TiO2/SnO2 Bilayer Electron Transport Layer for High Efficiency Perovskite Solar Cells

The electron transport layer (ETL) has been extensively investigated as one of the important components to construct high-performance perovskite solar cells (PSCs). Among them, inorganic semiconducting metal oxides such as titanium dioxide (TiO2), and tin oxide (SnO2) present great advantages in both fabrication and efficiency. However, the surface defects and uniformity are still concerns for high performance devices. Here, we demonstrated a bilayer ETL architecture PSC in which the ETL is composed of a chemical-bath-deposition-based TiO2 thin layer and a spin-coating-based SnO2 thin layer. Such a bilayer-structure ETL can not only produce a larger grain size of PSCs, but also provide a higher current density and a reduced hysteresis. Compared to the mono-ETL PCSs with a low efficiency of 16.16%, the bilayer ETL device features a higher efficiency of 17.64%, accomplished with an open-circuit voltage of 1.041 V, short-circuit current density of 22.58 mA/cm2, and a filling factor of 75.0%, respectively. These results highlight the unique potential of TiO2/SnO2 combined bilayer ETL architecture, paving a new way to fabricate high-performance and low-hysteresis PSCs.


Introduction
The high-efficiency, low-cost and facile fabrication process of halide perovskite solar cells (PSCs) have attracted tremendous attention in the field of photovoltaics in the past decade [1][2][3][4][5] and been regarded as the most promising substitute for traditional silicon (Si) and copper indium gallium selenide (CIGS) solar cells [6][7][8]. The sandwich structure of hybrid organic-inorganic based PSCs includes the electron transport layer (ETL), perovskite absorber layer, hole transport layer (HTL) and electrodes. Among them, ETL and HTL are used for the electron and hole extraction, respectively. However, the Spiro-OMeTAD are widely used as HTL in PSCs because of the simple synthesis, high carrier mobility and suitable valance band. The HTL are always fabricated by spin-coating on the top of a perovskite absorber layer with a dense and uniform film. In contrast, the ETL in PSCs is usually fabricated in a planar and/or mesoporous structure under the perovskite absorber layer [9][10][11]. The surface quality of ETL can substantially influence the deposition of perovskite film. Therefore, the electron transport layer and the corresponding interface of ETL/perovskite are significantly important parts to fabricate high-quality PSCs. Titanium dioxide (TiO 2 ) and/or tin oxide (SnO 2 ) thin films have been extensively investigated as an effective ETL in the PSCs, which can be fabricated by several different methods such as spin-coating, sputtering and chemical bath deposition (CBD) [12][13][14][15][16], to pursue a higher performance device.
Due to the facile planar configuration of PSCs, fabricating uniform, and compact ETL thin layer, it is imperative to pursue high performance. The conventional spin-coating Nanomaterials 2023, 13, 249 2 of 10 method shows a facile and efficient way to fabricate the TiO 2 -ETL. However, the uneven distribution of TiO 2 nanoparticles result in the carrier accumulation between perovskite (PVSK) and the ETL interface and an insufficient carrier extraction, leading to a low efficiency of resultant device [17,18]. Moreover, the large hysteresis of TiO 2 -ETL also impedes the further application of TiO 2 in the PSCs [19]. Alternatively, SnO 2 presents a reduced hysteresis, high carrier mobility and good energy level towards perovskite, which can greatly improve the performance of PSCs [20][21][22]. For example, You et al. proposed SnO 2 as a planar ETL in the PSCs, which not only reduces the energy barrier between ETL/PVSK, but also reduces the hysteresis of devices, resulting in a high performance PSC with a champion PCE of 20.5% [21]. However, uniformity of SnO 2 nanoparticles is still a concern for the device fabrication because of its uneven distribution by spin-coating technique. Therefore, high-quality ETL plays a crucial role in the fabrication of devices, which paves a promising way for high-efficiency PSCs. To address this issue, Xu et al. introduced a bilayer ETL of TiO 2 /ZnO thin layers into PSCs, which produces a compact interfacial layer to avoid direct contact between the FTO substrate and PVSK, leading to a reduced carrier accumulation at ETL/PVSK interface [23].
In this work, we propose a bilayer of ETLs that is composed of a CBD TiO 2 layer and a spin-coated SnO 2 layer. The presence of the SnO 2 thin layer on the top surface of CBD TiO 2 film can provide a higher current density and reduce the hysteresis of PSCs simultaneously. In addition, the diffusion of the K ion from SnO 2 can significantly improve the crystallinity of grains in the perovskite films. On the basis of this bilayer strategy, a higher power conversion efficiency (PCE) of 17.64% was achieved in comparison with the mono-TiO 2 ETL based PSCs with a PCE of 16.16%.
Device Fabrication: The cleaned fluorine-doped tin oxide (FTO) substrates are treated using UV-ozone for 60 min. Then, the TiO 2 thin layer was prepared by using the CBD method and the SnO 2 thin layer was fabricated with spin-coating technologies, as shown in Figure 1. First, 2 M aqueous TiCl 4 mother solution was prepared by dropping TiCl 4 into distilled water. During the preparation, the mother solution was continuously stirred at a low temperature of around 0 • C. The as-prepared TiCl 4 mother solution was stored in the refrigerator (<10 • C). Second, the as-prepared TiCl 4 mother solution was diluted to a 0.2 M TiCl 4 solution. The cleaned FTO substrates were placed vertically in the glassware. Then, 300 mL of 0.2 M TiCl 4 solution was poured into the glassware. The glassware was put into an oven with a temperature of 75 • C. After 1 h heating, the glassware was taken out followed by rinsing the FTO substrates several times using distilled water. Finally, the FTO substrates were annealed at a high temperature of 450 • C for 30 min. The FTO substrates were washed by the acetone, distilled water, and ethyl alcohol for 20 min, respectively. Before the deposition of TiO 2 thin films, the FTO substrates are treated by using UV-ozone for 60 min. SnO 2 films were prepared by spin-coating Alfa Aesar SnO 2 (diluted by H 2 O to 3%) at a speed of 3500 rpm for 30 s. The perovskite films were deposited by a two-step spin-coating method. Specifically, 1.35 M PbI 2 and 0.0675 M CsI were dissolved in organic solvent (DMF/DMSO = 19:1). The PbI 2 precursor solution was stirred at a temperature of 70 • C for 60 min. The mixed MAFA based organic cation precursor solution was prepared by dissolving 200 mg FAI, 100 mg MAI, 25 mg MABr and 25 mg MACl dissolved in 5 mL isopropanol. The PbI 2 precursor solution was first spin-coated at a speed of 3000 rpm for 30 s. The MA/FA cation solution was spin-coated at 3000 rpm for 30s. After annealing at Nanomaterials 2023, 13, 249 3 of 10 150 • C for 10 min, the perovskite film of Cs 0.05 FA 0.54 MA 0.41 Pb(I 0.98 Br 0.02 ) 3 was obtained. The hole transport layer of the spiro-OMeTAD film was deposited by spin-coating the spiro-OMeTAD solution at a speed of 3500 rpm for 25 s. Finally, 80 nm Au film was deposited as a counter electrode by thermal evaporation.
ing the spiro-OMeTAD solution at a speed of 3500 rpm for 25 s. Finally, 80 nm Au film was deposited as a counter electrode by thermal evaporation.
Device Characterization: The diffraction data of perovskites are collected by using a Bruker D8 Discover diffractometer (Bruker AXS) from 10° to 60°. Surface and cross-section morphology images are recorded by a scanning electron microscope (SEM) (Helios NanoLab G3). The TRPL results were collected by using the Hamamatsu equipment which can provide an excitation wavelength of 450 nm. The photoluminescence (PL) spectra were acquired by a JASCO FP-8500 spectrometer with an excitation wavelength of 450 nm. The current-voltage (J-V) measurements were performed under one sun illumination (AM1.5G, 100 mW/cm 2 ) by using a Keithley 2420. The devices were test by using a metal shadow mask with a dimension of 0.3 × 0.3 cm 2 . The EQE spectra of the devices were characterized by using Oriel IQE 200 equipment.   Device Characterization: The diffraction data of perovskites are collected by using a Bruker D8 Discover diffractometer (Bruker AXS) from 10 • to 60 • . Surface and crosssection morphology images are recorded by a scanning electron microscope (SEM) (Helios NanoLab G3). The TRPL results were collected by using the Hamamatsu equipment which can provide an excitation wavelength of 450 nm. The photoluminescence (PL) spectra were acquired by a JASCO FP-8500 spectrometer with an excitation wavelength of 450 nm. The current-voltage (J-V) measurements were performed under one sun illumination (AM1.5G, 100 mW/cm 2 ) by using a Keithley 2420. The devices were test by using a metal shadow mask with a dimension of 0.3 × 0.3 cm 2 . The EQE spectra of the devices were characterized by using Oriel IQE 200 equipment.

Results and Discussion
Figure 2a-d shows the top-view SEM images of the perovskite films fabricated on the TiO 2 and TiO 2 /SnO 2 substrates, which clearly shows a larger grain size of perovskite thin film based on the TiO 2 /SnO 2 substrates, compared with that on the TiO 2 substrates, with an average value changing from~380 nm to~540 nm, which can be verified by the statistics of perovskite grain size based on the TiO 2 and TiO 2 /SnO 2 substrates, as presented in Figure 2e,f. As is well-known, the commercial SnO 2 colloid precursor is stabilized by incorporating potassium hydroxide (KOH) [24]. The presence of K ion in the SnO 2 will diffuse into the perovskite thin film during the annealing process, which greatly enhances the crystallinity of perovskite grains, and reduces the hysteresis of resultant devices [25][26][27].
Nanomaterials 2023, 13, x FOR PEER REVIEW 4 of 10 an average value changing from ~380 nm to ~540 nm, which can be verified by the statistics of perovskite grain size based on the TiO2 and TiO2/SnO2 substrates, as presented in Figure 2e,f. As is well-known, the commercial SnO2 colloid precursor is stabilized by incorporating potassium hydroxide (KOH) [24]. The presence of K ion in the SnO2 will diffuse into the perovskite thin film during the annealing process, which greatly enhances the crystallinity of perovskite grains, and reduces the hysteresis of resultant devices [25][26][27].  Furthermore, the phase structure of perovskite thin film deposited on the TiO 2 and TiO 2 /SnO 2 substrates was investigated by X-ray diffraction (XRD), as presented in Figure 3a. The increase of XRD intensity (on the TiO 2 /SnO 2 substrate) verifies that the improved crystallinity of perovskites is accomplished with high absorption in a shortwavelength region (as shown in Figure 3b). Nanomaterials 2023, 13, x FOR PEER REVIEW 5 of 10 Furthermore, the phase structure of perovskite thin film deposited on the TiO2 and TiO2/SnO2 substrates was investigated by X-ray diffraction (XRD), as presented in Figure  3a. The increase of XRD intensity (on the TiO2/SnO2 substrate) verifies that the improved crystallinity of perovskites is accomplished with high absorption in a short-wavelength region (as shown in Figure 3b). In addition, the steady-state photoluminescence (PL) and time-resolved photoluminescence (TRPL) experiments were carried out to investigate carrier transport behavior. As seen in Figure 3c,d, the faster PL quenching of the perovskite thin film on the TiO2/SnO2 substrate indicates an enhanced electron extraction capability [28]. Moreover, the lifetimes of the corresponding perovskite thin films were fitted by a biexponential decay function [29,30]. The lifetime of the TiO2/SnO2-based sample is 15.4 ns, which is shorter than that of the TiO2-based sample (22.2 ns), indicating a faster carrier extraction from the perovskite thin film to TiO2/SnO2 electron transport layer [31]. Figure 4a,b shows the cross-section SEM images of devices fabricated on TiO2 and TiO2/SnO2 substrates. The uniform and dense perovskite absorber layers not only ensure the light harvest, but also effectively impede the carrier recombination in the devices. The current density-voltage (J-V) curves of the devices were measured under standard AM 1.5 G illumination and are shown in Figure 5a and Table 1  In addition, the steady-state photoluminescence (PL) and time-resolved photoluminescence (TRPL) experiments were carried out to investigate carrier transport behavior. As seen in Figure 3c,d, the faster PL quenching of the perovskite thin film on the TiO 2 /SnO 2 substrate indicates an enhanced electron extraction capability [28]. Moreover, the lifetimes of the corresponding perovskite thin films were fitted by a biexponential decay function [29,30]. The lifetime of the TiO 2 /SnO 2 -based sample is 15.4 ns, which is shorter than that of the TiO 2 -based sample (22.2 ns), indicating a faster carrier extraction from the perovskite thin film to TiO 2 /SnO 2 electron transport layer [31]. Figure 4a,b shows the cross-section SEM images of devices fabricated on TiO 2 and TiO 2 /SnO 2 substrates. The uniform and dense perovskite absorber layers not only ensure the light harvest, but also effectively impede the carrier recombination in the devices. The current density-voltage (J-V) curves of the devices were measured under standard AM 1.5 G illumination and are shown in Figure 5a and Table 1 [14,25]. The EQE spectra of the corresponding devices were presented in Figure 5b. The improved EQE in the short wavelength in terms of TiO 2 /SnO 2 -based device indicates faster carrier extraction and reduced recombination at the TiO 2 /SnO 2 /PVSK interface [32]. Similarly, the enhanced EQE at the long wavelength region also suggests that reduced defects and carrier recombination in the perovskite bulk film, which can be explained by the enlarged grain size and improved crystallinity of the perovskite grains [32]. As a result, the integrated J SC from EQE of the TiO 2 /SnO 2 -based device is 21.59 mA/cm 2 , which is higher than that of the TiO 2 based device (21.17 mA/cm 2 ). Furthermore, the TiO 2 /SnO 2 -based device exhibited a stable output (under initial maximum power point (MPP) voltage) with a PCE of 17.65%. In contrast, the TiO 2 based solar cell shows a poor output under MPP, yielding a low PCE of 15.74% (Figure 5c). More importantly, the hysteresis (hysteresis index (HI) = PCE RS /PCE FS ) of the TiO 2 /SnO 2 -based devices is also reduced, compared to TiO 2 -based devices [32][33][34]. The HI of the TiO 2 -based PSC is 1.56, which is decreased to 1.15 by incorporating SnO 2 into devices to construct the TiO 2 /SnO 2 bilayer ETL. Compared to TiO 2 based devices with a large hysteresis of 1.51, the improved efficiency and reduced HI of 1.18 for TiO 2 /SnO 2based PSCs indicates the bilayer ETL can improve the reproducible fabrication and the device performance. The improved efficiency of TiO2/SnO2-based solar cells can be attributed to a higher crystallinity of perovskite grains, which enhances light capture and reduces the defects at grain boundaries [14,25]. The EQE spectra of the corresponding devices were presented in Figure 5b. The improved EQE in the short wavelength in terms of TiO2/SnO2based device indicates faster carrier extraction and reduced recombination at the TiO2/SnO2/PVSK interface [32]. Similarly, the enhanced EQE at the long wavelength region also suggests that reduced defects and carrier recombination in the perovskite bulk film, which can be explained by the enlarged grain size and improved crystallinity of the perovskite grains [32]. As a result, the integrated JSC from EQE of the TiO2/SnO2-based device is 21.59 mA/cm 2 , which is higher than that of the TiO2 based device (21.17 mA/cm 2 ). Furthermore, the TiO2/SnO2-based device exhibited a stable output (under initial maximum power point (MPP) voltage) with a PCE of 17.65%. In contrast, the TiO2 based solar cell shows a poor output under MPP, yielding a low PCE of 15.74% (Figure 5c). More importantly, the hysteresis (hysteresis index (HI) = PCERS/PCEFS) of the TiO2/SnO2-based devices is also reduced, compared to TiO2-based devices [32][33][34]. The HI of the TiO2-based PSC is 1.56, which is decreased to 1.15 by incorporating SnO2 into devices to construct the TiO2/SnO2 bilayer ETL. Compared to TiO2 based devices with a large hysteresis of 1.51, the improved efficiency and reduced HI of 1.18 for TiO2/SnO2-based PSCs indicates the bilayer ETL can improve the reproducible fabrication and the device performance.

Conclusions
In summary, we developed a bilayer electron transport layer by combining CBD-TiO2 and spin-coated SnO2 in the perovskite solar cells. The TiO2/SnO2 bilayer ETLs provide not only a compact electron transport layer, but also accelerate the carrier transport in the

Conclusions
In summary, we developed a bilayer electron transport layer by combining CBD-TiO 2 and spin-coated SnO 2 in the perovskite solar cells. The TiO 2 /SnO 2 bilayer ETLs provide not only a compact electron transport layer, but also accelerate the carrier transport in the solar cells. Furthermore, the presence of K ion from SnO 2 can greatly improve the crystallinity of perovskite thin film and significantly reduce the hysteresis of resultant devices. Compared with the TiO 2 -based solar cells, the TiO 2 /SnO 2 -based solar cells demonstrate a higher PCE of 17.64% and a lower hysteresis index. These results highlight the potential fabrication of the TiO 2 /SnO 2 bilayer electron transport layers and will be a beneficial strategy to fabricate a high-quality perovskite thin film solar cell.

Data Availability Statement:
The data is available on reasonable request from the corresponding author.