UCLA

Electron microscopy has found wide application in material science and biology with the nanometer / angstrom resolution in planar images. Tomography has also made a revolutionary impact of non-destructively visualizing inner three-dimensional structures, especially in the field of clinical medical imaging. Digital signal processing is used in a broad range of electrical engineering to separate the signal from the noise; hence extracting true information embedded from the noisy data. In the past century, imperfections inside the crystalline structures have caught material scientists' eyes due to the capability of significantly changing physical properties of materials. In this dissertation, a remarkable stride in the field of electron tomography has been made by combining several novel techniques: scanning transmission electron microscopy to obtain high resolution two-dimensional images, the center of mass alignment method to solve the mis-alignment problem, the equally sloped tomography method to achieve best spatial resolution by dramatically alleviating the missing wedge problem, and the three-dimensional Fourier filtering method to enhance the signal-to-noise ratio. With these combinations, a 2.4 angstrom resolution in the tomographic reconstruction is demonstrated. Furthermore, atomic steps at the boundary between two grains, edge dislocations, and screw dislocations inside a 10nm platinum particle are observed, atom-by-atom.