RMN aliada a métodos estatísticos para controle de qualidade de sucos integrais e néctares de laranja

• Autor(a) principal: Wilhelms, Renan Ziemann
• Data de Publicação: 2021
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFG
• dARK ID: ark:/38995/0013000008wp3
• Texto Completo: http://repositorio.bc.ufg.br/tede/handle/tede/11633
• topic: RMN; Suco; Néctar; Laranja; O-PLS; RDA; PLS; NMR; Juice; Nectar; O-PLS; RDA; PLS; CIENCIAS EXATAS E DA TERRA::QUIMICA

Orange integral juice and nectar are highly consumed in the world and like any process is subject to production failures, however, with the use of 1H NMR spectroscopy is possible to identify these non-conformities. The present work aims to evaluate the chemical profile and identify chemical markers that can be used to build prediction models to distinguish integral juice from nectar and integral juice and nectar with and without the addition of apple. With the 1H NMR technique, it was possible to identify chemical markers to distinguish orange integral juice and nectar and the same ones with and without the addition of apple. The quantification of these chemical markers was done using Electronic REference To access In vivo Concentrations (ERETIC2), which were sucrose, α-glucose, β-glucose, malic acid, fructose, dimethylproline (DMP), citric acid, alanine, lactate and ethanol. Among these compounds, DMP stands out as a chemical marker for evaluating juices and nectars with and without apple addiction. The Redundancy Analysis (RDA) was of great importance as a filter for verifying the information on the labels of integral juices and nectars, where two samples with possible adulteration of apple addition were identified. With the results of the RDA, prediction models were constructed using the Orthogonal Projections to the Latent Structure (O-PLS) to differentiate integral juice from nectar and integral juice and nectar with and without the addition of apple. The constructed models were able to distinguish the types of drinks, with a Q2 equal to 0.706 for the integral juice and nectar model and Q2 of 0.681 for integral juice and nectar with and without added apple. Using partial least squares regression (PLS), an orange and apple content prediction model was built with good percentage prediction capacity with an R2 of 0.9989. Using the aforementioned tools, it was possible to develop models to support quality control and adulteration of orange integral juices and nectars.