Non-contact multi-channel terahertz refractive index detecting via focused orbital angular momentum

Abstract

Terahertz refractive index (RI) sensor is an important tool for characterization of material properties. However, traditional immersion RI sensors suffer from issues of liquid contamination and low detection efficiency. Therefore, this study designs a non-contact multi-channel terahertz RI sensor based on metasurface which can produce vortex beam with focused orbital angular momentum (FOAM). Computing Spatially Weighted Variance (SWV) of the contours and intensities of the images resulting from different RI demonstrates that the RI value correlates uniquely with the weighted variance. Specifically, in the near-field metasurface (NF metasurface), by employing 3-bit phase binary (PB) encoding and convolution operations on vanadium dioxide (VO₂) metasurfaces, we realized a high-purity FOAM with a focal length of f = 3500 μm. After fixing the sample under test at f = 3500 μm and analyzing the near-field FOAM amplitude images under various RI conditions using the SWV method, the RI detection sensitivity of the NF metasurface was calculated to be 15,775/RIU. To enhance detection efficiency and meet the requirements for far-field detection, we proposed a sensor capable of detecting multiple samples in the far-field. In the far-field metasurface (FF metasurface), when linearly polarized light is incident on this metasurface, the sensitivity of the FOAM to RI detection produced by this metasurface under the left circularly polarized (LCP) and right circularly polarized (RCP) components are 10,554/RIU and 13,292/RIU, respectively. The minimum change of the RI that can be detected by the near-field and far field sensors reaches 10⁻⁴ RIU. When VO₂ transitions to its dielectric state, the metasurface switches to specular reflection, thereby endowing the sensor with switching functionality for RI detection. This approach overcomes issues of liquid-contaminated surfaces and enables simultaneous detection of multiple substances, offering broad application prospects across various sensing scenarios.