Although extended gold surfaces are generally considered chemically inert [1,2] nanosized (< 5 nm) gold particles can be very effective catalysts for a number of oxidation reactions. [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17] There are reports of similar size effects for silver catalysts. [18,19] The origin of the nanoeffects in the catalytic properties of these metals is widely debated, [15] and no consensus has been reached. Based on a set of density functional theory calculations of the full reaction pathway for CO oxidation over extended surfaces as well as over small nanoparticles of a number of metals, we show that although platinum and palladium are the most active catalysts for extended surfaces at high temperatures, gold is the most active for very small particles at low temperature. The calculations capture the special catalytic properties of nanosized particles observed experimentally, which allows the origin of the effect to be analyzed.Herein, we focus on intrinsic metal effects; that is, we do not include additional possible effects that involve the support. It is not that such effects may not be important, [5,20,21] but it is useful to first establish the intrinsic metal effects, [15] in particular as it has been shown experimentally that nanostructured gold with no support is also catalytically active. [22,23] The key feature of our analysis is that we compare catalytic activities of different transition and noble metals for one specific reaction, the CO oxidation.The CO oxidation reaction on close-packed fcc(111) surfaces was considered initially, which will give a dominant contribution to the total catalytic rate over large metal particles. We consider the following elementary reactions:For the metals we consider herein, Reactions (R1) and (R2) are unactivated and fast, and we assume that these two reactions are in equilibrium. This means that we are limited to temperatures high enough that desorption is also fast. The possible formation of an oxide layer on the more reactive metals is neglected.The forward and reverse rate constants of the Reactions (R3) and (R4) are given by, where n i is a prefactor, E ai is the activation energy, k is the Boltzmann constant, and T is the absolute temperature. The activation energies are E a = max(E TS ÀE IS , 0) where E IS is the initial state energy and E TS is the transition-state energy. DS ai is the entropy difference between the transition state and the initial state. The entropy of adsorbed species are assumed to be zero, and the gas-phase entropies are taken from Ref. [24]. The adsorption energies of the different species E CO , E O 2 , and E O and the transition state energies are given with respect to the gas-phase molecules.Assuming the prefactors and adsorption entropies are independent of the metal, there are five metal-dependent parameters determining the kinetics: E CO , E O 2 , E O , E TS3 , and E TS4 . The transition-state energies are, however, found to scale linearly with the adsorption energies, as shown for E TS3 and E TS4 in F...