PROGRESS IN HAZELNUT QUALITY ASSESSMENT VIA ARTIFICIAL INTELLIGENCE (AI) SMELLING BASED ON GC×GC

The European hazelnut (Corylus avellana L.) is a tree nut that is mainly used by the confectionery industry in the form of raw or roasted kernels and in addition to chocolate pastes. The quality assessment of the raw materials has crucial repercussions in the supply chain costs; it is nowadays mainly based on the human examination, but to continue ensuring and fulfilling quality standards, fast, accurate, and objective quality control methods are of primary importance to support and sustain the human judgment. The whole volatile fraction of raw nuts, also referred to as volatilome, encrypts quality related information on cultivar/geographical origin, post-harvest treatments, bacteria/moulds contamination, oxidative stability, and overall sensory perception. The Artificial Intelligence (AI) smelling concept defined within sensomics is based on the detection of key odorant that have been proved to evoke a specific smell, it was realized in this research on a comprehensive two-dimensional gas chromatography system coupled with both mass spectrometer and flame ionization detector (GC×GC-MS/FID) to accurately quantify via multiple headspace solid-phase microextraction (MHS-SPME) more than 40 analytes including key-aroma compounds and spoilage markers.