Structural and compositional properties of semiconductor quantum dots and nanocrystals

摘要

The research carried out here employed analytical and imaging transmission electron microscopy and scanning transmission electron microscopy to gain a good understanding of local structure and composition of semiconductor nanocrystals and quantum dots for electronics and optoelectronics applications. One of the world's most advanced analytical scanning transmission electron microscopes in the field, the Daresbury SuperSTEM, was used to scrutinise the structure and composition of the samples. Three nanostructure systems are investigated in this thesis: 1. Structures consisting of Ge-nanocrystals (NCs) in alumina. Here HRTEM suggests relaxed and twinned smaller NCs grown annealed at lower temperature compared to elongated non-faulty bigger NCs annealed at higher temperature. HRTEM also suggests a polycrystalline structure of the matrix. 2. With regards to the InAs/GaAs quantum dots (QD) the study aims in particular at elucidating QD formation by investigating samples grown with and without growth interrupt (GI). Diffraction contrast TEM shows formation of buried dots in the sample prepared with GI whereas for the sample without GI the immediate growth of GaAs after InAs inhibits diffusion and segregation of In adotoms, and no footprint of buried dots has been observed. HRTEM and HAADF show coherent QDs in the sample with GI and abrupt InAs/GaAs interfaces in the sample without GI. In executing energy electron loss spectroscopy (EELS) and geometric phase analysis (GPA) the distribution of In in InGaAs/GaAs QDs has been obtained in samples grown in the critical thickness regime for quantum dot formation. The highest In percentage achieved in the dots grown with a nominal fraction of 100% was ~70%. EELS shows variations in the In concentration within the QD structure and wetting layer 3. In the case of Er-doped Si-NCs in silica this research tries to provide an understanding of structure, composition and position of excess Si and Er in the silica matrix of materials prepared under different growth conditions and to correlate this information with the PL emission, all with the aim to find preparation routes for optimum optical efficiency for applications of this materials system in silicon photonics. High spatial correlation between Si-NCs, Er and O in the Er and Si co-implanted sample with strong indication of an Er-oxide/Si core-shell structure had been found. The lack of an Er-oxide plasmon indicates, however, that the shell structure and its interface with the SiNCs is highly defective and a likely cause for non-radiative recombination. The sample with similar excess Er and Si concentrations but prepared in a two-stage implantation and annealing process shows a 10 times improvement in the optical emission. Here no spatial correlation between Er and Si-NCs was found in core loss EELS. EELS and HAADF evidenced more highly, near-atomically dispersed Er in the matrix with no formation of a core-shell structure as compared to the co-implanted sample. No footprint of Er-silicide plasmon was observed by low loss valence band EELS investigation in the co-implanted sample.