BACKGROUND: Coffee samples adulterated with roasted corn and roasted soybean were analyzed using a voltammetric electronic tongue equipped with a polypyrrole sensor array. METHODS: Coffee samples were adulterated in concentrations of 2%, 5%, 10% and 20% of roasted corn and roasted soybean; 5 replicates of each were used. The discrimination capacity of a voltammetric electronic tongue elaborated with a polypyrrole sensor array, was evaluated by principal component analysis and cluster analysis, while the capacity to perform quantitative determinations was carried out by partial least squares. RESULTS: The results obtained by the application of principal component analysis showed an excellent ability to discriminate adulterated samples. Additionally, the classifications obtained by cluster analysis was concordant with those obtained by principal component analysis. On the other hand, the evaluation of the ability to quantitatively analyze the adulterated samples showed that the polypyrrole sensor array provides sufficient information to allow quantitative determinations by partial least squares regression. CONCLUSIONS: It could be concluded that the voltammetric electronic tongue used in this work allows the suf- ficient analysis of coffee samples adulterated with roasted corn and roasted soybean.
BACKGROUND: Coffee samples adulterated with roasted corn and roasted soybean were analyzed using a voltammetric electronic tongue equipped with a polypyrrole sensor array. METHODS: Coffee samples were adulterated in concentrations of 2%, 5%, 10% and 20% of roasted corn and roasted soybean; 5 replicates of each were used. The discrimination capacity of a voltammetric electronic tongue elaborated with a polypyrrole sensor array, was evaluated by principal component analysis and cluster analysis, while the capacity to perform quantitative determinations was carried out by partial least squares. RESULTS: The results obtained by the application of principal component analysis showed an excellent ability to discriminate adulterated samples. Additionally, the classifications obtained by cluster analysis was concordant with those obtained by principal component analysis. On the other hand, the evaluation of the ability to quantitatively analyze the adulterated samples showed that the polypyrrole sensor array provides sufficient information to allow quantitative determinations by partial least squares regression. CONCLUSIONS: It could be concluded that the voltammetric electronic tongue used in this work allows the suf- ficient analysis of coffee samples adulterated with roasted corn and roasted soybean.