Authenticity control of pistachios based on 1H- and 13C-NMR spectroscopy and multivariate statistics

(1)H- and (13)C-NMR-spectra of pistachio oil from Iran, USA and Turkey were analysed with multivariate statistics in order to classify pistachios according to their countries of origin. The integrals of the resonances of the NMR-spectra contain information about the fatty acid composition of pistachio oil which is supposed to reflect the climatic conditions at the crop area. Consequently, NMR-spectroscopy was thought to be a suitable method for the distinction of those pistachio oils. Cluster analysis was found to lead to a distinction between pistachio samples from Turkey and other countries. Principle component analysis was used to extract components from the data, which primarily contain the information about unsaturated acyl chains in the oil and allowed to assign the samples very well to their respective countries of provenance. Linear discriminant analysis provides similar results with low error rate estimations. It was demonstrated that those data gained by (1)H- and (13)C-NMR-spectroscopy from samples with unknown provenance can be classified by principle component analysis as well as by discriminant analysis.