| New methods to study crystal surface dynamics were developed, within the research contexts of (1) pontaneous pattern formation on surfaces bombarded by ion beams, and (2) non-continuum surface relaxation below the thermodynamic roughening temperature.; First, theoretical and simulation models were developed to clarify the types of non-continuum behavior one might expect in the relaxation of nano-scale bumps and ripples. We used, for the first time, the "kinetic Monte Carlo" algorithm to study surface relaxation in simulation at virtual temperatures well below those previously studied. We developed a model that described the kinetics of "pinch-off," by which a surface becomes flatter when opposing steps at the high and low points of ripples thermally fluctuate into each other to create regions of high step curvature that emit atoms and absorb them, respectively.; Second, to make surfaces suitable for surface relaxation experiments, we used a spontaneous pattern forming instability on ion-bombarded Si(001) to make clean periodic rippled structures with wavelengths of 200--600 nm. The relationship between the ripple wavelength, sample temperature, and ion beam flux gave a sensitive measurement of the migration energy of dimers on Si(001) between 500 and 600°C, which was found to be 1.2 +/- 0.1 eV. We discovered that the concentration of mobile species on the surface is temperature and ion flux independent. We also discovered the mechanism for why the ripple amplitude does not grow without bound; the reason, in fact, is related to the pinch-off model for surface relaxation.; Third, using sputter rippled surfaces for samples, the relaxation behavior of Si(001) was studied in the range of 550 to 750°C. We discovered that the ripple amplitude did not decay exponentially. Instead, amplitude decay followed an inverse linear time dependence, indicating that relaxation was being driven by step-step interaction. By analyzing the "data collapse" of the amplitude decay curves, an activation energy of 1.6 +/- 0.2 eV was measured, consistently interpreted as the sum of the dimer creation energy (independently measured to be 0.35 +/- 0.05 eV) and our own measurement of 1.2 +/- 0.1 eV of the migration energy. |
