The main goal of the present study is the investigation of the sputtering of neutral particles from a metal surface under atomic and polyatomic ion bombardment using secondary neutral time-of-flight mass spectrometry (ToF SNMS). For postionization of neutral species, UV laser irradiation with wavelength 193 nm was utilized. For generation of polyatomic projectiles, a negative sputter cesium ion source suitable for ToF SNMS setup was developed and built. The ion source delivers negatively charged (m=1÷5) and polyatomic ions produced from a gold sputter target bombarded by positive Cs−mAu2AuCs+ ions. Mass separation of primary projectiles in the ion source is performed by a built-in compact Wien filter allowing to separate heavy ions in the energy range of several keV. In the experiment, an indium surface was bombarded by (m=1÷5) projectiles with total impact energy of 5 and 10 keV. The obtained mass distributions of sputtered indium species reveal that the partial yields of sputtered clusters increase under polyatomic ion bombardment. It is shown that the enhancement in total sputtering yield per constituent atom of the projectile ion is non-additively enhanced in the case of diatomic ion bombardment in comparison with monoatomic projectile ions impinging at the same velocity. The enhancement of partial yields observed for sputtered clusters is found to increase with increasing cluster size, reaching a factor of several ten for the largest detected cluster. −mAu Apart from sputtering yields, kinetic energy distributions (KED) of sputtered neutral indium atoms ejected under mono- and polyatomic projectile ion bombardment were measured. It is shown that In monomers sputtered by monoatomic projectiles with an impact energy of 5 keV are emitted mostly from linear collision cascades. At higher kinetic energy, or polyatomic projectile impact, it is revealed that in addition to the atoms sputtered from the linear cascade, a low energetic contribution of atoms sputtered from a collisional spike appears. This contribution in the KED increases with increasing impact energy and nuclearity of projectile. In the case of 10-keV projectiles, the sputtering process is shown to be dominated by the spike contribution. By subtracting the linear cascade contribution from the measured KED, the pure emission energy spectrum produced by the collisional spike is identified. It is found that the most probable emission energy of atoms emitted from the spike is more than one order of magnitude lower in comparison with the surface binding energy of indium. The obtained KED of indium monomers emitted from the spike were interpreted in terms of published theoretical models of the sputtering process from a spike. It is shown that the obtained data cannot be explained in terms of a thermal spike model. The obtained results are shown to agree more favorably with a thermodynamic gas flow model describing the particle emission process as a quasi-free expansion of a superheated near-surface volume. −1Au−3Au By comparing the partial sputtering yields of emitted secondary ions and their neutral counterparts, the ionization probabilities of indium atoms sputtered by atomic and polyatomic projectiles were measured. It is revealed that ionization probabilities of sputtered In monomers decrease when polyatomic projectiles are utilized. Data of this kind are of great interest both from a fundamental perspective and for practical applications of Secondary Ion Mass Spectrometry (SIMS) in surface analysis. The measured data are interpreted in terms of published theories of secondary ion formation. Our results indicate that the electronic excitation of the solid induced by the projectile impact decrease with increasing projectile nuclearity, a finding which reveals an opposite trend to that observed for the sputter yields. This surprising result cannot be understood in terms of published theory and has therefore motivated an ongoing study in our group to model excitation and ionization processes in the frame of a molecular dynamics computer simulation of sputtering process.