Abstract

In our previous works, we have made alumina template films and used it for nanowire, nanowhisk, and nanosphere fabrication and molecular aggregation studies. In the present paper, we have combined Al2O3 template and titania (TiO2) NT fabrication processes to achieve an Al2O3/TiO2 NT dye-sensitized solar cell (DSSC) devices. The DSSC structure includes glass substrate, transparent conductive film of ITO, Pt particles serving as the counter electrode, Al2O3/TiO2 NT film, dye, and ITO serving as the working electrode, and the electrolyte is injected into the counter-working interface. Al2O3 template was made by anodization and TiO2 NT was made by sol-gel deposition into Al2O3 template. Al2O3 template has a light, transparence, large surface, good mechanical strength, and flexibility, making it a candidate material for DSSC electrode template. TiO2 NT is a semiconductor with an energy gap that matches up very nicely with N3 sensitized dye.

1. Introduction

Due to increasing energy demands and concerns about global warming, scientists are looking for potential renewable energy sources. Because the sun is the most important inexhaustible and clean energy source, efficiently harvesting solar energy to generate electric power using photovoltaic technology beyond silicon systems has undergone rapid development over the past few years.

Presently there are several technical schemes for solar cell design, including monocrystalline/polycrystalline silicon solar cells, amorphous silicon solar cells, thin film solar cells, and wet type dye-sensitized solar cells (DSSCs). Of these, monocrystalline silicon solar cells currently have the leading position in the market due to their relatively high transformation efficiency (12–20%). However, since monocrystalline silicon wafers are expensive, manufacturing costs for these cells are high. DSSCs have gradually become more popular due to their lower cost and relatively simple manufacturing process. A DSSC consists of an anode, electrolytic solution, and a cathode. A semiconductor layer is formed on the surface of the anode and photosensitive dyes are absorbed therein.

Since the development of low-cost DSSC technology in 1991 by O’Regan and Grätzel [1], DSSC has been regarded as a promising candidate for next-generation solar cell design [2]. Traditionally, the electron-collecting layer (anode) of a DSSC is composed of randomly packed TiO2 nanoparticles (NPs). With sunlight irradiated from the transparent anode (front illumination), the best photovoltaic power conversion efficiency (η) of an NP-DSSC device has reached ~11% [35]. The great advantage of NP-DSSC is the large surface area of the nanoporous TiO2 films for dye adsorption. However, the trap-limited diffusion for electron transport in NP-DSSC is a limiting factor in achieving higher light-to-electricity conversion efficiency [6]. To improve charge-collection efficiency by promoting faster electron transport and slower charge recombination, several different methods have been established using TiO2 films constructed of oriented one-dimensional (1D) nanostructures. For example, the best cell performance of DSSCs based on 1D TiO2 nanowires (NWs) has reached η = 5.0% under front-side illumination [7, 8]. Also, 1D TiO2 nanotubes (NTs) have been synthesized using sol-gel [9] and potentiostatic anodization [10] methods.

NT-DSSCs using the TiO2 NT arrays on Ti foil as working electrodes have three important intrinsic features that allow it to outperform the conventional NP-DSSCs. First, the charge-collection efficiencies of NT films have been proved to be much better than those of NP films because of the 1D nature of the former, with a much slower charge recombination rate [11, 12]. This intrinsic advantage of NT-DSSC promotes its cell performance significantly with tube lengths up to 20 μm, as reported by Grimes and coworkers [13]. Second, the light-harvesting efficiencies of the NT films are much better than those of the NP films because of the stronger light scattering effect of the former. In fact, for a traditional high-efficiency NP-DSSC, adding an additional TiO2 layer with a larger particle size (~400 nm) is required to increase the light scattering effect, while this effect is a natural property for an NT-DSSC. Third, the anode fabrication of NT-DSSCs is much simpler and more cost-effective than that of NP-DSSCs. Direct anodization of a Ti foil in a one-step process produces the blank TiO2 film ready to make a working electrode in NT-DSSCs, whereas making the blank TiO2 film in NP-DSSCs requires a multiple coating process for at least two layers of TiO2 NPs coated on a relatively expensive transparent conducting oxide (TCO) substrate.

The use of a longer NT in the device causes certain problems. For example, cracking and a compact layer are produced on the surface of the NT films, and the longer the TiO2 NT, the weaker adhesion of NT on the Ti substrate. To reduce the cracks and eliminate the compact layer for a longer TiO2 NT, Grimes [13] and coworkers used 0.1 M hydrochloric acid (HCl) to dissolve the compact layer; Zhu and coworkers [12] used a supercritical CO2 drying technique to produce bundle-free and crack-free NT films. Chen et al. [14] used Al2O3 microparticles with the aid of ultrasonic vibration to remove the compact layer and to avoid crack formation. The mechanical strength of TiO2 NT is not great, so a longer tube can crack easily, leading to the peeling off of the NT film from the Ti substrate. Therefore, we propose to use anodic aluminum oxide (AAO or Al2O3 template) as a template, followed by deposition of TiO2 NT inside AAO to make a working electrode for NT-DSSC. AAO is fabricated by anodization process, and then the chemical etching method is used to achieve AAO nanotubes with open pores on both sides. Because AAO has better mechanical properties than TiO2 NT, the AAO morphology can also be controlled more easily than can that of TiO2 NT. For example, AAO pore diameter, length, and pore density can be controlled in the range of 10–500 nm, 0.1–200 μm, and 108–1012 tubes cm−2, respectively.

According to reports from Imai [1517], when TiF4 solutions have a pH below 1.0 or a TiF4 concentration below 0.03 M, neither precipitation nor film formation was observed. A large amount of precipitate was rapidly formed, and film was not deposited, on substrates above pH 3.1. Therefore, the concentration of TiF4 should be controlled above 0.03 M, and the pH value should be controlled between 1 and 3 so that TiO2 nanoparticles can present and be deposited on the substrate. For example, Chen et al. [18] used pH 1.8, 0.04 M TiF4 solution at 60°C for deposition on a TiO2 NT in AAO membrane.

In our previous work, we have made AAO film [19] and used this templates for the fabrication of nanowires [2023], nanowhisks [24], and nanospheres [2529] and the study of molecular aggregation [30, 31]. We have also made TiO2 NT film [3236] and used it for DSSC device fabrication, allowing the photovoltaic power conversion efficiency to reach η~7% [14, 37]. In this paper, we have combined our previous Al2O3 template, TiO2 NT, and DSSC fabrication technologies to achieve an Al2O3/TiO2 NT-DSSC device.

2. Experimental

2.1. Al2O3 Template Fabrication

Al2O3 templates with a pore size of 10 to 500 nm were generated by anodizing a commercial aluminum (Al) substrate (99.7%) in acid solutions of sulfuric acid (H2SO4), oxalic acid (COOH)2, or phosphoric acid (H3PO4). The Al substrate was first ground to no. 1000 by SiC waterproof paper and then annealed in an air furnace at 550°C for 1 hr. The sample was then electropolished in a bath consisting of 15 vol.% perchloric acid (HClO4, 70%), 70 vol.% ethanol (C2H6O, 99.5%), and 15 vol.% monobutyl ether ((CH3(CH2)3OCH2CH2OH), 85%) with a charge of 42 volts (DC) applied for 10 min, used platinum plate as a counter.

A 10 nm pore diameter template was then fabricated by anodizing the polished-Al substrate at 18 V in 10 vol.% H2SO4 at 15°C for 20 min, which was the first anodization. In order to obtain an orderly pattern on the substrate for the second anodization, the first anodization film was removed in 1.8 wt.% chromic acid (CrO3) + 6 vol.% H3PO4 solution at 60°C for 40 min. The resulting substrate, with a regular pattern on the surface, was used for the second anodization for several hours to form an AAO film with various thicknesses. After Al2O3 template formed, the Al substrate was removed by placing the sample in saturated copper chloride (CuCl2) + 10 vol.% hydrochloric acid (HCl) for 30 min. Finally, the sample was put in 5 vol.% H3PO4 at 25°C for 5 to 20 min. The nanotubes were widened to an ordered array and a good quality Al2O3 template film with 10 to 50 nm pore size was formed. Similar to the above process, for 40 to 90 nm pore diameter Al2O3 template, the electrolyte was 3 vol.% (COOH)2 at 25°C, and the applied voltage was 40 V. The time of pore widening was 10 to 90 min. For 180 to 500 nm pore diameter AAO template, the electrolyte was 1 vol.% H3PO4 at 0°C, and the applied voltage was 200 V. The time of pore widening was between 30 to 200 min [19, 20, 25, 38, 39].

2.2. DSSC Device Fabrication

Three key components are essential to construct a sandwich-type DSSC device: (1) a light-harvesting layer of TiO2/Al2O3 template film with a TiO2 and ITO coating to serve as a working electrode (anode); (2) a Pt-coated layer deposited on an ITO surface to serve as a counter electrode (cathode); (3) an iodine-based electrolyte filled into the space between the anode and the cathode to serve as a redox couple of the cell. The TiO2 NT was prepared by immersing Al2O3 template in 0.02 M titanium fluoride (TiF4) solution (pH = 3) at 25°C for 120 min, followed by annealing at 450°C for 1 hr to obtain anatase TiO2 NT in the Al2O3 template. To prevent short-circuits, on the template bottom was deposited a 300 μm compact TiO2 film by a 4-inch disk of Ti target (99.9% purity) sputtered in a gaseous mixture of 90% Ar and 10% O2. Sputtering power, gas pressure, substrate temperature, and sputter time were kept constant at 50 W, 5 0 × 1 0 3 , 50°C, and 80 min, respectively. Subsequently, a 120 μm ITO film was sputtered on TiO2 film to form a conducting electrode by sputter-deposition. An RF magnetron sputtering system was used for ITO preparation. A 4 inch disk of ITO target (99.9% purity) was sputtered on the TiO2 film. Sputtering power, gas pressure, substrate temperature, and sputter time were kept constant at 100 W, 5 0 × 1 0 3   torr, 30°C, and 30 min, respectively. The electrode fabrication process is shown in Figure 1. The electrode (Al2O3 template/TiO2 NT/TiO2 film/ITO film) was then soaked in ethanol containing 5 × 1 0 4  M RuL2(NCS)2 (N3 dye) for 7 hr to absorb N3 dye, forming the DSSC anode. The micromorphology of Al2O3 template, TiO2 NT, TiO2 compact film, and ITO film were determined by scanning electron microscope (SEM, JEOL 6500).

The Pt counter electrodes were prepared by sputter-deposition. A DC magnetron sputtering system was used for the Pt particles. A 4-inch disk of Pt target (99.9% purity) was sputtered. Sputtering power, gas pressure, substrate temperature, and sputter time were kept constant at 10W, 50 ×10−3, 30°C, and 5 sec, respectively.

To prevent short circuits, the cell was assembled with both anode and cathode sealed by using a hot-melt spacer of 25 μm thickness; small holes on the corners of the cell were reserved for the entrance of the electrolyte and the exhaust of the air. A thin layer of electrolyte was sucked into the interelectrode space by capillary force, and the holes were sealed by epoxy resin. The typical electrolyte solution consisted of 0.5 M (lithium iodide) LiI and 0.05 M iodine (I2) in acetonitrile (CH3CN, 99.9%). The photocurrent was produced using an HP model 4140 B measuring unit. The photocurrent conversion efficiency was tested under an AM 1.5 (300 W, 91160 Oriel Solar Simulator, 100 mw cm−2) on a 0.28 cm2 sample area.

3. Results and Discussion

The Al2O3 template was made by anodization, and the pore size was controlled by electrolyte and the pore-widening period. Figure 2 shows SEM images of Al2O3 template film with diameter of (a) 10 nm Al2O3 template, produced by 10 vol.% H2SO4 anodization and 5 vol.% H3PO4 pore widening for 5 min; (b) 50 and (c) 90 nm Al2O3 template produced by 3 wt.% C2H2O4 anodization and 5 vol.% H3PO4 pore widening for 30 min and 90 min; and (d) 300 nm, (e) 400 nm, and (f) 500 nm Al2O3 template produced by 1 vol.% H3PO4 anodization and 5 vol.% H3PO4 pore widening for 90 min, 180 min, and 250 min, respectively. Using the various electrolytes and pore-widening periods, pore sizes of 10 to 500 nm were obtained.

A barrier layer, which is a dense Al2O3 film on the bottom of Al2O3 template, and the electronic insulator film of the barrier resist electron transport. To make the Al2O3 template as an electrode, the barrier layer should first be removed and then replaced with a coating of conductance film on the template bottom. In our previous results [1821], the thickness of the barrier layer was proportional to the anodic applied voltage; the relationship between thickness and voltage can be denoted as nm/V. For example, the pore sizes of 10 to 50 nm, 40 to 90 nm, and 180 to 500 nm Al2O3 templates are anodized by 18, 40, and 200 V, producing barrier layers of 18, 40, and 200 nm, respectively. Therefore, the larger the pore sizes, the thicker the barrier layer and longer the pore-widening period needed. Figure 3 shows SEM images of the bottom of the Al2O3 template film: (a) a compact Al2O3 barrier layer covering the bottom of the Al2O3 template, (b) the barrier dissolved by 5% H3PO4 solution at 25°C for 10 min to etch grain boundaries, (c) 30 min, during which the barrier layer was partially dissolved, (d) 50 min, during which the barrier layer became a thin film, (e) 70 min, during which most of barrier layer was removed, and (f) 90 min, after which the barrier layer was removed completely.

The Al2O3 template had a larger surface area and transparent film. Figure 4 shows images of transparent Al2O3 template film: (a) original 4-inch transparent Al2O3 template film and (b) the color changing to light red after N3 sensitized dye was absorbed. As can be seen, an Al2O3 template structure with discontinuous pores but continuous pore walls gives Al2O3 template a light, transparent, large surface, and good mechanical strength and flexibility, making it a candidate material for DSSC electrode template. Before adding a coating of a conductive film of ITO on the bottom of the Al2O3 template, a semiconductor film of TiO2 should first be coated on the bottom of the template. A compact TiO2 film, called an under layer, which serves as an electron transport film (from dye to the conductance electrode) but prevents short circuits (between anode and cathode) is thus formed. Figure 5 shows SEM images of (a) a side view of a 300 μm TiO2 film on an Al2O3 template surface and (b) a top view of compact TiO2 film, which can be an under layer on the DSSC anode, and the rough film surface will enhance the following adherence properties of ITO on TiO2 film.

Figure 6 shows side-view SEM images of Al2O3 template made with 10% H2SO4 solution and pore widening for 8 min with (a) a tube diameter of 20 nm, and (b) TiO2 NT deposited inside Al2O3 template. Figure 7 also shows side-view SEM images of Al2O3 template made by 1% H3PO4 solution and pore widening for 80 min with (a) a tube diameter of 280 nm and (b) TiO2 NT deposited inside Al2O3 template. The thickness of the TiO2 NT pore wall depends on the deposition time. A longer deposition time causes a smaller TiO2 NT pore size, and a TiO2 rod presents after a long deposition time. However, the TiO2 NT surface area is reduced, with increased NT walls. Figure 8 shows SEM top view images of the interiors of 280 nm Al2O3 template with TiO2 NT deposited by 0.02 M TiF4 solution ( p H = 3 ) at 25°C for (a) 15 min, (b) 30 min, (c) 60 min, and (d) 120 min. TiO2 particles were deposited inside Al2O3 template and gradually formed a network film on Al2O3 template. Figure 9 shows a schematic diagram of electron (e) and hole (h+) transport in the TiO2 NT. The ITO film serves as a conductive film, and the TiO2 compact film severs as an under layer that resists electrolyte contact with the ITO film. Electrons are transported through the TiO2 NT to the TiO2 compact film and then conducting electrode; on the other hand, holes are transported through the electrolyte to counter. In this one-dimensional channel for carrier transportation, the amount of recombination of electron-hole (e/h+) is expected to be reduced. Figure 10 shows a schematic diagram of a DSSC device fabricated with Al2O3 and TiO2 tubes. The structure includes glass, ITO, Pt particles serving as a counter electrode, and Al2O3/TiO2 tubes film, dye, TiO2, and ITO serving as a working electrode. The electrolyte is injected into the counter-working interface. Based on the structure in Figure 10, the current-voltage characteristics of Al2O3/TiO2 NT-DSSC with N3 dye were measured as in Figure 11. The performance of the device was measured as having 2.7% conversion efficiency, 0.72 V of open circuit voltage, 5.5 mA cm−2 of short circuit current density, and 0.69 of fill factor. An awkward sealing technology with thin film DSSC device may cause short circuit effects between anode and cathode. Also, the device performance had a lower of open circuit voltage fill factor and short current density.

4. Conclusions

We propose the use of anodic aluminum oxide as a template followed by deposition of TiO2 NT inside the template to make a working electrode for flexible NT-DSSC. An alumina template was fabricated by anodization and chemical etching to achieve Al2O3 nanotubes with open pores on both sides. Al2O3 template with pore sizes of 10 to 50 nm, 40 to 90 nm, and 180 to 500 nm was anodized by 18, 40, and 200V. The thicknesses of the barrier layers were 18, 40, and 200 nm, respectively. Al2O3 template has better mechanical properties than TiO2 NTs and the Al2O3 template morphology can also be controlled easily than TiO2 NT. Based on the AAO and TiO2 NT as an electron transport film, the Al2O3/TiO2 NT-DSSC device was fabricated, and the performance of the device was measured at 2.7%.

Acknowledgments

Part of this study was supported by grants from the National Science Council, Taiwan (NSC-99-2221-E-239-025-). The authors would also like to thank Feng Chia University and National United University for financially supporting this work.