The effects of periodic textured substrate to control diffusion angle on the conversion efficiency of dye-sensitized solar cells

We suggested improving the conversion efficiency of dye-sensitized solar cells (DSSCs) by the micro-nano periodic textures to control the diffusion angle of the incident light for certain absorbed wavelengths of the used dye. A periodic texture (Prd-Tx) was designed to enhance the light path of the wavelength of DSSCs’ dye absorption with a wide process window by optical simulation (pitch: 1400 nm, pillar diameter: 460–560 nm, pillar height: more than 500 nm). The Prd-Tx was fabricated by photolithography processes and nanoimprinting (pitch: 1400 nm, pillar diameter: 500 nm, pillar height: 1000 nm). The Prd-Tx increased the DSSCs’ conversion efficiency (η of 3.13%), surpassing our previous best result (refabricated W-Tx, η of 3.08%). It was considered that the ohmic loss was suppressed owing to the Prd-Tx enhanced electrical conductivity at the interface between the transparent electrode, F-doped tin oxide (FTO), and TiO2.


Introduction
Recently, the contribution of renewable energy to society has surged to achieve the mission of decarbonization by the problems of global warming and continuable society.From 2010 to 2016, over 50% of global electricity capacity additions were from renewable electricity 1) and the shares of renewable energy sources are expected to be 30%-80% in 2100. 2)Solar cells, a prominent renewable energy source, have seen an increased share in the energy mix. 3)][6][7][8] The sensitization of TiO 2 with dyes narrows its bandgap (TiO 2 's original bandgap: 3.2 eV, which can be irradiated by light with a wavelength of 390 nm or less 9) ), enabling the conversion of light absorbed by the dye into electricity. 7)The schematic structure of DSSCs is depicted in Fig. 1(a).
DSSCs offer several advantages, including simplicity of design, 10) low material costs, 11) and the ability to customize through dye selection.This customization allows for applications resembling stained glass, as they can transmit light wavelengths other than those absorbed by the dye.However, despite these benefits, DSSCs have a lower energy conversion efficiency, with a maximum of around 13.0%, 6,12) compared to conventional silicon (Si) photovoltaic cells, which can exceed 26.0%. 12,13)This lower efficiency is partly due to DSSCs' limited ability to convert only UV light and the specific wavelengths absorbed by the dye.
To enhance the performance of DSSCs and address their limitations, several strategies have been explored.][16][17][18] This alteration aims to increase photon absorption efficiency and subsequent energy conversion.][21] This enhancement can lead to more effective electron transfer, thereby improving overall cell efficiency.
Additionally, the introduction of structural elements, such as a block layer, has been investigated. 22)Such layers are designed to inhibit the recombination of electrons with the photocatalyst, a process that reduces the efficiency of energy conversion.
Another structural modification in the realm of DSSCs pertains to the development of components engineered to optimize incident light management, effectively increasing, and confining it within the cells.This is achieved by integrating microscopic structures, known as Antireflection Textures (ART), at the interface of materials possessing varying refractive indices. 23)These textures, being subwavelength in scale, cause gradual changes in the refractive index, 24) as the light cannot discern these minute structures. 25)As a result, incident light absorption is enhanced.
Furthermore, micro-nano structures contribute to light diffusion, in transmission or reflection, through scattering or diffraction.[28] Among the various light management strategies, an approach involves introducing structures (such as pyramid-shaped, 26) porous, 27) or nanowire 28) configurations) that diffuse light upon transmission or reflection.While several methods exist for incorporating textures into DSSCs, techniques focusing on texturing under the electrode of the incident plane are less common.
Therefore, this study investigates the impact of microscopic texturing between the incident glass and the transparent electrode (fluorine-doped tin oxide, FTO) in DSSCs [Fig.1(B)], specifically regarding their conversion efficiency.By leveraging the effects of ART and light diffusion, coupled with the enhanced electron conductivity attributed to the expanded FTO electrode induced by texturing, this study aims to achieve a significant improvement in the conversion efficiency of DSSCs.
The results demonstrated that the incorporation of these textures enhanced the conversion efficiency of the DSSCs.Notably, different trends were observed between the DSSCs employing different dyes and texture shapes.These variations are hypothesized to stem from the distinct absorption wavelengths of the dyes.Consequently, it is suggested that a diffuse texture, capable of scattering the dye's absorption light with greater intensity and over a wider angle, could further amplify the conversion efficiency.
From these results, we fabricated a periodic pillar texture to diffract absorb wavelengths of dye with specific dimensions, featuring an 800 nm pitch, 400 nm pillar diameter, and 500 nm pillar height (Fig. 2).Despite exhibiting favorable optical properties, particularly in diffuse transmittance, this pillar texture adversely affected the conversion efficiency.The primary reason for this decrease in efficiency was attributed to poor electrical conductivity caused by the deposition state of the FTO electrode.
Specifically, during the formation of the FTO electrode, it covered each pillar texture, leaving voids among them because FTO film is formed by the commercial deposition method, Spray Pyrolysis deposition, which is PVD-like deposition, that preferentially deposits on convex tops.These voids created an insufficient connection between the FTO components, leading to compromised electrical conductivity.As a result, the promising optical properties of the pillar texture were overshadowed by the detrimental impact on the electrical aspects, ultimately hindering the overall conversion efficiency of the DSSCs.This finding underscores the critical importance of achieving a delicate balance between optical enhancements and electrical connectivity in the design and fabrication of textures for solar cell applications.
Receiving this result, the texture focusing on refining the deposition state of the FTO electrode to address these connectivity issues was oriented in this study.The texture structure was designed by optical simulation while considering the process friendly and fabricated by photolithography and PDMS NIL.The study aims to provide a deeper understanding of the interplay between textural modifications and the optical properties of DSSCs, contributing to the advancement of solar cell technology.

Optical simulation
The design of the Prd-Tx for DSSCs was conceptualized using an optical simulation Rigorous Coupled Wave Analysis (RCWA) (GSolver, Grating Solver Development).The model of the simulation was on a two-dimensional (2D) cross-section of a pillar Prd-Tx (Fig. 3).This simulation varied the pitch and convex width of the Prd-Tx's 2D crosssection to identify the most effective configuration.The pillar height was 500 nm because it was thought that a texture height variation of more than 400 nm would not contribute to the diffuse property so much. 29)These textures were evaluated based on an index at the absorption wavelength peaks of the N719 dye, specifically at 390 nm and 530 nm. 30)he evaluation index is mathematically represented as follows: In this equation, T n denotes the nth order diffraction transmittance, and represents the extension ratio of the optical path length due to diffusion.This formulation allows for the simultaneous assessment of both the intensity and angle of diffusion.The aim was to identify the ideal  periodic diffuse pillar texture by evaluating these characteristics.
It is important to note that this simulation focused on textures with a wide pitch (greater than 1000 nm) to ensure compatibility with process-friendly manufacturing approaches.

Fabrication method
The designed Prd-Tx for DSSCs was fabricated through a PDMS nanoimprinting at RT.The procedure involved several critical steps: 2.2.1.Master mold fabrication.The master mold was crafted from a 20 mm × 20 mm × 0.625 mm Si substrate, prepared by dicing (DAD522, DISCO).The dimensions of the mold's pillars were designed slightly larger than the intended Prd-Tx pattern to account for the shrinkage of the nanoimprint material.The process of creating the master mold is illustrated in Fig. 4, and the specific parameters of each step are detailed in Table I.
The lithography pattern employed in the fabrication process was a periodic hole pattern, chosen for its expedient lithography compared to pillar patterns when using Posiresist.Following the lithography step, a reversal of the pattern was achieved through the deposition of chromium (Cr) by sputtering, followed by a lifting-off process.
The Si substrate was coated with an EB resist (ZEP520A, ZEON) using a spin coater (MS-B150, MIKASA) after ultrasonic cleaning in Acetone and IPA each for 5 min and coated with HMDS primer.With this recipe, approximately 300 nm resist was coated.This was followed by lithography [Fig.4(1)] and development of the pattern.
Post-resist removal, the Si substrate underwent a mold release treatment.This involved UV/O 3 cleaning using a small excimer lamp (EX-mini, L12530-0, Hamamatsu Photonics) for 5 min, followed by the application of a release agent (DURASURF, HARVES) and linseed (NOVEC TM 7100, 3 M JAPAN), culminating in the completion of the Si master mold.2.2.2.Prd-Tx fabrication.Prd-Tx was fabricated by PDMS NIL at RT using an organosilesquioxane solution from a Si Master mold.A polydimethylsiloxane (PDMS) replica mold was created from the Si master mold, replicating the procedure from the previous study. 4,5)The specific parameters of PDMS replica mold fabrication are shown in Table II.
The Prd-Tx was imprinted onto a 20 mm × 20 mm × 0.4 mm Borosilicate glass (Eagle XG, Corning) substrate using the PDMS replica mold.The process also mirrored the previous study, 4,5) except for the spin-coating speed.While the prior study used a spin speed of 6000 rpm, this study employed a speed of 3000 rpm.The specific parameters of the nanoimprinting are shown in Table III.In this way, the texture made by SiO 2 that has a similar refractive index to Glass is fabricated, and it enables the incident light not to decay when passing the interface and contributes to the DSSCs' conversion efficiency well.2.2.3.DSSCs with Prd-Tx fabrication.We employed the Prd-Tx fabrication process for DSSCs as described in detail in the previous paper. 4,5)This process includes several key steps.On the SiO 2 -based Prd-Tx substrate, a transparent FTO electrode was deposited by SPD method.The FTO layer exhibited a resistance of 15.0 Ω/□ and a transmittance of 82.4%.TiO 2 nanoparticles were then coated onto the FTO-layered Prd-Tx.The TiO 2 -coated Prd-Tx was immersed in N719 dye (di-tetrabutylammonium cic-vis 8isothiocyanato) bis (2, 2′bipyridyl-4, 4′-dicarboxylato) ruthenium (II) (SIGMA-ALDRICH) and after that, it was assembled with Pt counter electrode with thermal adhesive gasket.
Finally, the device was filled with an electrolyte, and DSSCs were completed.Each of these steps follows the protocols in the previous study without modification.

Characterization
Characteristics of the Prd-Tx and DSSCs with Prd-Tx were evaluated, in comparison with a flat (non-textured) glass substrate and the W-Tx, which had shown the most significant improvement in conversion efficiency in our previous study.To evaluate the shape of the Master mold, field emission scanning electron microscopy (FESEM) (JSM-7500F, JEOL) was employed, and to analyze the top and cross-sectional view of Prd-Tx and FTO and TiO 2 deposition on it, atomic force microscopy (AFM) (AFM5100N, Hitachi High-Tech), FESEM and its EDS (EX-37001, JEOL) was employed.
The interface between TiO 2 and FTO on the textured glass was scrutinized to estimate electron conductivity, a crucial factor in DSSCs' performance.Cross-sectional images of the TiO 2 deposited on textured glass with FTO were captured using FESEM.These images, all scaled identically, were analyzed to measure the interface between the TiO 2 and FTO layers.This analysis aimed to reveal how the texture influences the electron transport mechanism within the cells.
To verify that the optical properties of the Prd-Tx aligned with those predicted by optical simulations, transmittance spectroscopy was employed (V-770 JASCO).The transmittance and diffuse transmittance were analyzed to estimate how much incident light was transmitted or diffused in the DSSCs.They are compared with the spectroscopy from optical simulation to confirm if the Prd-Tx truly has great optical properties.Also, the reflectance was analyzed to estimate the inner reflectance reduction by textures.
The impact of the textured glass on the conversion efficiency of DSSCs was quantified using a solar simulator (XES-155S1, Sanei Electronic).This measurement allowed for a direct comparison of the energy conversion capabilities of DSSCs with different textures (Prd-Tx, W-Tx, and flat glass).The efficiency data obtained from these tests are critical for understanding the effectiveness of the Prd-Tx in enhancing the overall performance of DSSCs.

Results and discussion
3.1.Results 3.1.1.Optical simulation results.Figure 5 shows the result of the optical simulation by RCWA.The horizontal axis of the graph is the concave width so as small as in horizontal axis, it is difficult to fabricate by nanoimprinting, and as big as, it is difficult to deposit FTO with a larger surface like the previous pillar texture [Fig.2(c)].
The texture of 1400 nm pitch and 460-560 nm pillar diameter showed higher EVALUATION INDEX in both the wavelength of 390 and 530 nm (N719 dye absorption peak), so the periodic pillar texture of 1400 nm pitch, 460-560 nm pillar diameter, and more than 500 nm pillar height were determined to be the design of Prd-Tx.
3.1.2.Microscopy of Master mold, Prd-Tx, and crosssection of TiO 2 deposited Prd-Tx with FTO. Figure 6 shows the FESEM image of the fabricated Master mold of Prd-Tx.The fabricated Master mold had approximately 1400 nm pitch, 600 nm pillar diameter, and 1100 nm pillar height; it had a slightly bigger pillar diameter than the design.From this Master mold, we fabricated the Prd-Tx.
3.1.3.Microscopy of Master mold, Prd-Tx, and crosssection of TiO 2 deposited Prd-Tx with FTO. Figure 7 shows the AFM image of the fabricated Prd-Tx.The fabricated Prd-Tx had approximately 1400 nm pitch, 500 nm pillar diameter, and 1000 nm pillar height, and it was the texture shape as our target.The Prd-Tx's pillar diameter and height shrank from the Master mold and the shape was acceptable for the shape we designed.
Figure 8 shows the image overlaid SEM image and EDX mapping of each texture with FTO, and TiO 2 .The image suggests that FTO was deposited with uniform thicknesses on the dimple (0.42 μm) or Flat (0.60 μm) texture, on the other hand, was deposited on the top of the pillar preferentially; the thickness on the top (0.42 μm) was thicker than on the bottom (0.15 μm).The image also shows that the TiO 2 layers adhere to the FTO layer on each texture, meaning the wider interface between TiO 2 and FTO would have good electron conductivity though only the FTO on the pillar texture had unpreferable electron property owing to its ununiform film formation.
To measure the contact area of FTO and TiO 2 , surface roughness analysis of AFM was employed at first.The measured areas of FTO's surface on each texture are shown in Table IV and the cross-sectional profiles are depicted in Fig. 9.The measured area on Prd-Tx was the largest of the three, but there is a possibility that AFM could not measure the surface area of the FTO on the Prd-Tx; given the sharp difference between the FESEM image in Fig. 8 and the obtained profile in Fig. 9, the profile does not trace all the FTO contours on the Prd-Tx.To compare all the surfaces of FTO on Prd-Tx, and others, the image analysis from the FESEM images was employed.
Table V shows the FTO/TiO 2 boundary line length measured from the image analysis of the cross-sectional image of FTO, TiO 2 deposited on Prd-Tx.The length of the interface per 4.03 μm width image was measured.The 03SP93-4 © 2024 The Author(s).Published on behalf of The Japan Society of Applied Physics by IOP Publishing Ltd interface on Prd-Tx was 17.3 μm on average, much longer than on W-Tx (9.25 μm) and Flat, so the interface on Prd-Tx was larger than others and it suppressed the ohmic loss between conductive layers of FTO, the transparent electrode and TiO 2 , the photocatalyst and electron generator.Figure 10(e) shows the reflectance of textures with FTO and TiO 2 .In the wavelength longer than 390 nm, Prd-Tx's reflection was more than W-Tx's.It might be because Prd-Tx had a weak anti-reflection effect owing to the larger texture size than W-Tx's.03SP93-6 © 2024 The Author(s).Published on behalf of The Japan Society of Applied Physics by IOP Publishing Ltd DSSCs with W-Tx, which had the highest conversion efficiency in our previous study.The Prd-Tx improved both the short-circuit current density and voltage and did not make the FF from Flat (DSSCs without textures) worse, while W-Tx only improved the short-circuit current and deteriorated FF greater than Flat.

Discussion
Prd-Tx was considered to have great diffuse transmittance properties by optical simulation and suppressed ohmic loss owing to great electron conductivity expected from image analysis of a cross-section of Prd-Tx with FTO and TiO 2 .
From these properties conversion efficiency of DSSCs with Prd-Tx was improved.
The actual optical properties could not be measured because the fatal optical property in the DSSCs with textures appears in the TiO 2 , the light absorption layer after the FTO layer, so any devices could not measure it.In the optical simulation, the designed texture had a great diffuse transmittance in the wavelength of more than 390 nm, so the Prd-Tx is considered to have a good diffuse transmittance property for the DSSCs using N719 dyes.
The FTO/TiO 2 interface on the Prd-Tx was expanded because the FTO on the Prd-Tx was deposited as if they follow the Prd-Tx's shape and there are blanks between each FTO on the Prd-Tx pillars.The FTO/TiO 2 interface on the W-Tx did not have a significantly wide area though the W-Tx had many nano-ordered dimples because they were buried with FTO and did not appear on the FTO's surface.The FTO/TiO 2 interface on Flat glass had a slightly wider area because the FTO layer was composed of grains with several tens of nanometers.

Conclusions
In this paper, the periodic pillar Prd-Tx which has great diffuse transmittance intensity and angle in the wavelength of N719 absorption was fabricated and applied to DSSCs.The Prd-Tx's structure was designed by RCWA optical simulation considering the process was friendly.The actual fabricated Prd-Tx had structures of 1400 nm pitch, 500 nm pillar diameter, and 1000 nm pillar height, which follows the designated structure.The FTO and TiO 2 interface on Prd-Tx had a wider area than ones on other textures because FTO was deposited as if they followed and covered the Prd-Tx's pillar structure.Owing to such an FTO film form and the wide interface between conductive layers, the DSSCs with Prd-Tx had great inner electrical conductivity and suppressed ohmic loss.The actual optical properties, especially for transmittance and diffuse transmittance of Prd-Tx with FTO, could not be measured because the fatal optical property that appears in the TiO 2 layer after the FTO layer could not be reproduced in the experimental environment, but Prd-Tx with FTO was expected to have approximately the desired diffuse transmittance property because the Prd-Tx glass texture followed the simulation result of only Prd-Tx glass texture with superstrate of air.With the suppression of ohmic loss owing to the inner structure that had a wide interface between the FTO layer and TiO 2 layer, and expected great diffuse transmittance, the periodic pillar Prd-Tx had succeeded in improving the conversion efficiency of DSSCs.
From the result of this paper, the periodic pillar of truncated cone texture which has a smaller diameter on the top than the bottom is expected to have the ability to improve the DSSCs' conversion efficiency more owing to the ability to make wide FTO/TiO 2 interface and anti-reflection because of the small structure on the top.

Fig. 3 .
Fig.3.RCWA optical simulation model, the area surrounded by dashed lines is the computed area, and in the simulation, the diffraction grid is considered as the repetition of the area in the dashed line.

Fig. 8 .
Fig. 8.The overlaid SEM and EDX image of the cross-section of textures with FTO and TiO 2 : (a) Prd-Tx, (b) W-Tx, and (c) Flat.

3. 1 . 5 .
Conversion efficiency of DSSCs with textured glass.Figure 11 shows the I-V curves and Table VI shows the properties of DSSCs with textures.The DSSCs with Prd-Tx had the highest conversion efficiency, surpassing the

Fig. 10 .
Fig. 10.Spectroscopies of textures, textures with FTO, textures with FTO and TiO 2 : (a) total transmittance of textured glasses, (b) diffuse transmittance of textured glasses, (c) total transmittance of textures with FTO, (d) diffuse transmittance of textures with FTO, and (e) reflectance of textures with FTO and TiO 2 .

Table I .
The processing conditions of Si master mold.

Table II .
The processing conditions of PDMS replica mold.

Table III .
The processing conditions of nanoimprinting.

Table VI .
The performances of DSSCs applied with textures.
©2024The Author(s).Published on behalf of The Japan Society of Applied Physics by IOP Publishing Ltd