Data Augmentation Empowering Simulation-Driven Design for Nanopore Based Single-Particle Sensing of Nanoplastics

Nanopore-based single-particle electrochemical sensing offers the possibility to determine the size, charge, type, and shape of nanoplastics simultaneously in a label-free manner. However, the rational simulation-driven design of nanopore sensing interfaces is hampered by the high computational cost of generating sufficient current traces for single-particle identification and condition optimization. To solve this issue, two data-augmentation strategies are proposed to accelerate simulation-guided nanopore design. The first one is integrated with finite-element simulations to rapidly produce raw single-particle current traces for nanoplastics of prescribed size and charge, enabling multi-feature classification that mirrors experimental data analysis. The second interpolation-based augmentation is adopted to explore the resolving power of a given nanopore. Two performance indicators-the resolution of nanoplastics' size and charge discrimination-are defined and applied to optimize the conical nanopore angle and electrolyte concentration. The resulting system (150 nm tip diameter, 20° cone angle, and 0.005 M KCl) resolves minimum differences of 0.732 nm in size and 1.885 mC m-2 in charge at >90% accuracy, for nanoplastics of 100-110 nm and-0.005 to-0.0025 C m-2. Crucially, the obtained nanopore design strategy is validated experimentally, highlighting the promise of data augmentation empowered simulation-driven nanopore engineering for single-particle nanoplastics sensing.