TiO₂ photoelectrodes for dye-sensitized solar cells (DSSCs) were fabricated by changing the thickness using screen printing method. The amount of dye adsorbed on TiO₂ increases as the thickness of the TiO₂ photoelectrode increases because of the increase in the dye molecule adsorption site. The performance of DSSCs increases up to a TiO₂ thickness of approximately 12 μm, indicating a tendency of dye adsorption that enables photocurrent generation. However, when TiO₂ is thicker than 12 μm, the TiO₂ films start to break, resulting in a decrease in the performance. The optimum thickness of the TiO₂ photoelectrode was 12 μm, which has a low charge transfer resistance

Keywords: dye-sensitized solar cells, photoelectrode, TiO₂, screen printing, dye adsorption

Dye-sensitized solar cells (DSSCs) are one of the most attractive solar cells because of their low manu- facturing cost, high conversion efficiency, and wide application compared with conventional silicon-based solar cells [1-4]. A conventional DSSC is composed of a photoelectrode, counter electrode, dye, and electrolyte [5]. The TiO₂ photoelectrode specifically generates photocurrent by absorbing the light, which is one of the most important factors of DSSCs. The material of the photoelectrode is an n-type semiconductor oxide with a wide energy band gap of 3–4 eV, such as TiO₂, ZnO, and SnO₂ [6-10]. TiO₂ is the most widely used ceramic material for photoelectrodes because of its large surface area in adsorbing dye molecules, suitable band edge levels for charge injection/extraction, and high chemical stability [11-16]. Various factors affect the open-circuit voltage (V_OC), short-circuit current density (J_SC), and fill factor (FF), which are parameters that determine the photoelectric conversion efficiency. In particular, the thickness of the TiO₂ photoelectrode is related to the transfer of electrons, diffusion, and generation of current, which are key factor that influence the overall perfor- mance of DSSCs [17-21, 22]. Therefore, the ceramic processing control to fabricate photoelectrode thin films is one of major issues to improve the performance of DSSCs.
In this study, TiO₂ photoelectrodes with various thicknesses were fabricated by a screen printing method. Commercial TiO₂ paste was also used because the efficiency and characteristics of the photoelectrode can be changed depending on the type of TiO₂ paste. The electrical properties and efficiency of the cells were measured, and the electrochemical process inside the DSSC was analyzed. Finally, the change in DSSC mechanism with the thickness of TiO₂ was confirmed using electrochemical analysis.

Fabrication of TiO₂ photoelectrode
Fluorine-doped tin oxide (FTO)-coated glass (TEC 8, Pilkington, UK) with a sheet resistance of 6–9 Ω/□ was sequentially cleaned in acetone (99.5%, Daejung Chemicals and Metals Co., Korea), 2-propanol (99+%, Alfa Aesar, USA) and deionized (DI) water for 10 min each and dried using an air gun. A screen printing method was used with TiO₂ paste (Ti-Nanoixde T/SP, Solaronix, Swiss) for mesoporous TiO₂ film deposition. First, TiO₂ paste was deposited using a screen printer. The screen-printed samples were dried at room temperature for 20 min and then dried again at 120 ^oC for 40 min in a dry oven to remove the organic solvent. As shown in Fig. 1(a), a series of steps constitutes one cycle, and the thickness is adjusted as the cycle repeats (1-6 cycles). After finishing the TiO₂ paste deposition, the films were sintered at 500 ^oC for 30 min in an ambient atmosphere box furnace and the heating rate was 10 ^oC/min (Fig. 1(b)). The final thickness of the TiO₂ film was controlled according to the number of screen printing cycles.

Assembly of DSSCs
After finishing the sintering step, the TiO₂ films were dipped with the prepared 0.3 mM (bis(tetrabutylammonium)-cis-di(thiocyanato)-N,N’–bis(4-carboxylato-4’-carboxylic acid-2,2-bipyridine) ruthenium (II) ethanol-based solution (N719, Solaronix, Swiss) in a dark field at room tempera- ture for 24 h. Before fabricating the counter electrode, two holes to inject the electrolyte were created in an indium-doped tin oxide (ITO)-coated glass (STN 10, UID, Korea) with a sheet resistance of 6-9 Ω/□, and then washed in the same process as for FTO/glass. Pt film as a counter electrode was deposited on the prepared ITO/glass using DC sputtering equipment (E-1030, Hitachi, Japan) at 25 mA and 30 s (the thickness of Pt: ~5 nm). For assembling the DSSC cells, the prepared photoelectrode and counter electrode were sealed using a 60 μm surlyn film (Solaronix, Switzerland). Electrolyte with the iodide/triiodide (I⁻/I³⁻) redox couple in acetonitrile solvent (AN-50, Solaronix) was injected through the holes. After finishing the electrolyte injection, the holes were enclosed by a cover glass or polyimide tape. The active area of TiO₂ photoelectrode and electrolyte injection was 0.25 cm² and 1.2 cm², respectively. The final DSSC cells were denoted as TSP×1, TSP×2, TSP×3, TSP×4, TSP×5, and TSP×6 according to the number of screen printing cycles. The schematic diagram of the structure and a photograph of the DSSC cells are shown in Fig. 2.

Characterization
The thickness of the TiO₂ films were analyzed by a high-resolution scanning electron microscope (HR-SEM, SU8230, Hitachi, Japan). Current-voltage (I–V) characteristics of the DSSCs were obtained under a standard illumination condition (100 mW/cm² AM 1.5G) using source meter (Keithely 2400, Tektronix, Inc. USA) in the range of ±1 V, at a scan step of 0.02 V and during a delay time of 0.50 s. Electrochemical impedance spectroscopy (EIS) analysis was evaluated with a frequency range of 9 mHz to 100 kHz in a similar standard illumination condition using a potentiostat (SP150, Biologic SAS, France) to confirm the change in the electrochemical component in the DSSCs. The results of the I–V curve and EIS were analyzed with the averages values of at least six individual cells. Electrochemical parameters were obtained from the equivalent circuit and AC impedance spectra using the Zsim mode in the EC-Lab^® analysis software.

In the case of the screen printing process, the thickness of the film can be changed depending on the organic binder, particle size of the deposition materials constituting the paste, and the distance between the screen mask and substrate. The thickness of the TiO₂ mesoporous film according to the number of screen printing cycles observed by HR-SEM is shown in Fig. 3(a). The thickness of the TiO₂ films increased as the number of printings increased. Fig. 3(b) also shows the variation of thickness of the TiO₂ films with screen printing. The thickness increases linearly by approximately 3 μm with each cycle.
As shown in Fig. 4(a), the color of the film after the dye loading process became darker as the thickness of the TiO₂ film increased. However, after five times screen printing (an approximately 12-μm thick film), a surface crack on the TiO₂ film was observed. In the sample with 6 cycles of screen printing, parts of the TiO₂ films were broken, as shown in Fig. 4(b).
UV-visible spectroscopy analysis was performed to confirm the change in the adsorption amount of the N719 dye according to the change in the thickness of the TiO₂ photoelectrode (Fig. 5). Similar to the color change tendency in Fig. 4, the absorbance in the 525-535 nm wavelength range had a tendency to increase as the thickness of the TiO₂ photoelectrode increased [23, 24]. Therefore, the amount of dye adsorption increased by increasing the TiO₂ mesoporous film thickness.
I–V characteristic measurements were performed to confirm the performance according to the change in the thickness of mesoporous TiO₂ films and the parameters are listed in Table 1. As shown in Fig. 6, the efficiency increases as the thickness increases for up to 4 screen printing cycles. Noticeably, the current density increases with increasing thickness corresponding to the tendency of efficiency. The increase in the concentration of electron generation is because of the increase in the adsorption amount of the dye as the TiO₂ thickness increases, resulting in an increase in Fermi level in the TiO₂ conduction band, following an increase in current density with up to four screen printing cycles (12-μm thick). However, after four cycles printing, the perfor- mance of DSSC tended to decrease. According to the previous studies, the DSSC performance decreased when the photoelectrode had a critical thickness because of the reduced light transmittance and increased recom- bination, leading to a reduction in the electron generation and current density [25,26]. However, in this study, surface cracks were observed in the samples with a thickness of 15 μm (five screen printing cycles) and 18 μm (6 screen printing cycles), as shown in Fig. 3 and Fig. 4(b). The decrease in current density and performance after the fourth screen printing cycle is attributed to the surface crack generated when the deposition is too thick. The V_OC has a tendency to decrease gradually as the thickness increased because the TiO₂ photoelectrode creates a surface trap level on the TiO₂ surface, resulting in an increase in the recombination with the electrolyte and consequent reduction in the V_OC [22, 27]. Moreover, our results demonstrate that the obtained performances are comparable with several literatures [28-31] (Table 1).
EIS was performed to confirm the electrochemical change in the DSSC according to the thickness change of the TiO₂ films. Fig. 7 shows the Nyquist plots of the electrochemical impedance spectra of the DSSC for changing the thickness of the TiO₂ photoelectrode. In a conventional DSSC, the three semicircles are drawn on the Nyquist plot in EIS data, which is designated to a redox reaction of 3I⁻/I³⁻ at the Pt counter electrode (the first semicircle in the high frequency range), charge transfer reaction at the interface between TiO₂/N719 and electrolyte (the second semicircle in the middle frequency range), and the Warburg diffusion process of iodide (the third semicircle in the low frequency range) [27, 32]. The electrochemical parameters obtained by the simulation of fitting between equivalent circuit and impedance spectra are listed in Table 2. The measured impedance data (dots) are well matched with the modeling data (lines) calculated by the equivalent circuit and electrochemical parameters. As shown in Fig. 7 and Table 2, the resistance of the charge transfer reaction in the TiO₂/dye-electrolyte interface decreased as the thickness of TiO₂ increased up to a thickness of 12 μm, and increased at thicker TiO₂. This result is consistent with the tendency of the current density and efficiency in I–V characteristics. Alternative to the tendency of the charge transfer resistance, the capacitance of the charge transfer increases up to a thickness of 12 μm and then decreases at thicker TiO₂. The tendency of impedance parameters in changing the thickness of the TiO₂ is in agreement with previous studies in that one of the factors that increases the performance as the thickness increases is an increase in current density because of a decrease in charge transfer resistance, and the increase in charge transfer capacitance is related to the total density of electrons [32-34].

TiO₂ mesoporous films were fabricated using a screen printing method with thickness of 3 μm/screen prints as determined by FE-SEM. As the thickness of TiO₂ films increased to 12 μm, the performance of the DSSCs increased. I–V measurements showed the highest efficiency in the 12-μm thick film. The increase in overall performance may be because of the increase in current density (ISC: 11.267 mA/cm² in the 12 μm TiO₂ films). However, the efficiency tends to decrease with TiO₂ thicker than 12 μm. The EIS demonstrated that the resistance of the TiO₂/dye-electrolyte charge transfer reaction corresponding to the second semicircle in the Nyquist plot decreases up to a thickness of 12 μm and then increases, which is consistent with the tendency in I–V measurements. As the thickness of the TiO₂ film increases, more current was generated because of an increase in the amount of dye loading, resulting in an increase in the performance. As the thickness increases, the amount of dye loading increased; however, the DSSC performance decreased at TiO₂ thicker than 12 μm. This performance decrease was attributed to the connection break among TiO₂ particles by surface cracking caused by stress differences between FTO/TiO₂ and the TiO₂ particle interface.

This research was supported by “Research Base Construction Fund Support Program” funded by Jeonbuk National University in 2020. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2018R1A4A1025528). This work was also supported by the Technology Development Program to Solve Climate Changes of the National Research Foundation (NRF) grant funded by the Korea government (Ministry of Science and ICT) (2017M1A2A2044930).