Fluorescence sensor array integrated with machine learning: robust discrimination of multiple metal ions using...

Area of Science:

• Analytical Chemistry
• Environmental Science
• Materials Science

Background:

• Accurate identification and monitoring of metal ions (MIs) are crucial due to their toxicity and environmental risks.
• Existing methods for MI detection can be complex and time-consuming.
• Trace concentrations of MIs pose significant public health and ecological threats.

Purpose of the Study:

• To develop a highly sensitive and selective fluorescence sensor array for simultaneous discrimination of multiple MIs.
• To integrate machine learning algorithms for enhanced MI identification and classification.
• To establish a cost-effective and rapid platform for environmental MI monitoring.

Main Methods:

• Utilized carboxyl-rich carbon dots (CR-dots) as the sensing material.
• Employed multiple buffer environments to achieve selective fluorescence responses.
• Evaluated five machine learning models (LDA, MLR, SVM, k-NN, RF) for data analysis.
• Tested the sensor array in real water samples and complex mixtures.

Main Results:

• The random forest model achieved the highest classification accuracy of 99.78% for MIs at concentrations as low as 0.05 µM.
• The sensor array demonstrated strong classification ability in diverse real-world samples, including tap water, lake water, soil leachates, and seawater.
• Successfully discriminated complex mixtures of binary, ternary, and quaternary metal ions with high accuracy.

Conclusions:

• The developed CR-dot-based fluorescence sensor array integrated with machine learning offers a robust and sensitive platform for MI detection.
• This cost-effective, label-free approach provides a promising strategy for rapid, high-throughput screening and monitoring of environmentally relevant metal ions.
• The technology has substantial potential for practical deployment in environmental protection, pollution assessment, and analytical chemistry.