🔬 optics

High-resolution mid-infrared single-photon upconversion ranging

This paper presents a high-resolution mid-infrared single-photon upconversion LiDAR system that utilizes nonlinear asynchronous optical sampling and time-correlated photon counting to achieve a 4 $\mu$ m ranging precision with single-photon sensitivity, effectively overcoming detection limitations in photon-starved scenarios.

Original authors: Shuhong Jiang, Kun Huang, Tingting Yu, Jianan Fang, Ben Sun, Yan Liang, Qiang Hao, E Wu, Ming Yan, Heping Zeng

Published 2026-05-27

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Shuhong Jiang, Kun Huang, Tingting Yu, Jianan Fang, Ben Sun, Yan Liang, Qiang Hao, E Wu, Ming Yan, Heping Zeng

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 measure the distance to a mountain in the dark, but you can only use a very dim flashlight. In the world of light-based measuring (LiDAR), scientists have been great at doing this with visible light (like a laser pointer) or near-infrared light. However, when they try to use mid-infrared light—which is like a "heat vision" spectrum that can see through fog, smoke, and even some solid materials like silicon—they hit a wall. The problem is that the "cameras" (detectors) needed to catch these specific infrared photons are usually slow, noisy, or require freezing temperatures to work.

This paper describes a clever new trick to solve that problem. The researchers built a system that can measure distances with mid-infrared light using just single photons (the tiniest possible packets of light) and with incredible precision, all while working at room temperature.

Here is how they did it, explained with some everyday analogies:

1. The Problem: The "Slow Camera" vs. The "Fast Runner"

Think of the mid-infrared light as a super-fast runner. To measure how far they ran, you need a camera that can take a picture fast enough to catch them. But the existing cameras for this type of light are like old, slow film cameras; they are too slow to catch the runner clearly, and they are very "grainy" (noisy).

2. The Solution: The "Magic Translator" (Upconversion)

Instead of trying to build a faster camera for the infrared runner, the team built a translator.

They take the invisible mid-infrared light and mix it with a second, very fast laser beam inside a special crystal.
This mixing process acts like a translator that instantly converts the "foreign language" of the infrared light into "English" (visible light).
Now, instead of needing a slow infrared camera, they can use a standard, high-speed silicon camera (the kind found in your phone or a regular digital camera) to catch the translated signal. This camera is fast, quiet, and works perfectly at room temperature.

3. The "Time-Stretch" Trick

Here is the most creative part. Even with the translator, the light pulses are still moving too fast for the camera to see the tiny details of the distance.

The Analogy: Imagine you are trying to read a book that is being flipped through at 1,000 pages per second. You can't read a single word.
The Fix: The researchers used a technique called asynchronous optical sampling. Imagine two people flipping through books, but one flips just a tiny fraction of a second slower than the other. Every time they flip, the page you see is slightly different.
Because of this tiny difference in speed, the "fast" light signal gets stretched out over time, like pulling a piece of taffy. A pulse that happened in a billionth of a second gets stretched out to last for a millisecond.
Now, the "slow" camera can easily read the stretched-out signal, seeing details that were previously invisible.

4. The Results: Seeing the Invisible

By combining this "translator" with the "time-stretch" trick, the team achieved two amazing things:

Extreme Sensitivity: They could detect the distance even when only one photon (or even less than one photon on average per pulse) was bouncing back. It's like being able to hear a whisper in a hurricane.
Microscopic Precision: They could measure distance changes as small as 4 micrometers. To put that in perspective, that is about the width of a single bacterium or a strand of spider silk.

5. What They Actually Tested

The paper doesn't just talk about theory; they tested it on real objects:

Moving Targets: They measured a mirror moving back and forth very quickly, proving the system can track fast motion.
Seeing Through Silicon: They shone the light through a stack of silicon wafers (which look like black glass to normal light but are transparent to this specific infrared light). The system successfully measured the thickness of the wafers and the layers inside them without cutting them open.

Summary

In short, the researchers created a "super-power" for mid-infrared light. They took light that is usually hard to measure, translated it into a language our standard cameras understand, and stretched it out so we can see the tiny details. This allows for ultra-precise, ultra-sensitive distance measuring that can see through fog and solid materials, all without needing expensive, freezing-cold equipment.

Technical Summary: High-Resolution Mid-Infrared Single-Photon Upconversion Ranging

Problem Statement
While single-photon laser ranging (LiDAR) has achieved widespread success in the visible and near-infrared (NIR) bands for applications such as autonomous driving and remote sensing, extending these capabilities to the mid-infrared (MIR) region remains a significant challenge. MIR operation is highly desirable due to reduced photon scattering in adverse weather (fog, haze) and inherent chemical selectivity for gas sensing. However, existing MIR detectors based on narrow-bandgap semiconductors, bolometers, or emerging low-dimensional materials typically lack the sensitivity required for single-photon detection (limited to ~pW/Hz $^{1/2}$ ). While superconducting nanowire detectors offer MIR photon-counting, they require cryogenic cooling, introducing system complexity. Furthermore, existing upconversion LiDAR systems face a trade-off: continuous-wave (CW) pumping offers wide range but is limited by detector timing jitter (sub-centimeter resolution), while pulsed pumping offers high resolution but is constrained by the short capture depth of mechanical delay lines.

Methodology
The authors propose and demonstrate a MIR upconversion LiDAR system based on nonlinear asynchronous optical sampling (NAOS) combined with time-correlated single-photon counting (TCSPC). The core architecture utilizes a dual-comb configuration with two mode-locked fiber lasers (Er-doped at 1.56 µm and Yb-doped at 1.03 µm) stabilized at slightly different repetition rates ( $f_r$ and $f_r - \Delta f_r$ ).

MIR Generation: The synchronized dual-color pulses undergo difference frequency generation (DFG) in a periodically poled lithium niobate (PPLN) crystal to generate a MIR probe pulse at 3.1 µm.
Asynchronous Upconversion: The MIR probe is split into a reference path and a target path. The reflected MIR signals are spatially mixed with a high-power 1.03 µm pump pulse train in a second PPLN crystal. Through sum-frequency generation (SFG), the MIR photons are upconverted to the visible band (0.77 µm).
Temporal Stretching: Because the pump and probe pulses have slightly different repetition rates, the pump pulse "sweeps" across the MIR probe pulse train in the time domain. This creates a time-stretch factor $M \approx f_r / \Delta f_r$ (approximately $2 \times 10^4$ in this experiment). Consequently, the fast temporal dynamics of the MIR signal are slowed down, allowing a low-bandwidth silicon photodiode to record the cross-correlation trace.
Photon Counting: For low-light scenarios, the upconverted signal is detected by a silicon single-photon counting module (SPCM) connected to a TCSPC device. The start signal is derived from the SFG pulse generated by the two lasers, and the stop signal comes from the SPCM.

Key Contributions

Novel Ranging Architecture: The paper introduces a pulsed-pumping upconversion scheme that overcomes the limited capture depth of mechanical scanners by using asynchronous optical sampling, enabling a wide non-ambiguous range (46.45 ns in the experiment, scalable to kilometers).
Jitter Mitigation via Time-Stretching: The system decouples the ranging resolution from the intrinsic timing jitter of the detector. By stretching the cross-correlation trace by a factor of $2 \times 10^4$ , the effective temporal resolution is determined by the optical pulse width and the stretching factor rather than the detector's jitter.
Room-Temperature MIR Photon Counting: The system achieves single-photon sensitivity in the MIR using a room-temperature silicon-based detector, bypassing the need for cryogenic cooling.
Incoherent Operation: Unlike coherent dual-comb schemes requiring phase-locking and carrier-envelope phase stabilization, this method relies on optical field intensity and nonlinear parametric processes, simplifying the laser source requirements.

Experimental Results

Resolution and Precision: The system achieved an effective temporal resolution of 9.8 ps (inferred from the cross-correlation width), which is two orders of magnitude better than the 830 ps timing jitter of the SPCM itself. This resulted in a ranging precision of 4 µm (Allan deviation) at an averaging time of 64 seconds.
Sensitivity: Ranging was demonstrated at an extremely low detected flux of $8 \times 10^{-5}$ photons per pulse. The total detection efficiency was approximately $3.2 \times 10^{-5}$ , limited primarily by collection efficiency and nonlinear conversion efficiency.
Dynamic Performance: The system successfully tracked a rapidly oscillating target (15 Hz) with a 1 kHz refresh rate, demonstrating inertia-free optical scanning superior to mechanical delay lines.
Multi-Layer Characterization: As a proof-of-principle, the system performed non-invasive depth profiling on stacked silicon wafers. It successfully resolved the geometric thickness (2.94 mm vs. 3 mm actual) and refractive index (3.49) of the silicon, as well as the reflectivity of the interfaces, leveraging the MIR transparency of silicon.

Significance and Claims
The paper claims to realize the first high-resolution MIR photon-counting LiDAR under low-light-level illumination. The authors emphasize that their approach alleviates the fundamental limitation imposed by the intrinsic timing jitter of photon-counting systems, enabling sub-micrometer ranging precision at the single-photon level without cryogenic infrastructure.

The work establishes a path toward sensitive and precise infrared sensing for "photon-starved" scenarios, such as long-standoff distance detection, environmental monitoring, and industrial quality control (e.g., non-destructive defect inspection of semiconductor materials). The authors note that while the current acquisition time for dynamic scenes is limited by the low photon flux, the integration of advanced computational algorithms could further enhance performance in photon-efficient regimes. They conclude that the technique is scalable to longer infrared or terahertz regions and that resolution could be further enhanced to the nanometer level using shorter optical pulses.