Abstract
Liquid chromatographic (LC) methodology potentially useful for the characterization of orange juice, with particular regard to detecting adulteration of orange juice by computer pattern recognition analysis, has been developed. After dilution with methanol the juice is extracted with hexane to remove the carotenoids, which are chromatographed on a C18 column with an acetonitrile-methanol-methylene chloride mobile phase and detection at 450 nm. Further extraction of the juice with methylene chloride isolates the methoxylated flavones, which are chromatographed by reverse phase LC with an acetonitrile-methanol-water mobile phase and detection at 280 nm. The flavanone glycosides remaining in solution are chromatographed on a C18 column with an acetonitrile-water mobile phase and detection at 280 nm. The precisions of the heights of the 32 LC peaks selected for pattern recognition analysis were determined from 5 replicate analyses of a single juice. Coefficients of variation of the replicates ranged from 0.3 to 4.5%, with an average of 2.1%. Adulteration of products with sodium benzoate-fortified pulpwash or grapefruit juice can be detected by this method. Pattern recognition analysis of the data obtained for 80 authentic and 19 adulterated orange juices showed that the method is potentially useful for distinguishing between authentic and adulterated products.
Application of artificial intelligence in pancreaticobiliary diseases
