Quantitative analysis by electron energy-loss spectroscopy

In this thesis, I present studies of several problems related to quantitative EELS, such as the spectrometer-system linearity, TEM-system chromatic aberration, absolute cross-section measurements and thickness measurements. If an indirect-exposure parallel-recording spectrometer is used, intensity nonlinearity of the detection-system is found to contribute more than 20% inaccuracy to quantitative measurements under commonly applied conditions. I suggest a convenient method to generate a correction curve for this nonlinearity; as an example, a correction curve to Gatan 666 Parallel EELS system is given and applied to cross-section and mean free path measurements.;I have analyzed the chromatic-aberration effect of TEM lenses both theoretically and experimentally, including its sources, its propagations, its influence on EELS energy resolution and on the accuracy of quantitative analysis. From these analyses, I determine conditions which eliminate the effect. It is shown that the chromatic-aberration contribution from weak intermediate lens(es) provides major contribution in TEM diffraction mode.;Measurements of some inner-shell cross-sections have been carried out for seven elements, filling a considerable gap in existing data which has arisen from edge-overlap problems encountered in relative (oxide-standard) measurements. In my experiments, chromatic aberration effects are avoided by carefully choosing experimental conditions, and intensity nonlinearity of the detecting system is corrected. I present the results in the form of dipole oscillator strength and differential oscillator strength, allowing them to be applied to a wide range of collection angles and electron incident energies. Finally, I have investigated several methods of measuring the local thickness of a TEM specimen, and suggest that the "Kramers-Kronig" sum-rule method has a high potential for obtaining accurate thickness.