|
|
|
Detailed Investigations on Hydrogen Formation by the Photocatalytic Splitting of WaterG. S. Herman, Y. Gao, J. Osterwalder,(a,b) C. S. Fadley,(a,c) C. H. F. Peden, Supported by PNNL Laboratory Directed Research and Development. Our program addresses some of the most poorly understood details of the photocatalytic splitting of water over TiO2 surfaces. These include the nature of the active catalyst surface site where the charge (electrons and holes) is trapped subsequent to electron transfer with water, and the origin of "structure sensitivity" in this catalytic reaction. In particular, we have focused our efforts towards investigating the anatase phase of TiO2, which has been found to be the most photoreactive. The recent availability at PNNL of anatase single crystal surfaces has allowed for a detailed determination of the crystal surface structure. Although considerable work has previously been done on powders, a consistent methodology involving well-characterized single crystal substrates has not yet been undertaken to determine the mechanism and structural specificity of photocatalytic hydrogen production from water. We have used anatase thin films grown at PNNL that have been bulk characterized with x-ray diffraction. These measurements indicate that the films grow as anatase on strontium titanate in a (001) orientation. The x-ray diffraction measurements are essentially a bulk structure probe. To address the available adsorption sites for water on anatase, more surface-sensitive probes are required. The effort to investigate the detailed geometric and electronic surface structure of these films has taken advantage of instrumentation and experimental techniques available in the EMSL, the University of Zurich (UZ) and Lawrence Berkeley National Laboratory (LBNL). We have used x-ray photoelectron diffraction (XPD) to confirm that the top 5-100 Å in the surface region is anatase and does not convert to rutile (the more stable polymorph) under our experimental conditions. There have been discrepancies in the literature whether rutile layers form on top of anatase particles (leading to enhanced photoreactivity), or whether anatase is stable under ultrahigh vacuum conditions. A comparison between experiment (EXP) and single-scattering calculations (SSC) is given in Figure 7.1. The Ti 2p intensity was collected for over 5000 data and are plotted at the top of Figure 7.1. The data are represented as a stereographic projection, and the high intensities are indicated by lighter values. There is very good agreement between experiment and calculations assuming a bulk-like anatase structure. Using a rutile structure in the calculations gave very poor agreement to the experimental results (not shown). These are the first measurements, with atomic specificity, that indicate anatase is stable under ultrahigh vacuum conditions. Furthermore, a direct comparison was made between rutile and anatase using x-ray photoelectron spectroscopy, and it was found that the core-level binding energies were identical for the polymorphs for Ti 2p and O 1s emission. Based on these measurements we have confirmed that the thin films are anatase (001). We have found that the anatase (001) surface undergoes a (1 x 4) reconstruction, as opposed to the expected (1 x 1) bulk termination. This reconstruction was found to be very stable and has not previously been reported. We have used angle-resolved mass-spectroscopy of recoiled ions (AR-MSRI), a technique that has recently been developed in the EMSL, to better understand the surface reconstruction, and the influence it may have on the photocatalytic hydrogen production from water. The experimental and theoretical data, for our best-fit model, are shown in Figure 7.2 as solid and dashed lines, respectively. The intensity of both oxygen and titanium recoiled ions are shown with respect to the sample azimuthal angle. There is considerable variation in intensity, which is directly related to the detailed atomic structure from the very top most surface layers. Theoretical simulations were performed for over thirty different structural models, and the model with the best-fit to experiment is shown at the top of Figure 7.2. This model is very complex and results from microfaceting that exposes lower energy (103) surface planes. Prior work on polycrystalline materials have shown that (101), (001), and (103) surfaces coexist. It is clearly evident that to fully understand the interaction of water with anatase surfaces it is essential to experimentally determine the detailed atomic coordinates of the surface. For example, prior theoretical studies on the interaction of water with anatase used a bulk terminated (1 x 1) anatase surface that apparently is not stable.
Finally, we have performed experiments to determine the electronic structure of anatase. No prior studies have measured the electronic band-structure for anatase, and this is an important first step in comparing experimental and theoretical investigations. Preliminary experiments were performed at the Advanced Light Source in Berkeley, California on beamline 9.3.2. The experimental data shown in Figure 7.3 indicate changes in the shape of the valence band spectra when obtained at several different photon energies (hn). Dispersion of specific peaks in the spectra will be mapped out with respect to their location in the Brillouin zone and then compared to theory.
To date these studies have provided new and unique information that will help clarify the structural sensitivity of the photocatalytic production of hydrogen from water. Both the geometric and electronic structure has been investigated in detail.
William R. Wiley Environmental Molecular Sciences Laboratory Feedback: webmaster@emsl.pnl.gov Revised: April 17, 2000 Security & Privacy PNNL-13147 |
