The photodissociation of aluminum clusters, Al+n (n=7–17), has been studied over a broad energy range (1.88–6.99 eV). Measurements of the lifetimes of the photoexcited clusters are described. Dissociation energies have been determined by comparing the measured lifetimes with the predictions of a simple RRKM model. The dissociation energies show an overall increase with cluster size, but there are substantial oscillations around n=7–8 and n=13–15. Cluster cohesive energies are derived from these results and from previous measurements of the dissociation energies of the smaller clusters. The cohesive energies of the larger clusters (n>6) are in good agreement with the predictions of a simple model based on the bulk cohesive energy and the cluster surface energy. However, the cohesive energies are substantially larger than the results of recent ab initio calculations. The photodissociation spectrum of Al+8 has been measured and shows a broad absorption feature with a maximum ∼470 nm.

The chemical reactions of Si+n (n=10–65) with O2 have been investigated using selected ion drift tube techniques. The smaller clusters are etched by O2 to give Si+n−2 (and two SiO molecules) and the larger clusters chemisorb oxygen forming an SinO+2 adduct. The transition occurs between n=29 and 36 under the conditions employed. There are large variations in the reactivity of the smaller clusters: Si+13, Si+14, and Si+23 are particularly inert. The variations in reactivity are rapidly damped with increasing cluster size and for clusters with 40–65 atoms the reactivity is nearly independent of size. However, these large clusters are ∼102 times less reactive towards O2 than most bulk silicon surfaces. Studies of the temperature dependence of the reactions reveal that they proceed through a metastable precursor state which is probably molecular O2 physisorbed to the cluster surface. Variations in the size of the activation barrier for dissociative chemisorption account for the changes in reactivity with cluster size. However, the difference between the cluster and surface reactivities is not due to the size of the activation barrier, but could be accounted for by the presence of only a few reactive sites on the clusters.

The chemical reactions of size selected Si+n (n=10–65) with D2O have been studied using injected ion drift tube techniques between temperatures of 258 and 404 K. The only products detected were a series of Sin(D2O)+m adducts. Large variations in reactivity were observed for the smaller clusters (n<40) that diminish with increasing cluster size. Si+11, Si+13, Si+14, Si+19, and Si+23 are particularly unreactive compared to their neighbors. At room temperature the larger clusters (n>40) are a factor of ∼10–1000 (depending on the bulk surface) less reactive towards water than bulk silicon. The reaction rates for all clusters exhibit an unusually strong negative temperature dependence but are independent of the buffer gas pressure. These results suggest that the reaction mechanism probably involves two steps. In the first step, a weakly bound molecularly adsorbed Si+n⋅⋅⋅D2O adduct is produced. The second step involves rearrangement to give a more strongly bound (and probably dissociatively adsorbed) SinD2O+ product. It appears that the reaction rates for some of the smaller clusters show a faster than linear dependence on D2O pressure. One possible explanation for this unusual observation is that a second D2O molecule solvates the transition state and significantly lowers the activation barrier for dissociative adsorption.

Interaction of silicon cluster ions with ammonia: The kinetics J. Chem. Phys. 93, 5709 (1990); 10.1063/1.459565Photothermal reflectance investigation of processed silicon. I. Roomtemperature study of the induced damage and of the annealing kinetics of defects in ionimplanted wafers

A crossed laser-molecular beam study of the one and two photon dissociation mechanism of bis (cyclopentadienyl) iron (ferrocene, FeCp2) has been performed at 193 and 248 nm. By combining electron bombardment mass spectroscopy with time-of-flight (TOF) measurements, the photodissociation mechanism at 193 nm is shown to have two distinct mechanisms. (1) FeCp2+hν→FeCp*+Cp; (2) FeCp+2hν→FeCp+Cp, FeCp→Fe+Cp. For the first mechanism, which accounts for less than 5% of the photodissociation events, the FeCp* velocity distribution is quantitatively consistent with a statistical dissociation producing FeCp in an excited, ligand field electronic state. The velocity distributions of the Cp and Fe fragments produced by the second mechanism (FeCp is an unstable intermediate) are also in excellent agreement with microcanonical calculations for both Cp elimination steps using the known metal–ligand bond energies of ferrocene. For the second mechanism, dissociation occurs on the lowest potential energy surface for each Cp elimination. Although one photon is energetically sufficient to remove one Cp ligand from ferrocene, RRKM calculations of the lifetime indicate that Cp elimination is extremely slow for dissociation along the ground electronic state potential energy surface. Hence, after internal conversion to the ground electronic state, the large photon absorption cross section (∼4 Å2) for the experimental irradiation conditions allows additional photons to be absorbed until the dissociation rate exceeds the up pumping rate. The large photon energy causes the dissociation rate to increase by many orders of magnitude for each additional photon absorbed. Consequently, there is strong selectivity for the total number of photons absorbed. Both mechanisms, occurring on two different electronic potential energy surfaces, suggest that dissociation induced by excitation of the ligand-to-metal charge transfer states accessed at 193 nm can be quantitatively described as a statistical, unimolecular decomposition. At 248 nm, the measured product velocity distributions are qualitatively consistent with the mechanism deduced from the 193 nm results, but the energy available for translation at this wavelength is too small to extract quantitative product translational energy distributions which are required to independently test the applicability of the statistical dissociation model.