HS HCC Techniques In Combination With Multi Cumulative Trapping and GC × GC qMS Followed by Machine Learning to Explore the Impact Of Packaging On Coffee Brew Aroma

The food analysis can use a large panel of different techniques to assess the authenticity and the quality of foodstuff. Volatile and semi-Volatile Organic Compounds (VOC and sVOC) are often monitored for assessing authenticity and quality of foods. In fact, VOCs and sVOCs may contain multi-level information. Their analysis is usually performed by the sampling of the headspace by high concentration capacity sorptive (HS-HCC) techniques, as most often solid-phase microextraction (SPME), followed by mono-dimensional gas chromatographic (GC) separation [1]. The use of comprehensive bi-dimensional GC (GC×GC), which enables the separation according to two different stationary phases, has been widely proved to increase the information extractable from the VOCs profile, allowing treating the generated two-dimensional chromatogram as a highly informative fingerprint [2].
On the other side, still a broad range of improvement is present in the upfront sample preparation step, different HS-HCC are now available and easily automated, as well as the possibility to add a trapping step to perform multiple-cumulative extractions [3].

This work investigates the impact of the packaging on the coffee brew aroma by using and comparing different tools (HiSorb and SPME) and approaches (mono extraction, single or multi-vial cumulative trapping extraction) through GC×GC-qMS and automated chemometric processing. Automatic alignment and tiles-based methods followed by machine learning and clustering was used to treat and reveal impacting features. The performance of the different tools and approaches have been compared as well as the associated clustering efficacy.