Elżbieta’s paper “Low-cost, syringe based ion-selective electrodes for the evaluation of potassium in food products and pharmaceutical” has recently been accepted in Electrochimica Acta. The article deals with the quantification of potassium in real samples using an electronic tongue based on low-cost syringe electrodes.
Ion-selective electrodes (ISE) are well known potentiometric sensors that allow rapid, non-invasive and simple disease monitoring and diagnosis. In this work, we developed low-cost ion-selective electrodes from disposable syringes equipped with plasticized PVC membranes. The electrodes produced were selective for potassium cations, as valinomycin was used as an ionophore in the ion-selective membranes. Although simple measurements in laboratory samples with ion-selective electrodes seem to be quite easy, quantification of potassium in real samples is more challenging. Real samples, such as beetroot soup, tomato-based foods or dried fruit, are more concentrated and consequently the ionic strength of the solutions is higher. The results were reproducible but subject to a relative error of up to 76%.
Fig.1 Fabrication of the syringe-based ion-selective electrodes: A. prepare a 2 mL syringe, B. remove the piston, C. pipette 50 µL of the membrane solution directly into the tip of the syringe, D. let the membrane dry, E. add internal electrolyte, remove any bubbles and close the syringe with a stopper equipped with an Ag/AgCl wire
To overcome this limitation, we constructed an electronic tongue based on the low-cost syringe electrodes. The arrays were composed of sets of ion-selective electrodes selective for different ions (K+, NH4+) and tested with a series of multivariate algorithms. When measuring more complex samples such as food or pharmaceutical supplements with variable ionic strength, recalibration is required. However, data from an ensemble random forest model showed that adding a single type of electrode to the array eliminated the need for calibration. Compared to quantification of potassium using a standard calibration curve, the root mean square error of prediction was almost six times lower.
Multivariate algorithms show that the composition of the e-tongue array is critical. PCA analysis showed that at least two different sensor types are required to achieve proper clustering of the sample arrays (e.g. NH4+; K+ Val). Although valinomycin is the most selective ionophore for potassium, it is not necessary to obtain the best prediction with the standard calibration curve. A dibenzo-18-crown-6-based electrode (also selective for potassium) can be used instead of the valinomycin-based electrode, allowing for affordable and more effective sensors for commercial purposes.
Research was funded through the NCN Sonata 2020/39/D/ST4/02256 project.