Robert J. Paruch

Computed Molecular Depth Profile for C60 Bombardment of a Molecular solid

Molecular dynamics (MD) simulations have been performed for 10 keV C60 bombardment of an octane molecular solid at normal incidence. The results are analyzed using the steady-state statistical sputtering model (SS-SSM) to understand the nature of molecular motions and to predict a depth profile of a δ-layer. The octane system has sputtering yield of ~150 nm(3) of which 85% is in intact molecules and 15% is fragmented species. The main displacement mechanism is along the crater edge. Displacements between layers beneath the impact point are difficult because the nonspherically shaped octane molecule needs a relatively large volume to move into and the molecule needs to be aligned properly for the displacement. Since interlayer mixing is difficult, the predicted depth profile is dominated by the rms roughness and the large information depth because of the large sputtering yield.

Seduction of Finding Universality in Sputtering Yields Due to Cluster Bombardment of Solids

Universal descriptions are appealing because they simplify the description of different (but similar) physical systems, allow the determination of general properties, and have practical applications. Recently, the concept of universality has been applied to the dependence of the sputtering (ejection) yield due to energetic cluster bombardment versus the energy of the incident cluster. It was observed that the spread in data points can be reduced if the yield Y and initial projectile cluster kinetic energy E are expressed in quantities scaled by the number of cluster atoms n, that is, Y/n versus E/n. The convergence of the data points is, however, not perfect, especially when the results for molecular and atomic solids are compared. In addition, the physics underlying the apparent universal dependence in not fully understood. For the study presented in this Account, we performed molecular dynamics simulations of Arn cluster bombardment of molecular (benzene, octane, and β-carotene) and atomic (Ag) solids in order to address the physical basis of the apparent universal dependence. We have demonstrated that the convergence of the data points between molecular and atomic solids can be improved if the binding energy of the solid U0 is included and the dependence is presented as Y/(E/U0) versus (E/U0)/n. As a material property, the quantity U0 is defined per the basic unit of material, which is an atom for atomic solids and a molecule for molecular solids. Analogously, the quantity Y is given in atoms and molecules, respectively. The simulations show that, for almost 3 orders of magnitude variation of (E/U0)/n, there are obvious similarities in the ejection mechanisms between the molecular and atomic solids, thus supporting the concept of universality. For large (E/U0)/n values, the mechanism of ejection is the fluid flow from a cone-shaped volume. This regime of (E/U0)/n is generally accessed experimentally by clusters with hundreds of atoms and results in the largest yields. For molecular systems, a large fraction of the total energy E is consumed by internal excitation and molecular fragmentation, which are energy loss channels not present in atomic solids. For small (E/U0)/n values, the cluster deforms the surface and the ejection occurs from a ring-shaped ridge of the forming crater rim. This regime of (E/U0)/n is generally accessed experimentally by clusters with thousands of atoms and results in the smallest yields. For the molecular systems, there is little or no molecular fragmentation. The simulations indicate, however, that the representation which includes U0 as the only material property cannot be completely universal, because there are other material properties which influence the sputtering efficiency. Furthermore, neither the Y/n nor Y/(E/U0) representation includes the energy loss physics associated with molecular fragmentation in the high (E/U0)/n regime. The analysis of the universal concept implies for practical applications that if the objective of the experiment is large material removal, then the high energy per cluster atom regime is applicable. If the objective is little or no molecular fragmentation in organic materials, then the low energy per atom regime is appropriate.

Physical basis of energy per cluster atom in the universal concept of sputtering

The interpretation of the variables, scaled by the number of projectile cluster atoms n, in the universal relation of the sputtering yield Y versus incident energy E, that is, Y/n vs E/n, is not necessarily obvious. Following on previous works, the objective of this study is to elucidate the physical basis of the energy per atom variable E/n. The authors employ molecular dynamics simulations of Arn cluster bombardment of Ag(111) metal samples for this study. The authors find that the energy per cluster atom quantity E/n is responsible for the fraction of the initial energy that is deposited in the solid, rather than energy per cluster mass E/m. The results show that even though there is an average loss of the energy for a cluster, each cluster atom loses a different fraction of its initial energy, thus yielding a distribution of energy loss by individual atoms. The analysis of these distributions indicates that the energy deposition process is more effective for clusters with higher E/n when compared to t...