Silicon Quantum Dots in a Silicon Dioxide Matrix

Abstract

The aim of the master work was to both produce silicon quantum dots in a SiO2 matrix, and to optically characterize them. The main goal of the optical characterization was to find the band gap of the produced quantum dots, as well as to find the relation between the size of the quantum dots and their bad gap.No quantum dots were produced within the master work, most likely because the thin substoichiometric SiOx layers were too thin to resist oxidation, or because the silicon content in the SiOx layers was too low. Samples with confirmed quantum dots have been investigated with photoluminescence, cathodoluminescence, fluorescence microscopy and ellipsometry. The photoluminescence with the first setup in room temperature seemed to only detect emission from defect states in a silicon oxide. Photoluminescence with the second setup had emission peaks located at 840-880 nm corresponding to band gaps of 1.41-1.48 eV. Emission was highest at 150 K, but still strong at room temperature. The measured mean quantum dot sizes were 5.2-6.0 nm. Ellipsometry models were made and fitted to the measured data. The silicon quantum dot layers and surrounding matrix were best modelled as an anisotropic medium with two dispersion models. The modelled band gap energy was 2.1 eV.A model for the relationship between the quantum dot size and its band gap has been made for silicon quantum dots in SiO2 and in Si3N4. The model adjusting for changes in effective masses corresponded well to silicon quantum dots larger than 3 nm in SiO2 when compared to measured data.A 95% confidence interval was made for the difference in mean quantum dot diameter due to annealing temperature. The results suggest that higher annealing temperatures produce larger quantum dots.The expected photoluminescence peaks calculated for each quarter sample from the mean quantum dot diameter, were at longer wavelengths than the measured photoluminescence peaks. This could be due to nitrogen in the matrix surrounding the quantum dots, or that the mean diameter and size distribution is not entirely correct.The size distribution for one specific sample was used in combination with the model that predicted the band gap according to the quantum dot diameter, to model the photoluminescence curve of said sample. The modelled curve corresponded fairly well to the measured photoluminescence curve. The measured emission peak was located approximately 50 nm lower than the modelled emission peak, suggesting the presence of nitrogen or a shifted size distribution.