![]() |
|
Low-Energy Cathodoluminescence Spectroscopy of Al-Doped TiO2 FilmsS. Goss,(a) L. J. Brillson,(a) and S. A. Chambers Supported by Office of Biological and Environmental Research, EMSL Operations, and OBES Division of Materials Science. Impurity doping of TiO2 enhances the catalytic and photocatalytic activity of this broadly utilized material. Nb doping increases the extent of photocatalytic destruction of chlorinated organic substances by ~100% by increasing the oscillator strength for interband electron-hole pair creation (Cui et al. 1995; Chambers et al. 1996). Nb is an n-type dopant in TiO2 when present in the parts per thousand level. However, when present at the few to several at. % level, substitutional Nb creates a nonbonding band that falls at an energy deep in the valence band region (Chambers et al. 1996). Nb substitutes for Ti at cation sites in the lattice (Chambers et al. 1996). As a Group V transition metal, Nb brings in one extra d electron per displaced Ti, thus increasing the oscillator strength for interband transitions. Al doping is used in industry to enhance the properties of TiO2 for a number of technologies, but the reasons for the effectiveness of Al doping are not known. Al, being a Group III metal, is expected to generate an acceptor state in the forbidden gap of TiO2. The goal of this project is to determine the electronic structure of Al-doped TiO2(110). Cathodoluminescence spectroscopy was chosen as a probe because of the likelihood that any new state(s) introduced by substitutional Al will be unoccupied. We have used electron-excited nanoscale luminescence spectroscopy (EENLS) to determine the dependence of electronic structure on chemical composition in Al-doped TiO2 epitaxial films. These films were grown by oxygen-plasma-assisted molecular beam epitaxy and characterized at PNNL, and then shipped to Ohio State University (OSU) for EENLS measurements. Al was shown by scanned-angle x-ray photoelectron diffraction to substitute for Ti in the lattice. Low-energy electron diffraction and reflection high-energy electron diffraction patterns reveal that excellent long-range order is maintained up to at least 6 at. % Al dopant level. Using incident electron beam energies from 0.6 to 4 kV to probe depths from 5 to >150 nm below the surface, we observe mid-gap state emission at 1.4 eV that may be due to Al doping, broad O vacancy emission centered at ~2.5 eV, and near-band-edge (NBE) transitions at 2.9-3.1 eV. The 1.4 eV emission appears specifically within a 20-nm-thick 6% Al-doped TiO2 layer grown on 4% and 2% doped layers of the greater thickness, all grown epitaxially on vacuum-reduced TiO2(110) substrates. No emission was seen from the deeper layers of lower dopant concentration. Recombination involving the 1.4 eV level increases dramatically with annealing at 600°C under 5 x 107 L oxygen treatment, while the 2.5 eV peak decreases. As mid-gap recombination increases, NBE emissions decrease strongly, indicating a pronounced decrease in free carrier concentration near the free surface. Mid-gap emission intensity shows no correlation with surface Al concentration, suggesting that the exact state of the near-surface region has a much larger influence on the emission intensity than does the Al concentration. The O vacancy peak intensity at ~2.5 eV decreases only gradually with increasing depth of penetration and this profile is affected only slightly by oxygen inter-diffusion. These results are summarized in Figure 2.3, and demonstrate the utility of EENLS to probe ultrathin wide band gap oxide films with nanometer scale depth resolution. The potential role of Al as the source of mid-gap states, and the effect of O inter-diffusion to further activate Al-derived mid-gap states as efficient recombination centers is being studied further. Furthermore, the dramatic changes in free carrier concentration with oxidation and deep-level activation suggest a role for recombination centers in controlling surface catalytic properties. We are currently carrying out secondary ion mass spectroscopy measurements for samples of varying Al dopant levels to make a more definitive connection between the presence of Al in the lattice and the 1.4 eV emission peak.
ReferencesChambers, S. A., Y. Gao, Y. J. Kim, M. A. Henderson, S. Thevuthasan, S. Wen, and K. Merkle, Surf. Sci. 365, 625 (1996). Cui, H., K. Dwight, S. Solad, and A. Wold, J. Solid State Chem. 115, 187 (1995).
William R. Wiley Environmental Molecular Sciences Laboratory Feedback: webmaster@emsl.pnl.gov Revised: June 12, 2001 Security & Privacy PNNL-13147 |