The Peculiar Case of Gold Nanoparticles

Prom the numerous results published on catalytic gold, it is clear that for particles larger than about 5nm the activity is very low. Haruta and co-workers [73] showed that for the CO oxidation at RT on gold catalysts prepared by chemical methods, the activity is maximum for sizes of around 3nm (see Fig. 3.16). This result was later confirmed by Goodman and coworkers [74] on Au particles grown by vacuum evaporation on TiO2(110). 

Many hypotheses have been proposed to explain the peculiar behaviour of gold nanoparticles. One of the first explanations attributes these observations to a quantum size effect. Indeed, Goodman and co-workers [74] observed on gold supported on Ti02 by STS the opening of a band gap for particles of 4nm or smaller (see Fig. 3.16). For size-selected clusters soft-landed on MgO 

The mechanism of the reaction for CO oxidation (the most studied reaction) is still an open question. Some papers conclude that the reaction occurs between adsorbed oxygen atoms resulting from the dissociation of an oxygen molecule [76,80,81] while others claim that adsorbed oxygen molecules react directly with adsorbed CO [21,75,78,79] or eventually that both mechanisms occur [82]. 

In a recent study, we have investigated the reactivity of gold nanoparticles prepared by UHV-deposition on in situ cleaved MgO surfaces [83,84]. Particles grown at 200° C or higher were in epitaxy on the MgO surface and presented truncated octahedron shapes, which is the equilibrium shape [85]. The profile of such a particle observed by HRTEM is presented in Fig. 3.18. The aspect ratio is 0.6. 

For the first time, a steady-state production of CO2 is observed at low pressure on a gold catalyst. The CO conversion is very high 50%. On 4-nm 

---