Weak Reaction Scatterometry of Plasmonic Resonance Light Scattering with Machine Learning

Area of Science:

• Nanotechnology
• Analytical Chemistry
• Machine Learning

Background:

• Weak chemical reactions are challenging to detect due to low signal intensity and noise interference.
• Accurate monitoring of weak reactions is crucial for understanding reaction mechanisms and applications.
• Existing methods like dark-field microscopy (DFM) struggle with instrument errors and detecting subtle changes.

Purpose of the Study:

• To demonstrate machine learning's capability in improving local surface plasmon resonance (LSPR) scatterometry for weak chemical reactions.
• To enhance the precision and reliability of plasmonic scattering imaging under DFM.
• To enable effective monitoring of unobvious or weak reactions using an automated calibration model.

Main Methods:

• Utilized machine learning algorithms to develop a calibration model for nanoprobe scattering signals.
• Applied dark-field microscopy (DFM) imaging for high-sensitivity LSPR scatterometry.
• Integrated machine learning calibration with DFM to automatically correct deviations and improve signal accuracy.

Main Results:

• Successfully monitored the weak oxidation of silver nanoparticles (AgNPs) in water by dissolved oxygen.
• Detected a reaction between AgNPs and mercury ions in a dilute solution (>1.0 × 10^-10 mol/L).
• Demonstrated significantly improved confidence and effectiveness of LSPR scatterometry for weak reaction detection.

Conclusions:

• Machine learning significantly enhances LSPR scatterometry for detecting and monitoring weak chemical reactions.
• The combined approach offers a powerful tool for sensitive imaging analysis and intelligent sensing.
• This method holds great potential for applications requiring the detection of trace-level reactions and analytes.