Researchers have developed a reusable magnetic sensing platform that combines surface-enhanced Raman scattering with machine learning to detect trace levels of uranyl ions, the most common soluble form of uranium in environmental waters. The system could support faster and more sustainable monitoring of uranium contamination in complex aquatic environments.
Uranium monitoring is important for environmental protection and the safe development of nuclear energy. However, detecting very low concentrations of uranyl ions can be difficult, particularly when samples contain many other dissolved substances. Conventional analytical techniques can provide high sensitivity, but they often depend on expensive instruments, complex sample preparation, and trained operators.
A research team led by Zhenli Sun at North China Electric Power University developed a magnetic SERS substrate called FA@tPF to address these limitations. The platform combines Fe₃O₄@SiO₂@Au microspheres with a polyarylene ether-based covalent organic polymer, known as tPF.
“Our goal was to create a sensing material that can not only capture trace uranyl ions efficiently, but also produce reliable spectral signals that can be interpreted automatically,” said Zhenli Sun, corresponding author of the study. “By combining selective enrichment, magnetic manipulation, Raman detection, and machine learning, the platform provides several functions within a single reusable system.”
The tPF layer plays a central role by providing sites that strongly interact with uranyl ions and enrich them near gold nanoparticles, where localized electromagnetic fields enhance their Raman signals. The magnetic core also allows the sensing material to be rapidly collected and concentrated using an external magnet.
Under a 20-minute enrichment condition, FA@tPF detected uranyl ions at concentrations as low as 1 × 10⁻⁷ mol·L⁻¹. The characteristic Raman band near 850 cm⁻¹ showed a strong relationship with uranyl concentration. In flow-through experiments designed to better represent moving water samples, the platform maintained the same detection limit after 20 minutes of enrichment.
The researchers also tested the sensor in the presence of several common ions, including magnesium, sodium, calcium, potassium, zinc, manganese, nitrate, and sulfate. The characteristic uranyl signal remained stable, indicating strong selectivity and resistance to interference.
A major feature of the platform is its reusability. Uranyl ions adsorbed onto FA@tPF could be removed using sodium carbonate solution, allowing the material to be regenerated. Clear SERS signals remained detectable after six adsorption and desorption cycles, while the material largely retained its structure and functional groups.
The team then added machine learning to automate interpretation of the Raman spectra. Principal component analysis showed highly consistent spectral clustering for FA@tPF. A convolutional neural network classified spectra before and after uranyl adsorption with 100% accuracy in the study dataset . Importantly, the researchers also used gradient-weighted class activation mapping, or Grad-CAM, to examine how the model reached its decisions.
The analysis showed that the model relied strongly on the uranyl-specific Raman feature near 850 cm⁻¹ rather than unrelated spectral noise. This interpretability provides an important link between automated classification and chemically meaningful information.
The study demonstrates how reusable functional materials, portable Raman spectroscopy, magnetic enrichment, and interpretable artificial intelligence can be integrated into a single sensing strategy. With further development, such systems could contribute to intelligent monitoring networks for uranium contamination and nuclear environmental safety.
===
Journal reference: Zhang W, Ma J, Zhang Y, Wang S, Wakeel M, et al. 2026. Reusable magnetic SERS platform functionalized with covalent organic polymers for trace-level and machine-learning-assisted uranyl detection. Sustainable Carbon Materials 2: e027 doi: 10.48130/scm-0026-0023
https://www.maxapress.com/article/doi/10.48130/scm-0026-0023
About Sustainable Carbon Materials :
Sustainable Carbon Materials (e-ISSN 3070-3557) is a multidisciplinary platform for communicating advances in fundamental and applied research on carbon-based materials. It is dedicated to serving as an innovative, efficient and professional platform for researchers in the field of carbon materials around the world to deliver findings from this rapidly expanding field of science. It is a peer-reviewed, open-access journal that publishes review, original research, invited review, rapid report, perspective, commentary and correspondence papers.
Follow us on Facebook , X , and Bluesky .
Sustainable Carbon Materials
Experimental study
Reusable magnetic SERS platform functionalized with covalent organic polymers for trace-level and machine-learning-assisted uranyl detection
27-Jul-2026