Artificial intelligence-motivated in-situ imaging for visualization investigation of submicron particles deposition in electric-flow coupled fields

Abstract

This study delves into the intricate deposition dynamics of submicron particles within electric-flow coupled fields, underscoring the unique challenges posed by their minuscule size, aggregation tendencies, and biological reactivity. Employing an operando investigation system that synergizes microfluidic technology with advanced micro-visualization techniques within a lab-on-a-chip framework enables a meticulous examination of the dynamic deposition phenomena. The incorporation of object detection and deep learning methodologies in image processing streamlines the automatic identification and swift extraction of crucial data, effectively tackling the complexities associated with capturing and mitigating these hazardous particles. Combined with the analysis of the growth behavior of particle chain under different applied voltages, it established that a linear relationship exists between the applied voltage and θ. And there is a negative correlation between the average particle chain length and electric field strength at the collection electrode surface (4.2×10⁵ to 1.6×10⁶ V·m^–1). The morphology of the deposited particle agglomerate at different electric field strengths is proposed: dendritic agglomerate, long chain agglomerate, and short chain agglomerate.