Determinação rápida e de baixo custo do potencial antioxidante em cascas de castanheira utilizando imagens digitais e modelo de regressão multivariada

Abstract

Brazil has a vast biodiversity and among the plants found in the Brazilian territory is the Bertholletia excelsa, popularly known as Brazil nut, well known for its almond and Brazil nut. Specialized metabolites are essential compounds for plants, which act as a defense to biotic and abiotic agents. They collaborate in the perpetuation of the species and its adaptation to the environment and climate. We can highlight phenolic compounds, terpenes and alkaloids as specialized metabolites. These compounds may have antioxidant potential. Antioxidants are compounds (endogenous or exogenous) that can neutralize free radicals, which are endogenous molecules or molecules generated by external factors with unpaired electrons, which can cause cell damage. The imbalance between free radicals and antioxidants can generate oxidative stress, responsible for certain degenerative diseases such as Alzheimer's, Parkinson's and even cancer. In order to extract compounds with antioxidant potential, it is important to carry out an experimental design to optimize the extraction of these bioactive compounds. Thus, tests such as total phenolic compounds (CFT), ABTS, DPPH and FRAP are useful to assess the antioxidant potential in plant extracts. To assist in this optimization and quantification, statistical tools can be used, in order to observe the factors that influence the extraction such as time, temperature and solvent concentration and lead us in a safe way to obtain the best extraction conditions. In the search for ease in the daily laboratory processes and the minimization of analysis costs, chemometric models were developed in order to predict the characteristics of interest. In this sense, in this work, models using digital images were used to evaluate the antioxidant potential in chestnut bark extracts. The central composite rotational design (CCR) used in the study suggested as the best condition for CFT extraction with antioxidant activity the one close to the central point, whose ethanol concentration was 50%. The results indicated an efficient modeling for CFT, FRAP and DPPH, while the model for the ABTS test showed that it needed other tests to search for a more efficient model.