Laurent J. Lewis

SELF-DIFFUSION OF ADATOMS, DIMERS, AND VACANCIES ON CU(100)

We use ab initio static relaxation methods and semiempirical molecular-dynamics simulations to investigate the energetics and dynamics of the diffusion of adatoms, dimers, and vacancies on Cu(100). It is found that the dynamical energy barriers for diffusion are well approximated by the static, 0 K barriers and that prefactors do not depend sensitively on the species undergoing diffusion. The ab initio barriers are observed to be significantly lower when calculated within the generalized-gradient approximation (GGA) rather than in the local-density approximation (LDA). Our calculations predict that surface diffusion should proceed primarily via the diffusion of vacancies. Adatoms are found to migrate most easily via a jump mechanism. This is the case, also, of dimers, even though the corresponding barrier is slightly larger than it is for adatoms. We observe, further, that dimers diffuse more readily than they can dissociate. Our results are discussed in the context of recent submonolayer growth experiments of Cu(100).

Diffusion of gold nanoclusters on graphite

We present a detailed molecular-dynamics study of the diffusion and coalescence of large (249-atom) gold clusters on graphite surfaces. The diffusivity of clusters is found to be comparable to that for single adatoms. Likewise, and even more important, cluster dimers are also found to diffuse at a rate that is comparable to that for adatoms and singleclusters. As a consequence, large islands formed by cluster aggregation are also expected to be mobile. Using kinetic Monte Carlo simulations, and assuming a proper scaling law for the dependence on size of the diffusivity of large clusters, we find that islands consisting of as many as 100 clusters should exhibit significant mobility. This result has profound implications for the morphology of cluster-assembled materials.

Kinetic activation-relaxation technique: An off-lattice self-learning kinetic Monte Carlo algorithm

Many materials science phenomena, such as growth and self-organisation, are dominated by activated diffusion processes and occur on timescales that are well beyond the reach of standard-molecular dynamics simulations. Kinetic Monte Carlo (KMC) schemes make it possible to overcome this limitation and achieve experimental timescales. However, most KMC approaches proceed by discretizing the problem in space in order to identify, from the outset, a fixed set of barriers that are used throughout the simulations, limiting the range of problems that can be addressed. Here, we propose a more flexible approach -- the kinetic activation-relaxation technique (k-ART) -- which lifts these constraints. Our method is based on an off-lattice, self-learning, on-the-fly identification and evaluation of activation barriers using ART and a topological description of events. The validity and power of the method are demonstrated through the study of vacancy diffusion in crystalline silicon.

Island morphology and adatom self-diffusion on Pt(111)

The results of a density-functional-theory study of the formation energies of (100)- and (111)-faceted steps on the Pt(111) surface, as well as of the barrier for diffusion of an adatom on the flat surface, are presented. The step formation energies are found to be in a ratio of 0.88 in favor of the (111)-faceted step, in excellent agreement with experiment; the equilibrium shape of islands should, therefore, clearly be nonhexagonal. The origin of the difference between the two steps is discussed in terms of the release of stress at the surface through relaxation. For the diffusion barrier, we also find relaxation to be important, leading to a

Critical heat flux around strongly heated nanoparticles.

20%

High frequency relaxation of o‐terphenyl

decrease of its energy. The value we obtain, 0.33 eV, however, remains higher than available experimental data; possible reasons for this discrepancy are discussed. We find the ratio of step formation energies and the diffusion barrier to be the same whether using the local-density approximation or the generalized-gradient approximation for the exchange-and-correlation energy.

Vibrational properties of nanoscale materials: From nanoparticles to nanocrystalline materials

We study heat transfer from a heated nanoparticle into surrounding fluid using molecular dynamics simulations. We show that the fluid next to the nanoparticle can be heated well above its boiling point without a phase change. Under increasing nanoparticle temperature, the heat flux saturates, which is in sharp contrast with the case of flat interfaces, where a critical heat flux is observed followed by development of a vapor layer and heat flux drop. These differences in heat transfer are explained by the curvature-induced pressure close to the nanoparticle, which inhibits boiling. When the nanoparticle temperature is much larger than the critical fluid temperature, a very large temperature gradient develops, resulting in close to ambient temperature just a radius away from the particle surface. The behavior reported allows us to interpret recent experiments where nanoparticles can be heated up to the melting point, without observing boiling of the surrounding liquid.

Stacking-fault energies for Ag, Cu, and Ni from empirical tight-binding potentials

Results of molecular dynamics simulations (MDS) of o‐terphenyl, a glass‐forming liquid, are analyzed in terms of the coupling model of relaxation. At low temperatures thermally activated relaxation processes are suppressed, whereby the density–density correlation function, C(t), obtained by MDS is determined entirely by vibrational modes. This enables the low temperature data to be used to deduce the vibrational density of states, g(ω). With g(ω) determined, the vibrational contribution, Cpho(t), is calculated at higher temperatures assuming that g(ω) is independent of temperature. At higher temperatures, relaxation makes its appearance and is modeled here by the fast dynamics of the coupling model. Assuming that vibration and relaxation contribute independently, the density–density self‐correlation function is given by the product Cpho(t)Crel(t), with the relaxation part obtained from the coupling model. There is good overall agreement between the calculated C(t) and the MDS data. Microscopic parameters,...

Optical properties of structurally relaxed Si / SiO 2 superlattices: The role of bonding at interfaces

The vibrational density of states (VDOS) of nanoclusters and nanocrystalline materials is derived from molecular-dynamics simulations using empirical tight-binding potentials. The results show that the VDOS inside nanoclusters can be understood as that of the corresponding bulk system compressed by the capillary pressure. At the surface of the nanoparticles the VDOS exhibits a strong enhancement at low energies and shows structures similar to that found near flat crystalline surfaces. For the nanocrystalline materials an increased VDOS is found at high and low phonon energies, in agreement with experimental findings. The individual VDOS contributions from the grain centers, grain boundaries, and internal surfaces show that, in the nanocrystalline materials, the VDOS enhancements are mainly caused by the grain-boundary contributions and that surface atoms play only a minor role. Although capillary pressures are also present inside the grains of nanocrystalline materials, their effect on the VDOS is different than in the cluster case which is probably due to the intergrain coupling of the modes via the grain boundaries.

Molecular dynamics simulation of a molecular glass at intermediate times

The intrinsic stacking-fault energies and free energies for Ag, Cu, and Ni are derived from molecular-dynamics simulations using the empirical tight-binding potentials of Cleri and Rosato [Phys. Rev. B 48, 22 (1993)]. While the results show significant deviations from experimental data, the general trend between the elements remains correct. This allows one to use the potentials for qualitative comparisons between metals with high and low stacking-fault energies. Moreover, the effect of stacking faults on the local vibrational properties near the fault is examined. It turns out that the stacking fault has the strongest effect on modes in the center of the transverse peak, and its effect is localized in a region of approximately eight monolayers around the defect.