Geometrically scalable narrowband perfect-reflection metasurfaces for mid-infrared molecular fingerprint barcode sensing

Abstract Conventional mid-infrared (MIR) sensing and spectroscopic detection systems suffer from bulky footprints, fixed operating wavelengths, and difficult on-chip integration. We propose a geometrically scalable all-dielectric metasurface based on hollow silicon (Si) nanosquare pillars, which achieves narrowband perfect reflection over a broad MIR range through structural scaling, and demonstrate its application in molecular fingerprint spectroscopic scanning and visualized barcode detection. By scaling the structure with a dimensionless scaling factor S from 1.0 to 1.55, the reflection peak is continuously redshifted from 1900cm -1 to 1350 cm -1 while the reflectivity remains above 99% and the quality factor (Q-factor) increases significantly with the feature size. Multipole decomposition reveals that the resonance is dominated by the magnetic dipole (MD) mode with highly localized electromagnetic fields. Based on a Lorentz dispersion model, we extract the characteristic absorption spectra of molecules including protein, 2,4-dinitrotoluene (2,4-DNT), and polyethylene (PE), and convert the absorption signals into molecular barcodes, enabling intuitive, label-free, and rapid trace detection. In single-molecule detection, each analyte exhibits a unique and distinguishable barcode; in multi-component mixtures, however, weak-absorption components may be masked by strong absorbers. The proposed metasurface offers polarization insensitivity, low loss, broad spectral scalability, and excellent integrability, providing a novel solution for MIR spectrometers, portable biochemical sensing, and environmental monitoring.