NADES-modified voltammetric sensors and information fusion for detection of honey heat alteration

Abstract

Natural Deep Eutectic Solvents (NADES) are considered the solvents of the moment. Indeed, it has been demonstrated that NADES could improve the electrochemical detection of some target analytes. In this work we present a smart approach taking advantage of different NADES features and using them to modify the electrolyte to obtain an array of sensors. Thus four NADES electrochemical sensors were applied for honey heat alteration detection. In this sense, local untreated honey samples were analyzed prior and after heat treatment (80 °C, 4 hs). At first, the data provided by each honey (unmodified and heated) were separately analyzed by partial least squares-discriminant analysis (PLS-DA). Then, to test whether data fusion strategies could detect heat alteration, NADES-voltammetric sensors data blocks were jointly analyzed merging two, three or four by applying sequential and orthogonalized partial least squares coupled with linear discriminant analysis (SO-PLS-LDA). The overall accuracy reached in the prediction stage ranged from 0.76 to 0.86 for individual models, meanwhile the model fusing the four sensors attained an optimal performance with a value of 0.97.