Influência de um tratamento com vapor e ácido nas características da bebida café conilon

Abstract

Coffee beans, species Coffea canephora variety conilon were submitted into a treatment with water steam and acid solution by an experimental design, which was generated using the methodology of the central composite rotational design (DCCR). The selected variables to the experimental design were steam pressure, acid solution concentration and time of coffee bean's contact with steam. In this design seventeen tests were proposed including the central point with three repetitions. The results obtained were processed in Statistica software version 10.0, by ANOVA tables that were used for assessing the level of significance and model, beyond the generation of response surfaces to indicate the influence of the variables in the final result. Coffee beans samples were roasted and analyzed for polyphenol content, pH, titratable acidity, caffeine and aromatic compounds. The aromatic compounds analyzed were divided into two groups, those that contribute positively or negatively to coffee flavor. Analysis of total polyphenols and caffeine were done by high performance liquid chromatography with a photodiode array detector (HPLC / DAD) and for analysis of aromatic compounds, gas chromatography -mass spectrometry (GC / MS). The results indicated that in higher values of steam pressure and contact time with the same, the titratable acidity decreased even with higher concentrations of acid solution applied. Caffeine showed maximum results for conditions of 0.8 bar to steam pressure, 12 minutes of contact with steam and acid solution concentration between 5.0 to 7.0 %. The higher results for the positive aromatic compounds were the conditions of acid concentration between 6 to 8 %, steam pressure between 0.4 to 0.6 bar and the contact time between 24 to 26 minutes, negative aromatic compounds showed minimum values in acid concentration between 0 to 1 %, steam pressure between 0.4 to 0.8 bar, and the contact time between 18 to 20 minutes. The statistical analysis showed that polyphenols and pH were significantly influenced by the variation of the roasting degree, not allowing to obtain predictive models by experimental design.