| Electronic and nanostructured materials have been investigated using advanced scanning transmission electron microscopy (STEM) techniques. The first topic is the microstructure of Ga and Sb-doped ZnO. Ga-doped ZnO is a candidate transparent conducting oxide material. The microstructure of GZO thin films grown by MBE under different growth conditions and different substrates were examined using various electron microscopy (EM) techniques. The microstructure, prevalent defects, and polarity in these films strongly depend on the growth conditions and substrate. Sb-doped ZnO nanowires have been shown to be the first route to stable p-type ZnO. Using Z-contrast STEM, I have showed that an unusual microstructure of Sb-decorated head-to-head inversion domain boundaries and internal voids contain all the Sb in the nanowires and cause the p-type conduction.;InGaN thin films and InGaN / GaN quantum wells (QW) for light emitting diodes are the second topic. Low-dose Z-contrast STEM, PACBED, and EDS on InGaN QW LED structures grown by MOCVD show no evidence for nanoscale composition variations, contradicting previous reports. In addition, a new extended defect in GaN and InGaN was discovered. The defect consists of a faceted pyramid-shaped void that produces a threading dislocation along the [0001] growth direction, and is likely caused by carbon contamination during growth.;Non-rigid registration (NRR) and high-precision STEM of nanoparticles is the final topic. NRR is a new image processing technique that corrects distortions arising from the serial nature of STEM acquisition that previously limited the precision of locating atomic columns and counting the number of atoms in images. NRR was used to demonstrate sub-picometer precision in STEM images of single crystal Si and GaN, the best achieved in EM. NRR was used to measure the atomic surface structure of Pt nanoacatalysts and Au nanoparticles, which revealed new bond length variation phenomenon of surface atoms. In addition, NRR allowed for measuring the 3D atomic structure of the nanoparticles with less than 1 atom uncertainty, a long-standing problem in EM. Finally, NRR was adapted to EDS spectrum images, significantly enhancing the signal to noise ratio and resolution of an EDS spectrum image of Ca-doped NdTiO3 compared to conventional methods. |
