov, D. C. Lupascu, L. M. Herz, A. Hinderhofer, F. Schreiber, A. Chernikov, D. A. Egger, O. Shargaieva, C. Cocchi, E. Unger, M. Saliba, M. M. Byranvand, M. Kroll, F. Nehm, K. Leo, A. Redinger, J. Höcker, T. Kirchartz, J. Warby, E. Gutierrez-Partida, D. Neher, M. Stolterfoht, U. Würfel, M. Unmüssig, J. Herterich, C. Baretzky, J. Mohanraj, M. Thelakkat, C. Maheu, W. Jaegermann, T. Mayer, J. Rieger, T. Fauster, D. Niesner, F. Yang, S. Albrecht, T. Riedl, A. Fakharuddin, M. Vasilopoulou, Y. Vaynzof, D. Moia, J. Maier, M. Franckevičius, V. Gulbinas, R. A. Kerner, L. Zhao, B. P. Rand, N. Glück, T. Bein, F. Matteocci, L. A. Castriotta, A. Di Carlo, M. Scheffler and C. Draxl, APL Mater., 2021, 9(10), 109202 CrossRef CAS .
6. S. K. Podapangi, F. Jafarzadeh, S. Mattiello, T. B. Korukonda, A. Singh, L. Beverina and T. M. Brown, Green Solvents, Materials, and Lead-free Semiconductors for Sustainable Fabrication of Perovskite Solar Cells, Royal Society of Chemistry, 2023, vol. 13, pp. 18165–18206 Search PubMed .
7. M. Ren, X. Qian, Y. Chen, T. Wang and Y. Zhao, J. Hazard. Mater., 2022, 426, 127848 CrossRef CAS PubMed .
8. Y. Jiang, L. Qiu, E. J. Juarez-Perez, L. K. Ono, Z. Hu, Z. Liu, Z. Wu, L. Meng, Q. Wang and Y. Qi, Nat. Energy, 2019, 4, 585 CrossRef CAS .
9. X. Wu, D. Zhang, X. Wang, X. Jiang, B. Liu, B. Li, Z. Li, D. Gao, C. Zhang, Y. Wang and Z. Zhu, Eco-friendly Perovskite Solar Cells: from Materials Design to Device Processing and Recycling, John Wiley and Sons Inc, 2023, Vol. 5 Search PubMed .
10. N.-G. Park, Nat. Sustain., 2020, 4, 192 CrossRef .
11. H. Hu, M. Singh, X. Wan, J. Tang, C.-W. Chu and G. Li, J. Mater. Chem. A, 2020, 8, 1578 RSC .
12. F. U. Kosasih, E. Erdenebileg, N. Mathews, S. G. Mhaisalkar and A. Bruno, Joule, 2022, 6, 2692 CrossRef CAS .
13. J. Han, R. H. Kim, S. Huang, J. Kim and J. S. Yun, Sol. RRL, 2024, 8, 2400262 CrossRef CAS .
14. H.-S. Kim, Y.-J. An, J. Il Kwak, H. J. Kim, H. S. Jung and N.-G. Park, ACS Energy Lett., 2022, 7, 1154 CrossRef CAS .
15. J. Jiao, C. Yang, Z. Wang, C. Yan and C. Fang, Results Eng., 2023, 18, 101158 CrossRef CAS .
16. H. J. Kim, Y. J. Kim, G. S. Han and H. S. Jung, Sol. RRL, 2024, 8(5), 2300910 CrossRef CAS .
17. A. J. Doolin, R. G. Charles, C. S. P. De Castro, R. G. Rodriguez, E. V. Péan, R. Patidar, T. Dunlop, C. Charbonneau, T. Watson and M. L. Davies, Green Chem., 2021, 23, 2471 RSC .
18. E. Radicchi, E. Mosconi, F. Elisei, F. Nunzi and F. De Angelis, ACS Appl. Energy Mater., 2019, 2, 3400 CrossRef CAS .
19. F. Ünlü, Green inks and Lead-free Compositions in Perovskite and Allied Materials for Photovoltaic Applications, Doctoral thesis, Dr. Hut, 2022 Search PubMed .
20. V. Gutmann, Coord. Chem. Rev., 1976, 18, 225 CrossRef CAS .
21. R. Stenutz, Gutmann acceptor and donor number, Stenutz, 2023, http://www.stenutz.eu/chem/solv21.php.
22. J. C. Hamill, J. Schwartz and Y.-L. Loo, ACS Energy Lett., 2018, 3, 92 CrossRef CAS .
23. H.-S. Yun, H. W. Kwon, M. J. Paik, S. Hong, J. Kim, E. Noh, J. Park, Y. Lee and S. Il Seok, Nat. Energy, 2022, 7, 828 CrossRef CAS .
24. N. K. Noel, S. N. Habisreutinger, B. Wenger, M. T. Klug, M. T. Hörantner, M. B. Johnston, R. J. Nicholas, D. T. Moore and H. J. Snaith, Energy Environ. Sci., 2017, 10, 145 RSC .
25. D. Zheng, F. Raffin, P. Volovitch and T. Pauporté, Nat. Commun., 2022, 13, 6655 CrossRef CAS PubMed .
26. L. Chao, T. Niu, W. Gao, C. Ran, L. Song, Y. Chen and W. Huang, Adv. Mater., 2021, 33(14), 2005410 CrossRef CAS PubMed .
27. X. Cao, L. Zhi, Y. Jia, Y. Li, K. Zhao, X. Cui, L. Ci, D. Zhuang and J. Wei, ACS Appl. Mater. Interfaces, 2019, 11, 7639 CrossRef CAS PubMed .
28. D. Kong and A. V. Dolzhenko, Sustain. Chem. Pharm., 2022, 25, 100591 CrossRef CAS .
29. C. Sullivan, Y. Zhang, G. Xu, L. Christianson, F. Luengo, T. Halkoski and P. Gao, Green Chem., 2022, 24, 7184 RSC .
30. Y. Zhong, Z. Liu, X. Luo, G. Liu, X. Wang, J. He, W. Sheng, D. Yu, C. Liang, L. Tan and Y. Chen, Energy Environ. Sci., 2024, 17, 5500 RSC .
31. G. S. Shin, J. Kim, S. Lee and N. Park, Bull. Korean Chem. Soc., 2021, 42, 1112 CrossRef CAS .
32. L. Xu, S. Che, J. Huang, D. Xie, Y. Yao, P. Wang, P. Lin, H. Piao, H. Hu, C. Cui, F. Wu, D. Yang and X. Yu, Appl. Phys. Lett., 2019, 115(3), 033101 CrossRef .
33. Q. Chen, J. C.-R. Ke, D. Wang, M. Z. Mokhtar, A. G. Thomas and Z. Liu, Appl. Surf. Sci., 2021, 536, 147949 CrossRef CAS .
34. Y. Rong, Z. Tang, Y. Zhao, X. Zhong, S. Venkatesan, H. Graham, M. Patton, Y. Jing, A. M. Guloy and Y. Yao, Nanoscale, 2015, 7, 10595 RSC .
35. A. A. Petrov, A. A. Ordinartsev, S. A. Fateev, E. A. Goodilin and A. B. Tarasov, Molecules, 2021, 26(24), 7541 CrossRef CAS PubMed .
36. N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu and S. Il Seok, Nat. Mater., 2014, 13, 897 CrossRef CAS PubMed .
37. N. Jaiswal, V. P. Singh, D. Kumari and S. K. Pandey, Energy Technol., 2024, 12(2), 2300639 CrossRef CAS .
38. D. A. Chalkias, A. Mourtzikou, G. Katsagounos, A. N. Kalarakis and E. Stathatos, Small Methods, 2023, 7(10), 2300664 CrossRef CAS PubMed .
39. J. Küffner, J. Hanisch, T. Wahl, J. Zillner, E. Ahlswede and M. Powalla, ACS Appl. Energy Mater., 2021, 4, 11700 CrossRef .
40. S. Lin, S. Wu, D. Guo, H. Huang, X. Zhou, D. Zhang, K. Zhou, W. Zhang, Y. Hu, Y. Gao and C. Zhou, Small Methods, 2023, 7(4), 2201663 CrossRef CAS PubMed .
41. Z. Ahmad, R. A. Scheidt, M. P. Hautzinger, K. Zhu, M. C. Beard and G. Galli, ACS Energy Lett., 2022, 7, 1912 CrossRef CAS .
42. F. Ren, H. Xiang, K. Zhao and C. Liu, J. Mater. Chem. C, 2023, 11, 13281 RSC .
43. S. M. Jain, B. Philippe, E. M. J. Johansson, B. Park, H. Rensmo, T. Edvinsson and G. Boschloo, J. Mater. Chem. A, 2016, 4, 2630 RSC .
44. H. Li, Y. Xia, C. Wang, G. Wang, Y. Chen, L. Guo, D. Luo and S. Wen, ACS Appl. Mater. Interfaces, 2019, 11, 34989 CrossRef CAS PubMed .
45. Y. Yang, J. Wu, T. Wu, Z. Xu, X. Liu, Q. Guo and X. He, J. Colloid Interface Sci., 2018, 531, 602 CrossRef CAS PubMed .
46. F. Zhao, J. Zhong, L. Zhang, P. Yong, J. Lu, M. Xu, Y. Cheng and Z. Ku, Sol. RRL, 2023, 7(11), 2300062 CrossRef CAS .
47. W. Xiang, J. Zhang, S. Liu, S. Albrecht, A. Hagfeldt and Z. Wang, Joule, 2022, 6, 315 CrossRef CAS .
48. C. Mu, J. Pan, S. Feng, Q. Li and D. Xu, Adv. Energy Mater., 2017, 7(6), 1601297 CrossRef .
49. M. Kim, G. H. Kim, T. K. Lee, I. W. Choi, H. W. Choi, Y. Jo, Y. J. Yoon, J. W. Kim, J. Lee, D. Huh, H. Lee, S. K. Kwak, J. Y. Kim and D. S. Kim, Joule, 2019, 3, 2179 CrossRef CAS .
50. M. Saliba, T. Matsui, J. Y. Seo, K. Domanski, J. P. Correa-Baena, M. K. Nazeeruddin, S. M. Zakeeruddin, W. Tress, A. Abate, A. Hagfeldt and M. Grätzel, Energy Environ. Sci., 2016, 9, 1989 RSC .
51. B. J. Kim, H. Choi, S. Park, M. B. Johansson, G. Boschloo and M. Kim, ACS Sustain. Chem. Eng., 2024, 12, 13371 CrossRef CAS .
52. G. Tumen-Ulzii, C. Qin, D. Klotz, M. R. Leyden, P. Wang, M. Auffray, T. Fujihara, T. Matsushima, J. Lee, S. Lee, Y. Yang and C. Adachi, Adv. Mater., 2020, 32, 1905035 CrossRef CAS PubMed .
53. C. Stavrakas, S. J. Zelewski, K. Frohna, E. P. Booker, K. Galkowski, K. Ji, E. Ruggeri, S. Mackowski, R. Kudrawiec, P. Plochocka and S. D. Stranks, Adv. Energy Mater., 2019, 9, 1901883 CrossRef .
54. T. P. A. van der Pol, K. Datta, M. M. Wienk and R. A. J. Janssen, Adv. Opt. Mater., 2022, 10, 2102557 CrossRef CAS .
55. N. Heshmati, M. R. Mohammadi, P. Abachi and S. O. Martinez-Chapa, New J. Chem., 2021, 45, 788 RSC .
56. S. Jeong, H. C. Yoon, N. S. Han, J. H. Oh, S. M. Park, B. K. Min, Y. R. Do and J. K. Song, J. Phys. Chem. C, 2017, 121, 3149 CrossRef CAS .
57. S. Wang, Z. Ma, B. Liu, W. Wu, Y. Zhu, R. Ma and C. Wang, Sol. RRL, 2018, 2, 1800034 CrossRef .
58. Y. Li, Y. Duan, J. Feng, Y. Sun, K. Wang, H. Li, H. Wang, Z. Zang, H. Zhou, D. Xu, M. Wu, Y. Li, Z. Xie, Z. Liu, J. Huang, Y. Yao, Q. Peng, Q. Fan, N. Yuan, J. Ding, S. Liu and Z. Liu, Angew. Chem., Int. Ed., 2024, 63, 202410378 CrossRef PubMed .
59. K. Lê, N. Heshmati and S. Mathur, Nano Convergence, 2023, 10, 47 CrossRef PubMed .
60. J. V. Patil, S. S. Mali and C. K. Hong, Nanoscale, 2019, 11, 21824 RSC .
61. X. Yang, H. Xiang, J. Huang, C. Zhou, R. Ran, W. Wang, W. Zhou and Z. Shao, J. Colloid Interface Sci., 2022, 628, 476 CrossRef CAS PubMed .
62. Q. Sun, B. Tuo, Z. Ren, T. Xue, Y. Zhang, J. Ma, P. Li and Y. Song, Adv. Funct. Mater., 2022, 32(51), 220885 CrossRef .
63. C. Fei, B. Li, R. Zhang, H. Fu, J. Tian and G. Cao, Adv. Energy Mater., 2017, 7(9), 1602017 CrossRef .
64. M.-R. Ahmadian-Yazdi, A. Rahimzadeh, Z. Chouqi, Y. Miao and M. Eslamian, AIP Adv., 2018, 8(2), 025109 CrossRef .
65. P. Boonmongkolras, D. Kim, E. M. Alhabshi, I. Gereige and B. Shin, RSC Adv., 2018, 8, 21551 RSC .
66. Y. Miao, M. Ren, Y. Chen, H. Wang, H. Chen, X. Liu, T. Wang and Y. Zhao, Nat. Sustain., 2023, 6, 1465 CrossRef .
67. F. Bisconti, A. Giuri, R. Suhonen, T. M. Kraft, M. Ylikunnari, V. Holappa, R. Po’, P. Biagini, A. Savoini, G. Marra, S. Colella and A. Rizzo, Cell Rep. Phys. Sci., 2021, 2, 100639 CrossRef CAS .
68. F. Bisconti, A. Giuri, N. Vanni, S. Carallo, S. Spera, R. Marrazzo, R. Po’, P. Biagini, B. Paci, A. Generosi, M. Guaragno, C. Esposito Corcione, A. Listorti, S. Colella and A. Rizzo, Nanoscale Adv., 2025, 7, 2145 RSC .
69. Y. Guo, H. S. Patanwala, B. Bognet and A. W. K. Ma, Rapid Prototyp. J., 2017, 23, 562 CrossRef .
70. C.-M. Hsieh, Y.-S. Liao, Y.-R. Lin, C.-P. Chen, C.-M. Tsai, E. Wei-Guang Diau and S.-C. Chuang, RSC Adv., 2018, 8, 19610 RSC .
71. B. Ding, Y. Li, S.-Y. Huang, Q.-Q. Chu, C.-X. Li, C.-J. Li and G.-J. Yang, J. Mater. Chem. A, 2017, 5, 6840 RSC .
72. S. Öz, A. K. Jena, A. Kulkarni, K. Mouri, T. Yokoyama, I. Takei, F. Ünlü, S. Mathur and T. Miyasaka, ACS Energy Lett., 2020, 5, 1292 CrossRef .
73. A. Farooq, M. R. Khan, T. Abzieher, A. Voigt, D. C. Lupascu, U. Lemmer, B. S. Richards and U. W. Paetzold, ACS Appl. Energy Mater., 2021, 4, 3083 CrossRef CAS .
74. Z. Liu, D. Liu, H. Chen, L. Ji, H. Zheng, Y. Gu, F. Wang, Z. Chen and S. Li, Nanoscale Res. Lett., 2019, 14, 304 CrossRef PubMed .
75. J. M. C. da Silva Filho, V. A. Ermakov and F. C. Marques, Sci. Rep., 2018, 8, 1563 CrossRef .
76. M. Vásquez-Montoya, J. F. Montoya, D. Ramirez and F. Jaramillo, J. Energy Chem., 2021, 57, 386 CrossRef .
77. S. Bae, S. J. Han, T. J. Shin and W. H. Jo, J. Mater. Chem. A, 2015, 3, 23964 RSC .
78. L. Zhu, Y. Xu, P. Zhang, J. Shi, Y. Zhao, H. Zhang, J. Wu, Y. Luo, D. Li and Q. Meng, J. Mater. Chem. A, 2017, 5, 20874 RSC .
79. J. Herterich, C. Baretzky, M. Unmüssig, C. Maheu, N. Glissmann, J. Gutekunst, G. Loukeris, T. Mayer, M. Kohlstädt, J. P. Hofmann and U. Würfel, Sol. RRL, 2022, 6(7), 2200195 CrossRef CAS .
80. S. Zhang, F. Ren, Z. Sun, X. Liu, Z. Tan, W. Liu, R. Chen, Z. Liu and W. Chen, Small Methods, 2024, 8(7), 2301223 CrossRef CAS PubMed .
81. T. Singh, S. Öz, A. Sasinska, R. Frohnhoven, S. Mathur and T. Miyasaka, Adv. Funct. Mater., 2018, 28(14), 1706287 CrossRef .
82. J. Haddad, B. Krogmeier, B. Klingebiel, L. Krückemeier, S. Melhem, Z. Liu, J. Hüpkes, S. Mathur and T. Kirchartz, Adv. Mater. Interfaces, 2020, 7, 2000366 CrossRef CAS .
83. B. Chaudhary, T. M. Koh, B. Febriansyah, A. Bruno, N. Mathews, S. G. Mhaisalkar and C. Soci, Sci. Rep., 2020, 10, 429 CrossRef CAS PubMed .
84. K. Dhivyaprasath and M. Ashok, Sol. Energy, 2023, 255, 89 CrossRef CAS .
85. M. E. O'Kane, J. A. Smith, T. I. Alanazi, E. J. Cassella, O. Game, S. van Meurs and D. G. Lidzey, ChemSusChem, 2021, 14, 2537 CrossRef .
86. B. Dou, L. M. Wheeler, J. A. Christians, D. T. Moore, S. P. Harvey, J. J. Berry, F. S. Barnes, S. E. Shaheen and M. F. A. M. van Hest, ACS Energy Lett., 2018, 3, 979 CrossRef CAS .
87. N. Li, S. Pratap, V. Körstgens, S. Vema, L. Song, S. Liang, A. Davydok, C. Krywka and P. Müller-Buschbaum, Nat. Commun., 2022, 13, 6701 CrossRef CAS PubMed .
88. D. P. McMeekin, P. Holzhey, S. O. Fürer, S. P. Harvey, L. T. Schelhas, J. M. Ball, S. Mahesh, S. Seo, N. Hawkins, J. Lu, M. B. Johnston, J. J. Berry, U. Bach and H. J. Snaith, Nat. Mater., 2023, 22, 73 CrossRef CAS .
89. Y. Yuan and J. Huang, Acc. Chem. Res., 2016, 49, 286 CrossRef CAS PubMed .
90. W. Tress, M. Yavari, K. Domanski, P. Yadav, B. Niesen, J. P. Correa Baena, A. Hagfeldt and M. Graetzel, Energy Environ. Sci., 2018, 11, 151 RSC .

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta01277e

This journal is © The Royal Society of Chemistry 2025