The water-gas shift (WGS) reaction rate per total mole of Au under 7% CO, 8.5% CO(2), 22% H(2)O, and 37% H(2) at 1 atm for Au/Al(2)O(3) catalysts at 180 °C and Au/TiO(2) catalysts at 120 °C varies with the number average Au particle size (d) as d(-2.2±0.2) and d(-2.7±0.1), respectively. The use of nonporous and crystalline, model Al(2)O(3) and TiO(2) supports allowed the imaging of the active catalyst and enabled a precise determination of the Au particle size distribution and particle shape using transmission electron microscopy (TEM). Further, the apparent reaction orders and the stretching frequency of CO adsorbed on Au(0) (near 2100 cm(-1)) determined by diffuse reflectance infrared spectroscopy (DRIFTS) depend on d. Because of the changes in reaction rates, kinetics, and the CO stretching frequency with number average Au particle size, it is determined that the dominant active sites are the low coordinated corner Au sites, which are 3 and 7 times more active than the perimeter Au sites for Au/Al(2)O(3) and Au/TiO(2) catalysts, respectively, and 10 times more active for Au on TiO(2) versus Al(2)O(3). From operando Fourier transform infrared spectroscopy (FTIR) experiments, it is determined that the active Au sites are metallic in nature. In addition, Au/Al(2)O(3) catalysts have a higher apparent H(2)O order (0.63) and lower apparent activation energy (9 kJ mol(-1)) than Au/TiO(2) catalysts with apparent H(2)O order of -0.42 to -0.21 and activation energy of 45-60 kJ mol(-1) at near 120 °C. From these data, we conclude that the support directly participates by activating H(2)O molecules.

Au/TiO(2) catalysts used in the water-gas shift (WGS) reaction at 120 °C, 7% CO, 22% H(2)O, 9% CO(2), and 37% H(2) had rates up to 0.1 moles of CO converted per mole of Au per second. However, the rate per mole of Au depends strongly on the Au particle size. The use of a nonporous, model support allowed for imaging of the active catalyst and a precise determination of the gold size distribution using transmission electron microscopy (TEM) because all the gold is exposed on the surface. A physical model of Au/TiO(2) is used to show that corner atoms with fewer than seven neighboring gold atoms are the dominant active sites. The number of corner sites does not vary as particle size increases above 1 nm, giving the surprising result that the rate per gold cluster is independent of size.

Platinum atomic layer deposition (ALD) using MeCpPtMe 3 was employed to prepare high loadings of uniformsized, 1-2 nm Pt nanoparticles on high surface area Al 2 O 3 , TiO 2 , and SrTiO 3 supports. X-ray absorption fine structure was utilized to monitor the changes in the Pt species during each step of the synthesis. The temperature, precursor exposure time, treatment gas, and number of ALD cycles were found to affect the Pt particle size and density. Lower-temperature MeCpPtMe 3 adsorption yielded smaller particles due to reduced thermal decomposition. A 300°C air treatment of the adsorbed MeCpPtMe 3 leads to PtO. In subsequent ALD cycles, the MeCpPtMe 3 reduces the PtO to metallic Pt in the ratio of one precursor molecule per PtO. A 200°C H 2 treatment of the adsorbed MeCpPtMe 3 leads to the formation of 1-2 nm, metallic Pt nanoparticles. During subsequent ALD cycles, MeCpPtMe 3 adsorbs on the support, which, upon reduction, yields additional Pt nanoparticles with a minimal increase in size of the previously formed nanoparticles. The catalysts produced by ALD had identical water-gas shift reaction rates and reaction kinetics to those of Pt catalysts prepared by standard solution methods. ALD synthesis of catalytic nanoparticles is an attractive method for preparing novel model and practical catalysts.