Controllable Crystallization of Perovskite Films during the Blade-Coating Fabrication Process for Efficient and Stable Solar Cells

Abstract

The scalable production of high-quality perovskite thin films is pivotal for the industrialization of perovskite thin film solar cells. Consequently, the solvent system employed for the fabrication of large-area perovskite films via coating processes has attracted significant attention. In this study, a solvent system utilizing a volatile solvent as the primary reagent has been developed to facilitate the rapid nucleation of volatile compounds. While adding the liquid Lewis base dimethylformamide (DMF) can help to improve the microstructure of perovskite films, its slow volatilization renders the crystal growth process uncontrollable. Based on the solvent system containing DMF and ethanol (EtOH), introducing a small amount of NH₄Cl increases the proportion of the intermediate phase in the precursor films. This not only results in a controllable growth process for the perovskite crystals but also contributes to the improvement of the film microstructure. Under the simulated illumination (AM1.5, 1000 W/m²), the photoelectric conversion efficiency (PCE) of the inverted solar cells has been improved to 20.12%. Furthermore, after 500 hours of continuous illumination, the photovoltaic device can retain 95.6 % of the initial, indicating that the solvent system is suitable for the scalable fabrication of high-quality FAPbI₃ thin films.

1. Introduction

Over the past decade, the laboratory efficiency record of perovskite solar cells (PSCs) has experienced a remarkable increase, advancing from an initial 3.8 % to a certified 26.1% [1,2,3]. This substantial improvement has attracted considerable interest from both the academic community and industry stakeholders. However, the deployment of this innovative photovoltaic technology is impeded by several critical challenges that necessitate further investigation and resolution [4,5,6]. Moreover, a significant disparity exists in comparison to crystalline silicon solar cells. The principal limitation is attributed to the quality of the perovskite thin film, which includes factors such as morphology, uniformity, crystalline structure, and defect density, all of which are inferior to those observed in spin-coated small-area perovskite thin films [7,8,9,10].

To date, the predominant methodologies for the fabrication of large-area perovskite thin films include solution-coating (such as blade- and slit-coating), inkjet printing, and vacuum evaporation. Among these techniques, solution-coating demonstrates significant advantages in terms of production efficiency, cost-effectiveness, and the ability to preserve the nonstoichiometric composition of the coated perovskite film. The recent scholarly literature has extensively documented the application of solution-coating methods for the preparation of large-area perovskite thin films, with the objective of improving the conversion efficiency of large-area modules [11,12]. Substantial progress has been made in this field of research. Irrespective of whether a small-area solution spin-coating technique or a large-area solution-coating method is utilized, both conform to the traditional LaMer crystal growth model [13,14,15,16]. Prior to attaining the critical nucleation concentration, the perovskite solute within the film undergoes rapid supersaturation facilitated by techniques such as anti-solvent extraction, air knife blowing, and rapid vacuum pumping. These post-processing methods induce instantaneous nucleation, thereby ensuring comprehensive substrate coverage. The formation of a defect-free film is of paramount importance. In the context of various post-processing methodologies, the regulation of the physical and chemical properties of the precursor solution is crucial for influencing the nucleation and crystallization processes of perovskite as well as determining the crystalline quality of the resultant thin film [17,18]. Solvents are generally categorized and optimized according to their volatility, solubility, and coordination properties to attain the optimal coating formulation. Deng et al. proposed the use of a solvent system consisting of high-volatility, low-coordination solvents in conjunction with a minor proportion of high-boiling-point, strong-coordination solvents. Through the application of air-knife-assisted film formation, they successfully prepared a large area of MAPbI3 perovskite film at a high speed of up to 99 cm/s [8,12]. However, a substantial quantity of acetonitrile (ACN) is utilized in the precursor solution, which can volatilize rapidly to facilitate high-speed film formation. Furthermore, in comparison to small-area perovskite films prepared via spin-coating, the current large-area perovskite films exhibit several significant issues: 1. The crystal phase and grain size distribution of perovskite crystals exhibit significant non-uniformity across a large area. 2. The surface roughness of the films varies considerably over an extensive area. 3. There is substantial variation in the thickness of the film across a large area. The aforementioned issues are intricately linked to the nucleation and growth mechanisms of perovskite crystals within thin films [14,19,20]. A significant challenge involves modulating the growth kinetics of perovskite thin film crystals to enable controlled fabrication over extensive areas [21,22,23]. Overcoming this challenge is crucial for the scalable production of high-quality perovskite thin films and the advancement of high-efficiency photovoltaic devices. In this study, we successfully prepared high-quality perovskite thin films through the incorporation of both solid and liquid Lewis base additives [24,25,26,27]. This can be achieved by employing a range of solvent systems and design strategies suitable for large-area fabrication. The coating process was conducted in dry air, adhering to the proposed design scheme.

In this study, FAPbI₃ thin films were fabricated using a vacuum flashing-assisted coating process with EtOH as the main solvent. The effects of DMF and ammonium chloride (NH₄Cl) on the nucleation and growth of FAPbI₃ perovskite crystals were systematically investigated. The addition of appropriate amounts of these Lewis base additives increased the size of the perovskite crystals. Notably, compared to the EtOH/DMF solvent system, the incorporation of a small amount of NH₄Cl not only facilitated an increase in grain size but also conferred advantages such as uniform grain size distribution and enhanced surface flatness. In conjunction with the analysis of the crystallization kinetics mechanism of perovskite crystals, the incorporation of a small amount of NH₄Cl into the precursor solution has been shown to effectively promote the formation of an intermediate structure in the unannealed film. This intermediate structure gradually transitions into crystalline form during the subsequent annealing process, thereby enhancing the controllable growth of the crystals. This method is advantageous for the scalable fabrication of high-quality FAPbI₃ films. Through the substantial enhancement of the microstructure of the FAPbI₃ perovskite thin film, the carrier transport properties in the associated photovoltaic device are significantly optimized. Consequently, this leads to marked improvements in both the photoelectric conversion efficiency (PCE) and the stability of the FAPbI₃ perovskite film solar cell.

2. Experimental Details

2.1. Materials

N, N-dimethylformamide (DMF, 99.8%), EtOH (99.8%), and IPA (99.5%) were brought from Sigma-Aldrich. NH₄Cl, lead(II) iodide(PbI₂) powers (99.8%), and formamidinium iodide (FAI) powers (99.8%) were acquired from Xi’an Yuri Solar Co., Ltd. (Xi’an, China). Nickel(II) acetylacetonate (99%) was purchased from Tokyo Chemical Industry. The metallic Ag particles were purchased from Sino-Platinum Metals Co., Ltd. (Kunming, China) Unless stated otherwise, all solvents and regents were stored and used without any purification. The C₆₀ and bathocuproine (BCP) powers (≥99.9%) were purchased from Advanced Election Technology Co., Ltd., Shenzhen, China.

2.2. Solution Preparation

The indium tin oxide (ITO)-coated glass substrates (Nippon Sheet Glass Co., Ltd., Osaka, Kitahama, Japan) were cleaned by sequential ultrasonication for 15 min with deionized water and ethanol. Then, the ITO substrates were dried by a high-speed N₂ gas flow before being stored. Then, a nickel(II) oxide (NiO) layer (hole transport layer) was deposited on the clean substrates. Next, NiO solution was prepared with a mixed solution of acetonitrile and ethanol (with a 90:10 volume ratio, 150 mL) of nickel acetylacetonate. The perovskite precursor solution (1.2 M) was prepared by mixing FAI and PbI₂ in mixed solvents (EtOH: DMF/Volume: Volume = 9:1).

2.3. Device Fabrication

The ITO glass should be washed with sequential ultrasonication for 15 min with detergent, deionized water, and ethanol and then subjected to UV ozone treatment (10 min). First, a NiO layer (hole transport layer, HTL) was deposited on the clean substrates. The prepared nickel acetylacetonate solution was automatically sprayed by an air nozzle on the hot ITO glass (200 °C) for 10 min, which was transferred into the glove box filled with N₂. The films were blade-coated with a precursor solution (coating speed of 20 mm/s), and then the wet films were transferred in a vacuum chamber to accelerate the evaporation of most solvents. Undergoing an annealing process (150 °C for 10 min), the FAPbI₃ films were prepared. Subsequently, the 20 nm-C₆₀ and 2 nm-BCP were sequentially deposited on the prepared FAPbI₃ films by thermal evaporation in a vacuum atmosphere (≤1 × 10⁻⁴ Pa), and the deposition rate was controlled below 0.2 nm/s. Finally, the Ag (120 nm) was deposited on the surface of the devices in a vacuum chamber (≤3 × 10⁻⁴ Pa) for the electrodes.

2.4. Characterizations

Morphologies of the prepared films and devices were carried out by scanning electron microscope (SEM) images acquired using Hitachi S-4800 (Tokyo, Japan). The XRD patterns of annealed and unannealed films were obtained by an Empyrean X-ray diffractometer with Cu Ka radiation (PANalytical B.V. Co., Almelo, The Netherlands). The absorption spectra of the perovskite films were carried out by a Hitachi U-4100 UV–Vis spectrophotometer (Tokyo, Japan). The photoluminescence (PL) spectra of the samples were recorded using the Edinburgh Instrument FLS920 (the excitation light source was a 450 W xenon lamp) (Livingston, Scotland). The electrochemical impedance spectroscopy (EIS) results of the devices were collected with a Zennium electrochemistry workstation (Zahner, Berlin, Germany). The photoluminescence (PL) spectra of the perovskite films prepared on glasses were recorded via the Edinburgh Instrument FLS920, and the excitation light source was a 450 W xenon lamp. Then the current density–voltage (J–V) curves and corresponding photovoltaic parameters of perovskite solar cells (PSCs, the active area was 0.09 cm²) were achieved with a black mask under a simulated illumination (standard AM1.5G, 1000 W/m²), and a silicon photodiode was used for calibrating the simulated illumination. The operational stability of the prepared PSCs was performed under continuous white-light LED array illumination in a dry atmosphere (RH ≤ 10%), and the test temperature was controlled at around 30 ± 5 °C.

3. Results and Discussion

Currently, the solvent system employed for the preparation of large-area perovskite films via vacuum/nitrogen knife-assisted crystallization blade-coating necessitates the use of volatile solvents at ambient temperature and pressure, such as EtOH and ACN. However, the nucleation and growth processes of perovskite crystals cannot be effectively regulated solely through the use of highly volatile solvents, thereby constraining their applicability in the scalable production of high-quality perovskite films. In this study, we employ EtOH with high volatility as the primary solvent, incrementally incorporating small quantities of DMF and NH₄Cl to maintain an adequate volatilization rate. The rapid evaporation of a substantial volume of solvent facilitates the swift and uniform nucleation of perovskite crystals. Concurrently, the formation of a minor DMF/NH₄Cl-related intermediate phase within the precursor films enhances the controllability of perovskite crystal growth during the post-annealing process. Based on previous research [4], three kinds of precursor solutions were prepared for this study: EtOH, EtOH + DMF (with a molar ratio of DMF/Pb²⁺ = 0.8), and EtOH + DMF + NH₄Cl (with a molar ratio of NH₄Cl/Pb²⁺ = 0.2). As illustrated in Figure 1a, wet films were initially formed by coating these precursor solutions onto a clean substrate at a coating speed of 20 mm/s in a controlled dry environment (relative humidity RH ≤ 10%). Subsequently, the coated substrates were placed in a vacuum chamber for a brief period of 5 seconds. Then, the samples were removed onto the heating table for a subsequent annealing treatment at 100 °C for a duration of 10 minutes. Figure 1b–d illustrate the surface morphologies of the perovskite thin films synthesized from three distinct precursor solutions.

The perovskite film synthesized from the precursor solution containing solely EtOH exhibits a flat surface and a dense microstructure [28,29]. The grain sizes within the film are uniformly distributed, albeit relatively small, ranging from 50 to 100 nm. This limited grain size is primarily attributed to the rapid volatilization of the solvent and the consequent increase in solution concentration, conditions under which nucleation predominantly occurs within the precursor films. Within a brief duration, a dense microstructure develops in the film, thereby inhibiting the crystal growth process. Concurrently, crystal growth ceases entirely once the solvent has fully volatilized. These factors collectively result in a smooth and compact microstructure of the film, albeit with a small grain size. During this process, the introduction of a minor quantity of DMF facilitates the gradual transformation of the mesophase structure present in the precursor film into a perovskite crystal structure during the post-annealing stage. Concurrently, DMF progressively volatilizes and dissolves the crystal boundaries, leading to the formation of larger crystal particles. Nevertheless, due to the low volatility of DMF, a limited number of voids are observed in the prepared film, as illustrated in Figure 1c. Subsequently, a minor quantity of NH₄Cl was introduced into the precursor solution, resulting in an enhancement of the crystal particles within the film. Concurrently, the surface morphology of the film exhibited significant improvement. To further investigate the dynamic process of crystal growth in the perovskite thin film, this study analyzed the crystal structure of three types of perovskite thin films both before and after annealing, utilizing X-ray diffraction spectra, as illustrated in Figure 2a,b.

Figure 2a presents the X-ray diffraction (XRD) spectra of various perovskite films, both before and after annealing at 100 °C for 10 minutes. In the sample devoid of DMF and NH₄Cl, the precursor film exhibits only a weak diffraction peak corresponding to the α-phase FAPbI₃ perovskite. This indicates the presence of a substantial number of fine α-phase perovskite crystal nuclei within the film. The full width at half maximum (FWHM) of the primary XRD peak associated with the α-phase crystal nuclei is 14.1°, and the crystal nucleus size, as calculated from the FWHM, is approximately 50 nm. With the addition of DMF to the precursor film, the relative integral area of the α-phase perovskite nucleus (100) correlated XRD diffraction peak (approximately 14.1°) progressively decreases. This observation indicates a gradual reduction in the proportion of the α-phase crystal (100) component within the film [4,30,31,32]. Concurrently, a secondary phase diffraction peak emerges between 8° and 9°, which, according to literature reports, is likely attributable to the formation of an intermediate phase resulting from the coordination of DMF with FAPbI₃. Upon the addition of a small quantity of NH₄Cl to the precursor solution, there is a progressive enhancement in the diffraction peak intensity of the secondary phase relative to the α-phase perovskite crystal nucleus. This indicates an increased relative content of the intermediate phase within the surface precursor film, corroborating the aforementioned analytical results. Post-annealing, the diffraction peaks corresponding to DMF and the intermediate phase are no longer observed, indicating a complete transformation of all products into α-phase perovskite crystals. After adding the NH₄Cl, the intensity of the corresponding 14.1° diffraction peak progressively increases, while the half-height width correspondingly decreases. This observation suggests that the size of the perovskite crystals within the film enlarges, corroborating the SEM test results. Consequently, the presence of the intermediate phase appears to facilitate the crystal growth of the perovskite film during the subsequent annealing process.

Subsequently, in light of the aforementioned conclusions, an analysis of the optical properties of perovskite thin films synthesized from the three precursor solutions was conducted. Figure 3a presents the absorption characteristic curve of the perovskite thin film. The absorption spectra of the perovskite thin films, fabricated using a precursor solution with pure EtOH as the primary solvent, predominantly span the range of 450 nm to 800 nm. The intersection of the absorption curve with the X-axis occurs at approximately 780 nm. This observation aligns closely with the absorption characteristic curve of perovskite thin films reported in the literature, suggesting that the thin film primarily comprises α-phase perovskite crystals [31]. Simultaneously, the absorption intensity of the film sample is significantly enhanced by the addition of DMF, primarily due to the increased crystal particle size within the film. Additionally, it was observed that the slope of the absorption edge (approximately 780 nm) in the film’s absorption curve is steeper compared to the aforementioned samples. This phenomenon is predominantly attributed to the effective improvement in the microstructure of the perovskite film following the incorporation of DMF. Furthermore, the absorption intensity of the thin film was significantly enhanced, and the slope of the absorption edge of the absorption curve was also improved following the addition of an appropriate amount of NH₄Cl to the precursor solution. This suggests that the incorporation of NH₄Cl further optimized the microstructure of the perovskite thin film [33]. Figure 3b presents the photoluminescence curves of the film samples prepared from the three different precursor solutions. The addition of a small amount of DMF and NH₄Cl to the precursor solution results in a gradual increase in the luminescence intensity of the thin film sample, suggesting an enhancement in the crystal structure of the synthesized FAPbI₃ thin film. This observation aligns with the findings from XRD and SEM analyses, which further corroborate that the incorporation of DMF and NH₄Cl in the precursor solution facilitates the improvement of the quality and controllability of FAPbI₃ thin films.

To further analyze the effects of DMF and NH₄Cl on the microstructure of FAPbI₃ films, Figure 4a presents the grain size distribution statistics obtained from SEM images of the thin film samples. It is evident that compared to the FAPbI₃ film prepared using a DMF precursor solution, the grain sizes of the sample prepared with an NH₄Cl precursor solution are generally larger and exhibit a more concentrated distribution. These findings indicate that the incorporation of DMF and NH₄Cl enhances the controllability of perovskite crystal formation, which is advantageous for subsequent scalability. Simultaneously, Atomic Force Microscopy (AFM) images of perovskite films (Figure 4b,c) reveal that the surface roughness of the FAPbI₃ film prepared using the NH₄Cl precursor solution is lower compared to that of the FAPbI₃ film prepared with the DMF precursor solution [34]. Consequently, the flatness of the film surface is enhanced with the addition of NH₄Cl, which establishes a foundation for further improving the carrier transport performance of photovoltaic devices. Additionally, Figure 4d presents the cross-sectional Scanning Electron Microscopy (SEM) images of the film samples. The presence of numerous voids at the interface between the FAPbI₃ film prepared using a DMF precursor solution and the substrate is evident, which is likely to result in suboptimal carrier transport within photovoltaic devices. Conversely, the incorporation of NH₄Cl significantly reduces the voids at the interface of the as-prepared perovskite thin film samples and the substrate. Numerous studies indicate that this phenomenon is primarily attributed to the volatilization of DMF film samples during the annealing process, which leads to the formation of numerous voids at the film’s base. Consequently, the introduction of a small quantity of NH₄Cl enhances the interface morphology between the film and the substrate. This improvement is crucial for optimizing carrier transport.

To further investigate the transport properties of carriers in photovoltaic devices, reverse-structure thin film photovoltaic devices were fabricated utilizing the aforementioned perovskite thin films. The architecture of these thin film photovoltaic devices comprises ITO glass/NiO layer (hole transport layer, HTL)/FAPbI₃/C₆₀ (electron transport layer, ETL)/BCP/Ag. Cross-sectional SEM images of the photovoltaic device are presented in Figure 5a, with an active area of 0.09 cm². Figure 5 illustrates the voltage–current characteristics of the photovoltaic device under dark conditions. The data indicate a decrease in the reverse saturation current density with the incorporation of NH₄Cl. This observation, when integrated with the analysis of the thin film microstructure, suggests that the addition of NH₄Cl mitigates defects within the perovskite thin film. Consequently, this reduction in defects diminishes carrier recombination within the device, thereby enhancing carrier transport properties. These improvements are anticipated to significantly elevate the photovoltaic performance of the perovskite thin film solar cell. Simultaneously, electrochemical impedance spectroscopy (EIS) measurements were conducted (Figure 5c) to evaluate the charge transfer and recombination processes in perovskite solar cells (PSCs) fabricated using various precursor solutions. The charge recombination resistance (Rrec) of the PSC prepared with DMF + NH4Cl was found to be higher than that of the sample prepared with DMF alone, suggesting that the interface characteristics of the thin film photovoltaic devices are substantially improved upon the incorporation of NH₄Cl. This result is consistent with the analytical results obtained from the cross-sectional SEM images of the aforementioned thin film samples. Moreover, enhanced charge extraction can contribute to improvements in both the short-circuit current density (J_sc) and the open-circuit voltage (V_oc) of photovoltaic devices.

Furthermore, the J–V characteristic curve of the photovoltaic device under AM1.5 illumination is presented in Figure 5d,e, with detailed parameters provided in

Table 1. Photovoltaic parameters of the champion PSCs fabricated with precursor solutions containing DMF and DMF + NH₄Cl.

		JSC (mA/cm2)	VOC (V)	FF (%)	PCE (%)
DMF	Forward scan	18.76	1.09	64.8%	13.25%
	Reverse scan	19.90	1.10	59.6%	13.05%
DMF + NH4Cl	Forward scan	22.86	1.11	79.3%	20.12%
	Reverse scan	22.46	1.10	79.9%	19.74%

. The J–V characteristic curve of the photovoltaic device sample incorporating DMF exhibits a pronounced S-shaped distortion, suggesting inferior photo-generated carrier transportation under illuminated conditions. Under illumination, the photovoltaic device exhibits an open-circuit voltage (V_OC) of 1.09 V, a short-circuit current density (J_SC) of 18.76 mA/cm², a fill factor (FF) of 64.80%, and a PCE of 13.25%. After the addition of NH₄Cl, the current–voltage (J–V) characteristic curve of the thin film photovoltaic device exhibits a progressive flattening. As a result, the open-circuit voltage increases to 1.1 V, the short-circuit current density rises to 22.86 mA/cm², and the PCE improves to 20.12%. Simultaneously, the relatively concentrated forward and reverse scanning results of photovoltaic devices with NH₄Cl addition suggest a high degree of repeatability in the process method. The observed decrease in photoelectric conversion efficiency may be attributed to an increase in the internal resistance of series cells. In subsequent stages, the photovoltaic efficiency can be enhanced by further adjusting the parameters of the precursor solution and optimizing the interface contact to minimize performance losses. Simultaneously, Figure 6a,b illustrate the statistical power conversion efficiency (PCE) of solar cells fabricated with and without the incorporation of NH₄Cl into the precursor solution, based on ten samples prepared under each set of conditions. The results indicate that the PCE of samples prepared without NH₄Cl ranges from approximately 4% to 7%, whereas the PCE of samples prepared with NH₄Cl ranges from approximately 16% to 21%, and the related parameters are shown in Table S1. These results further demonstrate that the addition of NH₄Cl significantly enhances the PCE and repeatability of perovskite thin film solar cells, thereby facilitating the scalable fabrication of PSCs.

Furthermore, we evaluated the operational stability of the photovoltaic device under continuous illumination. To enhance stability, a 20 nm layer of bismuth (Bi) was applied as an inert buffer layer before the evaporation of the silver (Ag) electrode [35,36]. This buffer layer serves to mitigate iodine-induced corrosion of the Ag electrode, which can result from the decomposition and diffusion of perovskite. The maximum power point (MPP) of the photovoltaic device was continuously monitored and assessed using a white LED as the aging light source, with an equivalent light intensity of 1 sun. The module temperature was maintained at 30 ± 5 °C, as depicted in Figure 6. After subjecting the aforementioned small photovoltaic device, with an effective area of 0.09 cm², to 500 hours of continuous illumination, the photoelectric conversion efficiency of the device incorporating NH₄Cl remains at 95.6% of its initial efficiency. In contrast, photovoltaic devices fabricated using DMF exhibit a more pronounced efficiency decay. Consequently, the incorporation of NH₄Cl enhances the stability of thin film photovoltaic devices, primarily due to the high-quality, low-defect-density perovskite thin film and the improved carrier transport mechanism at the NiO/C₆₀ stable interface.

4. Conclusions

In this study, high-quality FAPbI₃ perovskite thin films were synthesized using a coating method. The incorporation of DMF and NH₄Cl enhanced the stability and temporal window of the wet film and moderated the crystal growth rate within the film. This approach increased the content of the intermediate phase in the wet film, allowing perovskite grains to grow further during the annealing process. Consequently, the grain sizes were enlarged, the density of grain boundary defects was reduced, and the crystalline quality of the film was significantly improved.

Furthermore, the interfacial contact between the perovskite film and the substrate is significantly enhanced through the incorporation of liquid DMF and solid NH₄Cl. An analysis of the carrier transport mechanism under dark conditions reveals that the high-quality FAPbI₃ perovskite thin film, along with the improved interface performance, enhances the carrier transport properties of the corresponding photovoltaic device. Based on these findings, under simulated illumination conditions, the photovoltaic performance parameters of the thin film device were enhanced by the incorporation of a small amount of DMF and NH₄Cl, resulting in a photoelectric conversion efficiency of 20.12%. Furthermore, the decay curve of the photoelectric conversion efficiency under continuous irradiation with an equivalent 1-sun white-light source indicates that the improved microstructure of the perovskite thin film achieved through the addition of DMF and NH₄Cl also enhances the operational stability of the photovoltaic device. The results indicate that solid Lewis base additives enable the control and scalability of perovskite crystal growth, which is crucial for advancing the industrialization of perovskite thin film photovoltaic devices.