Detection of Plant-Derived Adulterants in Saffron ( Crocus sativus L.) by HS-SPME/GC-MS Profiling of Volatiles and Chemometrics

Gas chromatography with mass spectrometry detection (GC-MS) coupled with headspace-solid-phase microextraction (HS-SPME) was used to analyse the aroma profile of genuine saffron (Crocus sativus L.) and samples of this spice artificially adulterated with Calendula officinalis L. petals (calendula), Carthamus tinctorius L. petals (safflower) and Curcuma longa L. powdered rhizomes (turmeric). Preliminary analyses of genuine saffron and pure contaminants were performed to select the kind of SPME sorbent. Moreover, an experimental design combined with response surface methodology was applied to optimise the sample temperature and the fibre exposure time with the aim of enhancing the detection of the above adulterants in counterfeited saffron samples. The GC-MS chromatograms collected under the optimised conditions were finally handled by unsupervised and supervised multivariate statistical methods to differentiate the genuine saffron samples produced in three different Italian regions from artificially adulterated samples at 2–5% w/w contamination levels. Thirty genuine and 30 counterfeited (10 for each kind of adulterant) saffron samples were analysed. Principal component analysis was applied to assist the choice of the GC/MS data pre-treatment and classification of genuine and adulterated saffron samples was attempted by partial least square-discriminant analysis (PLS-DA). Predictive performance of PLS-DA models calibrated with 42 samples was finally tested on 18 saffron samples (9 genuine and 9 adulterated). All the external saffron samples were correctly classified regardless of the kind of contaminant, while in calibration, only a saffron sample contaminated with safflower was erroneously assigned to the group of genuine ones. Class modelling of genuine saffron performed by SIMCA (Soft Independent Model Class Analogy) exhibited a good sensitivity and 100% specificity for external adulterated samples.