Potential Sputtering of Insulators

Project number: Research project P 12388 Project title: Potential Sputtering of Insulators ABSTRACT OF THE PROJECT The proposed project deals with investigations of the recently discovered phenomenon of potential sputtering psi, i.e. the efficient removal of neutral and ionized target particles from certain insulator surfaces due to the potential (rather than the kinetic) energy of bombarding slow highly charged ions (HCI). By measuring total sputter yields with a sensitive quartz crystal micro-balance (as well as yields for secondary ion emission with as mass spectrometer) it will be investigated, which target materials show psi, what are the primary mechanisms for psi (Coulomb explosion- vs. defect mediated sputtering) and what is its practical relevance. To this goal a large number of insulator surfaces will be investigated. Some of the planned measurements require projectile ions in higher charge states than available from the Vienna ECR ion source. These measurements will be carried out at ion beam facilities in Berlin/Germany and Gaithersburg/USA.

Within this project we have studied the influence of projectile potential energy on ion-induced sputtering of solid surfaces. The two main results are: (a) We could prove that defect mediated desorption is the dominant mechanism for potential sputtering of alkali halides and (b) we have discovered the so far unrecognized mechanism of kinetically assisted potential sputtering. The value of these new findings can be estimated from the fact that our results have been published in a series of three papers in the leading physics journal Physical Review Letters as well as a number of papers in regular physics journals (see appendix). In more detail the following tasks have been successfully completed within the project. * Total sputter yields for impact of slow singly and multiply charged ions on metal- (Au), oxide- (Al2O3, MgO) and alkali-halide surfaces (LiF) have been measured as a function of projectile charge state and impact energy by using a highly sensitive quartz-crystal-microbalance technique. At very low kinetic energy of the projectiles the potential energy (i.e. the energy necessary to remove q electrons from the projectile, with q being the charge of the projectile) dominates the ion-surface interaction. A strong dependence of the total sputter yield on projectile charge state ("potential sputtering", PS) could be observed for Al2O3, MgO and LiF but not for Au. * Presently, several complementary models are being considered to explain potential sputtering. For the case of LiF we could unambiguously demonstrate that the "defect mediated desorption model" (DM model, originally developed for electron- and photon-stimulated desorption) is appropriate to explain all experimental facts. By measuring the total sputter yield for impact of various slow singly and multiply charged ions on LiF we found the threshold for potential sputtering (i.e. the minimum potential energy to induce PS) to be about 10 eV. This threshold coincides with the energy necessary to produce a cold hole in the valence band of LiF by resonant neutralisation in accordance with the DM model. * PS in Al2O3 can also be explained with the DM model. The sputtering yields measured for MgO, however, exhibit an unusually strong dependence on the ion kinetic energy. This new type of potential sputtering not only requires electronic excitation of the target material, but also the formation of a collision cascade within the target in order to initiate the sputtering process and has therefore been termed "kinetically assisted potential sputtering". In order to study defects induced by potential sputtering on the atomic scale we performed measurements of multiply charged Ar ion irradiated HOPG (highly oriented pyrolitic graphite) samples with scanning tunneling microscopy (STM). The only surface defects found in the STM images are protrusions. The mean diameter of the defects increases with projectile charge state while the height of the protrusions stays roughly the same indicating a possible "pre-equilibrium" effect of the stopping of slow multiply charged projectiles in HOPG. Individual Tasks A. Construction of a highly sensitive quartz crystal microbalance A new quartz-crystal microbalance for measuring total sputter yields in ion-surface collisions has been constructed (for details see ref. A09). The new electronic circuit to drive the quartz crystal ensures low noise and high frequency stability. By measuring total sputter yields for impact of singly charged ions on LiF target films a sensitivity limit of 0.5% of a monolayer per minute could be achieved. This is an improvement in sensitivity over previously used setups by at least a factor of 5. B. Investigations on the threshold for potential sputtering of LiF surfaces In recent studies on the impact of slow multiply charged ions on insulator surfaces, a dramatic increase of the yields for sputtering and secondary ion emission with projectile charge state has been observed for certain target species, e.g. for LiF and SiO2. In contrast to the well established process of kinetically induced sputtering, the ejection of target atoms and ions due to the potential energy of ions is largely unexplored. Currently available experimental evidence (refs. A01, A04) and theoretical considerations strongly favour a so-called ``defect- mediated desorption`` model over an explanation involving a ``Coulomb explosion`` mechanism (for a detailed discussion see ref. A05). To gain a more detailed understanding of the mechanisms responsible for potential sputtering (PS) we have measured total sputtering yields of LiF under impact of slow (20 eV, 100 eV, 500 eV and 1000 eV kinetic energy) singly and doubly charged ions. In these experiments the available potential energy has been varied from 5.1 eV (Na+ ) to 62.6 eV (Ne2+) and its influence on PS was investigated. Below 10 eV recombination energy, sputtering can only be observed for sufficiently fast projectiles (500 eV and 1000 eV) and is therefore ascribed to transfer of kinetic energy. The minimum potential energy necessary to induce PS from LiF was determined to be about 10 eV (ref. A02). This coincides with the energy to produce a ``cold hole`` in the F-valence band of LiF by resonant neutralization. At about 20 eV potential energy Auger capture becomes possible. The resulting electron-hole pairs localise as self-trapped excitons leading to a further increase in the sputtering yield (ref. A02). C. Supportive model calculations In cooperation with the group of Prof. J. Burgdörfer (Institut für Theoretische Physik der TU Wien) a theoretical model was developed that calculates the level shift of the incident ion and the deformation of the valence band under the influence of the projectile. The results of this model (ref. A10) are in full agreement with the experimental findings (ref. A02, A11). D. Measurements at Hahn Meitner Institut Berlin To perform PS experiments with HCI in high charge states (typically Xe25+ ) we had to move our apparatus to the 14 GHz ECR ion source at the Hahn Meitner Institute in Berlin (co-operation with Prof. N. Stolterfoht). In the first experimental campaign (Oct. - Nov. 1999) a new target surface (Al2O3) was studied. As expected from theoretical considerations also this target showed a continuing increase of the potential sputter yield with increasing charge state (ref. A13). These preliminary results were confirmed in a second campaign (Jan - Feb. 2000), where also a MgO target was studied. Sputtering yields measured for MgO showed an unusually strong dependence on the ion kinetic energy. This new type of potential sputtering not only requires electronic excitation of the target material, but also the formation of a collision cascade within the target in order to initiate the sputtering process and has therefore been termed "kinetically assisted potential sputtering" (for details see ref. A03). E. Individal ion impact events studied on an atomic scale by STM In order to study defects induced by potential sputtering on the atomic scale we performed measurements of multiply charged Ar ion irradiated HOPG (highly oriented pyrolitic graphite) samples with scanning tunneling microscopy (STM). The only surface defects found in the STM images are protrusions. The mean diameter of the defects increases with projectile charge state while the height of the protrusions stays roughly the same indicating a possible "pre-equilibrium" effect of the stopping of slow multiply charged projectiles in HOPG (for details see ref. A12).