Rare earth metals detection and recognition based on laser induced breakdown spectroscopy and machine learning

Area of Science:

• Materials Science
• Analytical Chemistry
• Environmental Science

Background:

• Electronic waste (e-waste) recycling is crucial for rare earth (RE) element recovery.
• Identifying RE-containing e-waste is challenging due to material similarities.

Purpose of the Study:

• To develop a rapid detection system for RE elements in e-waste.
• To classify rare-earth phosphors (REPs) using advanced analytical techniques.

Main Methods:

• Utilized laser-induced breakdown spectroscopy (LIBS) for elemental analysis.
• Applied machine learning algorithms: principal component analysis (PCA) for unsupervised classification and backpropagation artificial neural network (BP-ANN) for supervised identification.
• Analyzed spectra of three different phosphors, detecting Gd, Yd, and Y RE elements.

Main Results:

• LIBS successfully detected RE elements in phosphors.
• PCA effectively distinguished between the three phosphor types.
• The BP-ANN model achieved a 99.9% recognition rate for phosphor identification.

Conclusions:

• The integrated LIBS and ML system offers a highly accurate method for RE element detection.
• This innovative approach has significant potential for rapid, in situ classification of RE-containing e-waste.
• The system can enhance the efficiency of RE element recycling from e-waste streams.