Monosaccharide Sensing Based on Multivariate Analysis of Voltammetric Data Acquired From a Pt:Ru Electrode Array

摘要

Predictive models for concentration of mixed monosaccharide solutions were developed based on combinatorial electrochemistry and chemometric techniques. The columns on a 10×10 array of Pt wires were electrodeposited with 10 different Pt:Ru alloys. Cyclic voltammograms were performed in 1M solutions of glucose, fructose, and galactose. Principal component analysis was applied to the resulting data sets; a scores plot allowed classification of the pure solutions. Fifteen solutions containing the three sugars in concentrations ranging from 10-1000mM were used to train partial least squares regression models. Twelve independent test solutions were also prepared in similar concentration ranges. For these test samples, root-mean-squared-error-of-predictions (RMSEPs) of 142mM and 120mM were obtained for glucose and galactose. The RMSEP for fructose between 10-500mM was 128mM, but nonlinearities caused the model to fail at higher concentrations. These results demonstrate that with only a few electrode variants it is possible to differentiate monosaccharides in a semi-quantitative fashion.