Large-scale molecular dynamics investigations of ion-induced compositional dynamics leading to nanopattern formation at semiconductor surfaces

Nanopatterning;Ion irradiation;Molecular dynamics;Atomistic simulation;Ion-surface interactions;Semiconductors;Silicon;Gallium antimonide;Compositional phase dynamics

Lively, Michael Aaron ; Zhang, Yang ; Allain ; Jean Paul

关键词: Nanopatterning;    Ion irradiation;    Molecular dynamics;    Atomistic simulation;    Ion-surface interactions;    Semiconductors;    Silicon;    Gallium antimonide;    Compositional phase dynamics;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/98426/LIVELY-THESIS-2017.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Ion beam nanopatterning has been demonstrated to be a versatile method for obtaining a wide variety of surface features on a broad range of materials, with structures such as ripples, quantum dots, terraces, and ordered holes being obtained for various experimental conditions. However, theoretical modeling is well behind the experimental progress, and even for “simple” systems such as noble gas ion irradiation of silicon surfaces there exist several competing models proposing different pattern-forming mechanisms. For more complex systems, such as ion irradiation of binary alloys, the landscape of potential pattern-forming mechanisms remains very much a terra incognita, to the point where two models can predict the same surface morphologies while predicting diametrically-opposed surface compositional profiles. This knowledge chasm between experiments and theories requires a fundamental understanding of the ion-induced mechanisms that can lead to surface instabilities, to eliminate the dependence on simplifying assumptions and tunable parameters of existing modeling approaches. To close this gap, atomistic computational modeling is needed to allow direct observation of the ion-surface interactions at smaller length and time scales than can be accessed by experimental characterization techniques. At the same time, atomistic simulations must be able to account for changes over time in the surface structure or composition, which will influence the nature of the ion-surface interactions. The results from these atomistic simulations and the physical understanding gained can then be used as the basis for or as parametric inputs to multiscale models of nanopattern formation. Such a model has previously been developed, which is a hybrid molecular dynamics/kinetic Monte Carlo (MD/kMC) atomistic simulation that uses so-called “crater functions” obtained from MD simulations of single-ion impacts, combined with an atomistic kMC model of surface diffusion, to provide a complete description of the ion-surface interaction and the resulting surface nanopatterning without reliance on the assumptions and arbitrary parameters from other models. This simple computational model provides a well-tested starting point from which additional mechanisms can be implemented and used to study more complex material systems. Here, large-scale MD simulations are used to study the ion-induced compositional and phase dynamics, enabling the mechanisms that can cause patterning instabilities to be elucidated and characterized.The compositional evolution of GaSb under low-energy ion irradiation is studied by massive-scale MD simulations, which have been carried out on the Blue Waters high-performance computing platform at the University of Illinois. The first set of simulations consist of 500 eV Kr+ bombardment of a GaSb surface with a significantly-altered compositional profile designed to resemble experimental observations of the compositional depth profile at the onset of nanopattern formation. In regions of altered composition, thermodynamic phase separation is observed as the surface atoms rearrange themselves into clusters of the enriched component within 50/50 amorphous GaSb. Additionally, the pure Sb clusters in Sb-enriched regions self-organize into crystalline lattices, while the pure Ga clusters in Ga-enriched regions remain in an amorphous state. These results have demonstrated for the first time, using MD simulations, that the compositional depth variation observed from experiments can lead to a lateral compositional variation that may provide a potential pattern-forming instability. The second set of simulations consist of 500 eV ion irradiation of initially-pristine GaSb(110) by Ne+, Ar+, and Kr+ ion species up to the experimentally-relevant fluence of 7.5 × 1015 cm-2 with the goal of discovering how the ion-induced mechanisms leading to the formation of a compositional depth profile. While the surface quickly becomes amorphous under sustained ion bombardment, no ion species led to the emergence of a compositional depth profile. However, smaller “protoclusters” of Sb were formed in the subsurface, even in the absence of the compositional change necessary to drive thermodynamic phase separation. These protoclusters are conjectured to be formed from Sb precipitation out of the GaSb melt volume from ion-induced thermal spikes, and may function as the initial “seeds” that grow large enough to cause a compositional depth profile to form under the influence of additional mechanisms acting on timescales beyond the limits of MD simulations.The effects of implanted noble gas ions in Si are also studied with the use of high-fluence molecular dynamics simulations to reach cumulative ion fluences of ≥ 3 × 1015 cm-2. Ion species of Ne+, Ar+, Kr+, and Xe+ were studied with incident energies per ion ranging from 20 to 1000 eV and ion incidence angles ranging from 0° to 85°. The implanted ions tend to form clusters beneath the surface, which are formed purely by the kinetic motion of the ions and not due to diffusive processes. A cluster degassing mechanism is observed, which occurs when the Si surface above a cluster is eroded by ion sputtering and the gas atoms rapidly vacate the cluster. Immediately after the cluster has degassed, a rapid inflow of mass from the surrounding surface occurs to fill the resulting void. The combination of the cluster degassing and the resulting mass flow has a highly disruptive effect on the local surface morphology, which could destroy nanopattern “seeds” at the surface, which may be a missing mechanism from existing models of surface nanopatterning that can correct the quantitative inaccuracies of those models. Additionally, the shear stress distribution and elastic modulus were calculated for the ion-bombarded surfaces. While the shear stress distribution is in general agreement with expectation from previous computational studies, the strong variance in the stress depth profiles at different fluences suggests a highly-localized contribution from the implanted ion clusters which must be considered in stress-based models of ion beam nanopatterning. Comparing the elastic moduli for surfaces with and without ion clusters confirms that the presence of clusters within the surface has a significant influence on the mechanical properties of that surface.