Traffic-Light Fluorescent Probe Delivers Rapid Mercury Warnings via Deep Eutectic Chemistry and AI Vision

Mercury is one of the most notorious poisons in the environmental chemist’s catalogue, and catching it quickly can make the difference between a contained incident and a public health problem. A team of researchers in China has now unveiled a sensing platform that reads like something out of a science fiction briefing: a fluorescent material built inside a deep eutectic solvent, paired with a computer vision algorithm, that flashes red, yellow, or green depending on how much mercury is lurking in a sample. The work, published in Advanced Composites and Hybrid Materials, describes what the authors call a traffic-light intelligent early warning platform, and it delivers a detection limit of 60 nanomoles per liter for mercury ions while returning results for multiple samples in under half a second.

The heart of the innovation lies in an unusual class of liquids known as deep eutectic solvents, or DES. These materials form when hydrogen bond donors and acceptors are mixed in precise ratios, and the resulting eutectic mixture melts at a temperature far lower than either component alone. Chemists have increasingly turned to DES as greener alternatives to volatile organic solvents, but the Tianjin University of Traditional Chinese Medicine-led team exploited a subtler property: the dense, dynamic hydrogen bond network that permeates these liquids. That network acts as a molecular gathering ground, pulling mercury ions toward the fluorescent centers before they are even measured. The researchers describe this as a preconcentration effect, and it is the first stage of a cascade that amplifies the signal the probe ultimately produces.

The fluorescent component itself is equally carefully engineered. Rather than a conventional dye, the team constructed a partially rigid, delocalized conjugated fluorescent material featuring what they term an intramolecular coupled positive-negative junction. The delocalized conjugated system, spanning a large molecular framework, gives the material excellent luminescence performance, meaning it glows brightly and consistently enough to serve as a reliable analytical baseline. The embedded junction structure, reminiscent of the charge separation architecture found in semiconductor devices, helps organize the electronic behavior of the molecule so that its emission responds predictably to its chemical surroundings. When mercury ions arrive, they selectively quench the fluorescence, dimming the glow in a dose-dependent manner that can be traced back to concentration.

Selectivity is the perennial headache in metal sensing. Water samples, food extracts, and biological fluids contain a zoo of competing ions, and many fluorescent probes that respond to mercury also light up or dim in the presence of copper, iron, or calcium. The deep eutectic composite, according to the study, exhibits excellent selective fluorescence quenching against mercury ions, meaning the signal remains stable even when other metals are present. The combination of the DES preconcentration mechanism and the quenching response is what the authors describe as a synergistic preconcentration-response mechanism, and it is this two-step amplification that pushes the sensitivity to the 60 nanomolar detection limit, a level relevant to environmental monitoring needs.

But a sensitive probe is only half the story. Translating a fluorescence change into a decision that a field worker or clinician can act on has traditionally required bulky spectrometers and trained analysts. The researchers sidestepped that bottleneck by integrating their material with computer vision. Using the YOLOv8 object detection algorithm, a portable terminal captures images of the probe’s color response and interprets them in real time. The platform establishes a quantitative mapping relationship between the color information in each image and the actual concentration of mercury in the sample, effectively turning a smartphone-class camera into an analytical instrument. Detection results for multiple samples can be acquired within 472 milliseconds, a speed that opens the door to genuinely high-throughput screening.

The traffic light metaphor is not merely decorative. The platform’s software applies a grading logic that sorts every measurement into three tiers: green for safe, yellow for alert, and red for hazard. Built-in indicators mark these categories visually, so a user glancing at the screen sees an immediate verdict rather than a raw number. In validation experiments, the platform achieved overall recoveries ranging from 91.16 percent to 109.35 percent across multiple samples, a range that indicates the system reads true concentrations accurately even in the messy conditions of real sample matrices. That combination of speed, accuracy, and intuitive output is what distinguishes the work from conventional laboratory assays that depend on large instruments and low-throughput workflows.

The implications reach well beyond mercury. The authors argue that their strategy provides a new approach to point-of-care sensing of toxic contaminants, a field that has been dominated by expensive reagents and centralized laboratory infrastructure. A portable, low-cost platform that anyone can operate, and that renders its verdict in the universal language of traffic light colors, could be deployed at industrial discharge points, in agricultural water testing, in food safety inspections, or in clinics screening for heavy metal exposure. Because the underlying chemistry is modular, the same deep eutectic preconcentration concept could in principle be adapted to other target ions by swapping the fluorescent recognition unit, though the current study demonstrates the platform specifically for mercury.

There is also a materials science lesson embedded in the work. Deep eutectic solvents have mostly been cast in the role of reaction media or extraction agents, valued for their low toxicity and tunability. This study treats the DES as an active functional component of the sensor itself, its hydrogen bond network doing real analytical work by concentrating analytes at the sensing interface. That reframing, in which a green solvent becomes part of the signal amplification cascade, may inspire similar designs across the growing field of composite and hybrid materials, where the boundaries between matrix and active element are increasingly blurred.

The research, conducted by Kaining Yang, Hengmao Yue, and colleagues across Tianjin University of Traditional Chinese Medicine and Beihang University, was supported by funding from the National Natural Science Foundation of China, the National Key Research and Development Program of China, and several Tianjin municipal programs. Published open access, the study arrives as regulators and public health agencies worldwide continue to grapple with mercury contamination from artisanal gold mining, coal combustion, and industrial effluent. A tool that can turn a glowing composite into an instant red, yellow, or green verdict, at nanomolar sensitivity and in under half a second, offers a glimpse of how environmental monitoring might look when molecular design and artificial intelligence finally share the same bench.

Subject of Research: A deep eutectic solvent-based fluorescent composite with cascade signal enhancement for intelligent mercury ion detection

Article Title: Traffic light intelligent early warning platform for mercury: deep eutectic fluorescent composite with cascade signal enhancement function

Article References: Yang, K., Yue, H., Shi, J., Li, H., Peng, Y., Li, H., Li, F., Du, K., & Li, X. (2026). Traffic light intelligent early warning platform for mercury: deep eutectic fluorescent composite with cascade signal enhancement function. Advanced Composites and Hybrid Materials. https://doi.org/10.1007/s42114-026-02107-9

Image Credits: AI Generated

DOI: 10.1007/s42114-026-02107-9

Keywords: deep eutectic solvents, fluorescent probes, mercury detection, carbon quantum dots, computer vision, YOLOv8, early warning platform, heavy metal sensing, fluorescence quenching, environmental monitoring, point-of-care testing, hybrid materials