| Summary: | Sample clean-up methodologies based on supercritical fluid extraction (SFE) have been developed for determination of fat-soluble vitamins in foods. This thesis focuses on method development, stressing important issues concerning sample pre-treatment, extraction conditions and collection parameters. It was demonstrated that by using SFE followed by saponification and determination using RP-HPLC with UV and fluorescence detection, quantitative recoveries of vitamins A, E and b-carotene could be obtained in milk, milk powder, minced meat, liver paste, infant formula and canned baby food. The importance of modifying the sample matrix by adding an ethanolic entrainer prior to extraction was stressed. It was also shown that careful optimisation of the collection parameters is essential to accurately determine the vitamins. When solvent collection was employed, the type of solvent had an especially significant impact on the collection efficiency. In the work reported in this thesis, the possibility of replacing the rather harsh saponification step with on-line enzymatic hydrolysis using lipases was also investigated. Important parameters such as water content, temperature and extraction pressure were examined using several lipases and one esterase. It was shown that immobilised lipase from Candida antarctica type B was superior in terms of activity and operational stability. This on-line extraction/reaction methodology resulted in 100% conversion of retinyl palmitate to retinol (vitamin A), samples sufficiently clean for direct vitamin determination, and quantitative recoveries of vitamins A and E. Moreover, the methodology was fully automated and consumed much smaller amounts of organic solvents than conventional extraction methodologies. Instead of trying to obtain all the selectivity in the extraction step, work has also been done on using a solid phase trap in the collection step for selectivity enhancement. With a longer trap than is normally used in this type of equipment, it was found that the collection ...
|
|---|