| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 2.71 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
O presente trabalho consistiu na aplicação da voltametria cíclica como ferramenta analítica para a discriminação de amostras de méis, com o objetivo de se estabelecer uma técnica rápida, simples e barata em estudos futuros de análise de qualidade e autenticidade de méis. Para este propósito, o estudo envolveu a construção de eléctrodos de trabalho compósitos de grafite sem e com a adição de nanopartículas (NiO) que foram testados no seu desempenho analítico e comparados com os eléctrodos de trabalho de platina e carbono vítreo (eléctrodos comerciais), usando voltametria cíclica. Estes testes envolveram a análise de uma solução redox, de reação reversível, e a discriminação de cinco amostras de mel comerciais (caracterizadas no seu perfil polínico e cor). O método estatístico multivariado para comparação do desempenho analítico de cada eléctrodo na discriminação dos 5 méis selecionado foi o método de análise discriminante dos mínimos quadrados parciais (PLS-DA). A implentação deste método foi o segundo propósito deste estudo, tendo-se estabelecido critérios para a seleção do melhor modelo e avaliação da robustez na previsão da classificação das amostras. Neste estudo testaram-se também pré-tratamentos dos voltamogramas cíclicos para otimizar os modelos de previsão (escalonamento, 1ª e 2ª derivadas).
Os melhores modelos PLS-DA obtidos para os 4 eléctrodos testados mostraram resultados com sensibilidade e seletividade acima de 97%, indicando que apenas 1 amostra analisada no total de 40 análises (cada mel foi analisado 9 vezes) foi mal classificada nos modelos selecionados. O elétrodo de grafite foi o que permitiu obter o modelo PLS-DA mais simples (6 componentes) usando os dados originais dos voltamogramas cíclicos. Os resultados obtidos mostraram um cenário positivo para a análise voltamétrica cíclica, afirmando de que a mesma permite fazer a discriminação das amostras de mel e que poderá ser uma ferramenta adequada para a análise qualitativa deste produto.
The present work consisted in the application of cyclic voltammetry as an analytical tool for the discrimination of honey samples, in order to establish a fast, simple and inexpensive technique in future studies of honey quality and authenticity analysis. For this purpose, the study involved the construction of graphite composite working electrodes with and without the addition of nanoparticles (NiO) that were tested in their analytical performance and compared with platinum and vitreous carbon electrodes (commercial electrodes), using cyclic voltammetry. These tests involved the analysis of a redox solution, a reversible reaction, and the discrimination of five commercial honey samples (characterized in their pollen profile and color). The multivariate statistical method selected for comparing the analytical performance of each electrode on the 5 honeys’ discrimination was the method of partial least squares discriminant analysis (PLS-DA). The implentation of this method was the second purpose of this study, establishing criterias for the selection of the best model and evaluation of the prediction robustness in the samples’ classification. In this study, pre-treatments of the cyclic voltammograms were also tested to optimize the prediction models (scaling, 1st and 2nd derivatives). The best PLS-DA models obtained for the 4 electrodes tested showed results with sensitivity and selectivity above 97%, indicating that only 1 sample analyzed in the total of 40 analyzes (each honey was analyzed 9 times) was poorly classified in the selected models. The graphite electrode was the one that allowed to obtain the simplest PLS-DA model (6 components) using the original data of the cyclic voltamograms. The results obtained showed a positive scenario for the cyclic voltammetric analysis, stating that it allows to discriminate the honey samples and that could be an adequate tool for the qualitative analysis of this product.
The present work consisted in the application of cyclic voltammetry as an analytical tool for the discrimination of honey samples, in order to establish a fast, simple and inexpensive technique in future studies of honey quality and authenticity analysis. For this purpose, the study involved the construction of graphite composite working electrodes with and without the addition of nanoparticles (NiO) that were tested in their analytical performance and compared with platinum and vitreous carbon electrodes (commercial electrodes), using cyclic voltammetry. These tests involved the analysis of a redox solution, a reversible reaction, and the discrimination of five commercial honey samples (characterized in their pollen profile and color). The multivariate statistical method selected for comparing the analytical performance of each electrode on the 5 honeys’ discrimination was the method of partial least squares discriminant analysis (PLS-DA). The implentation of this method was the second purpose of this study, establishing criterias for the selection of the best model and evaluation of the prediction robustness in the samples’ classification. In this study, pre-treatments of the cyclic voltammograms were also tested to optimize the prediction models (scaling, 1st and 2nd derivatives). The best PLS-DA models obtained for the 4 electrodes tested showed results with sensitivity and selectivity above 97%, indicating that only 1 sample analyzed in the total of 40 analyzes (each honey was analyzed 9 times) was poorly classified in the selected models. The graphite electrode was the one that allowed to obtain the simplest PLS-DA model (6 components) using the original data of the cyclic voltamograms. The results obtained showed a positive scenario for the cyclic voltammetric analysis, stating that it allows to discriminate the honey samples and that could be an adequate tool for the qualitative analysis of this product.
Description
Mestrado de dupla diplomação com a Universidade Tecnológica Federal do Paraná
Keywords
Mel Voltametria cíclica Análise polínica Eléctrodos compósitos Método dos mínimos quadrados parciais Discriminação