Rapid and Green Volatilomics of Roasted Hazelnuts: Comparative Multiplatform Analysis and Advanced Data Mining

Volatile organic compounds (VOCs) are fundamental in determining the sensory quality of food products, with their profiles influenced by various factors. Specifically, the hazelnut volatilome is shaped by the origin (botanical, geographical) of the kernels, as well as by the industrial processing steps they undergo (storage, roasting). Volatilomics, therefore, is a valuable tool for quality assessment and process monitoring. While Gas Chromatography coupled with Mass Spectrometry (GC-MS) remains the reference technique, reducing analysis time and minimizing the environmental impact of analytical methods are becoming pressing priorities, especially in industrial contexts. The aim of this thesis was to evaluate the potential of Gas Chromatography coupled with Ion Mobility Spectrometry (GC-IMS) and Proton Transfer Reaction - Time of Flight - Mass Spectrometry (PTR-ToF-MS) as rapid and green analytical techniques for food volatilomics, focusing on their application to the roasted hazelnut volatilome characterization. Although PTR-ToF-MS is a well-established technique, its compliance with the principles of Green Analytical Chemistry had not been previously assessed. Therefore, this thesis begins by answering the question: “Is PTR-MS a green analytical tool?” GC-IMS, on the other hand, is an analytical platform that has only recently gained popularity in food volatilomics, and a significant part of the project was dedicated to assessing its potential for routine analysis of roasted hazelnut paste samples. Our investigation focused on various aspects: (i) optimizing an analytical protocol suitable for large-scale samplings (e.g., process monitoring and phenotyping); (ii) evaluating the IMS detector response to alkyl pyrazines, which are key compounds for hazelnut flavor and roasting markers; (iii) developing a quantitative approach for VOCs in hazelnut paste; (iv) implementing an automated peak detection workflow to extend GC-IMS application to untargeted fingerprinting. Furthermore, GC-IMS was employed for high-throughput analysis of large sample sets in two key applications: process monitoring and phenotyping. The results highlight its usefulness in both industrial research settings and as a support tool for agricultural breeding studies. Finally, to objectively compare the three techniques (GC-MS, GC-IMS, and PTR-ToF-MS), the JIVE (Joint and Individual Variation Explained) data fusion approach was applied to a dataset obtained from the multiplatform analysis of roasted hazelnut paste samples. JIVE allowed for the effective separation of shared from platform-specific analytical information. In conclusion, this thesis investigates the potential and limitations of applying rapid and green analytical techniques to food volatilomics, particularly in the case of roasted hazelnuts. It establishes PTR-ToF-MS as a green analytical technique and explores the potential of GC-IMS. Finally, by leveraging advanced data analysis techniques, it outlines an approach to objectively compare the three methods, highlighting their complementarity and proposing a strategy for selecting the most appropriate method based on study objectives and constraints.