Thermal Analysis coupled to fast cycling on-line Gas Chromatography-Photo Ionization Mass Spectrometry to study the flavor formation during the roasting process of coffee beans, cacao beans and nuts

摘要

The mild roasting of plant seeds such as green coffee beans, cocoa beans or nuts generate a multitude of different flavour compounds. Thus roasting represents a very important value-generating process for the food industry, inducing a lot of fundamental and applied research in this food chemical field. However, in order to foster understanding of the flavour generation processes, modern on-line analytical approaches need to be applied to follow the processes in realtime. One possibility to monitor the effluents of flavours and other volatile roasting processes is photo ionisation (PI) mass spectrometry. By using different PI approaches, trace compounds can be ionized selectively or a universal but soft ionisation is obtained. So PIMS has been applied successfully to investigate evolved roasting gases from the interior of individual single coffee beans as well as from industrial roasters with a capacity of more than one metric ton per hour [1]. A particularly well suited tool for laboratory investigation of food roasting processes in a thermal analysis device (e.g. thermogravimetry, TG) coupled to a PIMS system. In order to further increase the selectivity of TA-PIMS a newly developed optically heating fast-cycling gas chromatograph (fast OHGC) was developed. The fast OHGC is placed between the TA-oven and the PIMS-system. The new OHCG allows a very fast GC cycle time with heating up from 50 to 250 ℃ and cooling down again to 50 ℃ in less than 30 s. With this system e.g. the evolution of isomeric and isobaric key flavour compounds such as pyrazines, furan derivatives, phenols and carbonylic compounds from the mild roasting of different nuts (e.g. peanuts) or coffee can be followed with a time resolution of the cycle time (down to 20-30 seconds). However, in order to utilize the fundamental information of the flavour formation for an industrial roasting control, a statistical analysis of a large number of roasting experiments is required, using advanced multivariate methods such as non-negative matrix factorisation (NMF) followed by e.g. PCA analysis of obtained factors. The new OHGC technology and TA-coupling as well as applications of the new systems to different food processes (roasting of coffee, several different nut species and cocoa) are described. Finally the implications for industrial and academic food research are discussed.