The adsorption properties of O(2) molecules on anionic, cationic, and neutral Au(n) clusters (n=1-6) are studied using the density functional theory (DFT) with the generalized gradient approximation (GGA), and with the hybrid functional. The results show that the GGA calculations with the PW91 functional systemically overestimate the adsorption energy by 0.2-0.4 eV than the DFT ones with the hybrid functional, resulting in the failure of GGA with the PW91 functional for predicting the adsorption behavior of molecular oxygen on Au clusters. Our DFT calculations with the hybrid functional give the same adsorption behavior of molecular oxygen on Au cluster anions and cations as the experimental measurements. For the neutral Au clusters, the hybrid DFT predicts that only Au(3) and Au(5) clusters can adsorb one O(2) molecule.

We introduce chemical structures and reactivity of oxygen-centred radicals (O(-)˙) over transition metal oxide (TMO) clusters based on mass spectrometric and density functional theory studies. Two main issues will be discussed: (1) the compositions of TMO clusters that have the bonding characteristics of (or contain) the O(-)˙ radicals; and (2) the dependences (cluster structures, sizes, charge states, metal types, etc.) of the chemical reactivity and selectivity for the O(-)˙ radicals over TMO clusters. One of the goals of cluster chemistry is to understand the elementary reactions involved with complex heterogeneous catalysis. The study of the O(-)˙ containing TMO clusters permits rather detailed descriptions for how mono-nuclear oxygen-centred radicals may exist and react with small molecules over TMO based catalysts.

The nanoscale description of the reaction pathways and of the role of the intermediate species involved in a chemical process is a crucial milestone for tailoring more active, stable, and cheaper catalysts, thus providing “reaction engineering” capabilities. This level of insight has not been achieved yet for the catalytic hydrogenation of CO2 on Ni catalysts, a reaction of enormous environmental relevance. We present a thorough atomic-scale description of the mechanisms of this reaction, studied under controlled conditions on a model Ni catalyst, thus clarifying the long-standing debate on the actual reaction path followed by the reactants. Remarkably, formate, which is always observed under standard conditions, is found to be just a “dead-end” spectator molecule, formed via a Langmuir−Hinshelwood process, whereas the reaction proceeds through parallel Eley−Rideal channels, where hydrogen-assisted C−O bond cleavage in CO2 yields CO already at liquid nitrogen temperature.