Double-Coffee Ring Nanoplasmonic Effects with Convolutional Neural Learning for Sars-Cov-2 Detection

Abstract

We develop a sensing method based on the double-coffee ring phenomenon for the first time using localized surface plasmon resonance (LSPR) effect in gold nanoparticles (GNPs) to detect SARS-CoV-2 Nucleocapsid proteins with high sensitivity. Testing images are further analyzed via the convolutional neural learning for enhanced accuracy. The circular-shape hydrophilic PTFE porous membrane with a hydrophobic ring barrier is utilized as the sensing platform. When the virus proteins are interacting with antibody coated GNPs solution on the platform, a double-coffee ring image is observed and the convolutional neural network helps the differentiation for the first small protein-GNPs ring at the center and a second non-specific ring at the hydrophobic barrier. We use this double-coffee ring to detect viral infection and quantify the concentration of COVID-19 viruses in 5 ng/ml (LOD), similar to Abbott BinaxNOW® testing kit, to 1000 ng/ml. As such this detection scheme could open up a new class of biomolecular research in the field of micro/nano fluidics.