Since the initial report in 2009, perovskite solar cells have shown a wide range of applications in the optoelectronics field such as solar cells, light-emitting diodes, phototransistors, and photodetectors in just over a decade due to their high power conversion efficiency (PCE), ease of solution preparation and low cost (Kojima et al., 2009; Xie et al., 2020; Cao et al., 2022; Zhang et al., 2023a; Zhang et al., 2023b; Zhuang et al., 2023). The PCE of perovskite solar cells has increased significantly from the initial 3.8% to the current 26.1%, suggesting promising practical applications for perovskite solar cells in the near future (Kojima et al., 2009; Kim et al., 2012; Xiao et al., 2017; Cui et al., 2019; Cao et al., 2023; Laboratory, 2023; Park et al., 2023).

Currently, most of the initial studies on high-efficiency perovskite solar cells include a carrier transport layer that requires a high-temperature annealing process (>450°C). This process can effectively decrease the lattice defects of the carrier transport layer, enhance the carrier transport efficiency, and reduce the recombination rate. For example, in 2021, Changduk Yang et al. used two fluorinated isomeric analogues and commonly used TiO₂ with a high temperature annealing process as the hole and electron transport layers, respectively. As a result, a FAPbI₃ perovskite solar cell has achieved a certified PCE of up to 24.8% (Jeong et al., 2020). In 2023, Peng Gao et al. prepared a TiO₂ electron transport layer, followed by an annealing process at a high temperature of 450°C for 1 h to fully condense the inorganic network, crystallize the anatase phase, and remove all surfactant residues. As a result, the fabricated Cs_0.05 (MA_0.17FA_0.83)_0.95Pb(I_0.83Br_0.17)₃ based hole transport layer-free perovskite solar cells achieved a PCE of 11.08% (Lalpour et al., 2023). High temperature processes increase the production cost and complexity of perovskite solar cells, which hinders their industrial application. Additionally, most flexible devices are fabricated on polymer substrates, such as poly (ethylene 2,6-naphthalate (PEN) or polyethylene terephthalate (PET), which cannot withstand high temperatures. This seriously impedes the application of perovskite solar cells in the flexible electronic field.

Recently, tin oxide (SnO₂) has become one of the most promising materials for electron transport layers due to its excellent properties. These include wide band gaps (∼3.6–4.5 eV), suitable band gap positions, high electron mobilities (240 cm² V⁻¹ s⁻¹, which is about 100 times higher than that of TiO₂), and high carrier concentrations (>10¹⁵ cm⁻³)(Bai et al., 2023; Xu et al., 2022; zhou et al., 2023). Furthermore, SnO₂ materials are cost-effective, eco-friendly, and can be produced using various low-temperature solution preparation techniques, including the sol-gel, chemical bath deposition, spin coating, and spraying methods. These methods are highly suitable for creating efficient and expansive flexible devices (Yoo et al., 2021; Bai et al., 2023). By combining a hydrothermal method and a low-temperature spin coating method, Xu et al. prepared high-quality SnO₂ electron transport layers and a flexible device with a structure of PEN/ITO/SnO₂/Cs_0.05FA_0.81MA_0.14PbI_2.55Br_0.45/Spiro-OMeTAD/Au, achieving an efficiency of up to 18.1% and a certified efficiency of up to 17.27% (Liu et al., 2019). In 2023, Yixin Zhao et al. prepared a SnO₂ electron transport layer with adjustable performance using a low-temperature chemical bath deposition method. They successfully fabricated SnO₂ nanocrystals with high crystallinity by using Cl as dopants. These nanocrystals were then dispersed in water to create a dispersion for the fabrication of the SnO₂ electron transport layer using a low-temperature spin coating method. The electron transport layer of the perovskite solar cells was improved by the addition of Cl-doped SnO₂ nanocrystals. The highest PCE was −25% (effective area of 0.085 cm²) and 20% (mini-module with effective area of 12.125 cm²) (Wang et al., 2023). These findings suggest that SnO₂ has significant potential as electron transport layers in high-performance perovskite solar cells.

In this paper, we report a solvothermal method for fabricating highly crystalline SnO₂ nanoparticles at low temperatures, which are succussed in use of flexible perovskite solar cells. The solvothermal method produces SnO₂ nanoparticles with uniform particle size and high crystallinity. These nanoparticles are then dispersed in n-butanol solution to fabricate SnO₂ electron transport layers and devices. The J-V curves demonstrate that the rigid perovskite solar device reaches a maximum PCE of 15.61% when the SnO₂ dispersion concentration is 15 mg mL⁻¹. Finally, flexible perovskite solar cells by using 15 mg mL⁻¹ of SnO₂ dispersion is fabricate with a PCE of 14.75%.

Figure 1A displays the low-resolution TEM findings of SnO₂ particles. The results show the SnO₂ have an average size distribution of 4–6 nm. The selective area electron diffraction (SAED) of Figure 1A at the upper right corner indicates that the SnO₂ are well-crystallized, with four clear diffraction circles corresponding to the (110), (101), (220), and (311) planes of SnO₂ crystal (JCPDS Card: 41-1445) can be seen.

From the high-resolution TEM (HRTEM) image in Figure 1B, it is possible to distinguish two grains with different interplanar distances. Clear lattice fringes with an interplane d-spacing of 0.343 and 0.235 nm corresponding to the (110) and (101) planes of the SnO₂ crystal (JCPDS Card: 41-1445) can also be seen, indicating a highly crystalline property for the SnO₂ nanoparticles. Figure 1C shows the XRD diffraction pattern of the SnO₂ nanoparticles. Five distinct diffraction peaks appear at 2θ = 26.6^o, 34.0^o, 37.9^o, 51.6^o, and 64.7^o, which can be assigned to the (110), (101), (200), (211), and (112) planes of SnO₂, respectively. No noise peaks are found, indicating the good crystallinity of the SnO₂ nanoparticles. The high crystallinity of SnO₂ nanoparticles indicates that there are few defects in SnO₂, which will be beneficial in achieving high carrier mobility when used as electron transport layers. The crystal grain size is calculated to be −5 nm using the Debye-Scherrer formula, which is in good agreement with the TEM particle size shown in Figure 1A (Li et al., 2017). These well-fabricated SnO₂ crystalline nanoparticles can fill the valleys among the ITO grains, providing a smooth surface. Additionally, using these highly crystallized SnO₂ nanoparticles can result in electron transport layers with a high carrier extraction rate and without the need for a high-temperature annealing process.

Figure 2 compares the surface properties of the ITO substrate and the SnO₂ electron transport layer (abbreviated as SnO₂ 15) prepared with a 15 mg mL⁻¹ SnO₂/n-butanol solution as a precursor on a bare ITO substrate, using SEM and AFM images. It can be seen that the surfaces of ITO and SnO₂ 15 are both smooth and flat, and the surface of SnO₂ 15 shows a polished and compacted layer of particulate material, which is different from the flake morphology of ITO caused by the deposition of SnO₂ nanoparticles on the surface of ITO. The smooth, flat and dense SnO₂ electron transport layer is conducive to the deposition and carrier transport of perovskite films.

In order to understand the influence of the SnO₂ electron transport layers obtained by different dispersion concentrations on the light transmittance, UV-vis transmission spectrum analysis was performed on SnO₂ 5, SnO₂ 15, SnO₂ 30, SnO₂ 45, and SnO₂ 60 samples, as well as on the ITO substrate. As shown in Figure 3A, all of the samples show good transparency in the main absorption wavelength range of perovskite film (380–800 nm), which is beneficial to the light absorption of perovskite films.

The crystallinity, surface morphology, absorption and band gap of perovskite films have an important influence on the photovoltaic performance of planar heterojunction perovskite solar cells. Therefore, the effect of SnO₂ precursor concentration on the crystallinity of perovskite films was investigated. Perovskite films were spin-coated on the surface of the electron transport layers prepared by SnO₂ precursor solutions with different concentrations, which were then characterized by XRD, and the results are shown in Figure 3B. All the perovskite films show no impurities and are well crystallized, and the positions of the diffraction peaks of all the films are basically the same, with the three strongest diffraction peaks located at 2θ = 14.1°, 28.5°, and 31.9°, respectively, corresponding to the (110), (220), and (222) planes of the tetragonal perovskite structure, respectively (Guo et al., 2016). As the concentration of the SnO₂ precursor increases, the diffraction peak intensity of the perovskite films gradually decreases, indicating that the electron transport layers prepared at high concentration are not conducive to the crystallization of the perovskite films.

The surface morphology and UV-vis absorption spectra of perovskite films prepared on SnO₂ 15 sample are further characterized. Figure 3C shows the top view SEM image of perovskite. It can be seen that the perovskite film is uniform and smooth without any pinholes, which could reduce the recombination of photogenerated carriers and improve the photovoltaic performance of the corresponding devices. Figure 3D shows the UV-Vis absorption spectrum of the perovskite film. It can be seen that the MAPbI₃ perovskite film has a strong absorption in the short wavelength region. The band gap (E _g) of a semiconductor can be obtained from the Tauc curve (Huxter et al., 2008). As shown in Figure 3E, the band gap of the as-prepared MAPbI₃ is about 1.59 eV, which is in agreement with the value reported in the literature (Noh et al., 2013; Eperon et al., 2014).

The cross-sectional SEM image of the SnO₂ 15 electron transport layer-based device is used as a representative sample to characterize the device, and the result is shown in Figure 4A. The structure of the device can be listed as Glass/ITO/SnO₂/MAPbI₃/Spiro-OMeTAD/Ag, showing that the different layers of the device are tightly bonded without any no cracks. The thickness of the ITO, MAPbI₃, Spiro-OMeTAD and Ag is approximately 130, 450, 230, and 100 nm, respectively. It is important to emphasize that the SnO₂ electron transport layer is too thin to be observed. The schematic energy level diagram for the SnO₂ electron transport layer device is shown in Figure 4B, which shows a well-established energy level alignment (Jiang et al., 2010; Noh et al., 2013; Jaramillo-Quintero et al., 2015; Zhu et al., 2017).

In order to optimize the concentration of the SnO₂ nanoparticle dispersion, SnO₂ electron transport layers and corresponding rigid devices fabricated with different concentrations of SnO₂ nanoparticle dispersion are prepared, with the J-V curves of the highest efficiency shown in Figure 4C. “R” stands for reverse scanning and “F” stands for forward scanning. The parameter values for the relevant devices are summarized in Table 1. It can be seen that when the concentration of the dispersion was 5 mg mL⁻¹, the corresponding devices obtained efficiencies of 13.38% and 9.28% under reverse and forward scanning, respectively, and when the concentration was increased to 15 mg mL⁻¹, the two efficiencies were increased to 15.61% and 10.70%, respectively. However, as the concentration increases, the efficiency of the corresponding device in both forward and reverse scanning decreases, in particular, the open-circuit voltage (V _OC) and fill factor (FF) decrease significantly. This may be due to the fact that, on the one hand, the films obtained with a high concentration of SnO₂ are thicker, which is not favorable for fast charge carrier transport and, on the other hand, the crystallinity of the perovskite films on the thicker SnO₂ films is also poorer and recombination is more severe. Therefore, 15 mg mL⁻¹ was found to be the optimum concentration of SnO₂ dispersion.

The above results also show that the planar heterojunction devices based on SnO₂ obtained by the solvothermal method have a significant hysteresis, which is mainly due to the large number of defects in the interior of the perovskite film, in the region close to the surface and at the interface, These act as traps for the electrons and holes and are rapidly filled under reverse scanning conditions to form a good contact at the interface, whereas under forward scanning the efficiency obtained from the test at this point is low due to the fact that the transport of charge is partially trapped by the defects and hence hysteresis occurs.

Figure 4D shows the EQE spectra for the rigid device based on the SnO₂ 15 electron transport layer. It can be seen that the initial threshold wavelength of the EQE spectrum for the device is at −780 nm, which is consistent with the band gap of 1.59 eV for the perovskite film. The device has a high photoelectric response in the wavelength range between 380 and 760 nm, with a maximum EQE of −77% and an integrated current density of 18.31 mA cm⁻². There is some discrepancy between the integrated current density and the short-circuit current density obtained from the J-V curve, mainly due to insufficient filling of defects in the perovskite and the electron transport layer in the bulk and at the interface. As the EQE curve is tested under weak monochromatic light irradiation without optical bias, the photogenerated carrier concentration of the perovskite film is much lower than that tested under simulated sunlight irradiation. The lower carrier concentration makes the defects in the bulk and at the interface unable to be fully filled, and the photogenerated carriers produced by weak light are easily recombined at the interface or in the bulk, which reduces the efficiency of interfacial charge collection, resulting in the integrated current density being lower than the short-circuit current density.

Furthermore, 40 devices of rigid perovskite solar cells based on the SnO₂ 15 electron transport layer are fabricated and the statistical results of photovoltaic parameters are shown in Figure 5. Clearly, the device performance parameters distribute in a relative narrow range, showing a good reproducibility of our devices. Finally, flexible PEN/ITO was adopted to substitute rigid glass/ITO substrate to fabricate flexible perovskite solar cells at low temperature by using optimized SnO₂ dispersion (15 mg mL⁻¹). The best-performing flexible device has achieved a PCE of 14.75% with a V _OC of 1.082 V, a J _SC of 18.26 mA cm⁻² and an FF of 0.746 under reverse voltage scan (Figure 6). This study is expected to open a new way to prepare SnO₂ films at a low temperature for high performance flexible devices.

In this paper, SnO₂ nanoparticles with high crystallinity, good dispersibility and uniform particle size distribution were prepared by a solvothermal method. SnO₂ electron transport layers and perovskite solar cells were then prepared using different concentrations of SnO₂/n-butanol dispersions. Although the electron transport layer prepared with a high dispersion concentration shows a better light transmittance, the perovskite film on it has a worse film crystallinity. And a thicker SnO₂ film can lead to more carrier recombination and a higher internal resistance, resulting in a lower open-circuit voltage and fill factor. Ultimately, the optimal rigid and flexible device was obtained by using a SnO₂/n-butanol dispersion of 15 mg mL⁻¹ with a PCE of 15.61% and 14.75%, respectively. Our study paves a new way to prepare SnO₂ electron transport layers at a low temperature, making it suitable for flexible solar cells.