71074

SPME-GC-MS AND MULTIVARIATE ANALYSIS IN THE STUDY OF ROASTING VARIABLES FOR SPECIALTY COFFEES

Favorite this paper

The sensory analysis classifies specialty coffees and determines their market values. However, human and environmental factors can still affect the perception of coffee flavors. An alternate solution for the evaluation of coffee quality that this work employed is the use of solid-phase microextraction associated with gas chromatography and mass spectrometry (SPME-GC-MS) to determine the profile of volatile organic compounds (VOCs). Chemometrics was also important and helped in the interpretation of results. A scanning multivariate experimental planning in two levels framed the evaluation of the main operational variables (starting temperature, rotation speed, fuel flow increase, time after the first bean expansion – first cracking – and vent opening) for a small industrial roaster (Atilla 2 kg Gold Plus). Diverse roastings of the same specialty coffee provided 6 control roastings and 32 test roastings. Two professionals sensory qualified each roasting using cupping (SCAA method). Ground roasted beans provided samples in triplicate for the determination of VOC profile by SPME-GC-MS. NIST 08 library matches with 75% agreement or higher and corresponding Kovats indices helped identify the substances of higher influence in coffee discrimination. The first principal component in the Principal Component Analysis (PCA) of sensory evaluation provided 86% of total variance. The time after first cracking was the main variable in the segregation of two major coffee roasting sets. The first two components of PCA in SPME-GC-MS evaluation provided 79% of variance and segregated three major coffee groups. Again, the time after first cracking was the main influence in segregation, but the VOC profile presented a higher discriminatory power than cupping. These preliminary results show that VOCs profiles can help the investigation and control of roaster operational variables to raise the quality of the final product, since these profiles may minimize human and environmental errors.