Quantitative Spectroscopic Analysis Of Complex Heterogeneous Systems: Theory And Applications

Chemometrics is an interdisciplinary field, lying between the more established fields of chemistry, mathematics, statistics and computer sciences. The combination of instrument analysis and chemometric methodologies provides a new avenue to extract the information of analytes in complex systems. Spectral measurements of complex heterogeneous systems contain not only the spectral variations induced by the changes in the concentrations of chemical components but also those caused by the changes of the non-chemical factors, such as the fluctuations in the measurement conditions and samples’physical properties. It is rather challenging to extract useful chemical information from such complex spectral data with the application of traditional chemometric data analysis methods. Traditional multivariate calibration models are often affected by significant multiplicative effects resulting from light scattering, due to physical variations (e.g. particle size and shape, sample packing and sample surface, etc.) and variations in other non-chemical factors inherent within the individual samples. Therefore, the development of advanced methods capable of eliminating such effect and maintaining the robustness of the calibration model is the key to achieve accurate quantitative spectroscopic analysis of complex heterogeneous mixtures. This thesis focuses on accurate quantitative spectroscopic analysis of complex heterogeneous mixtures and mainly deals with the following aspects:In Chapter2, the influence of particle size distribution and sample compactness on Raman signals of powder mixtures was investigated, and an advanced calibration strategy was proposed and applied to the quantitative analysis of Raman measurements of a model system of powder mixtures consisting of barium nitrate and potassium chromate. The experimental results show that the proposed calibration strategy can provide comparatively accurate quantitative predictions for mass fractions of barium nitrate in powder mixtures with variations in particle size distribution as well as compactness. This technique holds promise for rapid, noninvasive and on-line monitoring of complex heterogeneous mixtures (e.g suspension system) using Raman spectrometry.In Chapter3, the effects of absorption and scattering properties of turbid medium on fluorescence spectra of biological samples, such as tissues or cell suspensions were explicitly studied. A novel fluorescence calibration model was proposed and applied to the quantitative analysis of a model system (Proflavine-Intralipid emulsion system). The experimental results show that the dual calibration strategy can provide comparatively accurate quantitative predictions for the concentration of Proflavine in turbid medium. This study indicates that accurate quantitative analysis of turbid samples by fluorescence spectrometry can be achieved through the combination of spectroscopic techniques with smart modeling methodology.In Chapter4, Raman imaging technique coupled with smoothed principal component analysis was applied to the analysis of tablets consisting of acetylsalicylic acid and magnesium stearate. The experimental results demonstrated that smoothed principal component analysis can significantly reduce the acquisition time of a Raman spectral image while retaining the accuracy of quantitative results, and hence promote the application of Raman imaging technique in area of rapid and noninvasive analysis of complex samples.