Improved terahertz time-domain spectroscopy via the extended Kalman filter

To implement terahertz time-domain spectroscopy (THz-TDS) systems that have short integration time and maintain performance, the extended Kalman filter (EKF) has been used. Performance of EKF with THz-TDS on vapor samples and solid samples is untested. Previous EKF methods were not adaptive for use on multiple THz-TDS systems. We provide an adaptive EKF method that can work on any THz setup, to improve maximum measurable absorption coefficient. Our adaptive EKF method responds to THz pulse properties. The adaptive EKF is exhaustively demonstrated for six different time-constant data sets spanning 1 ms to 300 ms for THz-TDS on vapor with electro-optic THz detection. The acquisition time saved is quantified and shown to significantly reduce acquisition time from 2.58 ms to 1 ms. This improvement of 2.58 times will have a significant effect to move spectral imaging from taking hours to taking minutes. Vapor absorption lines are maintained. We show the adaptive EKF maintain correct spectral information when compared to the high-resolution transmission molecular absorption (HITRAN) database. The adaptive EKF is also demonstrated for THz-TDS on vapor with photoconductive THz detection, thus demonstrating adaptability, and is applied for THz-TDS of a solid sample of polyethylene terephthalate glycol.