Nanoscale Pattern Self-Organization Under Ion Bombardment

Abstract

Surfaces spontaneously self-organize under ion bombardment into a range of morphologies, including ripples, nanodots, ultrasmoothness, and high-amplitude structures. In this dissertation, I present research into the causes of this evolution as well as ways to control the transient and final structures. The research is largely experimental, but motivated by and often performed alongside models and simulations. In the first chapter, I introduce the field of pattern self-organization under low-energy ion bombardment, and I present a history of the aspects of the field most relevant to this dissertation, including the model of sputter erosion developed by Bradley and Harper (BH) based on work by Sigmund, the model of mass redistribution developed by Carter and Vishnyakov (CV), and a crater function theory developed by Norris. In Chapter 2, I address the limitations of existing models. Using a morphology phase diagram and fits to grazing-incidence small-angle x-ray scattering (GISAXS) data, I show that the observed morphologies on Kr+-bombarded Ge are inconsistent not only with BH theory, but with the predicted transition angle from stability to instability predicted by the CV model. I then perform a model-independent self-consistency test of the truncated form of Norris crater function theory using curvature coefficients fit to GISAXS data from Ar+-bombarded Si and Kr+-bombarded Ge, and I show that the truncated form cannot reproduce experimental values of x-direction curvature coefficients. In Chapter 3, I present experimental and analytical methods that allow access to low-noise, high-wavenumber GISAXS data. I demonstrate that the linear behavior of high-wavenumber data strongly supports ion-impact-induced viscous flow rather than surface diffusion as the dominant relaxation mechanism governing surface evolution under ion bombardment. I describe a composite model that includes all physical parameters commonly thought to be important to pattern self-organization in the system, show fits of the model to high-wavenumber data for Ar+- and Kr+-bombarded Si, and discuss extracted values of fluidity, amorphous film thickness, and the coefficients describing stress accumulation and prompt atomic displacements. In Chapter 4, I show AFM topographs and GISAXS data on Ar+-bombarded Si samples evolving to their final, saturated structures in the nonlinear regime'' of pattern formation, at long times and high fluence. I identify the characteristics of these structures, including pattern wavelengths ranging from 10s to 100s of nm, the emergence of sawtooth structures and asymmetrical scattering profiles, and the emergence of harmonic peaks in the scattering data at integer multiples of the characteristic wavevector of the system. I present a simulation reproducing the topographical sawtooth structures using a surface evolution equation incorporating a cubic nonlinear term, and describe simulated GISAXS scattering that reproduces the harmonic peaks. Chapter 5 addresses the behavior of steep, sharp features under ion bombardment. I discuss two models that predict the evolution of these structures under ion bombardment and show their experimental validation using Ga+-bombarded Si. I then describe how the models are used to solve the inverse problem'' of predicting an initial structure that will evolve into a desired final morphology, and experimentally demonstrate this concept by observing the evolution of wavy pits into an array of knife-edge ridges under a rastered focused ion beam. Chapter 6 focuses on stress accumulation during ion bombardment for 250 eV Ar+-bombarded Si. I describe the construction of a multi-beam optical stress sensor vacuum chamber that allows measurements of stress accumulation during ion bombardment for a full range of ion incidence angles. I present data for x-direction stress accumulation and describe analysis techniques used to identify two stress responses in the system: an initial, compressive stress consistent with Swenson's model of beam-induced swelling, and a second stress response at longer times that switches from compressive to tensile in the range of 40°--60° ion incidence, consistent with Norris's model of pancake strain.