Terahertz waves occupy a niche between microwaves and infrared light. They penetrate clothing and plastic but do not ionize tissue, so they are being tested for security screening and medical diagnostics. Classical terahertz time-domain spectroscopy measures the shape of a short pulse reflected from an object. However, any movement of the object shifts the pulse in time, and it jumps out of the observation window. Existing fast methods, such as asynchronous optical sampling, scan the entire time range, which is slow and does not allow tracking a target. The work of the Japanese researchers fills this gap by offering a system that keeps the window locked onto the moving pulse.
The setup is based on two terahertz lasers with programmable frequency combs. The pulse repetition frequencies are tightly synchronized, but one laser can be phase-shifted to move the time window. Instead of scanning the whole range, the system rapidly oscillates the delay within a narrow window of 61 picoseconds, acquiring up to 154 full measurements per second. The position of the terahertz pulse peak within the window is measured in real time. If the peak shifts due to sample motion, feedback corrects the phase of the second laser, bringing the peak back to the center of the window. This approach allowed tracking a mirror moving at 4 centimeters per second over a distance of 12 centimeters.
The system held the terahertz peak within the 61-picosecond window while the target moved by 12 centimeters, which is 26 times the window width. The precision of peak position determination was 10 femtoseconds per single measurement, improving to 1.5 femtoseconds when averaged over 0.44 seconds, corresponding to a displacement of 225 nanometers. The maximum tracking speed was estimated at 10.8 centimeters per second. In a model experiment with a gold mirror on a vibration motor, they recovered oscillations with amplitudes of 2.1 millimeters (simulating breathing) and 0.21 millimeters (simulating heartbeat) on top of a linear displacement of up to 10 centimeters. The frequencies of 0.25 and 1.2 hertz were determined from the spectrum of the recovered signal.
This work stands apart from neighboring studies: they develop terahertz spectroscopy for static objects or slow processes, whereas here active tracking of a moving target with feedback is implemented for the first time. This is not just faster measurements but a paradigm shift: the observation window becomes mobile. The approach could be transferred to other time-of-flight spectroscopy methods where sample motion was considered an insurmountable nuisance.
The next step is to increase tracking speed by optimizing feedback parameters and reducing loop latency. Extending to two-dimensional scanning is also promising, so as to follow an object moving in a plane. In the longer term, the system could be combined with machine learning algorithms to automatically extract breathing and heartbeat from complex body motion.
The method could be used for contactless vital-sign monitoring, for example to observe patients in intensive care units. In industry, it could be applied to measure coating thickness or inspect product quality on a conveyor, where the distance to the sensor varies.
The researchers plan to increase tracking speed by reducing the delay in the feedback loop. They also intend to test the system on real biological tissues, not just a gold mirror.
In terahertz spectroscopy, measuring moving objects has long been a challenge, whether it is a person undergoing security screening or a product on a conveyor. This work directly addresses that problem by introducing feedback based on the peak position. It connects to the broader task of motion compensation in non-destructive testing and medical diagnostics.
This section is addressed to the author of the paper and their colleagues. Neighbouring works were found by meaning-based search across our archive, not by word overlap. We do not review: we show the field around the work.
What the work does
The work proposes a terahertz system that actively keeps the measurement window locked onto a moving object using a pair of frequency-programmable lasers and feedback. This made it possible to track displacements of up to 12 cm with an accuracy of 225 nm and, in a model experiment, to pick out oscillations that mimic breathing and heartbeat against a background of large motions.
Where its strength is
The strong point is the implementation of a moving observation window, which removes a fundamental limitation of time-domain terahertz spectroscopy: its sensitivity to sample displacement. This opens the door to real-time measurements on moving objects without sacrificing accuracy.
Where it could go next
Worth trying
Why this may matter
The result could matter for advancing contactless monitoring of vital signs in real-world conditions, and also for non-destructive testing in industry, where objects are often in motion.
Nearby works in our archive
The archive is sparse around this work: the closest review is only 0.58, against a usual 0.64. It looks like few have drilled nearby.
The precision of 225 nanometers is comparable to the size of an influenza virus. The system can detect a mirror displacement smaller than the wavelength of visible light.