Abstract
Solid contact ion-selective electrodes (SCISEs) offer many benefits over traditional liquid contact ion-selective electrodes. Their small size made them the default choice in many clinical analysis tools. Reproducibility of their production is crucial in achieving calibration-free sensors. Electrochemical impedance spectroscopy (EIS) is a versatile technique that can provide valuable information on many physico-chemical parameters of examined SCISEs and it can give results under 1 min. Discerning different phenomena that govern the EIS spectrum require the theoretical understanding of the processes (e.g., diffusion, heterogeneous kinetics etc.) that determine the time-dependent response of SCISEs. EIS simulations of SCISEs with Nernst-Planck-Poisson finite element method are applied to describe the experimental response of SCISEs. The numerical simulations are used to train a black-box supervised learning algorithm—a deep feedforward neural network—and a white-box symbolic regression algorithm to learn the underlying model of EIS spectra of SCISEs. The neural networks are used to significantly speed up the solution of the inverse problem of obtaining physico-chemical parameters from experimental data.
Funder
Nemzeti Kutatási Fejlesztési és Innovációs Hivatal
Emberi Eroforrások Minisztériuma
Publisher
The Electrochemical Society
Subject
Materials Chemistry,Electrochemistry,Surfaces, Coatings and Films,Condensed Matter Physics,Renewable Energy, Sustainability and the Environment,Electronic, Optical and Magnetic Materials
Cited by
1 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献