🔬 optics

Cost-effective time-stretch terahertz recorders using 1550 nm probes

This paper demonstrates that using 1550 nm probes with standard components enables cost-effective, single-shot time-stretch terahertz recording at high acquisition rates using low-bandwidth (1–3 GHz) oscilloscopes, thereby overcoming the traditional reliance on expensive high-speed electronics.

Original authors: Christelle Hanoun, Eléonore Roussel, Christophe Szwaj, Clément Evain, Marc Le Parquier, Jean-Blaise Brubach, Nicolas Hubert, Marie Labat, Pascale Roy, Marie-Agnès Tordeux, Serge Bielawski

Published 2026-04-14

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Christelle Hanoun, Eléonore Roussel, Christophe Szwaj, Clément Evain, Marc Le Parquier, Jean-Blaise Brubach, Nicolas Hubert, Marie Labat, Pascale Roy, Marie-Agnès Tordeux, Serge Bielawski

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to take a photograph of a hummingbird's wings. The wings move so fast that a normal camera just sees a blur. To capture the motion clearly, you need a camera with an incredibly fast shutter speed. But in the world of science, capturing "fast" things (like Terahertz waves, which are invisible light waves used in security scanners and medical imaging) is even harder. They move at speeds that make a hummingbird look like a snail.

For years, scientists have used a clever trick called "Time-Stretch" to solve this. Think of it like this: instead of trying to photograph the hummingbird directly, you take a photo of its shadow being cast on a wall that is moving away from you very fast. As the shadow moves away, it gets stretched out, turning a split-second blur into a long, slow-moving image that a regular camera can easily capture.

This paper is about a team of scientists who found a way to make this "stretching" trick much cheaper and much more powerful by changing the color of the light they use.

The Problem: The Expensive "Super-Camera"

Previously, to record these super-fast Terahertz waves, scientists had to use a very specific type of laser (usually red or infrared) and a "super-camera" (a high-speed oscilloscope) that cost a fortune.

The Analogy: Imagine trying to watch a movie. To see the action clearly, you needed a projector that cost $100,000. If you wanted to record a whole movie (thousands of frames), you needed an even more expensive projector. This made the technology too expensive for most labs to use.

The Solution: The "1550 nm" Switch

The researchers decided to switch their laser from the usual colors (800 nm or 1030 nm) to 1550 nm.

Why 1550 nm? This is the standard color used in fiber optic internet cables. Because it's the standard for the internet, the equipment (lasers, fibers, detectors) is mass-produced, cheap, and easy to find.
The Magic: By using this "internet color," they found that the "stretching" effect worked much better. It's like switching from a cheap rubber band to a super-elastic bungee cord. You can stretch the signal much further without it snapping.

How It Works (The Simple Version)

The Probe: They fire a laser pulse (the "camera flash") at a crystal.
The Interaction: A Terahertz wave (the fast event) hits the crystal and slightly changes the laser's shape.
The Stretch: The laser pulse is sent through a long, special fiber optic cable. This cable acts like a slow-motion machine, stretching the tiny, fast pulse into a long, slow wave.
The Recording: Because the wave is now stretched out, they can record it using a standard, relatively cheap oscilloscope (like the ones used in car repair shops or basic electronics labs) instead of a million-dollar super-computer.

The Big Wins

The paper highlights two major victories:

Cost Reduction:
- Old Way: Needed a $100,000+ camera to record the data.
- New Way: They used a standard $15,000 camera (an oscilloscope).
- Analogy: They went from needing a Formula 1 race car to do a grocery run, to using a reliable, affordable family sedan that does the job just as well.
More Data, Better Detail:
- Because the signal is stretched so much more effectively, they can capture way more details (samples) of the event.
- Analogy: Imagine recording a song. The old method gave you 100 pixels of audio data. The new method gives you 500 pixels. You can hear the music much more clearly and see the tiny nuances in the sound.

The Real-World Test

The team didn't just build this in a quiet lab; they took it to a massive particle accelerator (the SOLEIL synchrotron in France). They successfully recorded "giant" Terahertz pulses generated by electrons zooming around a ring.

They captured the pulses in "single-shot" mode, meaning they caught the event exactly as it happened, without needing to repeat it thousands of times to build a picture.
They recorded over 3,000 pulses in a row, watching how the energy burst and changed over time.

Why This Matters

This discovery is like finding a way to build a high-speed train using bicycle parts. It makes advanced scientific tools accessible to more people.

For Accelerators: It helps scientists monitor particle beams more efficiently.
For Medicine and Security: It could lead to cheaper, faster scanners that can see through clothes or detect diseases without radiation.
For Everyone: It proves that you don't always need the most expensive, cutting-edge gear to do world-class science; sometimes, you just need to use the right "off-the-shelf" parts in a clever way.

In short, these scientists took a complex, expensive technology and made it cheaper, simpler, and more powerful by simply changing the color of the light they used to the same color that powers the internet.

1. Problem Statement

Time-stretch electro-optic detection is a powerful technique for recording ultrafast Terahertz (THz) waveforms in a single shot at multi-Megahertz acquisition rates. This capability is critical for accelerator physics (e.g., monitoring electron bunches) and studying non-reproducible dynamics in spectroscopy.

However, the widespread adoption of this technology has been severely limited by two main factors:

High Cost: Traditional implementations require extremely high-bandwidth (and expensive) oscilloscopes or Analog-to-Digital Converters (ADCs) to read out the stretched signals.
Limited Sample Count: The number of data points (samples) that can be recorded per THz waveform is often constrained by the bandwidth of the readout electronics and the stretch factor achievable with standard laser wavelengths (typically 800 nm or 1030 nm).

The authors aim to overcome these bottlenecks by developing a cost-effective design that reduces the required readout bandwidth while simultaneously increasing the effective number of recorded samples.

2. Methodology

The research introduces a novel experimental design that shifts the probe laser wavelength from the conventional 1030 nm (Ytterbium-based) to 1550 nm (Erbium-based).

Experimental Setup:

Laser Source: A femtosecond Erbium fiber laser (Menlo C-Fiber) operating at 1550 nm with a 30 nm bandwidth (after amplification) and a repetition rate of 6.77 MHz.
Pulse Chirping: The laser pulses are first chirped using 40 meters of normal dispersion fiber.
THz Interaction: The chirped pulses interact with THz pulses (0.85 MHz repetition rate) emitted from the SOLEIL synchrotron radiation facility. The interaction occurs via the Pockels effect in a 3 mm-thick Gallium Arsenide (GaAs) crystal.
Time Stretching: The modulated pulses undergo a second, massive stretching phase using a low-cost commercial telecommunication Dispersion Compensation Module (DCM) based on Dispersion Compensating Fiber (DCF). This module provides a dispersion of 1360 ps/nm.
Detection: The stretched signals are detected using a balanced InGaAs photoreceiver (2.5 GHz bandwidth) and recorded using standard, low-cost electronics: either a 1 GHz oscilloscope or a 3 GHz ADC board.

Key Strategy:
The authors leverage the fact that the 1550 nm wavelength, combined with standard telecommunication fiber components (DCF), allows for significantly higher dispersion and longer stretch durations compared to 1030 nm systems, without needing complex nonlinear spectral broadening techniques.

3. Key Contributions

Wavelength Shift: The paper demonstrates that 1550 nm is a viable and superior alternative to 1030 nm for THz time-stretch detection, despite the scarcity of prior studies on single-shot 1550 nm THz electro-optic detection.
Cost Reduction: By utilizing the 1550 nm regime, the system achieves a high stretch factor using standard commercial components. This allows the use of 1–3 GHz bandwidth readout electronics, reducing the cost of the oscilloscope/ADC by more than an order of magnitude compared to previous systems requiring >10 GHz bandwidth.
Increased Effective Samples: The authors introduce the Time-Bandwidth Product ( $N_{eff,0}$ ) as a figure of merit. They show that the 1550 nm design yields a significantly higher effective number of samples per waveform compared to 1030 nm systems, even without additional spectral broadening.
Experimental Validation: The system was successfully tested at the SOLEIL synchrotron, recording coherent THz pulses from the microbunching instability at 0.85 MHz.

4. Results

Stretch Factor: The system achieved a temporal stretch factor of $M = 270$ , resulting in a stretched pulse duration of approximately 35–40 ns (FWHM).
Readout Performance: The system successfully recorded coherent THz pulses using a 1 GHz oscilloscope for development and a 3 GHz ADC board for systematic data collection.
Data Quality:
- The system captured the full dynamics of THz bursts, including carrier-envelope phase drift and spatio-temporal dislocations.
- The recorded spectra showed a broadband profile centered around 0.3 THz, consistent with Coherent Synchrotron Radiation (CSR).
Comparison Metrics (Table 1):
- 1030 nm System (Standard): With 40 nm bandwidth and 4 km of HI1060 fiber, the time-bandwidth product ( $N_{eff,0}$ ) was 21.
- 1030 nm System (Broadened): With 100 nm bandwidth (via nonlinear broadening), $N_{eff,0}$ was 53.
- 1550 nm System (This Work): With only 30 nm bandwidth and 80 km equivalent dispersion (via DCM), the $N_{eff,0}$ reached 101.
- Conclusion: The 1550 nm design achieved nearly 5 times the effective sample count of a standard 1030 nm system and 2 times that of a broadened 1030 nm system, despite using a narrower laser bandwidth and no nonlinear broadening.
Sensitivity: The noise-equivalent phase shift was measured at < 0.5 mrad (RMS), corresponding to an electric field sensitivity of < 7 V/cm inside the crystal. While the absolute noise floor is slightly higher than some previous 1030 nm works, the massive increase in sample count ( $N_{eff} \approx 100$ vs. $N_{eff} \approx 3$ ) provides a superior signal-to-noise ratio for reconstructing the waveform.

5. Significance

This work represents a paradigm shift in making high-speed THz diagnostics accessible:

Democratization of Technology: By reducing the hardware cost (specifically the readout electronics) by an order of magnitude, this technique becomes feasible for table-top THz time-domain spectroscopy and smaller accelerator facilities, not just large national labs.
Performance Enhancement: Contrary to the intuition that lower bandwidth implies lower performance, the 1550 nm approach actually increases the information content (number of samples) per measurement due to the superior dispersion properties of telecommunication fibers at this wavelength.
Future-Proofing: The use of 1550 nm aligns with existing infrastructure in many accelerator facilities that already utilize 1550 nm timing systems, facilitating easier integration.
Scalability: The authors note that further improvements are possible by adding nonlinear pulse broadening to the 1550 nm source, which could push the performance even further.

In summary, the paper proves that switching to 1550 nm probes allows for a "best of both worlds" scenario: lower cost and higher data resolution for single-shot THz recording.