Untitled

Funder

Organizational Unit

Publications

Application of an electronic tongue to detect gliadins in gluten-free foods

Publication . Peres, António M.; Dias, L.G.; Veloso, Ana C.A.; Meirinho, Sofia G.; Morais, Jorge Sá; Machado, Adélio A.S.C.

Celiac disease is an autoimmune-mediated disorder triggered in genetically susceptible individuals by the ingestion of some gluten proteins, namely gliadins. To prevent inadvertent gluten consumption, the Commission Regulation (EC) N°41/2009 will implement labeling foods as ‘gluten-free” or “low-gluten content.

2010Conference object

Open access

An electronic tongue for beer differentiation

Publication . Dias, L.G.; Peres, António M.; Barcelos, Tânia P.; Morais, Jorge Sá; Machado, Adélio A.S.C.

In this work an electronic tongue, based on a potentiometric solid-state mufti-sensor array, with 36 polymeric membranes, was built for developing an analytical tool to apply in process monitoring and quality control. As a first approach, this tool was applied together with a supervised pattern recognition tool to semi-quantitatively differentiate beers with different alcoholic levels.

2009Conference object

Open access

An electronic tongue taste evaluation: identification of goat milk adulteration with bovine milk

Publication . Dias, L.G.; Peres, António M.; Veloso, Ana C.A.; Reis, Filipa S.; Vilas-Boas, Miguel; Machado, Adélio A.S.C.

An electronic tongue with 36 cross-sensibility sensors was built allowing a successful recognition of the five basic taste standards, showing high sensibility to acid, salty and umami taste substances and lower performance to bitter and sweet tastes. The taste recognition capability was afterwards tested in the detection of goat milk adulteration with bovine milk, which is a problem for the dairy industry. This new methodology is an alternative to the classical analyticalmethods used to detect caprine milk adulterations with bovine milk, being a simpler, faster and economical procedure. The different signal profiles recorded by the e-tongue device together with linear discriminant analysis allowed the implementation of a model that could distinguish between rawskim milk groups (goat, cowand goat/cow) with an overall sensibility and specificity of 97% and 93%, respectively. Furthermore, cross-validation showed that the modelwas able to correct classify unknown milk samples with a sensibility and specificity of 87% and 70%, respectively. Additionally, the model robustness was confirmed since it correctly or incorrectly classified milk samples with, respectively, higher and lower probabilities than those that could be expected by chance.

2009Journal article

Open access

Análise semi-quantitativa e quantitativa de bebidas não alcoólicas com uma língua electrónica

Publication . Barcelos, Tânia P.; Dias, L.G.; Peres, António M.

Neste trabalho usou-se uma língua electrónica constituída por dois sistemas de multi-sensores, com 40 membranas poliméricas sensibilidade cruzada, com composições diferentes, e um eléctrodo de referência Ag/AgCl. O sistema foi aplicado na análise semi-quantitativa de bebidas não alcoólicas e na quantificação das concentrações de compostos orgânicos e açúcares existentes nessas bebidas. As concentrações de açúcares (frutose e glucose), do antioxidante (ácido ascórbico) e dos acidificantes (ácido málico e ácido cítrico) foram determinadas por HPLC. Os sinais das membranas poliméricas registados durante a análise das bebidas não alcoólicas com a língua electrónica foram tratados recorrendo a métodos quimiométricos não supervisionados (análise dos componentes principais) e métodos supervisionados (análise discriminante linear, regressão linear múltipla e regressão pelos mínimos quadrados parciais). Na análise semi-quantitativa, o objectivo foi diferenciar quatro grupos de bebidas não alcoólicas com diferentes concentrações de sumo de fruta adicionado: superiores a 30%, entre 14% e 30%, 5% e 10% e 0.1% e 2%. Foram analisadas, com a língua electrónica, 16 bebidas portuguesas adquiridas em supermercados. Os dados obtidos foram tratados com análise discriminante linear com o procedimento “stepwise”, o que permitiu classificar correctamente 100% dos dados originais e 93.8% no processo de validação cruzada. Para quantificar as concentrações dos acidificantes e açúcares foram usados dois métodos lineares para a obtenção de modelos de estimação/previsão: regressão linear múltipla e método dos mínimos quadrados parciais. Os modelos lineares obtidos apresentaram bons coeficientes de determinação (R2>0,84), tanto no grupo de calibração como no grupo de validação cruzada, para a estimação e previsão das concentrações dos açúcares. O modelo de previsão obtido para a quantificação do ácido cítrico não apresentou resultados satisfatórios. Para o ácido málico verificou-se não haver informação suficiente nos sinais obtidos da língua electrónica para obter qualquer modelo de quantificação. O ácido ascórbico não foi considerado no estudo quantitativo porque não foi encontrado em todas as amostras. In this work was used an electronic tongue built with two multi-sensors system, with 40 different polymeric membranes with cross-sensibility, and an Ag/AgCl reference electrode. These devices were applied in semi-quantitative analysis of soft drinks and in the quantification of organic acids and sugars present in these beverages. The beverage’s composition in sugars (fructose and glucose) and organic acids (ascorbic acid, malic acid and citric acid) was determined by HPLC. The signals of the electronic tongue polymeric membranes obtained in the analysis were treated with non-supervised multivariate methods such as, principal components analysis, and supervised methods such as, linear discriminant analysis, multiple linear regression and partial least square regression. In semi-quantitative analysis, the objective was to differentiate four nonalcoholic beverages with different added fruit juice contents: higher than 30%, between 14%-30%, 5%-10% and 0.1%-2%. A set of 16 portuguese beverages, purchased in supermarkets, were analysed with the electronic tongue device. The data obtained were treated by stepwise linear discriminant analysis, allowing a 100% overall correct classification using the original data and 93.8% for the cross-validation procedure. For the quantification of the organic acids and sugars were used two linear methods to obtain estimation/prediction models: multiple linear regression and partial least squares regression. All linear models obtained for the estimation and prediction of concentrations of sugars in the beverages, showed a good coefficient of determination (R2>0,84) for the original data and for the cross-validation procedure, respectively. The predictive model obtained for the citric acid quantification did not produced satisfactory results. For malic acid there was no sufficient information in the signals from the electronic tongue to obtain a quantification linear model. Ascorbic acid was not considered in the quantitative study because it was not found in all samples.

2009Master thesis

Open access

An electronic tongue for honey classification

Publication . Dias, L.G.; Peres, António M.; Vilas-Boas, Miguel; Rocha, Maria A.; Estevinho, Leticia M.; Machado, Adélio A.S.C.

An electronic tongue system was developed based on 20 all-solid-state potentiometric sensors and chemometric data processing, with polymeric membranes applied on solid conducting silver-epoxy supports and a Ag=AgCl reference electrode. The sensor array was applied to 52 commercial honey samples obtained randomly from different regions of Portugal. These samples were analysed independently for their pollen profiles by biological techniques and the data collected with the tongue were evaluated for discrimination of the samples with multivariate statistical methods (principal component analysis and linear discriminant analysis), to investigate whether the device may provide an analytical alternative for classification of honey samples with respect to pollen type, a task which is time consuming and requires skilled labour when performed by biological techniques. It was found that the tongue has a reasonable efficiency for classification of honey samples of the most common three types (with Erica, Echium and Lavandula as predominant pollens). With linear discriminant analysis, the honey samples yielded about 84% classification accuracy and 72% for crossed validation. In this study, the honey samples correctly classified for the different types of the dominant pollen were: 53% for Lavandula, 83% for Erica and 78% for Echium pollen.

2008Journal article

Open access