A Hybrid Model of Ion-Induced Surface Modification with Prompt Molecular Dynamics and Lattice-Free Kinetic Monte Carlo Diffusion
ABSTRACT: Ion beam nanopatterning is a robust, scalable technique for modification and fabrication of nanostructured surfaces. An ion beam incident on a surface acts as an athermal source of particles and energy, driving the surface far from equilibrium and leading to the emergence of metastable compositional and morphological phases. However, the fine process control needed for advanced nanotechnology applications remains elusive due to a lack of fundamental physical understanding of nanostructure formation and growth on complex, multi-component surfaces such as III-V semiconductors or metal-silicon systems. Recent experimental progress in the field has demonstrated the evolution of a three-dimensionally complex surface compositional profile during low-energy ion beam irradiation of III-V semiconductor systems such as GaSb, which is a necessary precursor to formation of ordered nanostructures. Novel analysis of x-ray scattering measurements indicates that the nanopatterning kinetics are described by highly nonlinear, nucleation-and-growth kinetics which are not included in any theoretical treatment of ion beam nanopatterning of compositionally complex materials. To elucidate the compositionally driven mechanisms which lead to these kinetics, large-scale computational simulations have been designed and carried out on the Blue Waters supercomputer at the University of Illinois and the Advanced Cyberinfrastructure (ACI) system at Pennsylvania State University.
Large-scale molecular dynamics (MD) simulations of 500 eV Kr+ ion irradiation of amorphous GaSb surfaces have connected the compositional depth profile observed in experiments to lateral compositional gradients via thermodynamically driven phase separation. This lateral compositional evolution is a necessary precursor for a pattern-forming surface instability. Other large-scale MD studies of GaSb(110) irradiation from initially-pristine surface conditions have shown the formation of Sb protoclusters due to prompt ion-induced collisional effects at high irradiation fluences exceeding 1015 cm-2. The protocluster formation is accompanied by significant structural transformation of the surface, including bulk amorphization, reduction of average per-atom bonding energies, and prevalence of non-tetrahedral bonding states which would otherwise characterize an amorphous semiconductor surface. However, a long temporal scale mechanism such as surface diffusion is necessary to connect these disruptive phenomena into a complete model of nanopattern formation. Finally, a large battery of single-ion impact MD simulations under a range of ion beam parameters into GaSb surfaces with variable compositional depth profiles has elucidated the connection between lateral variation of the compositional depth profile and local morphological instability. Specifically, the presence of a compositional phase interface near the surface leads to an increase in ion-induced energy deposition at or near the surface monolayer, which leads to enhanced surface erosion (i.e., sputtering) when the surface monolayer is Sb-enriched and/or when the sub-surface is Ga-enriched. Given this finding, the key physical mechanisms which remain to be deciphered are those which drive the three-dimensional compositional evolution of the ion-irradiated surface and activate this sputtering instability.
Accordingly, the culmination this work is a highly-parallelized, hybrid molecular dynamics/kinetic Monte Carlo (MD/KMC) model designed to simulate nanopattern formation on III-V and other compositionally complex surfaces. This diffusion model relies on a lattice-free point defect characterization method to analyze the defect distribution in the highly disordered ion irradiated surface. These defects then mediate diffusion events which are identified using per-atom neighbor lists and bond structural configurations to determine the activation energy. Therefore, the model is termed structural kinetic Monte Carlo (SKMC). The SKMC approach allows computational modeling to extend beyond the prompt temporal regime accessible by MD alone, addressing the three-dimensionally complex evolution of the surface beyond the microsecond scale. By alternating between MD ion irradiation steps and SKMC steps, a complete atomistic model of ion beam nanopatterning is therefore constructed. The application of this new modeling approach is demonstrated for the case of ion beam irradiation of cleaved GaSb surfaces up to a fluence of 4 × 1015 cm-2. Specifically, hybrid MD/SKMC simulations test the hypothesis that prompt cluster formation, diffusion-driven cluster growth, and compositional depth profile-modulated sputtering yields are the fundamental mechanisms driving nanopattern formation and growth. This modeling approach has broad applications beyond semiconductor surfaces to any class of complex nanomaterials under ion beam or plasma irradiation, such as high-entropy alloys currently under consideration as structural materials for fusion device applications.