Polymer bead size revealed via neural network analysis of single-entity electrochemical data

Single-entity electrochemistry methods for detecting polymer microbeads offer a promising approach to analyzing microplastics. However, conventional methods for determining microparticle size face challenges due to non-uniform current distribution across the surface of a sensing disk microelectrode. In this study, we demonstrate the utility of neural network (NN) analysis for extracting the size information from single-entity electrochemical data (current steps). We developed fully connected regression NN models capable of predicting microparticle radii based on experimental parameters and current-time data. Once trained, the models provide near-real-time predictions with good accuracy for microparticles of the same size, as well as the average size of two different-sized microparticles in solution. Potential future applications include analyzing various bioparticles, such as viruses and bacteria of different sizes and shapes.