Effect of SnO 2 Colloidal Dispersion Solution Concentration on the Quality of Perovskite Layer of Solar Cells

: The electron transport layer (ETL) is critical to carrier extraction for perovskite solar cells (PSCs). Moreover, the morphology and surface condition of the ETL could inﬂuence the topography of the perovskite layer. ZnO, TiO 2 , and SnO 2 were widely investigated as ETL materials. However, TiO 2 requires a sintering process under high temperature and ZnO has the trouble of chemical instability. SnO 2 possesses the advantages of low-temperature fabrication and high conductivity, which is critical to the performance of PSCs prepared under low temperature. Here, we optimized the morphology and property of SnO 2 by modulating the concentration of a SnO 2 colloidal dispersion solution. When adjusting the concentration of SnO 2 colloidal dispersion solution to 5 wt.% (in water), SnO 2 ﬁlm indicated better performance and the perovskite ﬁlm has a large grain size and smooth surface. Based on high efﬁciency (16.82%), the device keeps a low hysteresis index (0.23).


Introduction
ETLs in PSCs not only promote the separation of photogenerated electron-hole pairs but also improves the charge transport efficiency, thus avoiding the influence of charge accumulation on the device lifetime. At the same time, the material of ETL could significantly affect the topography of perovskite layers and device performance [1][2][3][4][5]. Our widely applied ETL materials are organic and inorganic materials. Although organic materials have good solubility in the treatment process, their low electron mobility and high cost hinder the commercialization of PSCs. Inorganic materials, especially metal oxides [6][7][8][9][10][11][12][13][14][15][16][17][18], are endowed with superior advantages of low cost, high stability, and excellent photoelectric property [12,19,20], are usually used as electron transport layer materials.
Metal oxides such as TiO 2 [6][7][8][9], ZnO [14,15], and SnO 2 [10][11][12][13][14]21], have become the mainstream materials of ETLs. Giordano and co-workers have demonstrated that the electronic trap states could be reduced and electron transport could be enhanced by introducing Li-ions into TiO 2 ETLs [22]. Finally, the power conversion efficiency (PCE) rocketed to 19% from 17%. However, the mesoporous TiO 2 ETL requires a sintering process under high temperature, which limits the development of low-temperature and highefficiency solar cells. Furthermore, PSCs prepared under low-temperature have aroused great interest in the past ten years. Ding and co-workers have prepared perovskite thin film with full coverage, high density, and good uniformity by increasing crystallizing speed on the ZnO electron transport layer [23]. However, the ZnO is not stable due to its chemical properties and the interface between ZnO and perovskite could degrade at high temperature, which demonstrates that ZnO ETL is not desirable for PSCs. However, SnO 2 has high electron mobility and high transparency [24]. Comparatively, SnO 2 has high electron mobility (10 −2 µS/cm) and a high conductivity (10 1 cm 2 V −1 S −1 ), as shown in Table S1 (Supplementary Materials). Therefore, SnO 2 is critical to the performance of PSCs prepared under low temperature.
The SnO 2 colloidal dispersion solution concentration used to prepare ETL has a significant effect on the perovskite film quality. Li et al. [25] reported that the high quality of ETLs can be produced by controlling the thickness of the film while it is treated by UV. The thickness is dependent on the concentration of SnO2. When the concentration of SnO2 was 20%, the PSCs obtained an optimal performance. Yang et al. [21] reported on EDTA-complexed SnO 2 ETLs by complexing EDTA with SnO 2 in planar-type PSCs, Yang et al. [26] and Qiang et al. [27] prepared the ETL material by dispersing the as-synthetic SnO 2 in the low boiling ethanol and the concentration of the SnO 2 spin-coating solution was optimized. However, the influence of the ETL formed by different SnO 2 concentrations on the quality of the perovskite film has not been carefully studied. Bahadur and coworkers [28] demonstrated that the PSC could exhibit a PCE of 8.51% when the volume ratio of SnO 2 colloidal solution and deionized water is 1:4. Therefore, SnO 2 could be applied in planer PSCs prepared under low-temperature and humid environments as superior roll-toroll-capable ETLs. However, few works about the quality of the perovskite layer deposited on tin dioxide ETL has been reported. Bu and co-workers [29] studied the morphology and grain size of CsFAMA perovskite deposited on Alfa-SnO 2 ETL and water Alfa-SnO 2 /KOH ETLs. However, the concentration of SnO 2 colloidal dispersion was not optimized. In recent years, Huang and co-workers [30] increased the substrate transmittance and reduced the density of trap-state of the perovskite layer by using the optimized SnO 2 layer. However, the optimal concentration interval of SnO 2 was relatively large.
In this paper, we systematically studied the single cation perovskite films formed on the ETL prepared with various concentrations of SnO 2 colloidal dispersion solution at room temperature. We found that the ETL formed with SnO 2 colloidal dispersion solution of different concentrations could affect the device performance as well as perovskite film morphology. When the SnO 2 colloidal dispersion solution concentration is set to be 5 wt.%, the perovskite film exhibits high quality with a large grain size. Moreover, the device efficiency is high and hysteresis is negligible. First, ITO substrate (PET/ITO and Glass/ITO) were cleaned with deionized water and detergent of suitable ratio in ultrasonic cleaner for 20 min. The ultrasonic cleaner was bought from Skymen Cleaning Technology Shenzhen Co., Ltd. The cleaner was manufactured in Shenzhen, China and the machine model is KQ-100E. Secondly, the processed ITO was treated with ethanol for twenty minutes in the ultrasonic cleaner.

Experiments
Thirdly, the ITO was treated with the mixed solvents of isopropyl alcohol, acetone, and deionized water for twenty minutes in the ultrasonic cleaner. The volume ratio of isopropyl alcohol, acetone and deionized water is 1:1:1. Finally, the ITO was treated with UV light for fifteen minutes in UV light cleaner, removing the organics. The UV light cleaner was purchased from Shenzhen Huiwo Technology Co., Ltd. The model was BZS250GF-TC and the machine was manufactured in Shenzhen, China.

Perovskite Absorption Layer Preparing Process
DMSO and DMF were used to disperse the PbI 2 powder to make 600 mg/mL precursor solution. The volume ratio of DMSO and DMF is 0.05:0.95. The CH 3 NH 3 I was dispersed in isopropyl alcohol to make the precursor solution. The precursor solution concentration is 70 mg/mL. Secondly, the lead diiodide layer was formed on the SnO 2 film by spin-coating lead diiodide precursor solution. The speed is 1500 rpm/s and the time is thirty seconds. Next, the perovskite layer was formed by spin-coating CH 3 NH 3 I precursor solution on the as-formed PbI 2 layer. The speed is 1500 rpm/s and the time is thirty seconds. Finally, the film was annealed at 150 degrees Celsius for twenty minutes.

Hole Transport Layer (HTL) Preparing Process
First, 5 mL of acetonitrile solution has been used to disperse 260 mg of Li-TFST. Then, 2 mL of chlorobenzene solution has been used to disperse 35 µL of TBP and 14.46 mg of Spiro-OMeTAD powder. The Spiro-OMeTAD precursor solution was finally formed by mixing the two solutions together. The HTL was made by spin-coating the prepared precursor solution on perovskite substrate. The speed is 3000 rpm/s and the time is thirty seconds. and annealing the as-formed film at sixty degrees Celsius for five minutes. Finally, the Spiro-OMeTAD layer was treated with oxidization for ten hours at ambient atmosphere.

Counter Electrode Preparation Process
The samples prepared by each layer were put into the vacuum thermal evaporation coating equipment to begin the preparation of the gold electrode. The gold (Au) electrode was prepared with a thermal vapor deposition method in high vacuum environment. The device area was defined by the overlap of ITO electrode and Au electrode, which was 0.2 cm × 0.2 cm.

Characterization
The morphological images of perovskite films were obtained with scanning electron microscope (SEM, Zeiss SIGMA, Kanagawa, Japan). A solar simulator equipped with air-mass (AM) 1.5 sunlight has been applied to conduct J-V test. The solar simulator model is Sol 3A and the machine is manufactured in Oriel, Newport, RI, USA. X-ray diffractometer was applied to obtain X-ray diffraction (XRD, D8 Focus, Bruker, Dresden, Germany) data of perovskite films formed on ITO/SnO 2 substrates. All the measurements are conducted in ambient conditions.

Results and Discussion
We have investigated the effects of different SnO 2 colloidal dispersion solution concentrations on perovskite films and the scanning electron microscope (SEM) measurements were performed. Figure 1 demonstrates the top-view SEM images of the perovskite layers spin-coated on an SnO 2 layer of different colloidal dispersion solution concentrations. As demonstrated in Figure 1a, the SnO 2 colloidal dispersion solution concentration was 10 wt.%, some holes appeared on the surface of perovskite. When the colloidal dispersion solution concentration of SnO 2 was 6.67 wt.% (Figure 1b), the holes on the surface of perovskite layer were significantly reduced. As shown in Figure 1d,e, when the SnO 2 colloidal dispersion concentration solution was 4 wt.% and 3.33 wt.%, Some holes appear on the surface of the perovskite and the grain size becomes larger. As demonstrated in Figure 1c, when the SnO 2 colloidal dispersion solution concentration was 5 wt.%, the perovskite film quality was significantly improved compared with the previous concentrations of 10, 6.67, 4, and 3.33 wt.%. From the perovskite film SEM images, we could observe that the films exhibit the best quality with larger grain size and few holes when the concentration of SnO 2 colloidal dispersion solution was 5 wt.%.
The XRD patterns for the perovskite films formed on the various SnO 2 ETLs are shown in Figure 2. The XRD patterns in Figure S1 [29]. The peak intensity of the (110) crystal plane of perovskite is the highest when the concentration of SnO 2 colloidal dispersion solution was 5 wt.%, this is because at the concentration the surface of SnO 2 has a certain roughness, as shown in Figure S2 (Supplementary Materials), which assists in the stress release during the annealing of the perovskite layer, resulting in a good quality of perovskite crystallization with almost no holes.
The relative intensity ratio of the PbI 2 peak and (110) of MAPbI 3 with the XRD patterns are shown in Table 1. By comparing the relative intensity ratio of the PbI 2 peak and (110) of MAPbI 3 , we determined that the relative intensity ratio gradually increased with the decrease of SnO 2 colloidal dispersion solution concentrations, indicating that there was PbI 2 remaining in the perovskite layers. The reported work shows that moderate excessive PbI 2 can passivate the defects of the perovskite film [31]. When the SnO 2 colloidal dispersion solution concentration was 5 wt.%, the content of PbI 2 is relatively moderate, which indicates that the quality of perovskite layer is relatively good. This experiment result is consistent with that of SEM.   Figure 3 demonstrates the J-V curves for PSCs prepared with various SnO 2 colloidal dispersion solution concentrations. When the SnO 2 colloidal dispersion solution concentration is 5 wt.%, the overall performance of the device is relatively higher, especially the open-circuit voltage (V OC ) and short-circuit current density (J SC ) is much higher than those of others. Moreover, PSCs prepared with 5 wt.% SnO 2 colloidal dispersion solution have a low hysteresis index (HI) based on high efficiency.  Table 2 shows the parameters of performance calculated from J-V curves in Figure 3. It can be seen from Table 2 that different SnO 2 colloidal dispersion solution concentrations have a significant impact on the PSC performance. When the SnO 2 colloidal dispersion concentration solution is 5 wt.%, the J SC of the PSC is 22.57 mA/cm 2 , the V OC is 1.03 V, and the filling factor (FF) is 73%. On this basis, the device has relatively high PCE and low HI. We have obtained the statistics of PCE of PSCs to justify the effect of concentration of SnO 2 colloidal dispersion solution on the device performance ( Figure 4). The data in Figure 4 represent reverse scanned data. Figure 4 demonstrates that the parameters of performance of 5 wt.% SnO 2 colloidal dispersion solution concentration-based devices exhibit relatively higher repeatability than the others. By optimizing the ETL preparation process, we obtained a relatively good process (SnO 2 colloidal dispersion concentration of 5 wt.%). Referring to this process, we prepared a flexible perovskite solar cell (F-PSC). Figure 5 demonstrates the reverse scanning J-V curves for an F-PSC prepared under optimized conditions. Table 3 shows the parameters of device performance calculated from reverse scanning J-V curves. It can be concluded from Table 3 that the flexible device exhibits a PCE of 7%. Based on the initial F-PSC, we made two bends with a curvature radius of 5 mm. After two bends with a curvature radius of 5 mm, the PCE is reduced to 46% of the initial F-PSC (The inset shows the photograph of F-PSC). This is caused by the brittle nature of polycrystalline perovskite layer, making the film vulnerable to damage under destructive bending experiments, which would deteriorate the device performance. However, the F-PSC can still work compared to rigid substrate-based PSCs.  To compare the morphology of perovskite film deposited on rigid substrate (Glass/ITO) and flexible substrate (PET/ITO), we conducted the SEM characterization of the perovskite film based on the same preparation process of the flexible substrate, as shown in Figure 6. We could find that the perovskite film is uneven and the grains are small. Moreover, grain stacks can be clearly observed from Figure 6, and the quality of the perovskite film prepared is poor. Perovskite films prepared on rigid substrate (as shown in Figure 1c) exhibit a better quality than flexible substrate. We speculate that the poor quality of the flexible substrate itself, resulting in the poor perovskite film quality. Figure 7 and Table 4 show the performance of PSCs prepared on two substrates. We could find that the PCSs prepared on rigid substrate obtained the V OC of 1.03 V and the PCE of 16.82%. The F-PSC prepared on a flexible substrate (PET/ITO) by using the same preparation process exhibited the V OC of 0.95 V and the PCE of 7.00%. It indicates that the quality of the perovskite layer prepared on a rigid substrate exhibits far higher than that on a flexible substrate, so the optimization of the preparation process of F-PSC remains to be studied.

Conclusions
In summary, we studied the influence of the concentration of SnO 2 colloidal dispersion solution on the topography of perovskite layer and performance of PSCs. We found that the perovskite film prepared with 5 wt.% SnO 2 colloidal dispersion solution has a large grain size and the morphology of perovskite layer could be modulated by controlling the SnO 2 colloidal dispersion solution concentration. Moreover, when our SnO 2 colloidal dispersion solution concentration was 5 wt.%, the content of PbI 2 is relatively moderate. The defects of perovskite films could be passivated with suitable PbI 2 residue, which favors the performance of PSCs prepared under high temperature. The PCE and FF of the device can reach 16.82% and 73%, respectively. We also found that the HI could be modulated by adjusting the concentration of SnO 2 colloidal dispersion solution. Based on high efficiency (16.82%), the HI was relatively low (0.23) when the SnO 2 colloidal dispersion solution concentration was 5 wt.%. Our work has provided a reference for low-temperature and high-performance optoelectronic device preparation process in the future.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest:
The authors declare no conflict of interest.