Data model for the elimination of matrix effects in enzyme-based flow-injection systems.

Abstract

This contribution presents a new conceptional enzyme-based flow injection analysis (FIA) system for the process and quality control of food processing and biotechnological systems. It provides the determination of different analytes in distinct process media on the base of a common experimental set-up. In contrast to known comparable systems, analysis is performed without the commonly used sample preparation and dilution steps. Instead, the adaptation to the necessary measurement range is realized by optimization of intrinsic system parameters. The central principle of the work presented is the elimination of occurring interferences by the heterogeneous matrix of the process sample. Based on a particular injection mode, the application of dehydrogenases as indicator enzymes and a specially developed data model using cognitive methods, cross sensitivities of the detector as well as disturbed reaction rates of the enzymes could be almost completely compensated. Two applications are presented, the analysis of ethanol in non-alcoholic beer and the online determination of D-/L-lactate during a lactic acid fermentation, which reveal the advantage of the developed system.