Determination of cobalt in wine and must by electrothermalatomization atomic absorption spectrometry: analyticalmethod development and validation

• Autor(a) principal: Catarino,S.
• Data de Publicação: 2009
• Outros Autores: Neves,A., Sousa,R. Bruno de
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://scielo.pt/scielo.php?script=sci_arttext&pid=S0254-02232009000200002
• topic: wine; cobalt; analytical method validation; electrothermal atomization atomic absorption spectrometry

An electrothermal atomization atomic absorption spectrometry method was optimized to quantify cobalt in musts and wines, especially within quality control scope. As far as wine samples are concerned, the method involves reduced risk of contamination by eliminating prior treatment other than dilution (1:2). For must samples a pre-treatment is needed, in order to destroy the organic matter that could interfere with the analytical determination. A conventional thermal program (with pyrolysis step), with and without matrix modifier [Mg(NO3)2], was developed and optimized. Pyrolysis and atomization curves were studied using a Co standard solution, must and wine samples. The analytical figures of merit of the method such as linearity of the calibration curve, specificity (standard additions test), accuracy, analytical limits and precision were studied using wine samples. The linear dynamic range under optimized conditions was 0.4-10.0 μg/L. The recoveries were between 87% and 98% for all of the wines and addition levels used in the test. The method presents a very satisfactory specificity for red wines, while matrix effect was observed for white wines. The limit of detection in undiluted wines was 0.8 μg/L. Repeatability was lower than 0.7 μg/L (RSD < 5%). For accuracy evaluation, Co content in six wine samples and one Co standard solution, from an international collaborative trial, was determined by ETAAS and ICP-MS with comparable results (in general the differences were lower than 5%). Bearing in mind that must samples are previously digested, it is expected that the method performance parameters are at least as favourable as those observed with wines.