Gold nanoparticles distinguish themselves from other nanoparticles due to their unique surface plasmon resonance properties that can be exploited for a multiplicity of applications. The promise of plasmonic heating in systems of Au nanoparticles on transition metal oxide supports, for example, Au/TiO2, rests with the ability of the surface plasmon in Au nanoparticles to effectively transfer energy into the transition metal oxide. Here, we report a critical observation regarding Au nanoparticle (Au55) surface plasmon excitations, that is, the relaxation of the surface plasmon excitation is very slow, on the order of several picoseconds. Starting from five plasmon states in Au55 nanoparticles using nonadiabatic molecular dynamics simulations, we find that the relaxation time constant resulting from these simulations is ∼6.8 ps, mainly resulting from a long-lived intermediate state found at around -0.8 eV. This long-lived intermediate state aligns with the conduction band edge of TiO2, thereby facilitating energy transfer injection from the Au55 nanoparticle into the TiO2. The current results rule out the previously reported molecular-like relaxation dynamics for Au55.

Superimposed STM images of the six stable rotation stations of a DMBI-P molecule during a step by step clockwise rotation induced by voltage pulses (V = 0.5 V, I = 0.5 nA, t = 10 s).

We assemble bistable silicon quantum dots consisting of four buckled atoms (Si4-QD) using atom manipulation. We demonstrate two competing atom switching mechanisms, downward switching induced by tunneling current of scanning tunneling microscopy (STM) and opposite upward switching induced by atomic force of atomic force microscopy (AFM). Simultaneous application of competing current and force allows us to tune switching direction continuously. Assembly of the few-atom Si-QDs and controlling their states using versatile combined AFM/STM will contribute to further miniaturization of nanodevices.

The computational simulation of photo-induced processes in large molecular systems is a very challenging problem. Firstly, to properly simulate photo-induced reactions the potential energy surfaces corresponding to excited states must be appropriately accessed; secondly, understanding the mechanisms of these processes requires the exploration of complex configurational spaces and the localization of conical intersections; finally, photo-induced reactions are probability events, that require the simulation of hundreds of trajectories to obtain the statistical information for the analysis of the reaction profiles. Here, we present a detailed description of our implementation of a molecular dynamics with electronic transitions algorithm within the local-orbital density functional theory code FIREBALL, suitable for the computational study of these problems. As an example of the application of this approach, we also report results on the [2 + 2] cycloaddition of ethylene with maleic anhydride and on the [2 + 2] photo-induced polymerization reaction of two C60 molecules. We identify different deactivation channels of the initial electron excitation, depending on the time of the electronic transition from LUMO to HOMO, and the character of the HOMO after the transition.