Modifying coffee quality by chemical manipulation

Chemical modification was applied to a coffee process by-product, silver skin, as well as raw coffee beans, with the aim to improve their aroma quality. Heat treatment in combination with sugar addition or enzyme treatment was applied to silver skin to encourage Maillard reactions. The manipulation applied to silver skin, however, did not give satisfactory results as the treatments neither caused significant increase in coffee aroma levels, nor yielded coffee aroma with quality resembling that of the real coffee. Chemical modification of raw Robusta coffee was carried out using fractionation and reconstitution approaches. The fractionation process involved the use of three types of solvent varying in polarity, dichloromethane (DCM), methanol (MeOH), and water, thus yielding four raw coffee fractions: DCM-soluble, MeOHsoluble, Water-soluble, and residue fractions. The reconstitution process involved wet mixing of the raw coffee fractions, vacuum drying and moisture content adjustment. Several reconstituted coffees were prepared with various proportions of the raw Robusta fractions, roasted and subjected to volatile analysis by Gas Chromatography-Mass Spectrometry (GC-MS). Statistical analysis by Principal Component Analysis (PCA) and the calculation of sum of normalized standard deviation (SNSD) of aroma compounds' odour activity values (OAVs) showed that the reconstituted Robusta that yielded the least variation, in term of aroma profile, from the higher quality coffee, Arabica, was the one composed of 70% d. b. MeOH fraction, 30% d. b. cell-wall material (residue), and 11 % w/w moisture content. The aroma profile of Arabica coffee was used as a reference due to its fine flavour that is commonly considered of better quality compared to that of Robusta (Briandet, Kemsley et al. 1996). Sensory evaluation (by sniffing) employing hedonic pairwise comparison technique confirmed the result from the GC-MS analysis that the aroma quality of the chosen reconstituted Robusta was improved since its aroma was significantly more preferred to that of the Robusta by the judges (30 people). Non-volatile compound analyses, however, suggested the need for further sensory study that involves tasting/drinking of the brews made with the new reconstituted Robusta for it contained significantly higher contents of bitter/astringent taste compounds, i.e. chlorogenic acids, caffeine and trigonelline, than the original coffees that could also affect the overall sensory quality of the coffee.EThOS - Electronic Theses Online ServiceGBUnited Kingdo