Optimizing the Aspect Ratio of Nanopatterned Mesoporous TiO2 Thin-Film Layer to Improve Energy Conversion Efficiency of Perovskite Solar Cells

The energy conversion efficiency (ECE) (η), current density (Jsc), open-circuit voltage (Voc), and fill factor (ff) of perovskite solar cells were studied by using the transmittance of a nanopatterned mesoporous TiO2 (mp-TiO2) thin-film layer. To improve the ECE of perovskite solar cells, a mp-TiO2 thin-film layer was prepared to be used as an electron transport layer (ETL) via the nanoimprinting method for nanopatterning, which was controlled by the aspect ratio. The nanopatterned mp-TiO2 thin-film layer had a uniform and well-designed structure, and the diameter of nanopatterning was 280 nm. The aspect ratio was controlled at the depths of 75, 97, 127, and 167 nm, and the perovskite solar cell was fabricated with different depths. The ECE of the perovskite solar cells with the nanopatterned mp-TiO2 thin-film layer was 14.50%, 15.30%, 15.83%, or 14.24%, which is higher than that of a non-nanopatterned mp-TiO2 thin-film layer (14.07%). The enhancement of ECE was attributed to the transmittance of the nanopatterned mp-TiO2 thin-film layer that is due to the improvement of the electron generation. As a result, better electron generation affected the electron density, and Jsc increased the Voc, and ff of perovskite solar cells.


Introduction
Renewable energy is a sustainable method for generating affordable electricity with a lower carbon footprint compared with other conventional resources [1]. Owing to environmental problems, the solar cell industry has developed rapidly in recent years. Therefore, many countries have invested in projects in order to discover better materials for solar cells [2,3]. Solar cells are devices that transform sunlight into electricity, and they have a very low environmental impact [2,4]. They have been a topic of interest in the green energy field, owing to the possibility of solar cells becoming an alternative energy source that can replace conventional energy sources [5]. Previous studies have aimed to improve the efficiency of solar cells [2,5,6]. These studies have achieved important improvements, and the dominance of silicon solar cells has been established in the market, exhibiting an energy conversion efficiency (ECE) higher than 20% [2]. However, as perovskite solar 2. Results and Discussion 2.1. Overall Scheme of the Process for Fabricating Perovskite Solar Cells with a Nanopatterned Mesoporous TiO 2 (mp-TiO 2 ) Thin-Film Layer Figure 1 shows the overall scheme of the fabrication process of perovskite solar cells with a nanopatterned mesoporous TiO 2 (mp-TiO 2 ) thin-film layer. Figure 1A shows the preparation of the perfluoropolyether (PFPE) mold from a Si master. The PFPE resin was coated on the Si master and covered with a polyethylene terephthalate (PET) film as a flexible backplane, as shown in Figure 1A(a). After UV irradiation, the PFPE resin underwent polymerization for hardening, as shown in Figure 1A(b). The PET film contained the PFPE mold, which was obtained from the PFPE resin using polymerization and was separated from the Si master, as shown in Figure 1A(c). Figure 1B shows the fabrication of perovskite solar cells with a nanopatterned mp-TiO 2 thin-film layer. The mp-TiO 2 thin-film layer was prepared on a fluorine-doped tin oxide (FTO) glass from a solution of TiO 2 nanoparticles (NPs) by using a spin coater, as shown in Figure 1B(a). It was then imprinted by using the PFPE mold for nanopatterning on the mp-TiO 2 thin-film layer as shown in Figure 1B(b). To imprint the mp-TiO 2 thin-film layer using the PFPE mold, the mp-TiO 2 thin-film layer was pre-baked at 70 • C for 1 min to evaporate the solvent. After imprinting, the mp-TiO 2 thin-film layer was sintered. The perovskite and hole transport materials, such as spiro-OMeTAD, were coated on the nanopatterned mp-TiO 2 thin-film layer, as shown in Figure 1B(c,d). Au was evaporated by using a metal evaporator, as shown in Figure 1B(e). The device structure was FTO/nanopatterned mp-TiO 2 /perovskite/sprio-OMeTAD/Au.  Figure 1 shows the overall scheme of the fabrication process of perovskite solar cells with a nanopatterned mesoporous TiO2 (mp-TiO2) thin-film layer. Figure 1A shows the preparation of the perfluoropolyether (PFPE) mold from a Si master. The PFPE resin was coated on the Si master and covered with a polyethylene terephthalate (PET) film as a flexible backplane, as shown in Figure 1A(a). After UV irradiation, the PFPE resin underwent polymerization for hardening, as shown in Figure 1A(b). The PET film contained the PFPE mold, which was obtained from the PFPE resin using polymerization and was separated from the Si master, as shown in Figure 1A(c). Figure 1B shows the fabrication of perovskite solar cells with a nanopatterned mp-TiO2 thin-film layer. The mp-TiO2 thinfilm layer was prepared on a fluorine-doped tin oxide (FTO) glass from a solution of TiO2 nanoparticles (NPs) by using a spin coater, as shown in Figure 1B(a). It was then imprinted by using the PFPE mold for nanopatterning on the mp-TiO2 thin-film layer as shown in Figure 1B(b). To imprint the mp-TiO2 thin-film layer using the PFPE mold, the mp-TiO2 thinfilm layer was pre-baked at 70 °C for 1 min to evaporate the solvent. After imprinting, the mp-TiO2 thin-film layer was sintered. The perovskite and hole transport materials, such as spiro-OMeTAD, were coated on the nanopatterned mp-TiO2 thin-film layer, as shown in Figure 1B(c,d). Au was evaporated by using a metal evaporator, as shown in Figure 1B(e). The device structure was FTO/nanopatterned mp-TiO2/perovskite/sprio-OMeTAD/Au.     mixed with a UV-curable resin, and the mixture was dropped on the Si master. The PET film covered the mixture as a flexible backbone. After the PET film was rolled and irradiated by UV, the PFPE mold was prepared. The pore size was 250 nm, and the interpore distance was 250 nm. The PFPE mold on the PET film was used for nanopatterning the mp-TiO 2 thin-film layer via the nanoimprinting method. mixed with a UV-curable resin, and the mixture was dropped on the Si master. The PET film covered the mixture as a flexible backbone. After the PET film was rolled and irradiated by UV, the PFPE mold was prepared. The pore size was 250 nm, and the interpore distance was 250 nm. The PFPE mold on the PET film was used for nanopatterning the mp-TiO2 thin-film layer via the nanoimprinting method.  Figure 3 shows the FE-SEM and optical images of non-nanopatterned and nanopatterned mp-TiO2 thin-film layers. The non-nanopatterned mp-TiO2 thin-film layer shows only TiO2 NPs (Figure 3a). The nanopatterned mp-TiO2 thin-film layer shows uniform and well-designed pores on the surface, such as the moth-eye nanostructure, which was formed using the nanoimprinting method. The pore size was 280 nm, and the interpore distance was 220 nm, as shown in Figure 3b.   Figure 4 shows the FE-SEM images of the mp-TiO2 thin-film layer for different nanopatterning depths. The diameter of the pore was 280 nm, and the depths of the pore were 75, 97, 127, and 167 nm. The aspect ratio defined by the depth and diameter of the nanopatterned mp-TiO2 thin-film layer was controlled by different depths. Perovskite solar cells were fabricated with the nanopatterned mp-TiO2 thin-film layer at different depths to optimize the transmittance by using the aspect ratio.  Figure 3 shows the FE-SEM and optical images of non-nanopatterned and nanopatterned mp-TiO 2 thin-film layers. The non-nanopatterned mp-TiO 2 thin-film layer shows only TiO 2 NPs (Figure 3a). The nanopatterned mp-TiO 2 thin-film layer shows uniform and well-designed pores on the surface, such as the moth-eye nanostructure, which was formed using the nanoimprinting method. The pore size was 280 nm, and the interpore distance was 220 nm, as shown in Figure 3b. mixed with a UV-curable resin, and the mixture was dropped on the Si master. The PET film covered the mixture as a flexible backbone. After the PET film was rolled and irradiated by UV, the PFPE mold was prepared. The pore size was 250 nm, and the interpore distance was 250 nm. The PFPE mold on the PET film was used for nanopatterning the mp-TiO2 thin-film layer via the nanoimprinting method.  Figure 3 shows the FE-SEM and optical images of non-nanopatterned and nanopatterned mp-TiO2 thin-film layers. The non-nanopatterned mp-TiO2 thin-film layer shows only TiO2 NPs (Figure 3a). The nanopatterned mp-TiO2 thin-film layer shows uniform and well-designed pores on the surface, such as the moth-eye nanostructure, which was formed using the nanoimprinting method. The pore size was 280 nm, and the interpore distance was 220 nm, as shown in Figure 3b.   Figure 4 shows the FE-SEM images of the mp-TiO2 thin-film layer for different nanopatterning depths. The diameter of the pore was 280 nm, and the depths of the pore were 75, 97, 127, and 167 nm. The aspect ratio defined by the depth and diameter of the nanopatterned mp-TiO2 thin-film layer was controlled by different depths. Perovskite solar cells were fabricated with the nanopatterned mp-TiO2 thin-film layer at different depths to optimize the transmittance by using the aspect ratio.  Figure 4 shows the FE-SEM images of the mp-TiO 2 thin-film layer for different nanopatterning depths. The diameter of the pore was 280 nm, and the depths of the pore were 75, 97, 127, and 167 nm. The aspect ratio defined by the depth and diameter of the nanopatterned mp-TiO 2 thin-film layer was controlled by different depths. Perovskite solar cells were fabricated with the nanopatterned mp-TiO 2 thin-film layer at different depths to optimize the transmittance by using the aspect ratio.

XRD Images of Perovskite Solar Cells with a Nanopatterned Mesoporous TiO2 (mp-TiO2) Thin-Film Layer
The crystallinity values of the nanopatterned mp-TiO2 thin-film layer and perovskite were confirmed by X-ray diffraction (XRD) patterns as shown in Figure 6. The nanopatterned mp-TiO2 thin-film layer shows the anatase phases from

XRD Images of Perovskite Solar Cells with a Nanopatterned Mesoporous TiO2 (mp-TiO2) Thin-Film Layer
The crystallinity values of the nanopatterned mp-TiO2 thin-film layer and perovskite were confirmed by X-ray diffraction (XRD) patterns as shown in Figure 6. The nanopatterned mp-TiO2 thin-film layer shows the anatase phases from

XRD Images of Perovskite Solar Cells with a Nanopatterned Mesoporous TiO 2 (mp-TiO 2 ) Thin-Film Layer
The crystallinity values of the nanopatterned mp-TiO 2 thin-film layer and perovskite were confirmed by X-ray diffraction (XRD) patterns as shown in Figure 6. The nanopatterned mp-TiO 2 thin-film layer shows the anatase phases from

Transmittance Data of Perovskite Solar Cells with a Nanopatterned Mesoporous TiO2 (mp-TiO2) Thin-Film Layer
The transmittance of the nanopatterned mp-TiO2 thin-film layer with different aspect ratios of diameter 280 nm and depths 75 nm, 97 nm, 127 nm, and 167 nm were measured using UV-vis spectra, as shown in Figure 7A. The schematic function of the transmittance and reflectance of the nanopatterned mp-TiO2 thin-film layer is shown in Figure 7B. The transmittance of the non-nanopatterned mp-TiO2 thin-film layer is higher than that of the nanopatterned mp-TiO2 thin-film layer at a wavelength of 380 nm. However, the transmittance of the non-nanopatterned mp-TiO2 thin-film layer is lower than that of the nanopatterned mp-TiO2 thin-film layer for wavelengths higher than 380 nm. The nanopatterned mp-TiO2 thin-film layer increases surface roughness and changes optical effects. Furthermore, it induces changes in the refractive index that affect the transmittance of light through a medium. The decrease in the refractive index increases the transmittance and vice-versa. For 75 nm, 97 nm and 127 nm depths, the transmittance increased, but it decreased at 167 nm depth [51]. This result may suggest that the refractive index of the nanopatterned depths of 75 nm, 97 nm, and 127 nm decreases but increases at 167 nm. This result shows that the aspect ratio of the nanopatterned structure affects the transmittance. In addition, if the angle of the reflected/diffracted light is greater than the critical angle, it causes a total internal reflection due to a decrease in the value of the refractive index of the TiO2 layer to air (refractive index of TiO2 (≈2.1), FTO (≈1.9), Glass (≈1.5), Air (=1.0)). The light is then reflected towards the mp-TiO2, and thus the transmittance to the perovskite layers increases [52]. Considering the perovskite materials, whose absorbance wavelength varies from 300 nm to 800 nm, the increment in transmittance of the nanopatterned mp-TiO2 thin-film layer affects the Jsc, Voc, ff, and the ECE of perovskite solar cells.

Transmittance Data of Perovskite Solar Cells with a Nanopatterned Mesoporous TiO 2 (mp-TiO 2 ) Thin-Film Layer
The transmittance of the nanopatterned mp-TiO 2 thin-film layer with different aspect ratios of diameter 280 nm and depths 75 nm, 97 nm, 127 nm, and 167 nm were measured using UV-vis spectra, as shown in Figure 7A. The schematic function of the transmittance and reflectance of the nanopatterned mp-TiO 2 thin-film layer is shown in Figure 7B. The transmittance of the non-nanopatterned mp-TiO 2 thin-film layer is higher than that of the nanopatterned mp-TiO 2 thin-film layer at a wavelength of 380 nm. However, the transmittance of the non-nanopatterned mp-TiO 2 thin-film layer is lower than that of the nanopatterned mp-TiO 2 thin-film layer for wavelengths higher than 380 nm. The nanopatterned mp-TiO 2 thin-film layer increases surface roughness and changes optical effects. Furthermore, it induces changes in the refractive index that affect the transmittance of light through a medium. The decrease in the refractive index increases the transmittance and vice-versa. For 75 nm, 97 nm and 127 nm depths, the transmittance increased, but it decreased at 167 nm depth [51]. This result may suggest that the refractive index of the nanopatterned depths of 75 nm, 97 nm, and 127 nm decreases but increases at 167 nm. This result shows that the aspect ratio of the nanopatterned structure affects the transmittance. In addition, if the angle of the reflected/diffracted light is greater than the critical angle, it causes a total internal reflection due to a decrease in the value of the refractive index of the TiO 2 layer to air (refractive index of TiO 2 (≈2.1), FTO (≈1.9), Glass (≈1.5), Air (=1.0)). The light is then reflected towards the mp-TiO 2 , and thus the transmittance to the perovskite layers increases [52]. Considering the perovskite materials, whose absorbance wavelength varies from 300 nm to 800 nm, the increment in transmittance of the nanopatterned mp-TiO 2 thin-film layer affects the J sc , V oc , ff, and the ECE of perovskite solar cells.

Characterization of Perovskite Solar Cells with and without the Nanopatterned Mesoporous TiO2 (mp-TiO2) Thin-Film Layer
The I-V characteristics of perovskite solar cells were characterized under one condition of the sun, as shown in Figure 8A. The results of current density (Jsc), open-circuit voltage (Voc), fill factor (ff), and ECE ( ) are summarized in Table 1. The ECE of perovskite solar cells without the nanopatterned mp-TiO2 thin-film layer is 14.07%, while for perovskite solar cells with the nanopatterned mp-TiO2 thin-film layer, it increased from 14.50% to 15.83%, with an increase in depth from 75 nm to 127 nm. This is due to an increment in Jsc from 23.50 mA/cm 2 to 24.62 mA/cm 2 , as the nanopatterned mp-TiO2 thin-film layer leads to a better transmittance that enhances the electron generation. Voc increases with an increase of Jsc. This is because, as the electron generation increases, as shown in Figure 8B, the electron density and Fermi level of the nanopatterned mp-TiO2 thin-film layer increase. Voc is determined by the conduction band of the nanopatterned mp-TiO2 thin-film layer and the valence band of perovskite. Thus, Voc also increases from 0.869 V to 0.896 V.

Characterization of Perovskite Solar Cells with and without the Nanopatterned Mesoporous TiO 2 (mp-TiO 2 ) Thin-Film Layer
The I-V characteristics of perovskite solar cells were characterized under one condition of the sun, as shown in Figure 8A. The results of current density (J sc ), open-circuit voltage (V oc ), fill factor (ff ), and ECE (η) are summarized in Table 1. The ECE of perovskite solar cells without the nanopatterned mp-TiO 2 thin-film layer is 14.07%, while for perovskite solar cells with the nanopatterned mp-TiO 2 thin-film layer, it increased from 14.50% to 15.83%, with an increase in depth from 75 nm to 127 nm. This is due to an increment in J sc from 23.50 mA/cm 2 to 24.62 mA/cm 2 , as the nanopatterned mp-TiO 2 thin-film layer leads to a better transmittance that enhances the electron generation. V oc increases with an increase of J sc . This is because, as the electron generation increases, as shown in Figure 8B, the electron density and Fermi level of the nanopatterned mp-TiO 2 thin-film layer increase. V oc is determined by the conduction band of the nanopatterned mp-TiO 2 thin-film layer and the valence band of perovskite. Thus, V oc also increases from 0.869 V to 0.896 V. To confirm the results, the perovskite solar cells with and without a nanopatterned mp-TiO 2 thin-film layer were characterized by the electrochemical impedance spectroscopy (EIS) and photoluminescence (PL) spectroscopy, as shown in Figure 9A,B. In electrochemical impedance spectroscopy, the recombination resistance (R rec ) of perovskite solar cells with the nanopatterned mp-TiO 2 thin-film layer was suppressed compared to those without the nano-patterned mp-TiO 2 thin-film layer. The decrement in recombination was affected by the increment in electron density and it was also affected the increment in V oc . Thus, the V oc of perovskite solar cells with nanopatterned mp-TiO 2 thin-film layer increased. Moreover, in PL spectra, the perovskite materials with the nanopatterned mp-TiO 2 thin-film layer quenched more compared to those without the nanopatterned mp-TiO 2 thin-film layer. The generated electrons from perovskite solar cells with nanopatterned mp-TiO 2 thin-film layer quenched more, which means that the generated electrons were transferred from perovskite to the nanopatterned mp-TiO 2 thin-film layer effectively before recombination. It means that the transferred electrons increase the electron density of the nanopatterned mp-TiO 2 thin-film layer. There is also a change in the Fermi level of the nanopatterned mp-TiO 2 thin-film layer due to its conduction band. Thus, the V oc changed because V oc is determined by the conduction band of TiO 2 and the valence band of perovskite materials. However, the ECE (η) of perovskite solar cells with a 167 nm depth nanopatterned mp-TiO 2 thin-film layer decreased to 14.24% with a decrease in V oc and ff. From these results, the aspect ratio of nanopattern between the diameter and depth should be optimized for better ECE of perovskite solar cells.

Characterization of Perovskite Solar Cells with and without the Nanopatterned Mesoporous TiO2 (mp-TiO2) Thin-Film Layer
The I-V characteristics of perovskite solar cells were characterized under one condition of the sun, as shown in Figure 8A. The results of current density (Jsc), open-circuit voltage (Voc), fill factor (ff), and ECE ( ) are summarized in Table 1. The ECE of perovskite solar cells without the nanopatterned mp-TiO2 thin-film layer is 14.07%, while for perovskite solar cells with the nanopatterned mp-TiO2 thin-film layer, it increased from 14.50% to 15.83%, with an increase in depth from 75 nm to 127 nm. This is due to an increment in Jsc from 23.50 mA/cm 2 to 24.62 mA/cm 2 , as the nanopatterned mp-TiO2 thin-film layer leads to a better transmittance that enhances the electron generation. Voc increases with an increase of Jsc. This is because, as the electron generation increases, as shown in Figure 8B, the electron density and Fermi level of the nanopatterned mp-TiO2 thin-film layer increase. Voc is determined by the conduction band of the nanopatterned mp-TiO2 thin-film layer and the valence band of perovskite. Thus, Voc also increases from 0.869 V to 0.896 V.  To confirm the results, the perovskite solar cells with and without a nanopatterned  To demonstrate the repeatability of perovskite solar cells with and without a nano patterned mp-TiO2 thin-film layer, the 36 perovskite solar cells with and without a nano patterned mp-TiO2 thin-film layer were measured, and the data are illustrated in the his tograms as shown in Figure 10. The pattern of the IPCE spectra of perovskite solar cells is similar to the transmittance of a nanopatterned mp-TiO2 thin-film layer because transmittance is affected by Jsc, Voc, ff and the ECE of perovskite solar cells. The intensity of the IPCE spectra on perovskite sola cells with the non-nanopatterned mp-TiO2 thin-film layer is higher than that with the na nopatterned mp-TiO2 thin-film layer at a wavelength of 430 nm. However, the intensity of the IPCE spectra on perovskite solar cells with the non-nanopatterned mp-TiO2 thin film layer is lower than that with the nanopatterned mp-TiO2 thin-film layer at wave lengths ranging from 430 nm to 800 nm. Higher intensity of the IPCE spectra indicates To demonstrate the repeatability of perovskite solar cells with and without a nanopatterned mp-TiO 2 thin-film layer, the 36 perovskite solar cells with and without a nanopatterned mp-TiO 2 thin-film layer were measured, and the data are illustrated in the histograms as shown in Figure 10. To demonstrate the repeatability of perovskite solar cells with and without a nanopatterned mp-TiO2 thin-film layer, the 36 perovskite solar cells with and without a nanopatterned mp-TiO2 thin-film layer were measured, and the data are illustrated in the histograms as shown in Figure 10. The pattern of the IPCE spectra of perovskite solar cells is similar to the transmittance of a nanopatterned mp-TiO2 thin-film layer because transmittance is affected by Jsc, Voc, ff and the ECE of perovskite solar cells. The intensity of the IPCE spectra on perovskite solar cells with the non-nanopatterned mp-TiO2 thin-film layer is higher than that with the nanopatterned mp-TiO2 thin-film layer at a wavelength of 430 nm. However, the intensity of the IPCE spectra on perovskite solar cells with the non-nanopatterned mp-TiO2 thinfilm layer is lower than that with the nanopatterned mp-TiO2 thin-film layer at wavelengths ranging from 430 nm to 800 nm. Higher intensity of the IPCE spectra indicates The pattern of the IPCE spectra of perovskite solar cells is similar to the transmittance of a nanopatterned mp-TiO 2 thin-film layer because transmittance is affected by J sc , V oc , ff, and the ECE of perovskite solar cells. The intensity of the IPCE spectra on perovskite solar cells with the non-nanopatterned mp-TiO 2 thin-film layer is higher than that with the nanopatterned mp-TiO 2 thin-film layer at a wavelength of 430 nm. However, the intensity of the IPCE spectra on perovskite solar cells with the non-nanopatterned mp-TiO 2 thin-film layer is lower than that with the nanopatterned mp-TiO 2 thin-film layer at wavelengths ranging from 430 nm to 800 nm. Higher intensity of the IPCE spectra indicates that more electrons are generated from 430 nm to 800 nm. More electron generation improves the electron density on the mp-TiO 2 thin-film layer, which, in turn, increases J sc , ff, and V oc . Thus, the ECE of perovskite solar cells with the nanopatterned mp-TiO 2 thin-film layer increases. The integrated J sc can be determined from Equation (1).
Here, F is the Faraday constant, E eλ is the solar spectral irradiance, and N A is the Avogadro constant. J sc is calculated by Equation (1), as shown in Figure 11. The integrated J sc is matched with the measured J sc of perovskite solar cells with the non-nanopatterned or nanopatterned mp-TiO 2 thin-film layer. that more electrons are generated from 430 nm to 800 nm. More electron generation improves the electron density on the mp-TiO2 thin-film layer, which, in turn, increases Jsc, ff, and Voc. Thus, the ECE of perovskite solar cells with the nanopatterned mp-TiO2 thin-film layer increases. The integrated Jsc can be determined from Equation (1).
Here, F is the Faraday constant, Eeλ is the solar spectral irradiance, and NA is the Avogadro constant. Jsc is calculated by Equation (1), as shown in Figure 11. The integrated Jsc is matched with the measured Jsc of perovskite solar cells with the non-nanopatterned or nanopatterned mp-TiO2 thin-film layer.

Preparation of Perfluoropolyether (PFPE) Mold
The UV-curable resin was prepared by mixing the PFPE resin (Fluorolink MD700, Solvay Solexis, Milan, Italy) with 3% w/w photoinitiator (Darocur 1173, Sigma-Aldrich, St. Louis, MO, USA). The UV-curable resin was dropped on the patterned silicon substrate (i.e., Si master) and covered with polyethylene terephthalate (PET) film as a backplane. To spread the UV-curable resin between the PET film and Si master, the PET film was rolled over the resin and cross-linked by UV irradiation at 365 nm for 5 min. After curing, the PFPE mold was detached from the Si master.

Preparation of Perfluoropolyether (PFPE) Mold
The UV-curable resin was prepared by mixing the PFPE resin (Fluorolink MD700, Solvay Solexis, Milan, Italy) with 3% w/w photoinitiator (Darocur 1173, Sigma-Aldrich, St. Louis, MO, USA). The UV-curable resin was dropped on the patterned silicon substrate (i.e., Si master) and covered with polyethylene terephthalate (PET) film as a backplane. To spread the UV-curable resin between the PET film and Si master, the PET film was rolled over the resin and cross-linked by UV irradiation at 365 nm for 5 min. After curing, the PFPE mold was detached from the Si master.

Preparation of Nanopatterned mp-TiO 2 Thin-Film Layer
The fluorine-doped tin oxide (FTO) glass was washed with water, ethanol, and acetone. The compact TiO 2 precursor, which comprises 1 mL of titanium diisopropoxide bis(acetylacetonate) (75 wt% in isopropanol, Sigma-Aldrich, St. Louis, MO, USA) in 6 mL of butanol (99%, Daejung Chemicals, Siheung-si, Korea) was spin-coated on FTO at 4000 rpm for 20 s, followed by annealing at 250 • C for 10 min. This process was repeated twice. To form the mesoporous TiO 2 (mp-TiO 2 ) layer, 1 g of TiO 2 paste (Ti-Nanoxide T/SP, Solaronix, Aubonne, Switzerland) was diluted in 12 mL of anhydrous ethanol (99.9%, Daejung Chemicals, Siheung-si, Korea). The dispersion solution was spin-coated on a compact TiO 2 /FTO. After prebaking on a hotplate at 70 • C for 1 min to evaporate the solvent slightly, the PFPE mold was placed on the mp-TiO 2 /compact TiO 2 /FTO substrate. The substrate was heated at 70 • C, and then a pressure of 2 bar was applied for 5 min. After removing the mold, the nanopatterned mp-TiO 2 /compact TiO 2 /FTO substrate was annealed at 500 • C for 1 h.

Fabrication of Perovskite Solar Cells
The perovskite precursor solutions were spin-coated on the nanopatterned mp-TiO 2 layer by using the hot-casting technique at 90 • C and then annealed at 130 • C for 1 h. The hole transport materials were prepared by stirring 73. at room temperature for 2 h. The spiro-OMeTAD was spin-coated at 4000 rpm for 60 s, followed by drying in a glovebox overnight. The above procedures were implemented inside a dried air-filled glovebox at a dew point of approximately −40 • C. Finally, the Au electrode was deposited by the thermal evaporator.

Characterization of Perovskite Solar Cells
The morphology and thickness were characterized by field emission scanning electron microscope (FE-SEM, Hitachi S4700). The photovoltaic properties of perovskite solar cells were characterized by the solar simulator (Newport Corp., Model 94022A, Irvine CA, USA) at the Future Energy Convergence Core Center (FECC) with a source meter (Keithley Instruments Inc., Keithley 2400, Cleveland, OH, USA) under AM 1.5 illumination (100 mW/cm 2 ). The incident photon to electron conversion efficiencies were obtained by the solar cell QE measurement system (PV Measurements Inc., QEX7, Point Roberts, WA, USA). The formation of crystalline TiO 2 and perovskite was confirmed using an X-ray diffractometer (Rigaku Corp., D/max-2500, Tokyo, Japan). The transmittance measurements were obtained using a UV-vis spectrophotometer (Jasco International Co., Ltd., V-730, Tokyo, Japan). The electrochemical impedance spectroscopy (EIS) was measured at 250 mV applied bias under 1 sun light (1 Hz-300 kHz). The photoluminescence (PL) was measured with an excitation light source at 325 nm and an applied laser power of 50 mW.

Conclusions
We prepared nanopatterned mesoporous TiO 2 (mp-TiO 2 ) thin-film layers with different aspect ratios (the diameter was 280 nm while the depths were varied: 75 nm, 97 nm, 127 nm, or 167 nm) by using the imprinting method. Perovskite solar cells were then fabricated with the nanopatterned mp-TiO 2 thin-film layers with different aspect ratios for better energy conversion efficiency (ECE). The ECE of perovskite solar cells without the nanopatterned mp-TiO 2 thin-film layer is 14.07%. With the nanopatterned mp-TiO 2 thin-film layer, the ECE of the perovskite solar cells increased from 14.50% (depth: 75 nm) to 15.83% (depth: 127 nm) because the transmittance of the nanopatterned mp-TiO 2 thin-film layer improved at a wavelength greater than 400 nm that affected the electron generation. Greater electron generation increased the J sc and electron density on the active layer of perovskite solar cells, which affected ff. Moreover, an increase in electron density shifted the Fermi level of TiO 2 to a more negative potential that reflects to the increment of V oc . The IPCE spectra also shows similar tendency with the transmittance spectra of the nanopatterned mp-TiO 2 thin-film layer at a wavelength above 430 nm, which indicates that high transmittance is related to electron generation on perovskite solar cells. However, when the perovskite solar cells were fabricated with a 127 nm depth nanopatterned mp-TiO 2 thin-film layer, the ECE decreased as the depth of nanopattern went beyond a certain depth. The non-optimized nanopatterned mp-TiO 2 (75 nm, 97 nm, and 167 nm) shows relatively low efficiency due to the decrease in transmittance. Therefore, in our study, the aspect ratio with a width of 280 nm and a depth of 127 nm was optimized. Our research suggests that the optimized nanopatterned mp-TiO 2 thin-film layer can be controlled by the imprinting method and can be applied on perovskite solar cells for better ECE with better transmittance. This study expects the optimized nanopatterned mp-TiO 2 thin-film layer to be implemented in other types of solar cells. New features can be developed using the best aspect ratio, which can be optimized by using machine learning algorithms and AI for a better performance.