A Raman Spectroscopic Study of Paint and Dairy Samples

Abstract

This thesis investigates the use of Raman spectroscopic techniques for analysing paint and dairy samples (milk powder and model cheese systems).

Several samples of paint from a painting called ‘Hope’ by Edward Burne-Jones were obtained from the Dunedin Public Art Gallery. The purpose of this study was to determine whether the pigments and/or binder could be identified using Raman microscopy. The technique conclusively identified several pigments in two of the samples including: cerulean blue, Prussian blue, Vermilion and zinc white.

FT-Raman spectroscopic data was investigated with respect to its ability to be used to identify milk powder samples spiked with calcium carbonate and whether quantification of the adulterant is possible. Use of the spectroscopic data with principal components analysis (PCA) was useful for identifying samples which had been spiked and for distinguishing samples with larger differences in calcium carbonate concentration. Partial least squares (PLS) and univariate analysis were able to be used with spectroscopic data to create standard curve plots which could be used in the quantification of calcium carbonate in milk powder. However, limit of quantification (LOQ) values indicated that neither technique would be effective in quantifying calcite concentrations as low as those used for calcium fortification in retail milk powder samples.

Model cheese systems were the primary focus of this research. Raman spectroscopic techniques (primarily Raman microscopy) were investigated with respect to their uses in studying cheese. FT-Raman spectroscopy with PCA was used determine whether model bi-component cheese mixtures could be distinguished based on constituent concentrations. PCA was found to be very effective for achieving this. Mixtures were then analysed with respect to whether interactions between components were observable in the spectra (as band broadening or shifting). Interactions were not found. Morphology of the model cheese mixtures was investigated using confocal Raman microscopy. This technique was effective for analysis of morphology of model bi-component cheese mixtures in both two and three dimensions. Overlap between peaks of different components is the main limitation of this method when analysing morphology. Raman microscopy was compared to confocal laser scanning microscopy (CLSM) with respect to production of 3D images. While CLSM produced images in much less time, Raman microscopy has the advantage of being able to identify carbohydrates.