Reactive Anti-Solvent Engineering via Kornblum Reaction for Controlled Crystallization in (FA_0.83MA_0.17Cs_0.05)Pb(I_0.85Br_0.15)₃ Perovskite Solar Cells

Abstract

Regulating the crystallization dynamics of perovskite films is key to improving the efficiency and operational stability of (FA_0.83MA_0.17Cs_0.05)Pb(I_0.85Br_0.15)₃ perovskite solar cells (PSCs). However, precise regulation of the crystallization process remains challenging. Here, we introduce a reactive anti-solvent strategy based on the Kornblum reaction to modulate crystallization via in-situ chemical transformation. Specifically, trans-cinnamoyl chloride (TCC) is employed as a single-component anti-solvent additive that reacts with dimethyl sulfoxide (DMSO) in the perovskite precursor solution. The resulting acylation reaction generates carbonyl-containing products and sulfur ions. The carbonyl oxygen coordinates with Pb²⁺ ions to form Pb–O bonds, which retard rapid crystallization, suppress heterogeneous nucleation, and facilitate the growth of larger perovskite grains with improved film uniformity. Additionally, the exothermic nature of the reaction accelerates local supersaturation and nucleation. This synergistic crystallization control significantly enhances the film morphology and device performance, yielding a champion power conversion efficiency (PCE) of 23.02% and a markedly improved fill factor (FF). This work provides a new pathway for anti-solvent engineering through in-situ chemical regulation, enabling efficient and scalable fabrication of high-performance PSCs.

1. Introduction

The organic–inorganic hybrid perovskite, as an extremely high-performance material, shows remarkable and enormous potential. Perovskite materials have a high absorption coefficient [1] and a long carrier diffusion length [2], which together enable outstanding power conversion efficiencies (PCEs) [3]. Among the various fabrication approaches, solution-processed methods dominate due to their low cost, scalability, and compatibility with large-area production [4]. Nevertheless, achieving uniform, compact, and defect-minimized perovskite films remains a major challenge. A critical factor in this process is the regulation of perovskite crystallization, which is highly sensitive to solvent–antisolvent interactions and additive engineering.

In solution-based fabrication, the careful selection and timing of anti-solvent application are essential to ensure controlled nucleation and high film quality [5]. Over the past decade, researchers have made significant strides in improving film morphology by tuning the composition of perovskite precursors, solvents, anti-solvents, and additives [6,7]. Among these strategies, the use of additives in either the precursor or anti-solvent phase has proven particularly effective in modulating crystallization kinetics and passivating defects.

For instance, Cong et al. introduced 4-morpholine formamidine hydrochloride (MFC) into the anti-solvent phase, where strong hydrogen bonding promoted crystal growth and reduced deep-level defects, resulting in larger grains and improved carrier transport. [8] Similarly, the multifunctional additive o-TB-GDY, featuring high π-conjugation, was shown to enhance crystallization and surface passivation by serving as nucleation seeds and interacting with undercoordinated Pb(0) sites [9]. Wu et al. utilized 1-(2-methoxyphenyl)piperazine hydrochloride (2MPCl) to passivate Pb²⁺ and halide vacancies via ionic and hydrogen bonding while also tuning the energy-level alignment between perovskite and PCBM for improved charge extraction and stability [10]. Additionally, Li et al. proposed an in-situ functional group conversion strategy using Bis-PEG4-NHS ester, which hydrolyzed during annealing to generate carboxylic acids that enhanced interactions with PbI₂ and FAI, thereby reinforcing structural integrity and defect passivation [11]. These efforts underscore the importance of anti-solvent engineering in guiding the crystallization process and improving film quality. However, a more controllable and reaction-driven approach remains desirable.

In this work, we report a novel anti-solvent additive strategy based on the Kornblum reaction, wherein trans-cinnamoyl chloride (TCC) reacts with DMSO in the precursor solution during the annealing step. This acylation reaction produces carbonyl compounds and sulfur ions, which coordinate with Pb²⁺ and modulate the nucleation kinetics. The exothermic nature of the reaction further accelerates local supersaturation, promoting homogeneous and rapid nucleation. The combined effects lead to the formation of dense, large-grained perovskite films with improved morphology and reduced defects. Devices fabricated using this strategy exhibit significant improvements in both short-circuit current and fill factor, achieving a PCE of 23.02%. Our findings demonstrate a powerful and scalable approach for in-situ chemical regulation of perovskite crystallization, advancing the performance and manufacturability of PSCs.

2. Results

2.1. The Selection of Chemical Reactions

In this work, TCC was chosen to react with DMSO. Studies have shown that the acyl chloride group can react with DMSO. Previous research has indicated that the rate of this reaction is extremely fast [12]. After adding TCC to DMSO, a large amount of heat is rapidly released. It is envisioned that when TCC is added to chlorobenzene as an anti-solvent and then annealed at 105 °C for 30 min, TCC will react rapidly upon contact with DMSO in the perovskite solution, and then the Pummerer rearrangement occurs [13], releasing a large amount of heat. This accelerates the rapid nucleation of perovskite, preventing excessive contact with oxygen and water during the annealing process [14]. Meanwhile, the carbonyl group slows down the nucleation rate of perovskite, which is beneficial for improving the quality of the perovskite film [15]. The operation process is shown in Figure 1.

2.2. The Influence of TCC on the Morphology of Perovskite

To investigate the influence of TCC on the film quality of (FA_0.83MA_0.17)Pb(I_0.85Br_0.15)₃ perovskite, the XRD pattern was used to characterize the films treated with different concentrations (0, 5, 10, 20, 30 μL/mL). As shown in Figure 2a, all samples exhibited obvious perovskite diffraction peaks of the (110) plane located at 14.1°, with an intensity of 6188 [16]. As shown in Figure 2b, the (110) plane intensity increased to 6866 when the TCC concentration was 10 μL/mL and then decreased rapidly to 5898 when the concentration was increased to 30 μL/mL. This indicated that the introduction of TCC changed the orientation of the film, and the perovskite films with 10 μL/mL TCC showed the preferred orientation at the (110) plane. In addition, the full width at half maximum (FWHM) of the (110) plane was also recorded. It was clear that the films with 10 μL/mL TCC had the lowest FWHM among those films, which illustrated the enhanced crystalline quality of the perovskite micro-crystals and corresponded to the planar morphology changes discussed later. In combination with solvent chemical analysis, it is speculated that TCC may react with DMSO in the perovskite precursor solution, thereby affecting the formation and growth of crystal nuclei and influencing the crystallization quality of the film.

To further investigate the effect of TCC on the morphology of perovskite films, SEM was employed to characterize the surface morphology of films modified with different TCC concentrations (0, 5, 10, 20, 30 μL/mL) (Figure 3). At a magnification of 50,000, the SEM images clearly revealed the grain distribution of the films, and the grain size was quantitatively analyzed using Nano Measurer 1.2. For the control group (0 μL/mL), the average particle size and standard deviation were 0.66 ± 0.24 μm. When treated with 5 μL/mL TCC, the average particle size and standard deviation reached 0.63 ± 0.18 μm. After treatment with 10 μL/mL TCC, the average particle size and standard deviation reached 0.70 ± 0.25 μm, indicating an increase in the average particle size. As the TCC concentration further increased, the perovskite grain size began to decrease. When the TCC concentration was 20 μL/mL, the average particle size and standard deviation were 0.65 ± 0.26 mm. At a TCC concentration of 30 μL/mL, the average particle size and standard deviation were also 0.59 ± 0.19 μm. The particle size distribution diagram of the device is shown in Figure S1. This trend was consistent with the XRD results. Combined with XRD analysis, it is hypothesized that at lower concentrations, TCC may interact with DMSO in the perovskite precursor, accelerating nucleation and delaying the crystallization process of perovskite, thereby leading to a smoother perovskite surface and ultimately influencing the film’s microstructure and optoelectronic properties. Figure S2a,b show cross-sectional images of the perovskite films, indicating that the thickness of the perovskite layer did not change significantly after TCC treatment. AFM and contact angle measurements were conducted to systematically characterize the changes in the perovskite films before and after treatment. The AFM results (Figure S3) revealed that the untreated control film exhibited a relatively smooth surface with a root mean square (RMS) roughness of 26.62 nm. In contrast, after 10 μL/mL TCC treatment, the RMS roughness decreased to 24.32 nm. This observation is consistent with the SEM results: the introduction of TCC regulated the perovskite crystallization process, resulting in a denser and more uniform grain size, which led to a smoother surface morphology. Such structural modifications may improve the interfacial contact between the perovskite layer and charge transport layers, thereby enhancing the charge extraction efficiency in the device [17].

2.3. Effect of TCC on the Photovoltaic Performance of PSCs

To investigate the effect of TCC treatment on the electrical performance of PSCs, devices with the structure FTO/SnO₂/PVK/Spiro-OMeTAD/Ag were fabricated. Different concentrations of TCC solution (10 μL/mL, 20 μL/mL, 30 μL/mL) were incorporated into chlorobenzene as an antisolvent, and all devices were prepared under the same conditions. Figure 4a shows the J-V curves of each group of devices measured under standard AM 1.5 G illumination (100 mW/cm²) using a solar simulator. Table S1 lists the performance parameters of the control device and those treated with different concentrations (5 μL/mL, 10 μL/mL, 20 μL/mL, 30 μL/mL) of TCC solution. From the table, the control device exhibited a PCE of 22.53%, an open-circuit voltage (V_OC) of 1.12 V, an FF of 80.94%, and a short-circuit current density (J_SC) of 24.83 mA/cm². After treatment with 10 μL/mL TCC antisolvent, the device showed an improved FF of 82.38% and a J_SC of 25.22 mA/cm², indicating enhanced charge carrier extraction efficiency. However, the V_OC decreased to 1.11 V, while the overall PCE increased to 23.04%. When the TCC concentration was further increased to 20 μL/mL, the device exhibited a J_SC of 25.12 mA/cm², an FF of 83.15%, and a further decreased V_OC of 1.09 V, resulting in a maximum PCE of 22.75%. Upon further increasing the concentration, both the FF and V_OC continued to decline, as shown in Table S1. The J_SC, V_OC, and FF of these devices showed negligible differences (see Supporting Information, Figure S4). The improvement in the FF and J_SC may stem from optimized film crystallinity, effectively enhancing charge transport. The decrease in the V_OC can be attributed to the exothermic reaction between acyl—chloride and DMSO. When TCC comes into contact with DMSO, a large amount of heat is released, which leads to an excessively high annealing temperature. Previous studies have clearly indicated that an overly high annealing temperature can cause a decline in the open-circuit voltage of the device [18].

Additionally, Mott–Schottky (M–S) analysis, transient photocurrent (TPC) decay, and transient photovoltage (TPV) decay measurements were conducted on the devices. The results are shown in Figure 4b. The built-in electric field of the standard device was approximately 0.88 V, while that of the TCC-treated device was around 0.86 V, indicating a negligible change in the V_OC [19]. The fitting formulas for TPC and TPV are as follows.

$$\mathrm{TPC}/TPV={\displaystyle \frac{A_1{r}_1^2+A_2{r}_2^2}{A_1{r}_1+A_2{r}_2}}$$

After fitting the TPC data, the photocurrent decay response time of the standard device was 3.87 ms, while that of the TCC-treated device was 4.11 ms, showing no significant difference. However, fitting the TPV decay curves revealed that the decay time of the standard device was 4.2 ms, whereas that of the treated device was 3.0 ms (Figure 4c,d). The significantly shorter photovoltage decay time suggests reduced carrier recombination at the interface of the treated device, which is beneficial for improving device efficiency [20].

To further investigate the passivation effect of TCC on defects in the perovskite layer of PSCs, the dependence of the device’s J_SC and V_OC on light intensity (I) was tested. In Figure 4e, the slope of the linear relationship between the V_OC and Ln(I) for the control PSC was 1.42, while that of the TCC-treated device was 1.39, indicating a slight reduction in non-radiative recombination in the treated device. Meanwhile, the dark current curves in Figure 4f further confirmed the reduction in non-radiative recombination [21].

2.4. Mechanism of Action of TCC on Perovskite Films

For the device treated with the anti-solvent doped with TCC, the color of the perovskite precursor wet film changes to brown more rapidly, which reflects an accelerated nucleation process. This is beneficial for achieving more uniform crystallization of the perovskite film and improving the film quality. To further investigate the influence of TCC on perovskite nucleation, SEM was used to observe the annealing process of the perovskite, as shown in Figure 5. For the control-group device, the average sizes of the perovskite grains after annealing for 5 and 10 min are approximately 0.62 ±0.16 and 0.68 ± 0.18 μm, respectively. In contrast, for the perovskite treated with the anti-solvent doped with TCC, the grain sizes are about 0.64 ± 0.17 and 0.70 ± 0.16 μm. The influence on the perovskite grain size during the crystallization process is not significant. From Figure 5a,d, it can be observed that there are many small grains in the perovskite treated with the TCC-doped anti-solvent before annealing. It is speculated that TCC reacts rapidly with DMSO to generate reaction products containing carbonyl groups, and a large amount of heat is released during the reaction. The remaining DMF and DMSO are further evaporated, which promotes the precipitation of perovskite grains. During the subsequent growth process, the oxygen atom in the carbonyl group, with a partial negative charge due to its high electronegativity, can form coordinate bonds with metal cations (such as Pb²⁺) in the perovskite through electrostatic interaction. This interaction can stabilize the chemical environment around the metal cations, slow down the nucleation rate, make the formed crystal nuclei more uniform, and is beneficial for the growth of high-quality perovskite crystals.

Regarding the effect of TCC treatment on the optical properties of perovskite films, UV-Vis tests were conducted on both the TCC-treated films and the control group films, with the results shown in Figure 6a. Both the treated and untreated perovskite films exhibited excellent light absorption in the wavelength range of 500–800 nm. Compared to the control group, the perovskite film treated with 10 μL/mL of TCC demonstrated a significantly increased absorbance, which will generate more carriers [22].

To investigate whether the introduction of TCC affects the energy band of the perovskite film, ultraviolet-visible (UV-vis) absorption spectroscopy was performed on the perovskite film. Figure 6b shows the Tauc plot derived from the UV-vis results. After TCC incorporation, the bandgap of the film was 1.551 eV, while that of the control film was 1.567 eV.

Meanwhile, steady-state photoluminescence (PL) and time-resolved photoluminescence (TRPL) tests were conducted on the perovskite thin films treated with 10 mL/mL TCC and the control perovskite thin films, respectively. The results are shown in Figure 6c,d. Figure 6c shows the PL spectra of the perovskite thin films before and after TCC treatment. It can be seen that the PL peak intensity of the perovskite thin film modified by TCC is significantly higher than that of the control film. This is because the TCC treatment improves the crystallization quality of the perovskite and reduces the defects [23].

The measured TRPL data were fitted using a double exponential function, and the obtained fitting parameter values are listed in Table S2. After modification with 10 μL/mL of TCC, the average carrier lifetime (τ_ave) of the perovskite film was 193.26 ns, significantly longer than that of the control film (τ_ave = 68.37 ns). The prolonged carrier lifetime is attributed to the influence of TCC treatment on the nucleation process of perovskite, leading to a reduction in film defects and consequently suppressing defect-induced non-radiative recombination.

To further investigate the impact of the reaction on perovskite, X-ray photoelectron spectroscopy (XPS) was used to analyze the perovskite, as shown in Figure 7. It can be seen that after the treatment, a small peak of the oxygen element appears at 530 eV, while the characteristic peak of Pb shows no obvious change. The small peak of oxygen at this position may originate from a certain interaction between the carbonyl group and the Pb inside the perovskite. However, due to the relatively low degree of this interaction, the Pb 4f peak shows no obvious change, indicating that the passivation effect on the perovskite is not significant.

2.5. The Impact of TCC on the Stability of PSCs

The stability of perovskite films is closely related to their surface morphology and hydrophobicity. A dense and smooth film structure can effectively block the penetration of environmental moisture, thereby delaying the degradation of perovskite materials and improving the long-term stability of PSCs. [24] To investigate the effect of TCC on the surface properties of the films, water contact angle measurements (Figure S5) show that the control film has a contact angle of approximately 51°, while the film treated with 10 μL/mL TCC shows no significant change in contact angle. This indicates that TCC modification does not significantly alter the surface chemistry of the perovskite film, and its influence on film hydrophobicity is negligible. Combined with AFM data, it is inferred that TCC primarily influences the physical morphology of the film by regulating the crystallization process rather than modifying its wettability through surface chemical alterations.

At room temperature, the prepared devices with/without TCC treatment were placed in an air environment with a humidity of approximately 60% ± 10% and a temperature of 25 °C to test their air stability. The results are shown in Figure 8a. After being placed for 150 h, the standard device could still maintain 84.2% of its initial efficiency. However, after TCC treatment, the device only retained 71.6% of its initial efficiency. After TCC treatment, the stability of the device decreased slightly.

The PSCs were placed in a heating environment (75 °C; N₂ atmosphere) in a glove box to study their thermal stability. The results are shown in Figure 8b. It can be observed that the PSCs treated with TCC have poorer thermal stability than the untreated PSCs. The efficiency of the untreated PSCs remains around 70% of its initial value after 175 h, while that of the PSCs treated with TCC drops to 51% of its initial value after 175 h. The thermal stability of the PSCs after TCC treatment decreases significantly.

Figure 8. (a) The air stability and (b) thermal stability of the control and TCC-modified devices.

Figure 8. (a) The air stability and (b) thermal stability of the control and TCC-modified devices.

The reason for the decreased stability after treatment may be that the anti-solvent strategy introduces stress into the perovskite. The residual lattice strain in the perovskite film can affect the lattice stability [25]. The doping strategy commonly used in the FAPbI₃ lattice inevitably changes the local lattice strain [26]. On the other hand, the reaction between the acyl chloride group and DMSO releases a large amount of heat, which also introduces thermal stress into the perovskite and has an adverse effect on the device stability [27]. Moreover, excessively high temperatures may also cause excessive growth of crystal grains, which actually reduces the stability of the device [18].

3. Materials and Methods

3.1. Materials

In this experiment, all reagents were used directly without additional purification. Formamidine acetate (FAAc, 99%), hydroiodic acid solution (57% aqueous solution), methylamine ethanol solution (8% aqueous solution), hydrobromic acid solution (33% aqueous solution), and Trans-cinnamyl chloride (TCC, 95%) were purchased from Aladdin(Wuhan, China). Lead iodide (PbI₂, 99.99%), 2-isopropanol (IPA, 99.8%), acetonitrile (ACN, 99.9%), Lithium bis((trifluoromethyl)sulfonyl), azanide (Li-TFSI, 99.5%) Tributyl phosphate (t-Bp, 99.99%), lead bromide (PbBr₂, 99.99%), cesium iodide (CsI, 99.99%), N,N-dimethylformamide (DMF, 99.9%), dimethyl sulfoxide (DMSO, 99.9%), and chlorobenzene (CB, 99.8%) were purchased from Sima-Aldrich(Shanghai, China). 2,2′,7,7′-tetra[N, n-bis (4-methoxyphenyl) amino]-9,9′-spirodifluorene (99.9%) (Spiro-OMeTAD) was purchased from Derthon(Guangdong, China). All the chemical substances were used in the way they were received, without any further purification.

3.2. Preparation of the PSCs

The fluorine-doped tin oxide (FTO) glass substrate was successively cleaned with ultrapure water, acetone, and ethanol in ultrasonic baths. Each step takes 15 min, and then it is placed in a 70 °C drying oven to dry for 1 h. Before spin coating, the FTO substrate is further cleaned with ultraviolet plasma for 10 min. To prepare compact SnO₂ electron transport layers (ETLs), SnO₂ nanoparticles were diluted with deionized water to 2.67% (volume ratio of 1:4) to obtain a hydrated tin dioxide solution. Then, at a spin coating speed of 3000 rpm for 30 s with a spin coating acceleration of 2000 revolutions per second², a layer of SnO₂ was deposited. After spin coating, the substrate was placed on a heating plate at 150 °C for annealing for 30 min. For perovskite precursor solutions, 0.15 mM (16.8 mg) of methylamine bromide (MABr), 0.05 mM (12.9 mg) of cesium iodide (CsI), 0.3 mM (20.3 mg) of methylamine chloride (MACl), 0.85 mM (146.2 mg) of FAI, 1.1 mM (507.1 mg) of lead iodide (PbI₂, 5%mol excess), and 0.15 mM (55.1 mg) of PbBr₂ were dissolved in 1 mL of a mixed solvent (DMF:DMSO = 9:1, by volume). To prepare perovskite films, 50 μL of perovskite precursor solution was spin-coated onto the tin dioxide (SnO₂) ETL. The coating was first carried out at 1000 rpm for 10 s and then at 5000 rpm for 25 s. Five seconds before the end of spin coating, 100 μL of countersolvent was slowly added, and then annealing treatment was carried out at 105 °C for 30 min. The above-mentioned countersolvents refer to the mixed countersolvents of pure CB and TCC solution (TCC dissolved in CB, volume ratios of 100:1, 100:2, and 100:3, respectively). Before preparing perovskite films, the electron transport layer was subjected to 15 min of ultraviolet ozone treatment to enhance the wettability of the perovskite precursor solution. After cooling to room temperature, 50 μL of 2,2′,7,7′-tetra-[N, n-bis (4-methoxy phenyl) amino]-9,9′-Spiro-OMeTAD solution was spin-coated onto the treated perovskite film at a speed of 4000 revolutions per second for 30 s. The preparation method of Spiro-OMeTAD solution is to dissolve 60 mM (72.3 mg) of Spiro-OMeTAD in 1 mL of CB and then add 28.8 μL of tert-butylpyridine (t-Bp) and 17.5 μL of Li-TFSI solution (520 mg Li-TFSI dissolved in 1 mL of ACN) to the solution. The preparation process of perovskite films and hole transport layers was carried out at room temperature (21 ± 2 °C) in a nitrogen atmosphere glove box. Finally, 120 nm of Ag electrode was deposited through vacuum thermal evaporation.

3.3. Characterization and Measurements

X-ray diffraction (XRD) measurements were conducted using Rigaku’s Miniflex600 model instrument. The ultraviolet-visible absorption spectrometer uses the Cary5000 model instrument from Agilent. Under the standard AM 1.5 simulated solar irradiation (SS-F5-3A), using a metal mask with an aperture area of 0.1 cm², the photocurrent density-voltage (J-V) curve of the device was measured using a Tektronic (Guangdong, China) 2400A source instrument. Ultraviolet-visible (UV-Vis) absorption spectra were recorded on a Cary 5000 spectrophotometer. The water contact angle was measured by a DataPhysics contact angle tester. The top-view and cross-section morphologies of the perovskite films were investigated using field-emission scanning electron microscopy (SEM) (Apreo S LoVac, Thermo, Thermo, Waltham, MA, USA). The surface roughness by atomic force microscopy (AFM) was carried out using Oxford Jupiter XR. Steady-state photoluminescence (PL) (excitation wavelength 460 nm) and time-resolved photoluminescence (TRPL) were conducted with Edinburgh Instruments Ltd.(Livingston, Edinburgh, UK) (FLS 980). The light intensity dependence of the V_OC, dark current, transient photovoltage (TPV), and transient photocurrent (TPC) and Mott–Schottky measurements were tested in the dark using an electrochemical workstation (Zennium Zahner, Kronach, Germany) at room temperature.

4. Conclusions

In summary, this study proposes an anti-solvent reaction strategy by introducing TCC molecules to react with DMSO in the perovskite precursor solution, thereby influencing the nucleation process of perovskite, passivating the internal defects of perovskite, and significantly improving the filling factor and efficiency of PSCs. This work proposes a novel and effective strategy to enhance device efficiency, thereby achieving highly efficient PSCs. In addition, this strategy also has great application potential in other perovskite-based optoelectronic devices.