Moisture Effect on CO Oxidation over Supported Gold Catalysts
Mitsutaka Okumura, Susumu Tsubota, Masatake Haruta,
Supported gold catalysts are featured by their preference of the ambient conditions: low temperatures, as well as the coexistence of oxygen and moisture. Since residual moisture is inevitable in the reactant gas in the atmospheric pressure, it is important to know how moisture affects the catalytic activities, from the viewpoint of industrial applications. On the other hand, moisture concentration is extremely low under the ultra-high vacuum conditions, where model catalysts are investigated by the techniques of surface science. Thus, the effect of moisture concentration should be taken into account in comparing the activities measured under the various conditions, if it is substantially large.
It has been noticed that the activities for CO oxidation by gold catalysts are greatly influenced by moisture in the reactant gas, so that water molecule often appears in the proposed reaction mechanisms. There have been several reports regarding the moisture effect on the activities of gold catalysts. In most cases, however, the effect of moisture is qualitatively discussed above `3 ppm and the reported results are somewhat confusing . Here we report the results of quantitative investigation of moisture effect on the CO oxidation over Au/TiO2, Au/Al2O3 and Au/SiO2, raging from `0.1 ppm to `6,000 ppm H2O.
0.9 wt% Au/TiO2 and 0.4 wt% Au/Al2O3 are prepared by the deposition-precipitation method, while 10 wt% Au/SiO2 catalyst was prepared by the gas-phase grafting method. The catalyst samples were calcined in air at 673 K for 4 h. Mean diameters of gold particles calculated from the transmission electron micrographs were 3.0, 3.9, 8.2 nm for Au/TiO2, Au/Al2O3 and Au/SiO2, respectively. Catalytic activities above `3 ppm H2O were measured with an ordinary fix-bed flow reactor, whereas an ultraclean reactor line  was employed for the measurements below `3 ppm H2O. The reactant gas was 1 vol% CO in air. H2O concentration was monitored with cryoptical dew-point meters, as well as a electrostatic capacitance dew-point meter. Prior to the measurements, the catalyst samples were heated at 523 K for 30 min, except for the thorough drying at `0.1 ppm H2O (673 K for 3 days). Reaction rates are represented by the turnover frequencies per surface Au atom, assuming the 0th-order kinetics.
For all the tested catalyst samples, moisture generally shows a positive effect (Fig. 1). Activity enhancement was no less than two orders of magnitude. Moisture has relatively large effects on both Au/Al2O3 and Au/SiO2. But the effect of moisture is remarkable above `100 ppm for Au/Al2O3, while the activities below `3ppm were considerably decreased for Au/SiO2. In the case of Au/TiO2, the moisture effect below `100 ppm is not so large as the time-dependent change in activity, which often disturb the reproducibility of data. The activity for Au/TiO2 at `3,000 ppm H2O is so high that the accurate reaction rate could not be determined from the saturated CO conversion (100 %).
It is noted that the apparent activation energies for Au/TiO2 and Au/Al2O3 are not so much influenced by moisture. This indicates that moisture does not significantly change the reaction mechanisms for these catalysts. However, the conversion curve for Au/SiO2 is drastically altered by the moisture concentration (Fig. 2). It is not too much to say that moisture in the reactant gas is a prerequisite for the activity of Au/SiO2.
These results show that the moisture effects on CO oxidation over three catalyst samples are clearly different, which suggests that the reaction mechanisms depend on the nature of support. This confirms our hypothesis that the reaction takes place at the perimeter interface between gold particles and the support surface. Since the abrupt switching in H2O concentration is always followed by the gradual change in activity, the amount of H2O adsorbed on the catalyst surface seems to determine the catalytic activity.
There are several possibilities for the origin of moisture effect. Water-gas shift reaction can be excluded, because the concentrations for CO and H2O are extremely different in this study. H2O-derived species, e.g. ÐOH group, may participate in the reaction intermediates, although such species have not been spectroscopically evidenced. Another possibility is the modification of electronic properties near the perimeter interface by moisture, while the decomposition of carbonate species accumulated during the reaction may takes place. Surface characterisation under the ambient conditions are required to obtain further information, which might lead to the elucidation of the reaction mechanisms for the practical gold catalysts.
 G.C. Bond and D.T. Thompson, Gold Bull. 33(2) (2000) 41.
 M. DatEand M. Haruta, J. Catal. 201 (2001) 221.
moisture, Silica, Alumina, CO oxidation, nanoparticle, Titania