Effect of UV-Light Treatment on Efficiency of Perovskite Solar Cells (PSCs)

We employed ultra-violet (UV) light treatment on the TiO2 layer prior to coating the perovskite layer to improve the solar conversion efficiency of perovskite solar cells (PSCs). A laboratory-made UV treatment system was equipped with various UV light sources (8 W power; maximum wavelengths of 254, 302, and 365 nm). The UV light treatment improved the power conversion efficiency (PCE) while coating the uniformity layer and removing impurities from the surface of cells. After the PSCs were exposed to UV light, their PCE developed approximately 10% efficiency; PBI2 decreased without changing the structure.


Introduction
In recent years, renewable energy sources have attracted increasing attention due to oil depletion and other environmental problems. Several studies have been conducted to address the problems. Converting solar energy into heat, fuel, and electricity is one of the most effective way of obtaining renewable energy [1]. Therefore, silicon-based solar cells now dominate the photovoltaic market due to the higher efficiencies of up to 25% compared to other solar cell devices. Owing to the high production cost of silicon-based solar cells, interest in organic material-based solar cell systems is increasing [2]. Compared to inorganic photovoltaic cells, organic material-based solar cells such as organic solar cells, dye-sensitized solar cells (DSSCs), and perovskite solar cells (PSCs) have low production costs, are light-weight, and have flexible device fabrication [3][4][5][6]. Despite their excellent advantages, organic material-based solar cells have low conversion efficiencies and unstable small-size devices (~1 cm 2 ) [7]. In the majority of works that have been reported, research groups have focused on higher efficiency in solar cell devices. The efficiency has been improved mainly by determining new perovskite material composition and modified film deposition processes [8].
Currently, PSCs have many layers, such as a transparent electrode (fluorine-doped thin oxide, FTO), a compact electron transport layer (ETL), a perovskite layer, a hole transport layer (HTL), and a metallic electrode (Au, Ag, and Cu). The active layers between two electrodes are coated by a two-step method. Despite performing the PSC fabrication process in a well-controlled glove box under low-level humidity, moisture, and oxygen, defects and poor adhesion between layers appeared [9,10]. Moreover, after a compact TiO 2 layer and a mesoporous TiO 2 layer were deposited on the FTO electrode, the temperature increased to over 450 • C. Therefore, to achieve a high energy conversion performance of PSCs, the layers should be smooth and uniform [11][12][13].
Ultraviolet (UV) irradiation has been mainly used for photocatalytic applications on the surface or for improving the surface coating energy. UV light is classified as A (400-315 nm), B (315-280 nm), and C (280-200 nm). UV radiation has shorter wavelengths than visible light. Unfortunately, all UV light is not helpful to humans, and UV-A and UV-B are related to skin cancer [14]. Among them, UV-C Energies 2020, 13 is used to remove impurities on the surface by induced ozone (O 3 ) in the air. Some research groups have tried to enhance the power conversion efficiency (PCE) of solar cells [15,16]. Until now, other surface science fields have reported on whether oxide materials, such as TiO 2 , SnO 2 , and ZnO, have photocatalysis effects in a range of light with various wavelengths [17,18]. Work comparing various ultraviolet light has not yet been in conducted in the field of PSCs. Moreover, based on previous studies, we confirmed that UV light irradiation (UV-C) treatment improved the surface of a dye-sensitized solar cell (DSSC) [19]. However, there have been very few studies on other UV light sources (A and B).
In this study, we investigated the effect of the UV light source and the light irradiation time of the PSC fabrication process to enhance its PCE.

Materials and Preparation of Solar Cells
Chemical reagents were purchased from Sigma-Aldrich chemical company and were used without further purification. The two-step spin-coating method was performed for the PSC layer and thermal evaporation for the Ag and Au electrode (Kojundo Co. Ltd, 99.95%). The FTO substrate (thickness < 2.2 mm, sheet resistance < 8 ohm/sq) was purchased from Wooyang GMS Company. Prior to coating the chemical solutions, we cleaned the FTO substrate with acetone, ethanol, and deionized water using ultrasonication. The substrates were then blown by nitrogen gas to remove any remaining impurities or contamination. Three kinds of solutions were required to form the PSC layer-TiO 2 paste, CH 3 NH 3 PbI 3 , and spirobifluorene (Spiro-MeOTAD). First, laboratory-made TiO 2 paste was used to form a compact layer (C-L) and a mesoporous layer (M-L). The TiO 2 paste was prepared with TiO 2 powder (anatase < 25 nm, 99.7%, Sigma-Aldrich). Second, CH 3 NH 3 PbI 3 was formed during the spin-coating process with CH 3 NH 3 I and PbI 2 solutions, according to a previous study [20]. Finally, a spiro-MeOTAD solution was dissolved in a mixture of chlorobenzene, 4-tert-butyl pyridine and lithium bis(trifluoromethanesulfonyl)imide (Li-TFSI) (520 mg LI-TSFI in 1 ml acetonitrile, Sigma-Aldrich) and used without any additional processing.
The schematics of the PCS fabrication procedure are shown in Figure 1. Prior to the fabrication of the solar cells, fluorine-doped tin oxide (FTO)-coated glass substrates (sheet resistance: 8 ohm/sq, films thickness: 400 nm) were cleaned with acetone, ethanol, and de-ionized water in a bath-type sonicator for 20 min each. Any impurities on the cleaned substrates were removed by nitrogen gas before drying in an oven at 100 • C for 30 min.  light treatment (DL-S = distance from light to sample).

Measurements
The structure and morphology of the prepared samples were observed using field-emission scanning electron microscopy (FE-SEM, Hitachi S-4500) and X-ray diffraction (XRD, Rigaku corporation SmartLab) using Cu-Kα radiation (λ = 1.54056 Å, 9 kW) in the 2 θ range of 20-80°. The contact angles of the water droplets on the solid surface were measured using a contact angle meter. The conversion efficiencies of the prepared solar cell devices were measured a using current-voltage (J-V) solar simulator (K3400, McScience Company) adjusted to AM 1.5G (100 mW/cm 2 ).

Results and Discussion
Firstly, to investigate whether UV treatment on the film surface influenced the structural properties and surface morphology, we measured the water contact angle (WCA) of an m-TiO2/c- The PSC fabrication procedures were carried out in a nitrogen gas filled groove box. The compact TiO 2 (c-TiO 2 ) layer was coated at 3000 rpm for 30 s at 120 • C for 10 min on a hot-plate, separately. Then, the mesoporous TiO 2 (m-TiO 2 ) layer was deposited on the c-TiO 2 -coated FTO substrate at 3000 rpm for 30 s at 450 • C for 10 min in a furnace. The CH 3 NH 3 PbI 3 layer was formed by combining CH 3 NH 3 I and PbI 2 solutions on the m-TiO 2 /c-TiO 2 film by the spin-coating method [21][22][23]. After the spiro-MeOTAD layer was coated at 4000 rpm for 30 s, the samples were loaded on a thermal evaporator machine. Approximately 100 nm of Ag was thermally evaporated on the spiro-MeOTAD-coated film in a vacuum (under 10 -6 torr).

UV Surface Treatment of Samples
Laboratory-made UV irradiation equipment consisted of an aluminum reflector and 8 W UV light lamps with maximum wavelengths of 254, 302, and 365 nm. More detailed equipment and setup conditions are shown in Figure 1. Prior to depositing the CH 3 NH 3 PbI 3 layer on the m-TiO 2 /c-TiO 2 -coated FTO substrate (sample size: 20 mm × 20 mm), the UV light treatment was carried out on the surface at room temperature in a sealed box.

Measurements
The structure and morphology of the prepared samples were observed using field-emission scanning electron microscopy (FE-SEM, Hitachi S-4500) and X-ray diffraction (XRD, Rigaku corporation SmartLab) using Cu-Kα radiation (λ = 1.54056 Å, 9 kW) in the 2 θ range of 20-80 • . The contact angles of the water droplets on the solid surface were measured using a contact angle meter. The conversion efficiencies of the prepared solar cell devices were measured a using current-voltage (J-V) solar simulator (K3400, McScience Company) adjusted to AM 1.5G (100 mW/cm 2 ).

Results and Discussion
Firstly, to investigate whether UV treatment on the film surface influenced the structural properties and surface morphology, we measured the water contact angle (WCA) of an m-TiO 2 /c-TiO 2 /FTO substrate and images of samples coated with the CH 3 NH 3 PbI 3 layer, as shown in Figure 2. The samples were treated for 1 h by UV light A, B, and C, separately. As shown in Figure 2, compared to the reference sample without surface treatment (approximately 20 • ), the WCA of all samples with UV light surface treatment decreased due to the hydrophilic surface of the substrate. Among them, it was thought that UV-C light was the most effective to obtain a high-energy, for example, hydrophilic surface [24]. As light has a longer wavelength, transmittance increased and energy was low. It is assumed that developing surfaces have more effect.   Therefore, we selected UV-C light with the shortest wavelength and investigated the properties of samples treated under other conditions. Subsequently, we reduced 60 min to 30 min, and the change in the surface wettability of the film surface was confirmed by the WCA, as shown in Figure 3a,b. Despite the reduction of the surface treatment time, we were able to obtain a hydrophilic surface. The median of three measurements of the WCA values was calculated as the median of three determinations on recorded images of the samples. Following UV-C treatment of the m-TiO 2 /c-TiO 2 /FTO substrate, the WCA of the surface decreased from 17 • to approximately 0 • . The UV-C light from specific wavelength ranges was introduced to enhance the surface energy, and its surface temporarily became hydrophilic. Moreover, owing to the enhanced hydrophilicity of the sample surface and the strong hydrogen bonds to water molecules, the dye-liquid solution could uniformly cover the m-TiO 2 /c-TiO 2 /FTO substrate in the spin coating process after the liquid solutions were dropped on the sample [25][26][27].
Energies 2020, 13, 1069 4 of 11 Therefore, we selected UV-C light with the shortest wavelength and investigated the properties of samples treated under other conditions. Subsequently, we reduced 60 min to 30 min, and the change in the surface wettability of the film surface was confirmed by the WCA, as shown in Figures  3a,b. Despite the reduction of the surface treatment time, we were able to obtain a hydrophilic surface. The median of three measurements of the WCA values was calculated as the median of three determinations on recorded images of the samples. Following UV-C treatment of the m-TiO2/c-TiO2/FTO substrate, the WCA of the surface decreased from 17° to approximately 0°. The UV-C light from specific wavelength ranges was introduced to enhance the surface energy, and its surface temporarily became hydrophilic. Moreover, owing to the enhanced hydrophilicity of the sample surface and the strong hydrogen bonds to water molecules, the dye-liquid solution could uniformly cover the m-TiO2/c-TiO2/FTO substrate in the spin coating process after the liquid solutions were dropped on the sample [25][26][27]. We observed the sample surface after dropping the CH3NH3PbI3 solution onto it and the photo images of the samples are shown in Figures 3c,d. After the m-TiO2/c-TiO2/FTO substrate was exposed to UV-C light, we obtained samples with better film-coating uniformity. Further, the sample had no defects unlike samples without UV-C light surface treatment (film defects: dotted lines in samples, as shown in Figure 3d.
We investigated the effect of the crystal structure on the UV light surface treatment.  We observed the sample surface after dropping the CH 3 NH 3 PbI 3 solution onto it and the photo images of the samples are shown in Figure 3c,d. After the m-TiO 2 /c-TiO 2 /FTO substrate was exposed to UV-C light, we obtained samples with better film-coating uniformity. Further, the sample had no defects unlike samples without UV-C light surface treatment (film defects: dotted lines in samples, as shown in Figure 3d. We investigated the effect of the crystal structure on the UV light surface treatment. Figure 4a (100 nm) were sequentially stacked on the substrate as shown in Figure 4b. The thickness of the layers was consistent with the configuration previously reported in a similar fabrication procedure [20].
Energies 2020, 13, 1069 5 of 11 nm) were sequentially stacked on the substrate as shown in Figure 4b. The thickness of the layers was consistent with the configuration previously reported in a similar fabrication procedure [20].  The optical absorbance spectra of coated PbI 2 are shown in Figure 5. After the UV-C treatment of the sample, the height of the absorbance peak of PbI 2 detected at around 481 nm decreased. The UV-C treatment was conducted by a combination of two wavelengths, i.e., 84 and 254 nm. 1) Ozone (O 3 ) gas was generated from oxygen in the air by light irradiation at 184 nm. 2) Impurity particles such as organic contaminants attached to the surface were removed by oxidation with O 3 after irradiation by light at 254 nm. Furthermore, Saekow et al. [28] reported that after UV-C treatment, the light could generate O atoms as well as OH-groups from the H 2 O molecules in the air. In addition, the O atoms on the surface of the TiO 2 layer were replaced by OH − groups. Owing to an increase in the number of O − and O 2− species, more Ti 3+ ions were produced compared with Ti 4+ ions. The increase in the Energies 2020, 13, 1069 6 of 11 number of Ti 3+ ions on the surface could attract and combine the carboxylic groups with the PbI 2 , and the PbI 2 concentration decreased as the absorbance peak of PbI 2 was reduced, as shown in Figure 5. The optical absorbance spectra of coated PbI2 are shown in Figure 5. After the UV-C treatment of the sample, the height of the absorbance peak of PbI2 detected at around 481 nm decreased. The UV-C treatment was conducted by a combination of two wavelengths, i.e., 84 and 254 nm. 1) Ozone (O3) gas was generated from oxygen in the air by light irradiation at 184 nm. 2) Impurity particles such as organic contaminants attached to the surface were removed by oxidation with O3 after irradiation by light at 254 nm. Furthermore, Saekow et al. [28] reported that after UV-C treatment, the light could generate O atoms as well as OH-groups from the H2O molecules in the air. In addition, the O atoms on the surface of the TiO2 layer were replaced by OH − groups. Owing to an increase in the number of O − and O 2− species, more Ti 3+ ions were produced compared with Ti 4+ ions. The increase in the number of Ti 3+ ions on the surface could attract and combine the carboxylic groups with the PbI2, and the PbI2 concentration decreased as the absorbance peak of PbI2 was reduced, as shown in Figure 5.
We prepared 4 reference samples (without UV light surface treatment) and 18 UV-light treated samples (UV-A, UV-B, and UV-C). We used Ag as the metal electrode. When the wavelength was shorter, the efficiency of PSCs improved, as shown in Figure 6a. We prepared 4 reference samples (without UV light surface treatment) and 18 UV-light treated samples (UV-A, UV-B, and UV-C). We used Ag as the metal electrode. When the wavelength was shorter, the efficiency of PSCs improved, as shown in Figure 6a. The optical absorbance spectra of coated PbI2 are shown in Figure 5. After the UV-C treatment of the sample, the height of the absorbance peak of PbI2 detected at around 481 nm decreased. The UV-C treatment was conducted by a combination of two wavelengths, i.e., 84 and 254 nm. 1) Ozone (O3) gas was generated from oxygen in the air by light irradiation at 184 nm. 2) Impurity particles such as organic contaminants attached to the surface were removed by oxidation with O3 after irradiation by light at 254 nm. Furthermore, Saekow et al. [28] reported that after UV-C treatment, the light could generate O atoms as well as OH-groups from the H2O molecules in the air. In addition, the O atoms on the surface of the TiO2 layer were replaced by OH − groups. Owing to an increase in the number of O − and O 2− species, more Ti 3+ ions were produced compared with Ti 4+ ions. The increase in the number of Ti 3+ ions on the surface could attract and combine the carboxylic groups with the PbI2, and the PbI2 concentration decreased as the absorbance peak of PbI2 was reduced, as shown in Figure 5.
We prepared 4 reference samples (without UV light surface treatment) and 18 UV-light treated samples (UV-A, UV-B, and UV-C). We used Ag as the metal electrode. When the wavelength was shorter, the efficiency of PSCs improved, as shown in Figure 6a. As can be seen from the EQE data of Figure 6b, the longer wavelength of the UV light source resulted in smaller quantum efficiency, which is expected to be the result of the recombination at the surface. A short-wavelength indicates that the grain size does not increase significantly but that the roughness increases resulting in a lowered fill factor. Longer the wavelengths indicated lower efficiency due to the decrease in the current density. Among the samples with UV surface treatment, the efficiency of that with UV-C treatment was the highest. It was confirmed that a UV-C light sources have the greatest effect on improving the performance of PSCs. Therefore, we investigated the effect of light exposure time on PSCs.
UV light at 245 nm was irradiated onto the TiO 2 layer on the FTO substrate for 10, 20, and 30 min. An absorbance graph, Figure 7 was obtained, and the absorbance of the whole region, except the PbI 2 peak, was further increased compared with the conventional PbI 2 peak. The TiO 2 layer is Energies 2020, 13, 1069 7 of 11 cleaned by UV-C light irradiation and it has an appropriate structure, which could be allowed to be better perovskite formation on the surface. Moreover, the formation of perovskite, which enables well bonded at the perovskite/TiO 2 interface, normally transfers electrons to the TiO 2 layer, disturbing the open-circuit voltage (V OC ). The EQE measurements were employed to provide independent corroboration of the short-circuit current-density (J SC ) values collected with the solar simulator. The most significant influence on PCE was given by improvement in the J SC values, as a consequence of overall improved EQE behavior for all samples.
surface. A short-wavelength indicates that the grain size does not increase significantly but that the roughness increases resulting in a lowered fill factor. Longer the wavelengths indicated lower efficiency due to the decrease in the current density. Among the samples with UV surface treatment, the efficiency of that with UV-C treatment was the highest. It was confirmed that a UV-C light sources have the greatest effect on improving the performance of PSCs. Therefore, we investigated the effect of light exposure time on PSCs.
UV light at 245 nm was irradiated onto the TiO2 layer on the FTO substrate for 10, 20, and 30 min. An absorbance graph, Figure 7 was obtained, and the absorbance of the whole region, except the PbI2 peak, was further increased compared with the conventional PbI2 peak. The TiO2 layer is cleaned by UV-C light irradiation and it has an appropriate structure, which could be allowed to be better perovskite formation on the surface. Moreover, the formation of perovskite, which enables well bonded at the perovskite/TiO2 interface, normally transfers electrons to the TiO2 layer, disturbing the open-circuit voltage (VOC). The EQE measurements were employed to provide independent corroboration of the short-circuit current-density (JSC) values collected with the solar simulator. The most significant influence on PCE was given by improvement in the JSC values, as a consequence of overall improved EQE behavior for all samples. The UV-C light exposure time increased the efficiency of the manufactured solar cell when the perovskite was applied with the VOC reduction, as confirmed in Table 1.  The UV-C light exposure time increased the efficiency of the manufactured solar cell when the perovskite was applied with the V OC reduction, as confirmed in Table 1. The V OC of the UV-C treated perovskite solar cell increased from 0.953 to 1.008 and the mean efficiency increased from 14.142% to 15.076%. Figure 8a shows the J-V characteristics of the fabricated cells after UV treatment under AM 1.5G illumination (100 mW/cm 2 ). Details of the device parameter values measured from the solar stimulator are listed in Table 1. The PSCs without UV radiation had a J SC of 24.239 mA/cm 2 , an F.F. of 61.370%, a V OC of 1.284 V, and a PCE of 14.142%. The PSC samples exposed to UV light had a J SC of 24.239 mA/cm -2 , an F.F. of 64.602%, a V OC of 1.008 V, and a PCE of 15.076%. The improvement in the values was caused by the variation of J SC .
The parameter J SC was attributed to the number of photo-generated electrons transferred to the perovskite layer through the TiO 2 layer. Therefore, it is important to improve the photovoltaic conversion efficiency of PSCs to improve the interfacial properties between TiO 2 and perovskite layers. The TiO 2 film surface irradiated by the internal resistance and charge transfer kinetics of prepared PSCs was measured using an electrochemical impedance spectroscopy (EIS) analyzer in light-dark conditions [29,30]. Figure 8b shows the EIS results of the samples prepared under the condition of UV-light irradiation. The diameter of the second semicircle is a measure of the recombination resistance (R rec ) in the PSC samples, which systematically increased after UV-light exposure, thereby indicating that the UV treatment was characterized by an increased recombination of UV light irradiation compared to no light-exposure. Higher RCT was induced for lower current loss and an increased F.F.
To confirm the effect of UV-C treatment on the PCSs, seven samples were selected, and their average PCE was calculated. After UV treatment, their maximum efficiency increased by 14.026% to 15.583%, as shown in Figure 8a. Energies 2020, 13, 1069 9 of 11 prepared PSCs was measured using an electrochemical impedance spectroscopy (EIS) analyzer in light-dark conditions [29,30].  Figure 8b shows the EIS results of the samples prepared under the condition of UV-light irradiation. The diameter of the second semicircle is a measure of the recombination resistance (Rrec) in the PSC samples, which systematically increased after UV-light exposure, thereby indicating that the UV treatment was characterized by an increased recombination of UV light irradiation compared to no light-exposure. Higher RCT was induced for lower current loss and an increased F.F.
To confirm the effect of UV-C treatment on the PCSs, seven samples were selected, and their average PCE was calculated. After UV treatment, their maximum efficiency increased by 14.026% to 15.583%, as shown in Figure 8a

Conclusions
In this study, we investigated the effect of UV-light source and irradiation time on the fabrication of PCSs. Their power energy conversion efficiency was investigated for a simulated solar irradiation power density of 100 mW/cm 2 (AM 1.5G). After UV treatment on the c-TiO/m-TiO2/FTO substrate, the film surface became hydrophilic; the perovskite layer was easily and uniformly deposited on the

Conclusions
In this study, we investigated the effect of UV-light source and irradiation time on the fabrication of PCSs. Their power energy conversion efficiency was investigated for a simulated solar irradiation power density of 100 mW/cm 2 (AM 1.5G). After UV treatment on the c-TiO/m-TiO 2 /FTO substrate, the film surface became hydrophilic; the perovskite layer was easily and uniformly deposited on the film. The PCE of PSCs treated by UV-C light increased more than those treated with UV-A or UV-B. Samples irradiated with UV light had a low series resistance, high recombination resistance inside the device, and a high PCE of approximately 10%, compared with PSCs without UV-light irradiation.