| Resumo: | Volatile fingerprinting by Headspace-Solid-Phase Microextraction followed by Direct Injection Mass Spectrometry (HS-SPME DIMS) has been assessed for the first time as an alternative approach to gas chromatographic methods (HS-SPME GC-MS) for the rapid classification of honey botanical origin. In order to fully evaluate the potential of SPME for honey authentication, samples from different botanical sources (eucalyptus, citrus, acacia, rosemary and honeydew) were analyzed by both approaches using either carboxen/polydimethylsiloxane (C/PDMS) or polyacrylate SPME fibers. These datasets were further subjected to different supervised and unsupervised chemometric techniques, including powerful machine learning methods such as FreeViz, not previously applied for this purpose. The best overall classification rate (99% success), outperforming that provided by GC-MS data (98.25% success), was obtained when partial least squares-linear discriminant analysis was applied to mass spectral fingerprints gathered by C/PDMS fiber. The HS-SPME DIMS approach here optimized is shown as an advantageous alternative in terms of analysis time over chromatographic methods (2 vs 50 min), as well as a reliable and cost-efficient approach for honey source authentication issues both in research and food industry fields.
|
|---|
