Optimized Terahertz Hyperspectral Analysis in the Frequency- and Time-Domains

We optimize a fully automated method to differentiate unique sample components via hyperspectral imaging using terahertz time-domain spectroscopy (THz-TDS). Standard methods for THz hyperspectral image analysis rely on extracting amplitude data from the frequency-domain THz transmission spectra. We show benefits in analysis made possible by the simultaneous use of both amplitude and phase components of the THz transmission data. Incorporating full-field THz information improves the integrity of sample component differentiation with respect to sample composition. Furthermore, we present on the ability to achieve comparable sample component separation by performing analysis on the time-domain THz data. Forgoing the Fourier transformation reduces computational time and costs while maintaining high quality data interpretability. As a demonstration of methods, we automatically distinguish the structure of a split-ring resonator metamaterial sample.