Fabrication of Dye-Sensitized Solar Cells with a 3D Nanostructured Electrode

Abstract

A novel Dye-Sensitized Solar Cell (DSSC) scheme for better solar conversion efficiency is proposed. The distinctive characteristic of this novel scheme is that the conventional thin film electrode is replaced by a 3D nanostructured indium tin oxide (ITO) electrode, which was fabricated using RF magnetron sputtering with an anodic aluminum oxide (AAO) template. The template was prepared by immersing the barrier-layer side of an AAO film into a 30 wt% phosphoric acid solution to produce a contrasting surface. RF magnetron sputtering was then used to deposit a 3D nanostructured ITO thin film on the template. The crystallinity and conductivity of the 3D ITO films were further enhanced by annealing. Titanium dioxide nanoparticles were electrophoretically deposited on the 3D ITO film after which the proposed DSSC was formed by filling vacant spaces in the 3D nanostructured ITO electrode with dye. The measured solar conversion efficiency of the device was 0.125%. It presents a 5-fold improvement over that of conventional spin-coated TiO₂ film electrode DSSCs.

1. Introduction

Dye-sensitized solar cells (DSSCs) that belong to the group of thin film solar cells are currently the most efficient second-generation solar technology available. The advantages of these promising solar cells are numerous including low cost, requiring no expensive manufacturing steps, large-scale feasibility [1], capable of working in low-light conditions, and higher efficiencies at higher temperatures,. The European Union Photovoltaic Roadmap forecasts that DSSCs will be a potential significant renewable power source by 2020.

The DSSC consists of a transparent and conductive anode made of fluorine-doped tin dioxide deposited on a glass substrate, a thin layer of titanium dioxide (TiO₂) on the back of the conductive plate, photosensitive ruthenium-polypyridine dyes covalently bonded to the surface of the TiO₂, a thin iodide electrolyte layer, and a conductive counter electrode (typically platinum). The photocurrent in the DSSC is attributed to the photoelectrons that are produced when photons of sunlight (with enough energy to be absorbed) strike the dye on the surface of the TiO₂. Photoelectrons move directly from the LUMO (Lowest Unoccupied Molecular Orbital), that is the excited state of the dye, to the conduction band of the TiO₂, and from there they diffuse to the conducting electrode. The dye molecules that have lost an electron strip one from the iodide electrolyte, which is oxidized into triiodide. The triiodide then regains its lost electron by mechanically diffusing to the counter electrode, where the electrons, after flowing through the external circuit, are reintroduced. The key factors that determine the photoelectric efficiency are the structure of individual elements and the type of dye used in the cell. The former affects the separation of the electron-hole pairs, the migration of photoelectrons and holes, and the recombination of photoelectrons and holes. The latter controls the short-circuit current (𝐽_sc) and the open-circuit voltage (𝑉_oc) of the cell.

The original DSSC was invented by O'Regan and Grätzel [2] in 1991. The use of a porous nanocrystalline TiO₂ thin film as the electrode greatly enhanced the binding area for the dye, resulting in a tremendously enlarged photoreaction area. The absorption capability of the dye to visible light captures 46% of the illumination from the sun. The conversion efficiency of the original solar cell was 7.1%. Since then, intensive efforts have been devoted to the improvement of the solar conversion efficiency. Recent progress in nanotechnology, particularly in porous nanostructured TiO₂, has further increased the efficiency to 12%, close to that of traditional low-cost commercial silicon panels.

The main focus in the development of new TiO₂ nanostructures has been how to enhance the efficiency of the photoreaction and the electron migration processes. Chou et al. [3] used the sol-gel method to fabricate TiO₂ nanostructures followed by spreading them on a TiO₂ thin film. An et al. [4] utilized reactive electrodeposition to produce nano grains of TiO₂. Different TiO₂ nanotube arrays [5–8] were invented to enable the multidirectional transfer of the photoelectrons. The total amount of photon absorption could also be enhanced because a nanotube array is thicker than a thin film. The photocurrent could be increased by adding various metallic oxides to a TiO₂ thin film [9, 10].

Since the average traveling distance of a photoelectron is only about 10 nanometers, most do not have enough momentum to reach the transparent conducting thin film electrode, resulting in a relatively low efficiency compared to silicon based solar cells. Therefore, the conversion efficiency could also be improved by increasing the conductivity of the transparent conducting thin film. Takaki et al. [11] used DC magnetron sputtering to fabricate transparent indium tin oxide (ITO) whiskers on a glass substrate. However, the efficiency of electron migration in the DSSC electrode was degraded by the irregular structure of the ITO whiskers. Wan et al. [12] used the self-catalytic vapor-liquid-solid growth method to synthesize one-dimensional ITO nanowires and whiskers on an yttrium stabilized zirconia (YSZ) substrate; however, their synthesizing procedure was complex, leaving room for further improvement. Wang et al. [13] used a template assisted method to grow high aspect ratio ITO nanotube arrays. However, the irregular arrangement of the ITO nanotubes would likely retard the transfer of photoelectrons. A more conductive electrode element for better solar conversion efficiency is desired.

The purpose of this study is to develop a novel scheme for fabrication of DSSCs with better solar conversion efficiency. In this novel scheme, the conventional thin film electrode is replaced by a 3-dimentional nanostructured ITO electrode, fabricated using RF magnetron sputtering, with the barrier layer of an anodic aluminum oxide (AAO) membrane being the template. The crystallinity and conductivity of the 3D ITO film are further enhanced by annealing [14]. Titanium dioxide nanoparticles are electrophoretically deposited on 3D ITO film. The proposed DSSC is formed after filling up vacant spaces in the 3D nanostructured ITO electrode with dye.

2. 3D Nanostructured Electrode Fabrication

2.1. Electrode Fabrication

The distinguishing characteristic of the novel DSSC is the replacement of the conventional thin film electrode by a 3D nanostructured ITO electrode. This is followed by electrophoretic deposition of TiO₂ on the 3D ITO film after which vacant spaces in the electrode are filled by the dye. Figure 1 shows a schematic illustration of the proposed 3D nanostructured electrode for DSSC.

Figure 1 

Schematic illustration of the proposed 3D nanostructured TiO2/ITO/AAO electrode.

The fabrication procedure includes preparation of the AAO membrane, modification of the barrier-layer surface modification, sputtering of the ITO thin film, deposition of the TiO₂ nanoparticles, and deposition of the dye. The process is itemized in detail below.

(1) Preparation of the AAO Membrane [15, 16]
The AAO films were prepared using a well-known anodizing process. Aluminum foils were cleansed and electropolished before anodization. AAO films, with a nanopore diameter of around 60 nm and a thickness of 50 𝜇m were obtained by anodizing the polished aluminum foil in a 0.3 M phosphoric acid solution under an applied voltage of 90 V at 0^∘C for 2 hours.

(2) Modification of the Barrier-Layer Surface
During the first stage, the remaining aluminum beneath the barrier layer was dissolved in an aqueous CuCl₂⋅HCl solution that was prepared by dissolving 13.45 g of powdered CuCl₂ into 100 mL of a 35 wt% hydrochloric acid solution. The removal of the remaining aluminum reveals the honey-comb-like surface of the barrier-layer. The honey-combs have an average convex diameter of 80 nm. Then, the barrier-layer surface was immersed in a 30 wt% phosphoric acid at room temperature for 35 minutes to modify the surface structure.

(3) Deposition of the ITO Thin Film
The modified barrier-layer surface functioned as a template for deposition of the 3D nanostructured ITO thin film by radio frequency (RF) magnetron sputtering. The experimental conditions were as follows: pressure =4.5×10⁻⁵ torr; temperature = 150^∘C; argon = 30 sccm; power = 50 W; processing time = 90 minutes. A 3D nanostructured ITO film with a thickness of about 30 nm was obtained after the sputtered film was cooled to room temperature.

(4) Annealing
The electrical properties and crystalline structure of RF magnetron sputtered thin films usually are not good enough [17–19]. An additional annealing process was utilized to further modify the surface structures of the 3D nanostructure of the ITO film and increase the conductance of the sample [20, 21]. The annealing procedures included: heating the sample to 150^∘C and 500^∘C, respectively at a rate of 9^∘C/sec where it remained for 10 minutes, then cooling the sample in air to room temperature.

(5) Deposition of TiO₂ Nanoparticles
The electrophoretic deposition method [22] was used to deposit TiO₂ nanoparticles uniformly on the ITO thin film. The deposition processes included the following.

(i) Sol Preparation [23]. In the sol preparation procedure, 30 mL Ti (IV) of isopropoxide [Ti(OCH(CH₃)₂)₄] (Alfa Aesar, Ward Hill, MA) were first added to 60 mL of glacial acetic acid (Fisher Scientific, Fair Lawn, NJ) after which the mixture was stirred for 5–10 minutes. Then 30 mL of DI-water where stirred into the mixture. White precipitates formed when the DI-water was added. The chemical reaction for this process is

Ti{OCH(CH₃)₂}₄+2H₂O→TiO₂+4(CH₃)₂CHOH.

After the formation of white precipitates, stirring continued until approximately 30 minutes after the solution became clear. Since the PH value of the final solution was about 2 and the isoelectric point (IEP) of TiO₂ is 6.2, the TiO₂ nanoparticles in the solution were positive charged. The final clear sol was stored at −20^∘C when not in use.

(ii) Electrophoretic Deposition. Electrophoretic deposition was conducted with a picoammeter (Sversa Stat II, Princeton Applied Research). The ITO thin film deposited sample was first placed at the working electrode (WE). The counter electrode was a Pt film; the reference electrode (RE) was Ag/AgCl. An electric potential of DC 1.4 V was applied during the electrophoretic deposition process. The deposition duration ranged from 180 seconds to 5400 seconds.

Titanium dioxide, which is found in natural minerals such as rutile, anatase, and brookite, acts as a photocatalyst under ultraviolet light (when in the form of anatase). A sintering process is usually applied to ensure that the TiO₂ nanoparticles are in the form of anatase. In this study, the sample was heated to the crystallization temperature of anatase (500^∘C) at a rate of 2^∘C/minute where it remained for 1 hour, then cooled in air to room temperature.

(iii) Spin-Coating. Since the TiO₂ nanoparticle film was not thick enough, an additional spin-coated TiO₂ layer was added. A 3 𝜇m thick TiO₂ layer could be obtained using a first-stage rotating speed of 500 rpm followed by a second-stage rotating speed of 2000 rpm, both for 20 seconds.

(6) Dye Deposition
The device was then immersed in a solution of sensitized dye N3 (Dyesol) for 24 hours to allow the dye molecules to covalently bond to the surface of the TiO₂. Unbound particles were rinsed away in an ethanol solution. The absorption spectrum of N3 in the visible light scope ranges from 400 nm to 800 nm, having two peaks at 538 nm and 398 nm.

2.2. DSSC Assembly

A schematic illustration of the assembly of the DSSC is shown in Figure 2, including the fabrication of the counter electrode fabrication, preparation of the electrolyte, and assembly of the parts.

Figure 2 

DSSC assembly.

(1) Counter Electrode Fabrication
Platinum foil was chosen as the counter electrode material. The platinum foil was electropolished before being attached to a transparent conductive oxide (TCO) thin film to ensure a better efficiency of light reflection into the cell body.

(2) Electrolyte Preparation
It is desirable that the electrolyte used in DSSCs has the following properties: high diffusion coefficient, fast oxidation-reduction reaction, and low intrinsic resistance. The recipe for the electrolyte employed in this study which satisfies these specifications is 0.5 M LiI + 0.05 M I2 + 0.5 M TBP + 0.6 M BMII in Acetonitrile + Valeronitrile (1:1).

(3) Parts Assembly Procedure
The parts assembly procedure: parafilm is cut into a rectangle. A 2×2cm² square is punched into the center of the rectangular parafilm. A thin film of AB glue is applied to the bottom surface of the parafilm to attach it to the electrode. Fine clamped the electrode and the counter electrode. The electrolyte is injected into the cell body using a syringe. The photoreaction area can be well controlled by the square punched in the parafilm.

2.3. Fabrication Results and Discussion

(1) Barrier-Layer Surface Modification Results
Figure 3 shows SEM images of the original and the modified barrier-layer surface after being etched in a 30 wt% phosphoric acid for 35 minutes. Due to the stress concentration effect during anodization, the phosphoric acid etched out more alumina at the borders between the cells than from the cell surfaces, resulting in an orderly hemispheric barrier-layer surface.

(a)

(b)

Figure 3 

SEM images of the barrier-layer surfaces: (a) original, (b) after 35 minutes of etching.

(2) ITO Resistance Measurement
Since the ITO thin film on the AAO template was a 3D nanostructure, its electrical properties could not be represented in terms of the sheet resistance. Instead, the I-V curve of the ITO film was measured using a Keithley 2400 Digital Source Meter for a convenient estimation of its resistance. The resistance of the 3D ITO film was calculated by Ohm's law.
The resistances of annealed ITO films subjected to two annealing temperatures are illustrated in Figure 4. The curve indicated by AC denotes the resistances measured between locations A and C on the device as shown in Figure 5, while the BC curve indicates the resistances measured between locations B and C. Location B is the middle point between location A and location C. It is observed that the resistance of the 3D ITO film was reduced from 6.5 M Ω to 0.5 M Ω. Therefore, an annealing temperature of 500^∘C was employed in this study.

Figure 4 

Resistances of ITO film annealed at different temperatures.

Figure 5 

Schematic illustration of the reference locations for resistance measurement.

(3) TiO₂ Nanoparticles Deposition Results
Figure 6 shows SEM images of nano-hemispheric TiO₂/ITO/AAO electrodes subjected to TiO₂ deposition for various durations. The morphology of the TiO₂ film varied with the deposition duration, starting from scattered nanoparticles to a complete thin film. It can be observed that after a 250-second deposition, the TiO₂ nanoparticles were uniformly deposited on the nano-hemispheric ITO array. The measured transparency of this device obtained using an optical power meter (model 1830C, Newport) was 70%.

(a) 180 seconds

(b) 250 seconds

(c) 1200 seconds

Figure 6 

SEM images of nano-hemispheric TiO2/ITO/AAO electrodes after various deposition durations.

The relative energy-dispersive X-ray spectroscopy (EDS) spectrum for the 3D TiO₂ film shown in Figure 6(b) is illustrated in Figure 7. It can also be seen that TiO₂ molecules were successfully deposited on the 3D ITO film.

Figure 7 

EDS spectrum for the TiO2 film shown in Figure 6(b).

In addition to heating to the crystallization temperature of anatase, 500^∘C, sintering was conducted at various temperatures. Figure 8 illustrates the X-ray diffraction (XRD) spectra deposited TiO₂ nanoparticles sintered at various temperatures. The XRD spectra indicate that the TiO₂ nanoparticle sintered at temperatures higher than 500^∘C possesses better crystal lattices of anatase. However, the higher sintering temperature will induce a degradation in the ITO’s conductivity so sintering at 500^∘C is suggested.

Figure 8 

XRD spectra for the deposited TiO2 nanoparticles sintered at various temperatures.

(4) Dye Deposition Result
Figure 9 shows an SEM image after N3 soaking of the device shown in Figure 6(b).

Figure 9 

SEM image after N3 soaking.

3. Solar Conversion Efficiency Measurement

3.1. Apparatus Setup

An Oriel Xe-lamp was used as the light source, and a Keithley 2400 picoammeter was employed to measure the dark-light and illuminated I-V curves of the DSSC during the efficiency measurement experiments. The position of the light source was adjusted such that AM 1.5 (100 mW/cm²) of power was delivered to the surface of the measured DSSC.

The solar conversion efficiency (𝜂) of a DSSC can be estimated using the conversion efficiency formula 𝑃𝜂=^max𝑃_in, (1) where 𝑃_max, and 𝑃_in denote the maximum output power and the input power, respectively. Since a DSSC usually contains a series resistance and a shunt resistance, the fill factor (FF) is introduced to count both effects. 𝑃FF=^max𝐽_sc×𝑉_oc, (2) where 𝑉_oc is the open-circuit voltage and 𝐽_sc is the short-circuit current. The solar conversion efficiency of a DSSC can be calculated by 𝐽𝜂=^sc×𝑉_oc×FF𝑃_in. (3)

3.2. Measurement Results and Discussions

Efficiency measurements for three kinds of electrode were conducted.

(1) Electrophoretically Deposited TiO₂ Nanoparticle Electrode
The open-circuit voltage and the current density were measured to be 0.59 V and 0.0176 (mA/cm²), respectively. The fill factor and the solar conversion efficiency were calculated to be 36.3% and 0.005% according to (2) and (3).

(2) Spin-Coated TiO₂ Film Electrode
The spin-coated TiO₂ film was 3 𝜇m thick. Measurement results were as follows: open-circuit voltage = 0.79 V; current density = 0.083 (mA/cm²); fill factor = 33.3%; solar conversion efficiency = 0.0257%.

(3) Deposited TiO₂ Nanoparticle and Spin-Coated TiO₂ Film Composite Electrode
An additional spinning-coated TiO₂ layer with a thickness of 3 𝜇m was added to the deposited TiO₂ nanoparticles. The measurement results are shown in Figure 10. The horizontal part of the dark current was about 0 𝜇A, approaching the standard. The open-circuit voltage and the current density were measured to be 0.66 V and 0.436 (mA/cm²), respectively. Accordingly, the fill factor and the solar conversion efficiency were calculated to be 37% and 0.125%.

Figure 10 

Conversion efficiency measurement of the proposed DSSC.

Although the solar conversion efficiency of the proposed DSSC is only 0.125%, this is a 5-fold improvement over that of the conventional DSSC fabricated using a spin-coated TiO₂ film electrode. The low conversion efficiency can be attributed to the following factors.

(1) The Deposited TiO₂ Nanoparticle Layer Was Not Thick Enough
The thickness of the deposited TiO₂ nanoparticle layer was about several tens of nm. Compared to the 10 𝜇m thickness of the conventional TiO₂ thin film electrode, this did not seem thick enough. As a result, light only had a short retention period in the cell.

(2) The ITO Film Was Not Conductive Enough
The resistance measurement results shown in Figure 4 indicate that the resistance of the ITO film produced by annealing at 500^∘C was still on the scale of MΩ. It can be presumed that electrons could not be conveyed efficiently through the ITO electrode.

4. Conclusion

A novel DSSC scheme for better energy conversion efficiency is proposed. The characteristic feature is the replacement of the conventional thin film electrode by a sputtered 3D nanostructured ITO electrode fabricated using an AAO template. The electrophoretic deposition of TiO₂ nanoparticles on the 3D ITO film was followed by soaking in N3 dye to fill vacant spaces in the electrode.

The measured open-circuit voltage and the current density of the proposed scheme were 0.66 V and 0.436 (mA/cm²), respectively. Accordingly, the fill factor and the solar conversion efficiency were calculated to be 37% and 0.125%. It improved 5-fold over that of the conventional spin-coated TiO₂ film electrode DSSC. We expect to improve the conversion efficiency in future works.

Acknowledgment

The authors would like to address their thanks to the National Science Council of Taiwan for their financial support of this work under grant NSC-98-ET-E-005-004- ET.