Search

Highlights • Chemometric techniques employed for optimization in food analysis are shown. • Description, advantages, and drawbacks of the experimental designs are presented. • A comparison between the response surface methodologies for food analysis has been performed. Abstract This work presents a critical review of multivariate techniques employed for optimization of methods developed in food analysis. A comparison between the response surface methodologies has been performed, it evidencing advantages and drawbacks of these. Applications of the main chemometric tools (central composite and Box Behnken designs and Doehlert matrix) often utilized for optimization of sample preparation procedures and also instrumental conditions of analytical techniques for determination of organic and inorganic species in food samples are shown. Also, a brief discussion on the use of multiple responses and robustness test in food analysis has been presented. [ABSTRACT FROM AUTHOR]

Highlights: [•] A CE-DAD method to separate 17 phenolic compounds from olive oil was optimised. [•] Resolution, runtime and migration time RSD values were evaluated as responses. [•] This was the first optimisation study to evaluate 37 responses simultaneously. [•] Derringer’s desirability function was used to simultaneously optimise the responses. [•] The method was validated and applied to 15 samples found in Brazilian supermarkets. [ABSTRACT FROM AUTHOR]

• Copper determination by digital images analysis using smartphone and PhotoMetrix® application. • Different region of interest (ROI) sizes, cameras and ways of processing data were evaluated. • Accurate results were obtained using low-cost apparatus and a small reagent volumes. • The multivariate calibration by Partial Least Squares presented better results. Analysis of digital images by smartphone was used for copper quantification in sugarcane spirit (cachaça) samples through the formation of blue complex between copper and cuprizone. An experimental design was carried out to evaluate the best complexation reaction conditions. Moreover, different image acquisition procedures (external camera coupled to a smartphone or the smartphone camera) with different regions of interest sizes, distances in image acquisition, and concentration ranges of the calibration curve and the influence of processing the curve in univariate and multivariate modes, by PLS, were evaluated. The results obtained in three real samples and two spikes were compared with those of UV–Vis spectrophotometry, used as a reference method, and they show the potential of the proposed method for the accurate determination of copper. When compared to traditional techniques, the proposed method has the advantages of portability and low cost in addition to requiring a smaller amount of reagents. [ABSTRACT FROM AUTHOR]