We investigate the unimolecular dissociation dynamics of energy-rich sodium cluster ions, Na+n (5≤n≤40) by measuring the time evolution of their sequential monomer or dimer evaporative cooling. The experimental technique, tandem time-of-flight mass spectroscopy, measures the relative rate of competing dissociation channels from metastable ion clusters selected during an initial sampling time interval immediately following the creation of the ion cluster ensemble. Pulsed laser UV photoionization converts the distribution of neutral clusters emerging from a free-jet expansion to the distribution of ion clusters from which the initial selection takes place. For the smaller clusters, 3≤n≤14, we compare the measured dissociation rates with those calculated from a modified version of the RRK theory of unimolecular dissociation. In applying the theory we use monomer and dimer binding energies determined from theoretical calculation. For larger clusters, 15≤n≤40, the binding energies are not known, and we invert the calculation, using measured dissociation fractions, to determine the binding energies of the cluster ions.

The unimolecular dissociation of energy rich lithium cluster ions shows that Li+n dissociate by sequential atom or dimer loss. The binding energies of Li+n (n=4–42) generated in an evaporative ensemble are determined from unimolecular decay, within a well defined time window, and energy constraint. They present a sawtooth behavior vs cluster size less pronounced that it should be from a simple metal model. Odd–even alternation is superimposed on the sawtooth behavior, with odd sized cluster ions being more stable. Cohesive energies per atom of Li+n are deduced from these dissociation energies up to n=40 and from extended photo-induced measurements up to n=95. Cohesive energies per atom of neutral clusters Lin are derived by combining these ionic cohesive energies with the literature ionization potentials. The linearity of the neutral cluster cohesive energy vs the cluster surface to volume ratio permits a volume and a surface energy to be deduced. These values are compared to the bulk values.

Multistep photon absorption has been used to measure the collective excitation of free lithium clusters having up to 1500 atoms. The blueshift of the Mie resonance energy, as cluster size increases, probes the surface effects. Its absolute value is consistent with the dielectric constants of the bulk down to a 100 atom cluster. The comparison with calculations in random-phase approximation in the local approximation demonstrates that the jellium model is no longer valid for lithium clusters. PACS numbers: 78.20.Dj, 61.46.+w The optical response of small metal particles as a function of size offers the possibility to follow the development of collective effects in metallic systems. Of these, alkali clusters, and in particular sodium and potassium, are regarded as a prototype of simple metallic particles. Their dipole absorption spectra are consistent with the optical excitation predicted by the classical Mie theory treatment of collective dipole oscillation in spherical or ellipsoidal distorted metallic droplets [1][2][3][4]. A more elaborate approach was obtained from time-dependent density-functional-theory calculations to quantitatively interpret the experimental data on sodium and potassium clusters having less than 40 atoms [5]. However, it was of fundamental interest to explore large cluster sizes to bridge the gap between the free small particles and the bulk in order to know to what extent the macroscopic dielectric function describes the optical response of metallic clusters.We have developed recently an experimental procedure for extending optical absorption measurements [6] to large masses, i.e., a few thousand atoms. For large potassium clusters we have shown that the experimental value of the resonance energy evolves toward the infinite limit h(DM deduced either from the experimental volume plasmon energy hcop [7] or from the surface plasmon energy hcos [8]: hwp^=^ h(OM^^^ hcDs^.Those values are close to the free-electron values. The small difference, less than 5%, which still remains between experimental values and free-electron values is greatly reduced when core polarization and effective mass are included [7,9].Lithium metal differs from the other alkali metals. The measured volume plasmon energy is greatly shifted down from the free-electron value and its linewidth is very broad as compared to other alkali metals [10]. The deviation of the experimental plasmon-peak position from the free-electron value was mainly interpreted in terms of optical effective mass and is due to intraband transitions as pointed out by Paasch [ll]. The volume plasmon linewidth in lithium was interpreted as a plasmon decay via interband transitions [10,12]. So the question arises as to whether such a deviation still exists for clusters. The measurement of the plasmon resonance of lithium clusters is a real challenge to know to what extent the dielectric constants of the bulk are pertinent parameters to interpret the collective excitation of lithium clusters. If so, the microscopic calculation of the dynamic response o...