Paper SS-TuP30
UV Induced Photodesorption of O₂ on Rutile TiO₂(110): An Angular Imaging Study

Tuesday, November 1, 2011, 6:00 pm, Room East Exhibit Hall

The binding states of oxygen on TiO₂(110) are important for many reactions, including the degradation of organic compounds and photodesorption. Experimental data and theoretical calculations have shown many different configurations for O₂ on TiO₂(110), including O₂^-, O₂^2-, and O₄^-. Defect sites in titania (bridging oxygen vacancy, Ti³⁺ interstitial), while necessary for the adsorption of oxygen, also play a role in the initial binding states. In this experiment, the angular distributions and images of oxygen photodesorption from a rutile TiO₂(110) surface are studied under UHV conditions using a pump-probe Time-of-Flight (TOF) detection scheme to help determine the initial binding states of oxygen on this surface. Excitation occurs via exposure to 3.7 eV photons followed by one-photon ionization using 13.05 eV photons. The delay time between the lasers can be varied according to the maximum desorption velocity of the oxygen molecules. Ions were detected using a dual microchannel plate and a phosphor screen. A CCD camera positioned behind a phosphor screen captured the light emission of the phosphor, allowing for the imaging of the desorbing neutral O₂ molecules. SimION 3D was used to simulate the motion of the ions through a Time-of-Flight (TOF) mass spectrometer to generate a probability distribution for detection that was used to compare images at different delays.

Previous experiments on the oxygen velocity distribution on TiO₂ (110) showed 3 different “channels” for desorption, with two being “fast” and one being “slow”. The velocity distribution for the slow channel tracked with surface temperature, indicating that a trapping desorption mechanism dominated this channel. The two “fast” channels, however, did not depend on temperature and were attributed to two different oxygen binding states on TiO₂ (110). Many different binding states of oxygen have been predicted and it is unclear which state is responsible for which channel. Images of the fastest channel and comparisons at different coverage and photon flux are shown.

The methyl radical velocity distributions from 4 different ketones (acetaldehyde, acetone, butanone, and acetophenone) were also investigated. For ketones where methyl radical desorption is not the preferred pathway (butanone), a 2+1 REMPI scheme was used for signal enhancement. Using these model systems allows for direct comparison of different properties of each molecule.