High-speed and single-shot terahertz spectroscopy using a wavelength-tunable source [Invited]

Area of Science:

• Terahertz (THz) spectroscopy
• Non-linear optics
• Spectroscopic instrumentation

Background:

• Practical applications of terahertz (THz) waves require high-speed, real-time spectroscopic techniques.
• Existing THz time-domain spectroscopy (THz-TDS) has limitations in speed and tunability.
• Fast spectroscopy using wavelength-tunable THz sources is an emerging research area.

Purpose of the Study:

• To overview wavelength-tunable THz sources.
• To introduce advancements in high-speed, single-shot THz spectroscopy using parametric sources.
• To present a real-time sample identification method using machine learning.

Main Methods:

• Overview of wavelength-tunable THz sources, focusing on parametric sources.
• Exploration of three high-speed spectroscopy approaches: wavelength sweeping/switching, simultaneous multi-wavelength generation, and quasi-single-shot spectroscopy.
• Development and application of a machine learning-based real-time spectral identification method.

Main Results:

• Demonstration of high-speed and single-shot THz spectroscopy capabilities using parametric sources.
• Successful implementation of a machine learning algorithm for real-time spectral identification and visualization.
• Achieved high dynamic range in spectral identification and visualization of reagent distributions.

Conclusions:

• The developed high-speed THz spectroscopy methods significantly enhance real-time analysis capabilities.
• Machine learning integration enables rapid identification and visualization of sample properties.
• This methodology is poised to broaden the application scope of THz spectroscopy into challenging domains.