Métodos qualitativos para detecção de adulterantes em leite cru: validação intra e interlaboratorial

• Ano de defesa: 2016
• Autor(a) principal: Carina de Souza Gondim
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais
• Palavras-chave em Português: Métodos qualitativos; Validação de método; Alimentos Análise; Leite Adulteração e inspeção; Espectroscopia de infravermelho
• Link de acesso: https://hdl.handle.net/1843/BUOS-BAPJJS

Classical and instrumental qualitative methods have been used in the detection of adulterants in milk, although few studies related to their performance are available in the literature. In this context, the main objectives of this work were: i) to validate qualitative classical tests for detection of milk adulterations by in-house and collaborative processes, proposing improvements in the analytical procedure and strategys for their interlaboratorial validation; and ii) to evaluate the use of mid infrared spectroscopy by attenuated total reflectance (ATR) also in the detection of milk adulterations by different classification approaches. The classical methods evaluated were those described in Brazilian legislation for the detection of starch, chlorides, sucrose, chlorine and hypochlorite, hydrogen peroxide and neutralizers (rosic acid method). In general, these tests presented satisfactory performance, considering the characteristic parameters of qualitative methods, being the modified versions of the methods of detection of starch, chlorides and sucrose, more efficient in the detection of adulterants than the official versions, that was confirmed by interlaboratorial validation processes. The application of infrared spectroscopy technique was performed through a sequential strategy that involved different classification approaches. In the untargeted approach, the determination of the adulterants water, sodium bicarbonate, sodium chloride and sodium hypochlorite was considered unsuitable. In the targeted approach, the two-class Soft Independent Modeling of Class Analogy (SIMCA) models performed better than the SIMCA model established for all the adulterants. The instrumental method was considered suitable as a screening tool for adulterated milk samples. The evaluation of both methodologies provided a deeper knowledge about techniques widely used in the control and monitoring of milk frauds.