Characteristics of Dye-Sensitized Solar Cells with TiO2 Stripes

A TiO2 strip array with a thickness of 90 nm was fabricated by photolithography and physical vapor deposition. This work utilized the chemical and physical methods to fabricate the TiO2 strip array. A porous semiconductor layer made of TiO2 nanoparticles was coated on the TiO2 strip array. The TiO2 strip array has a one-dimensional protrusive structure. The energy conversion efficiency (4.38%) of a dye-sensitized solar cell (DSSC) with the TiO2 strip array exceeded that (3.20%) of a DSSC without a TiO2 strip array by 37%. In addition, this result was verified by the electrochemical impedance spectra of the two DSSCs. Therefore, the TiO2 strip array can be used to increase the energy conversion efficiencies of DSSCs. The large energy conversion efficiency of the DSSC with the TiO2 strip array arises from the large surface area of the one-dimensional protrusive structure and its specific electron transport paths. The DSSC with the TiO2 strip array has advantages of economical production cost, easy fabrication, and boosting energy conversion efficiency.

A normal DSSC has a structure of substrate/transparent conductive oxide (TCO) layer/porous semiconductor layer with dye molecules/electrolyte layer/counterelectrode layer/substrate. As the dye molecules absorb solar energy, they generate electrons and holes. The electrons are injected into the conduction band of the porous semiconductor layer because the lowest unoccupied molecular orbital (LUMO) of the dye has a higher energy level than the conduction band. The electrons are transported from the porous semiconductor layer to an external circuit via the TCO layer because the porous semiconductor layer has a higher conduction band than the TCO layer. The transportation causes the loss of the electrons of the dye molecules, so the dye molecules are oxidized. The electrolyte molecules cause the reduction of the oxidized dye molecules, so the electrolyte molecules regenerate the dye molecules. The electrolyte molecules lose their electrons due to the regeneration of the dye molecules, so the electrolyte molecules are oxidized. The electrons from the external circuit cause the reduction of the oxidized electrolyte molecules, so the electrons regenerate the electrolyte molecules. The electrons are recycled in the DSSC and external circuit, so the DSSC generates electrical energy.
A conventional porous semiconductor layer is made of TiO 2 nanoparticles. The TiO 2 nanoparticles have large surface areas, so a large number of dye molecules are adsorbed to their surfaces. The dye molecules that absorb solar energy can generate many electrons. However, the nanoparticles are a zero-dimensional structure [11]. Because the electrons in the zero-dimensional structure have no specific transporting paths, they are easy to recombine with holes. The recombination of the electrons and holes reduces the energy conversion efficiency of a DSSC. A good porous semiconductor layer allows electrons to be transported in specific paths, increasing the energy conversion efficiency of a DSSC. Pavasupree et al. used a porous semiconductor layer that is made of TiO 2 nanoparticles and TiO 2 nanorods to fabricate a DSSC [12]. The energy conversion efficiency of the DSSC is larger than that of a DSSC with a porous semiconductor layer that is made of only TiO 2 nanoparticles by 22%. This result reveals that the TiO 2 nanorods allow electrons to be transported in specific paths, increasing the energy conversion efficiency of the DSSC. Asagoe et al. utilized a porous semiconductor layer that is made of 90% TiO 2 nanoparticles and 10% TiO 2 nanowires to fabricate a DSSC [13]. The energy conversion efficiency of the DSSC is larger than that of a DSSC with a porous semiconductor layer that is made of 100% TiO 2 nanoparticles by 17%. This result reveals that the TiO 2 nanowires with a concentration of 10% in the TiO 2 nanoparticles facilitate the electron transport in the DSSC. However, the energy conversion efficiency of the DSSC decreased significantly as the concentration of the TiO 2 nanowires in the TiO 2 nanoparticles increased from 10% to 100%. This result arises from the fact that the TiO 2 nanowires with a concentration that exceeds 10% in the TiO 2 nanoparticles form the entangled structure that obstructs the electron transport in the DSSC. Therefore, TiO 2 nanostructures that are fabricated only by chemical synthesis cannot further increase the energy conversion efficiencies of DSSCs.
Choi et al. reported that a counterelectrode with a patterned TiO 2 layer enhances the photon harvesting in a DSSC due to the light reflection from the patterned TiO 2 layer [14]. Baba et al. reported that a patterned substrate that is coated with a gold film increases the energy conversion efficiency of a DSSC due to the light scattering from the grating substrate [15]. It is interesting to use a patterned photoelectrode to increase the energy conversion efficiency of a DSSC.
A TiO 2 strip array with a thickness of 90 nm was fabricated using photolithography and physical vapor deposition in this work. The TiO 2 strip array exhibited a onedimensional protrusive structure. The protrusive structure increased the surface area of the TiO 2 strip array, and offered the specific paths for electron transport. Therefore, the energy conversion efficiency of a DSSC with the TiO 2 strip array was larger than that of a DSSC without a TiO 2 strip array by 37%. The TiO 2 strip array has potential in developing next-generation solar cells. This work proposed the chemical and physical methods for increasing the energy conversion efficiencies of DSSCs using TiO 2 strip arrays. The methods have many advantages such as various TiO 2 structures, economical production cost, and easy fabrication. Figure 1a presents the design of a TiO 2 stripe array that is deposited on a TiO 2 compact layer. Each of the TiO 2 stripes in the array is designed to have height (h), linewidth (t), period (p), and length (l) of 90 nm, 15 µm, 60 µm, and 5 mm, respectively. The TiO 2 stripe array was created by photolithography and physical vapor deposition. A TiO 2 compact layer with a thickness of 22 nm was deposited on a glass substrate with a fluorine-doped tin oxide (FTO) layer (Ruilong, Miaoli County, Taiwan) using a magnetron sputter. The FTO layer functions as a TCO layer in this work. The TiO 2 compact layer in a DSSC was used to not only prevent the contact of dye molecules and a TCO layer but also facilitate the contact of a porous semiconductor layer and the TCO layer [16,17]. An ENPI202 photoresist (Everlight Chemical Industrial, Tainan, Taiwan) was coated on the TiO 2 compact layer using a spin coater. The time and speed of the spin were 40 s and 4500 rpm for the coating of the photoresist, respectively. Stripe patterns in a photomask were transferred onto the ENPI photoresist layer by the irradiation of UV light. The ENPI photoresist was developed using its developer. The photoresist stripe patterns were obtained following the development. TiO 2 was deposited on the photoresist stripe patterns using the magnetron sputter. The FTO glass was immersed into acetone to remove the photoresist stripe patterns. The TiO 2 stripe array was obtained following the immersion. Figure 1b presents the optical microscope image of the TiO 2 stripe array. Therefore, the TiO 2 stripe array was reliable in this work.

Materials and Methods
rine-doped tin oxide (FTO) layer (Ruilong, Miaoli County, Taiwan) using a magnetron sputter. The FTO layer functions as a TCO layer in this work. The TiO2 compact layer in DSSC was used to not only prevent the contact of dye molecules and a TCO layer but als facilitate the contact of a porous semiconductor layer and the TCO layer [16,17]. An EN PI202 photoresist (Everlight Chemical Industrial, Tainan, Taiwan) was coated on the TiO compact layer using a spin coater. The time and speed of the spin were 40 s and 4500 rpm for the coating of the photoresist, respectively. Stripe patterns in a photomask wer transferred onto the ENPI photoresist layer by the irradiation of UV light. The ENP photoresist was developed using its developer. The photoresist stripe patterns were ob tained following the development. TiO2 was deposited on the photoresist stripe pattern using the magnetron sputter. The FTO glass was immersed into acetone to remove th photoresist stripe patterns. The TiO2 stripe array was obtained following the immersion Figure 1b presents the optical microscope image of the TiO2 stripe array. Therefore, th TiO2 stripe array was reliable in this work.  Figure 2a,b present the schematic configuration of DSSCs with and without a TiO strip array, respectively. A TiO2 paste layer made of TiO2 nanoparticles was coated on th TiO2 stripe array using a blade coater. The TiO2 paste layer functioned as a porous sem iconductor layer in this work. The TiO2 paste layer was heated from room temperature to 500 °C with a step of 5 °C/min, and held at a temperature of 500 °C for 45 min. The tem perature of the TiO2 paste layer was decreased from 500 °C to room temperature using natural cooling. The TiO2 paste layer was immersed into a N719 mixture of 50 mL ace tonitrile, 50 mL tert-Butanol, and 59.4 mg D719 dye (Everlight Chemical Industrial, Tai nan, Taiwan) for 24 h. The TiO2 compact layer, TiO2 stripe array, TiO2 paste layer, and N719 mixture with a concentration of 0.5 mM compose a photoelectrode in this work. A Pt layer with a thickness of 50 nm was deposited on a FTO glass using the magnetron sputter. The FTO glass with the Pt layer functioned as a counterelectrode in this work. An empty cell was obtained by separating the photoelectrode and counterelectrode using Surlyn film with a thickness of 20 μm. The empty cell was filled with an EL-200 electro lyte (Everlight Chemical Industrial, Tainan, Taiwan). The DSSC with the TiO2 stripe array was obtained following the filling of the electrolyte. Figure 2b displays that a DSSC ha the same geometrical structure as the design of Figure 2a, but has no TiO2 stripe array The DSSC without a TiO2 stripe array was used to evaluate the energy conversion effi ciency of the DSSC with the TiO2 stripe array.  Figure 2a,b present the schematic configuration of DSSCs with and without a TiO 2 strip array, respectively. A TiO 2 paste layer made of TiO 2 nanoparticles was coated on the TiO 2 stripe array using a blade coater. The TiO 2 paste layer functioned as a porous semiconductor layer in this work. The TiO 2 paste layer was heated from room temperature to 500 • C with a step of 5 • C/min, and held at a temperature of 500 • C for 45 min. The temperature of the TiO 2 paste layer was decreased from 500 • C to room temperature using natural cooling. The TiO 2 paste layer was immersed into a N719 mixture of 50 mL acetonitrile, 50 mL tert-Butanol, and 59.4 mg D719 dye (Everlight Chemical Industrial, Tainan, Taiwan) for 24 h. The TiO 2 compact layer, TiO 2 stripe array, TiO 2 paste layer, and N719 mixture with a concentration of 0.5 mM compose a photoelectrode in this work. A Pt layer with a thickness of 50 nm was deposited on a FTO glass using the magnetron sputter. The FTO glass with the Pt layer functioned as a counterelectrode in this work. An empty cell was obtained by separating the photoelectrode and counterelectrode using a Surlyn film with a thickness of 20 µm. The empty cell was filled with an EL-200 electrolyte (Everlight Chemical Industrial, Tainan, Taiwan). The DSSC with the TiO 2 stripe array was obtained following the filling of the electrolyte. Figure 2b displays that a DSSC has the same geometrical structure as the design of Figure 2a, but has no TiO 2 stripe array. The DSSC without a TiO 2 stripe array was used to evaluate the energy conversion efficiency of the DSSC with the TiO 2 stripe array. All the experiments in this work were carried out using twenty DSSCs with TiO2 strip arrays and twenty DSSCs without TiO2 strip arrays. The experimental data in this work are the averages of the measured values of the DSSCs with and without the TiO2  All the experiments in this work were carried out using twenty DSSCs with TiO 2 strip arrays and twenty DSSCs without TiO 2 strip arrays. The experimental data in this work are the averages of the measured values of the DSSCs with and without the TiO 2 strip arrays. Figure 3a,b present the scanning electron microscope images of the TiO 2 stripe array that was deposited on the TiO 2 compact layer at 45 • and 80 • angles of incidence, respectively. The TiO 2 stripe array and its design have similar geometrical dimensions. The TiO 2 stripe array was successfully fabricated by the photolithography and physical vapor deposition. The TiO 2 stripe array has a one-dimensional protrusive structure, so the layer has a large surface area and specific paths for electron transport. Figure 3c,d present the top and side views of the atomic force microscope (AFM) images of the TiO 2 strip array that was deposited on the TiO 2 compact layer, respectively. The AFM images provide reliable data on the actual relief of the surface of the TiO 2 strip array. All the experiments in this work were carried out using twenty DSSCs with TiO2 strip arrays and twenty DSSCs without TiO2 strip arrays. The experimental data in this work are the averages of the measured values of the DSSCs with and without the TiO2 strip arrays. Figure 3a,b present the scanning electron microscope images of the TiO2 stripe array that was deposited on the TiO2 compact layer at 45° and 80° angles of incidence, respectively. The TiO2 stripe array and its design have similar geometrical dimensions. The TiO2 stripe array was successfully fabricated by the photolithography and physical vapor deposition. The TiO2 stripe array has a one-dimensional protrusive structure, so the layer has a large surface area and specific paths for electron transport. Figure 3c,d present the top and side views of the atomic force microscope (AFM) images of the TiO2 strip array that was deposited on the TiO2 compact layer, respectively. The AFM images provide reliable data on the actual relief of the surface of the TiO2 strip array.   Figure 4a,b displays the setup of the experimental transmission (reflection) spectra of the FTO glass substrates with and without the TiO 2 strip array. The spectra were measured using a transmission and reflection measurement system (LSRT-R, LiveStrong Optoelectronics, Kaohsiung, Taiwan). The system includes an optical fiber, white-light source, photometric integrating sphere, and spectrometer. The wavelength of the system ranges from 300 nm to 900 nm, and its spectral resolution is 1 nm. The optical fiber was connected with the white-light source, and guided the light from the source. The light was incident from the side of the FTO glass substrates (compact TiO 2 layers) for measuring the transmission (reflection) spectra of the FTO glass substrates with and without the TiO 2 strip array. The transmissions (reflections) of the FTO glass substrates with and without the TiO 2 strip array were normalized with that of air (a white broad). source, photometric integrating sphere, and spectrometer. The wavelength of the system ranges from 300 nm to 900 nm, and its spectral resolution is 1 nm. The optical fiber was connected with the white-light source, and guided the light from the source. The light was incident from the side of the FTO glass substrates (compact TiO2 layers) for measuring the transmission (reflection) spectra of the FTO glass substrates with and without the TiO2 strip array. The transmissions (reflections) of the FTO glass substrates with and without the TiO2 strip array were normalized with that of air (a white broad).  Figure 4c presents that the FTO glass substrate with the TiO2 strip array had a lower transmittance at a wavelength than that without a TiO2 strip array because the TiO2 strip array scattered the light that was incident from the side of the FTO glass substrate. This result implies that the TiO2 strip array may hinder incident photons from going into a DSSC, so the DSSC has a small energy conversion efficiency. Figure 4d displays that the FTO glass substrate with the TiO2 strip array had a larger reflectance at a wavelength than that without a TiO2 strip array because the TiO2 strip array scattered the light that was incident from the side of the compact TiO2 layer. This result implies that the TiO2 strip array may confine the photons that are reflected by the Pt layer of a DSSC in its TiO2 compact layer, so the DSSC has a large energy conversion efficiency. The results in Figure  4c,d reveal that the TiO2 strip array caused the competition between the entrance of the incident photons and the confinement of the reflected photons. Therefore, it is interesting to study the effect of a TiO2 strip array to the energy conversion efficiency of a DSSC.   Figure 4c presents that the FTO glass substrate with the TiO 2 strip array had a lower transmittance at a wavelength than that without a TiO 2 strip array because the TiO 2 strip array scattered the light that was incident from the side of the FTO glass substrate. This result implies that the TiO 2 strip array may hinder incident photons from going into a DSSC, so the DSSC has a small energy conversion efficiency. Figure 4d displays that the FTO glass substrate with the TiO 2 strip array had a larger reflectance at a wavelength than that without a TiO 2 strip array because the TiO 2 strip array scattered the light that was incident from the side of the compact TiO 2 layer. This result implies that the TiO 2 strip array may confine the photons that are reflected by the Pt layer of a DSSC in its TiO 2 compact layer, so the DSSC has a large energy conversion efficiency. The results in Figure 4c,d reveal that the TiO 2 strip array caused the competition between the entrance of the incident photons and the confinement of the reflected photons. Therefore, it is interesting to study the effect of a TiO 2 strip array to the energy conversion efficiency of a DSSC. Figure 5 displays the dependences of the short-circuit currents (I SC ) of the DSSCs with and without the TiO 2 strip array on their open-circuit voltages (V OC ). The I SC -V OC curves in Figure 5 were obtained using a source meter (2400, Keithley, Beaverton, OR, USA) and AM 1.5 solar simulator (XES-40S2, San-EI Electric, Osaka, Japan). The DSSCs with and without the TiO 2 stripe array had an identical active area of 5 mm × 6 mm in this work.

Results and Discussion
The averaged photovoltaic parameters of the DSSCs with and without the TiO 2 strip array are summarized in Table 1. The DSSC with the TiO 2 strip array had a larger opencircuit voltage, short-circuit current, and fill factor than that without a TiO 2 strip array. This result reveals that the energy conversion efficiency of the DSSC with the TiO 2 strip array was larger than that of the DSSC without a TiO 2 strip array by 37%. The large energy conversion efficiency of the DSSC with the TiO 2 strip array arises from the one-dimensional protrusive structure of the TiO 2 strip array and specific electron transport paths of the structure. The DSSCs with and without the TiO 2 strip array had low energy conversion efficiencies because the thickness of the TiO 2 paste layers in the DSSCs was not optimized. Efforts are being made at the authors' laboratory to optimize the thickness of the TiO 2 paste layer in a DSSC with a TiO 2 strip array, the results of which will be published in the near future. curves in Figure 5 were obtained using a source meter (2400, K USA) and AM 1.5 solar simulator (XES-40S2, San-EI Electric, O with and without the TiO2 stripe array had an identical active are work. The averaged photovoltaic parameters of the DSSCs with an array are summarized in Table 1. The DSSC with the TiO2 s open-circuit voltage, short-circuit current, and fill factor than t array. This result reveals that the energy conversion efficiency of strip array was larger than that of the DSSC without a TiO2 strip energy conversion efficiency of the DSSC with the TiO2 strip one-dimensional protrusive structure of the TiO2 strip array transport paths of the structure. The DSSCs with and without t low energy conversion efficiencies because the thickness of the DSSCs was not optimized. Efforts are being made at the authors the thickness of the TiO2 paste layer in a DSSC with a TiO2 st which will be published in the near future.   This work studied the effect of photoelectrodes with TiO 2 strip arrays on the energy conversion efficiencies of DSSCs. The DSSC cell comprising the photoelectrode with the TiO 2 strip array was 37% higher than that comprising the photoelectrode without a TiO 2 strip array because the TiO 2 strip array in the former facilitates the electron transport. Choi et al. studied the effect of counterelectrodes with TiO 2 strip arrays on the energy conversion efficiencies of DSSCs [14]. The DSSC cell comprising the counterelectrode with the TiO 2 strip array had a 22% higher energy conversion efficiency than that comprising the counterelectrode without a TiO 2 strip array because the TiO 2 strip array in the former enhanced the photon harvesting. The results in our and Choi's works revealing that the improvement of photoelectrodes and counterelectrodes can increase the energy conversion efficiencies of DSSCs. The incident photon-to-current efficiency (IPCE) spectra of the DSSCs with and without the TiO 2 strip array were measured using a measurement system (LSQE-N, LiveStrong Optoelectronics, Kaohsiung, Taiwan). Figure 6 presents the IPCE spectra of the DSSCs with and without the TiO 2 strip array. The DSSC with the TiO 2 strip array had larger IPCEs at the frequencies of its spectrum than that without a TiO 2 strip array. This result verifies the DSSC with the TiO 2 strip array had a larger short-circuit current than that without a TiO 2 strip array due to J SC = q ϕ(λ) × IPCE(λ) × dλ, where J SC is the short-circuit current density of a DSSC; q is the charge of an electron; ϕ(λ) is the wavelength-dependent photon fluxes of an incident light; and IPCE(λ) is the wavelength-dependent IPCEs of the DSSC.
the DSSCs with and without the TiO2 strip array. The DSSC with the TiO2 strip array had larger IPCEs at the frequencies of its spectrum than that without a TiO2 strip array. This result verifies the DSSC with the TiO2 strip array had a larger short-circuit current than that without a TiO2 strip array due to SC ( ) ( ) where JSC is the short-circuit current density of a DSSC; q is the charge of an electron; φ(λ) is the wavelength-dependent photon fluxes of an incident light; and IPCE(λ) is the wavelength-dependent IPCEs of the DSSC. The electrochemical impedance spectra of the DSSCs with and without the TiO2 strip array were measured using a potentiostat/galvanostat (WonATech, ZIVE SP1, Seoul, Korea) at a solar intensity of 100 mW/cm 2 and bias voltage of 10 mV, and the frequency in the measurement ranged from 10 mHz to 1 MHz. Figure 7a,b present the Nyquist and Bode plots of the DSSCs with and without TiO2 strip array. The plots were obtained from the electrochemical impedance spectra of the DSSCs with and without the TiO2 strip array. The electrochemical impedance spectra of the DSSCs with and without the TiO 2 strip array were measured using a potentiostat/galvanostat (WonATech, ZIVE SP1, Seoul, Korea) at a solar intensity of 100 mW/cm 2 and bias voltage of 10 mV, and the frequency in the measurement ranged from 10 mHz to 1 MHz. Figure 7a,b present the Nyquist and Bode plots of the DSSCs with and without TiO 2 strip array. The plots were obtained from the electrochemical impedance spectra of the DSSCs with and without the TiO 2 strip array.
short-circuit current density of a DSSC; q is the charge of an electron; φ(λ) is the wavelength-dependent photon fluxes of an incident light; and IPCE(λ) is the wavelength-dependent IPCEs of the DSSC. The electrochemical impedance spectra of the DSSCs with and without the TiO2 strip array were measured using a potentiostat/galvanostat (WonATech, ZIVE SP1, Seoul, Korea) at a solar intensity of 100 mW/cm 2 and bias voltage of 10 mV, and the frequency in the measurement ranged from 10 mHz to 1 MHz. Figure 7a,b present the Nyquist and Bode plots of the DSSCs with and without TiO2 strip array. The plots were obtained from the electrochemical impedance spectra of the DSSCs with and without the TiO2 strip array.  Table 2 displays the averaged series resistances (R s , R ct , and R ec ), averaged peak frequencies (f p ) of the Bode plots, and averaged electron lifetimes (τ) for the DSSCs with and without the TiO 2 strip array. R s is a series resistance related to the electron transport in the FTO/photoelectrode interface; R ct is a series resistance related to the electron transport in the electrolyte/Pt layer interface; and R rec is a series resistance related to the electron transport in the photoelectrode/electrolyte interface. f p is a peak frequency of the second semi-circle of a DSSC in its Bode plot. τ is given by τ = 1/(2π × f p ). The DSSC with the TiO 2 strip array had a smaller R s than the DSSC without a TiO 2 strip array. This result reveals that the electron transport from the DSSC to an external circuit was more efficient in the DSSC with the TiO 2 strip array than in the DSSC without a TiO 2 strip array. The DSSC with the TiO 2 strip array had a larger R rec than that without a TiO 2 strip array. This result displays that the electron injection from the LUMO of the dye to the conduction band of the photoelectrode was more efficient in the DSSC with the TiO 2 strip array than in that without a TiO 2 strip array. The DSSC with the TiO 2 strip array had a larger τ than that without a TiO 2 strip array. This result reveals that the recombination of the electrons and holes had a smaller probability in the DSSC with the TiO 2 strip array than in that without a TiO 2 strip array. The small R s , large R rec , and large τ of the DSSC with the TiO 2 strip array reveal that the TiO 2 strip array increased the energy conversion efficiency of the DSSC because its one-dimensional protrusive structure has a large surface area and specific electron transport paths. The DSSCs with and without the TiO 2 strip array had similar R ct due to their identical electrolyte and Pt layer.

Conclusions
This work developed a design through photolithography and physical vapor deposition to fabricate the TiO 2 strip array with a thickness of 90 nm. The TiO 2 paste layer made of the TiO 2 nanoparticles was coated on the TiO 2 strip array. The TiO 2 strip array had a one-dimensional protrusive structure. The protrusive structure increased the surface area of the TiO 2 strip array, and offered specific paths for the electron transport in the DSSC with the TiO 2 strip array. Therefore, the energy conversion efficiency of the DSSC with the TiO 2 strip array was larger than that of the DSSC without a TiO 2 strip array by 37%. The TiO 2 strip array has potential in developing next-generation solar cells. This work utilized chemical and physical methods to fabricate the TiO 2 strip array. The methods have advantages of various TiO 2 structures, low production cost, and easy fabrication.

Conflicts of Interest:
The authors declare no conflict of interest.