The principal drive toward fabricating nano-dimensional materials lies in the promise of achieving unique properties and superior performance due to these materials’ inherent nano-architectures. Titanium dioxide is a particularly versatile material with many technological applications, including a photocatalyst, photovoltaic material, gas sensor, optical coating, structural ceramic, electrical varistor, biocompatible material for bone implants, and a spacer material for magnetic spin valve systems. Therefore, the ability to control nanoscale TiO2 architectures can be expected to positively impact many technologies.
In conjunction with a suitable organic dye or semiconductor, TiO2 is a key material for low cost, large area, excitonic photovoltaic cells.1 Titania nanotubes fabricated by anodization are ordered, high aspect ratio structures with nanocrystalline walls oriented perpendicular to the substrate. These nanotubes have a well-defined and controllable pore size, wall thickness, and tube-length (see Figure 1). For a 20μm nanotube array, the effective surface area is approximately 3000 times that of a planar unstructured surface. Like most one-dimensional structures, this high-aspect ratio nanotube array architecture promotes efficient harvesting of photons by orthogonalizing the processes of light absorption and charge separation.2 The tube geometry is unique in that it allows for an active material to be either conformally coated on the walls of the nanotubes (where the large surface area is utilized), or filled into the pores of the semiconductor, which promotes efficient exciton harvesting.
Figure 1. Field-emission scanning electron microscopy (SEM) image of a 46.6μm long TiO2 nanotube array. The insets show the top view [top right] and cross-section [bottom right] of the arrays at higher magnification.
We have fabricated two varieties of nanotube arrays, so-called transparent and non-transparent. Non-transparent nanotube arrays are grown on an opaque metallic Ti substrate, while transparent nanotubes are formed by anodizing a Ti film sputtered onto a conductive glass substrate. We have optimized processes to fabricate opaque nanotubes up to 1mm in length, although calculations indicate that ∼30–40μm long nanotube arrays are optimal to maximally harvest light without suffering losses. The optimization of our processes to produce transparent nanotubes is ongoing, and we are currently able to reproducibly make nanotubes 4μm long. Transparent nanotubes lend themselves to frontside illumination, which avoids photonic losses due to absorption by the platinized counter electrode and the redox electrolyte in dye-sensitized solar cells. Backside illumination geometry, where the aforementioned photonic losses occur, is the only mode of illumination possible for non-transparent nanotube arrays.
Figure 2. Schematics of illumination geometries. [left] Frontside geometry with a transparent TiO2 nanotube array. [right] Backside geometry with a non-transparent TiO2 nanotube array.
Our research group pioneered the development of transparent and non-transparent titanium dioxide nanotube arrays.3,4 We were the first to report functioning dye-sensitized solar cells using TiO2 nanotube arrays in both frontside and backside illumination geometries (see Figure 2).5 We have demonstrated liquid junction dye-sensitized solar cells with an efficiency of up to 7% using commercially available ruthenium bipyridine containing sensitizer N-719.6 More recently, we have been investigating other sensitizers, including donor antenna dyes and carboxylated polythiophenes.
A principal limitation of the widely used nanoparticulate TiO2 electrodes is that electron transport occurs through trap-limited diffusion, a slow process that allows for back-electron transfer. The concomitant recombination losses limit device efficiencies, especially at longer wavelengths. Preliminary studies of electron transport and recombination in TiO2 nanotube arrays indicate that the recombination lifetimes of photogenerated charge carriers are longer than those of nanoparticulate electrodes. However, the transport times are similar due to the nanocrystalline nature of the tube walls (typical grain size ∼100nm). We are, therefore, pursuing various strategies to increase the grain size and improve the crystallinity of the nanotube arrays.
We are also actively investigating the possibility of using TiO2 nanotubes to improve the performance of bulk heterojunction polymeric solar cells. One strategy involves confining the semiconducting polymer blend in the nanotubes. This confinement alleviates the problem of segregation of the polymer phases, provides a second heterojunction (n-type TiO2) for charge separation, and uses the nanotube architecture as an electron-accepting network to limit electrical dead ends in the polymeric device. As shown in Figure 3, by infiltrating our nanotubes with a blend of a hole transporting polymer, poly(3-hexylthiophene), and a methanofullerene, [6,6]-phenyl-C61-butyric acid methyl ester (PCBM), we have obtained 4.1% efficient solid-state double heterojunction solar cells.7
Figure 3. [left] Schematic illustration and [right] electrical characteristics of a double heterojunction solid state solar cell incorporating TiO2 nanotubes.
Craig Grimes, Karthik Shankar
Pennsylvania State University
University Park, PA
Craig Grimes is a professor of electrical engineering at Penn State University.
Karthik Shankar is a member of the Center for Solar Nanomaterials at Penn State University.