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...

We present a set of benchmark calculations for the Kohn-Sham elastic transmission function of five representative single-molecule junctions. The transmission functions are calculated using two different density functional theory (DFT) methods, namely an ultrasoft pseudopotential plane wave code in combination with maximally localized Wannier functions, and the norm-conserving pseudopotential code Siesta which applies an atomic orbital basis set. For all systems we find that the Siesta transmission functions converge toward the plane-wave result as the Siesta basis is enlarged. Overall, we find that an atomic basis with double-zeta and polarization is sufficient (and in some cases even necessary) to ensure quantitative agreement with the plane-wave calculation. We observe a systematic down shift of the Siesta transmission functions relative to the plane-wave results. The effect diminishes as the atomic orbital basis is enlarged, however, the convergence can be rather slow.

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 k i = n i exp[ÀDG ai /k T] = n i exp[À(E ai ÀTDS ai )/k T], 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...

Stable, single-molecule conducting-bridge configurations are typically identified from peak structures in a conductance histogram. In previous work on Pt with H 2 at cryogenic temperatures it has been shown that a peak near 1G 0 identifies a single-molecule Pt-H 2 -Pt bridge. The histogram shows an additional structure with lower conductance that has not been identified. Here, we show that it is likely due to a hydrogen decorated Pt chain in contact with the H 2 molecular bridge. DOI: 10.1103/PhysRevLett.98.146802 PACS numbers: 73.63.Rt, 63.22.+m, 73.23.ÿb, 85.65.+h The interest in chains of single metal atoms bridging between two electrodes is largely due to their unique properties as ideal one-dimensional systems [1]. For clean metals, only Au, Pt, and Ir form atomic chains [2,3]. However, atomic or molecular adsorption on metal surfaces can widen this scope. Recently, 2 nm long Ag atomic chains have been created in the presence of oxygen at ultralow temperature, while clean Ag only forms short chains [4]. Atomic chains have been imaged by transmission electron microscope for the noble metals Cu, Ag, and Au [5][6][7][8].In the first experiments contacting molecules by Pt atomic leads it was shown that a single hydrogen molecule H 2 can be contacted and there appeared to be no indication for atomic chain formation [9,10]. By use of point contact spectroscopy and shot noise measurements, the system was characterized in detail and close agreement with atomistic model calculations was obtained [9,10]. The Pt-H 2 -Pt junction was first identified by its conductance. It shows up as a recurring plateau in the conductance when controllably breaking a contact, and in a histogram of conductance values collected for many such breakings it gives rise to a sharp peak near 1G 0 , where 1G 0 2e 2 =h is the conductance quantum. This main peak at 1G 0 for the Pt=H 2 system is therefore well understood. However, there is more structure in the conductance histogram for the Pt=H 2 junctions, which has not been explained. In particular, a strong peak is found at about 0:1-0:2G 0 , suggesting that other configurations of hydrogen between Pt leads may be formed. In the present study we focus on those structures, having a conductance below 1G 0 , and we present evidence that they can be attributed to the formation of a hydrogen decorated Pt atomic chain that forms one of the leads contacting a hydrogen molecule.The measurements have been performed using the mechanically controllable break junction technique (see Ref.[11] for a detailed description). Once under vacuum and cooled to 4.2 K a fine Pt wire was broken. Atomicsized contacts between the wire ends can be formed using a piezoelement for fine adjustment. H 2 was admitted via a capillary. dc two-point voltage-biased conductance measurements were performed by applying a voltage in the range from 10 to 150 mV. Every statistical data set was built from a large number (over 3000) of individual digitized conductance traces. ac voltage bias conductance measurements were performed ...

The electrical properties of single-molecule junctions, consisting of an organic molecule coupled to metal electrodes, are sensitive to the detailed atomic structure of the molecule-metal contact. This, in turn, is determined by the anchoring group linking the molecule to the metal. With the aim of identifying and comparing the intrinsic properties of two commonly used anchoring groups, namely thiol and amine groups, we have calculated the atomic structure and conductance traces of different Au-S-Au and Au-NH(2)-Au nanojunctions using density functional theory (DFT). Whereas NH(2) shows a strong structural selectivity towards atop-gold configurations, S shows large variability in its bonding geometries. As a result, the conductance of the Au-NH(2)-Au junction is less sensitive to the structure of the gold contacts than the Au-S-Au junction. These findings support recent experiments which show that amine-bonded molecules exhibit more well-defined conductance properties than do thiol-bonded molecules.