Terahertz metamaterial liquid detector optimized by deep learning

This study proposes a label-free ethanol liquid detection method based on the absorption peak shift of a metamaterial terahertz detector. The designed metamaterial liquid detector consists of a five-layer structure, including interdigitated copper metal wires, a polyimide dielectric layer, an open square ring with nested double-word-shaped VO₂, and a copper metal plate. By optimizing structural parameters using a deep neural network (DNN), high absorption rates and high shifts are achieved under different operating conditions. CST simulation results show that the absorption peak of the detector shifts reward as the droplet volume increases and shifts blueward as the ethanol concentration increases, and the absorption peak spacing narrows under high-temperature conditions. Additionally, through magnetic field strength and surface current analysis, the working principle of the detector is revealed: at room temperature, VO₂ barely participates in absorptive wave regulation, but as the temperature rises, the conductivity of VO₂ increases, providing a dual guarantee for the accuracy of ethanol detection. This study not only provides an efficient and precise new method for ethanol liquid detection but also holds broad application prospects in fields such as food safety, environmental monitoring, and biomedicine.