Depolarization characteristics of scattered light from metal surfaces based on near-infrared polarized supercontinuum

Traditional polarization detection techniques are constrained by the limited dimensionality of single-wavelength information or their dependence on sunlight in passive remote sensing. To overcome these limitations, this study first employed the linearly polarized supercontinuum (SC) in the active metal surfaces polarization detection. By developing a polarized SC source with a polarization extinction ratio (PER) exceeding 25 dB in the near-infrared wavelength range of 1020-1200 nm and employing a spectral subtraction calculation method, we achieved efficient full-band measurement of the PER. The results indicated that the PER decays exponentially with increasing surface roughness when the roughness is much smaller than the incident wavelength, while an anomalous enhancement of PER occurs in specific bands when the roughness approaches the wavelength scale. It means that the SC has higher accuracy in measuring sample roughness compared to the single-wavelength laser. Furthermore, the measured PER exhibits a non-monotonic dependence on the incident angle, initially increasing to a peak value and then decreasing. Notably, the angle corresponding to the maximum PER shifts toward smaller values with increasing surface roughness. A physical model based on Kirchhoff approximation and microfacet theory was established, providing a consistent explanation for the experimental observations. The study also indicates that the wavelength dependence of the complex refractive index enables the use of linearly polarized SC for metal samples identification, providing both the "spectral fingerprint" and "polarization fingerprint" of the sample in a single step, significantly improving the differentiation ability between metals compared with the single-wavelength polarized laser. This work offers what believe to be a new paradigm for polarization remote sensing and industrial surface inspection.