THz Multidimensional Spectroscopy Sensing: A Novel Approach for High-Precision Metal Slag Recognition Using Packed Gratings

The accurate differentiation of slag types is pivotal for the metal market, given its implications for safety, environmental impact, and economic recovery of precious metals. However, conventional analytical techniques, such as XRD and scanning electron microscope (SEM)-EDS, are fraught with high costs, lengthy procedures, and potential dangers. In this investigation, we unveil an innovative method that harnesses the power of multidimensional spectroscopy, the linear discriminant analysis (LDA) algorithm, and THz time-domain spectroscopy (THz-TDS), obviating these challenges. Through a unique "mask filling" fabrication technique, we create devices laden with metallic particles. These devices, when interrogated with the aforementioned spectroscopic and analytical tools, yield a wealth of information, due to their integrated external controls. By distilling the essence of the spectral data via LDA, we craft a sophisticated multidimensional dataset that elevates the precision of slag classification from a lackluster 75.0% to an impressive 95.8%. Our strategy stands as a beacon of progress, offering a safer, swifter, and more economically viable pathway to the characterization of metal slags.