A four-point probe correlation of oxygen sensitivity to changes in surface resistivity of TiO2(001) and Pd-modified TiO2(001)
John R. McCormick, John R. Kitchin, Mark A. Barteau, and Jingguang G. Chen, “A four-point probe correlation of oxygen sensitivity to changes in surface resistivity of TiO2(001) and Pd-modified TiO2(001)”, Surface Science 545(1-2), L741-L746 (2003). https://doi.org/10.1016/j.susc.2003.08.041
This summary was written by a language model from the paper and has not been checked line by line. Rely on the published paper for details.
Question
Semiconducting metal oxides such as TiO2 serve as gas sensors because adsorbing or reacting gases change their surface resistivity, and dispersed metal particles are known to improve the response. A rational design of such sensors calls for an atomic-level picture of how surface stoichiometry, gas adsorption and supported metal particles each affect the measured resistivity. This short letter asks whether a linear four-point probe (FPP), used in ultrahigh vacuum alongside standard surface science tools, can relate the resistivity of a single-crystal rutile TiO2(001) surface to its Ti/O stoichiometry, to oxygen exposure, and to the size of deposited Pd particles (McCormick et al. 2003).
Approach
Two UHV chambers were used. One (base pressure in the 10-10 Torr range) had Auger electron spectroscopy (AES), a mass spectrometer, a Pd evaporation source and a commercial linear FPP with four tungsten tips 0.25 mm apart; a constant current of 1 to 1000 microamps was passed through the outer tips and the voltage drop measured across the inner ones, with the crystal electrically isolated. The second chamber had AES, a similar Pd source and a variable-temperature STM operated at room temperature in constant current mode (about 2 V, 0.1 nA). The TiO2(001) crystal was cleaned by Ar sputtering and reoxidized by annealing at 773 to 1000 K in roughly 1 to 5 x 10-5 Torr O2, and the surface was assumed to be electrically isotropic. Because some near-surface reduction was unavoidable and starting resistivities could not be reproduced exactly from run to run, results are reported as percent changes from the initial resistivity. Pd was deposited by physical vapor deposition at room temperature and the Pd/TiO2(001) surfaces were annealed for 2 h at different temperatures to sinter the particles, with a freshly cleaned and oxidized crystal for each FPP experiment (McCormick et al. 2003).
Key findings
Sputtering a fully oxidized TiO2(001) surface lowered both its O/Ti ratio (estimated by AES) and its surface resistivity in parallel. After 2500 s of sputtering the surface composition reached about TiO1.5, and the experiment was stopped there to avoid reducing the bulk. The authors interpret the similar trends as confirming that the resistivity tracks the surface TiOx stoichiometry, consistent with the band gap of titanium oxides rising from about 0.9 eV for TiO1.3 to 3.1 eV for TiO2 (McCormick et al. 2003).
Exposing a sputtered surface to 1.1 x 10-5 Torr O2 at 300 K increased the resistivity over time, as expected if adsorbed oxygen removes conduction electrons from the near-surface region. The resistivity stayed constant after the dosing valve was closed, so adsorption was irreversible at 300 K. Assuming the resistivity change is proportional to oxygen coverage, the authors fit the transient with first-order (molecular) and second-order (dissociative) Langmuir kinetics using the same asymptote of 3.35 percent change; the second-order model fit much better, pointing to dissociative O2 adsorption on TiO2(001) at room temperature (McCormick et al. 2003).
STM showed Pd sintering with annealing: typical particles were 1 to 6 nm at 300 K, 5 to 12 nm after 573 K, and 10 to 23 nm after 773 K. On exposure to 2.5 x 10-7 Torr O2 at 300 K, freshly deposited Pd gave little improvement over bare TiO2(001), the 573 K annealed surface was clearly more sensitive, and the 773 K annealed surface fell back to nearly the bare-oxide response. The authors attribute the enhancement to Pd particles acting as sites that adsorb and dissociate O2 and then spill oxygen onto the oxide, with the 573 K treatment giving the most effective number of such sites and further sintering reducing them. They note that spectroscopic work is still needed to quantify the link between particle size and the oxygen adsorption rate (McCormick et al. 2003).
Significance
The letter shows that an FPP can be integrated with AES and STM in UHV to follow surface electrical transport on a model oxide single crystal, providing a bridge between surface science and the response of chemiresistive gas sensors. It links resistivity to surface reduction, uses the resistivity transient as a probe of adsorption kinetics, and shows that the promoting effect of a supported metal on sensing is size dependent, with an intermediate particle size performing best (McCormick et al. 2003).