Discriminação de espécies de café (Coffea arabica e Coffea canephora) pela composição de diterpenos

AUTOR(ES)
DATA DE PUBLICAÇÃO

2008

RESUMO

Brazil is the major coffee producing and exporting country and the most important species are Coffea arabica and Coffea canephora (robusta). Because of its lower commercial value the robusta coffee is often blended to arabica, but presents inferior quality and acceptability. Visual indicators are eliminated after roasting and grinding, so an alternative method to identification of the species is required. Chromatographic techniques have been used, particularly High Performance Liquid Chromatography (HPLC), due to the versatility and the ability to quantify thermolabile compounds. The diterpenes kahweol and cafestol are reported as potential indexes because they are present in different contents in species and they have thermal stability greater than the majority of the water-soluble compounds. The evaluation of the roasting effect and the composition of diterpenes in defective beans are important in order to use these compounds to discriminate species in commercial coffees, in which the level of defects and roasting degree are unknown. The aim of the work was to discriminate arabica and robusta coffee species by their contents of diterpenes. Samples of arabica and robusta coffees and blends of these species with different roasting degree (light, medium and dark) and levels of defective beans were used. Direct saponification, extraction with tert-Butyl methyl ether and clean up with water were applied for extraction. Reverse-phase HPLC, isocratic elution (55% acetonitrile in water) and UV detection were used. The results were used to study the stability of diterpenes (ANOVA) and to generate and validate a predictive model to estimate the proportion of robusta added to arabica coffee (multiple regression). The content of kahweol varied between 661 and 933 mg/100 g in arabica coffee, but it was not found in robusta. Cafestol ranged from 275 to 478mg in arabica and from 163 to 275 mg/100 g in robusta coffee. The addition of robusta coffee reduced the diterpenes content. However the effect was differently depending on the roasting degree and level of defects. The degradation of diterpenes did not increase with the increasing severity of roast. The model which used cafestol as a single variable had low predictive power. The used of the content of kahweol in the model allowed a good fit (R2 = 0.91), which was improved (R2 = 0.93) by the combined use of the compounds kahweol and cafestol. Considering samples of each roasting degree independently, it was observed that medium and dark roasting models had better fit. This is interesting because coffees are usually commercialized with darker roast in Brazil. Validation using external samples shows the necessity to add more samples to increase the limits the model and its predictive power. There were no differences between kahweol and cafestol levels comparing defective or regular beans. The parameters kahweol and cafestol showed potential for discrimination of species on samples with different roasting degrees and level of defects.

ASSUNTO(S)

alimentos - avaliação sensorial avaliação sensorial café - torrefação diterpenos coffee food - sensory evaluation sensorial evaluation diterpenes

ACESSO AO ARTIGO

Documentos Relacionados