Multivariate Optimization of Electrochemical Biosensors for the Determination of Compounds Related to Food Safety-A Review.

We summarize the application of multivariate optimization for the construction of electrochemical biosensors. The introduction provides an overview of electrochemical biosensing, which is classified into catalytic-based and affinity-based biosensors, and discusses the most recent published works in each category. We then explore the relevance of electrochemical biosensors for food safety analysis, taking into account analytes of different natures. Then, we describe the chemometrics tools used in the construction of electrochemical sensors/biosensors and provide examples from the literature. Finally, we carefully discuss the construction of electrochemical biosensors based on design of experiments, including the advantages, disadvantages, and future perspectives of using multivariate optimization in this field. The discussion section offers a comprehensive analysis of these topics.

Development of a voltammetric electronic tongue for the simultaneous determination of synthetic antioxidants in edible olive oils.

A voltammetric electronic tongue (E-tongue) is "a multisensor system, which consists of a number of low-selective sensors and uses advanced mathematical procedures for signal processing based on pattern recognition and/or data multivariate analysis such as artificial neural networks (ANNs), principal component analysis (PCA), among others". Thus, E-tongues in combination with chemometrics tools result in more accurate and selective analytical methods. In this work, we report results of a simple and reliable electroanalytical method to determine butyl hydroxyanisole (BHA), butyl hydroxytoluene (BHT) and propyl gallate (PG) in edible olive oils (EOO). Therefore, the square wave voltammetry (SWV) was used on platinum and carbon fiber disk ultramicroelectrodes (E-tongue configuration) combined with chemometrics tools to perform these studies. On the other hand, two data fusion strategies were used in order to combine electrochemical data obtained for each working electrode in the E-tongue: low-level data fusion (LLDF) and mid-level data fusion (MLDF). In addition, to reduce the dimensionality of the dataset in MLDF, the discrete wavelet transform (DWT) was used. Finally, to assert the predictive capability of the method for BHA, BHT, and PG determination in real samples, a recovery study for the antioxidants in EOO samples was performed, demonstrating the analytical accuracy of the proposed method. Moreover, from the comparison between the proposed electrochemical method with the AOAC reference method and others found in the literature in terms of the quality of the model (REP %) and the percent recovery assays (%) in different samples, our results were better than other reported previously for the simultaneous determination of BHA, BHT, and PG in real samples. Moreover, the percent recovery assays obtained with the proposed electrochemical method were in good agreement with those obtained by the chromatographic method.

Determination of kinetic parameters of the enzymatic reaction between soybean peroxidase and natural antioxidants using chemometric tools.

The oxidation of eugenol, isoeugenol and vanillin natural antioxidants catalyzed by the soybean peroxidase enzyme was studied using uv-vis spectroscopy. An experimental design was used to optimize the different variables. The multivariate curve resolution method was used to obtain the profiles of antioxidant absorbance's as a function of time due to uv-vis absorption bands of both antioxidants and the enzymatic reaction product/s show a strong overlap. From these results, apparent Michaelis-Menten constants as well as the kinetic parameters k1 and k3 involved in the catalytic cycle of peroxidases were calculated. The antioxidant apparent acidity constants were also determined at different pH's from uv-vis spectrophotometric measurements. Values of k1 were (0.6 ±â€¯0.1) × 105 M-1 s-1, (2.0 ±â€¯0.2) × 105 M-1 s-1 and (7.0 ±â€¯0.5) × 106 M-1 s-1 and k3 (4.0 ±â€¯0.2) × 105 M-1 s-1, (6.0 ±â€¯0.6) × 105 M-1 s-1 and (6.0 ±â€¯0.9) × 106 M-1 s-1 for eugenol, isoeugenol and vanillin, respectively.

Development of an electrochemical method to determine phenolic monoterpenes in essential oils.

A simple, rapid and non-expensive method is proposed to determine phenolic monoterpenes such as thymol and carvacrol in essential oils of thyme and oregano. The linear sweep voltammetry based on glassy carbon electrodes was the electrochemical technique used. Thymol and carvacrol have one main oxidation peak in non-aqueous media centered at about 1.3 V vs. Ag/AgCl. The electron transfer process is mainly diffusion controlled. The calibration plots generated using the commercial standards of thymol and carvacrol were used to estimate the total content in real samples. The calibration plots were linear in the concentration range from 8.5 × 10-5 to 1.3 × 10-3 mol L-1 and 7.9 × 10-5 to 1.2 × 10-3 mol L-1 for thymol and carvacrol, respectively. Results obtained with the electrochemical method are in good agreement with those of the official method (gas chromatography). In addition, the analytical procedure does not require previous preparation of the sample or modification of the electrode surface. The electrochemical technique used is very simple to apply. Under these conditions, the methodology proposed is a good, simple and fast option to perform a quality control of essential oils.

Square wave voltammetry with multivariate calibration tools for determination of eugenol, carvacrol and thymol in honey.

The simultaneous determination of eugenol (EU), thymol (Ty) and carvacrol (CA) in honey samples, employing square wave voltammetry (SWV) and chemometrics tools, is informed for the first time. For this purpose, a glassy carbon electrode (GCE) was used as working electrode. The operating conditions and influencing parameters (involving several chemical and instrumental parameters) were first optimized by cyclic voltammetry (CV). Thus, the effects of the scan rate, pH and analyte concentration on the electrochemical response of the above mentioned molecules were studied. The results show that the electrochemical responses of the three compounds are very similar and that the voltammetric traces present a high degree of overlap under all the experimental conditions used in this study. Therefore, two chemometric tools were tested to obtain the multivariate calibration model. One method was the partial least squares regression (PLS-1), which assumes a linear behaviour. The other nonlinear method was an artificial neural network (ANN). In this last case we used a supervised, feed-forward network with Levenberg-Marquardt back propagation training. From the accuracies and precisions analysis between nominal and estimated concentrations calculated by using both methods, it was inferred that the ANN method was a good model to quantify EU, Ty and CA in honey samples. Recovery percentages were between 87% and 104%, except for two samples whose values were 136% and 72%. The analytical methodology was simple, fast and accurate.

Quantitative determination of tocopherols in edible vegetable oils using electrochemical ultra-microsensors combined with chemometric tools.

We have developed an electroanalytical method to quantify different isomers of tocopherols in edible vegetable oils. The method uses the square wave voltammetry on a carbon fiber disk ultramicroelectrode in benzene/ethanol+0.1 mol L(-1)H2SO4. Because the oxidation peaks of these natural antioxidants show an important overlapping, we have used two chemometric tools to obtain the multivariate calibration model. One method was the multivariate curve resolution-alternating least square (MCR-ALS), which assumes a linear behavior, i.e., the total signal is the sum of individual signals of components, and another nonlinear method such as artificial neuronal networks (ANNs). From the accuracy and precision analysis between nominal and estimated concentrations by both methods, we could infer that the ANNs method was a good model to quantify tocopherols in edible oil samples. Recovery percentages were between 94% and 99%. In addition, we found a difference of 1.4-6.8% between the total content of tocopherols in edible oil samples and the vitamin E content declared by the manufacturers.

Development of an amperometric biosensor based on peroxidases to quantify citrinin in rice samples.

An amperometric biosensor based on horseradish peroxidase (EC1.11.1.7,H2O2-oxide-reductases) to determine the content of citrinin mycotoxin in rice samples is proposed by the first time. The method uses carbon paste electrodes filled up with multi-walled carbon nanotubes embedded in a mineral oil, horseradish peroxidase, and ferrocene as a redox mediator. The biosensor is covered externally with a dialysis membrane, which is fixed to the body side of the electrode with a Teflon laboratory film, and an O-ring. The reproducibility and the repeatability were of 7.0% and 3.0%, respectively, showing a very good biosensor performance. The calibration curve was linear in a concentration range from 1 to 11.6nM. The limits of detection and quantification were 0.25nM and 0.75nM, respectively. For comparison, the citrinin content in rice samples was also determined by fluorimetric measurements. A very good correlation was obtained between the electrochemical and spectrophotometric methods.

Electrochemical oxidation of butein at glassy carbon electrodes.

The electrochemical oxidation of flavonoid butein is studied at glassy carbon electrodes in phosphate and citrate buffer solutions of different pH values, and 1M perchloric acid aqueous solutions by cyclic and square wave voltammetries. The oxidation peak corresponds to the 2e(-), 2H(+) oxidation of the 3,4-dihydroxy group in B ring of butein, given the corresponding quinone species. The overall electrode process shows a quasi-reversible behavior and an adsorption/diffusion mixed control at high butein bulk concentrations. At low butein concentrations, the electrode process shows mainly an adsorption control. Butein surface concentration values were obtained from the charge associated with the adsorbed butein oxidation peaks, which are in agreement with those values expected for the formation of a monolayer of adsorbate in the concentration range from 1 to 5µM. Square wave voltammetry was used to perform a full thermodynamic and kinetics characterization of the butein surface redox couple. Therefore, from the combination of the "quasi-reversible maximum" and the "splitting of the net square wave voltammetric peak" methods, values of (0.386±0.003) V, (0.46±0.04), and 2.7×10(2)s(-1) were calculated for the formal potential, the anodic transfer coefficient, and the formal rate constant, respectively, of the butein overall surface redox process in pH4.00 citrate buffer solutions. These results will be then used to study the interaction of butein, and other flavonoids with the deoxyribonucleic acid, in order to better understand the potential therapeutic applications of these compounds.