Explainable artificial intelligence for predicting rare earth elements leaching from secondary resources

Area of Science:

• Materials Science
• Chemical Engineering
• Artificial Intelligence

Background:

• Growing global demand for rare earth elements (REEs) necessitates sustainable extraction methods.
• Secondary resources like e-waste and mine tailings offer a viable alternative to primary mining.
• Optimizing leaching processes for REE recovery from these complex matrices is challenging.

Purpose of the Study:

• To develop an explainable AI system for predicting REE leaching efficiency.
• To identify critical factors influencing REE extraction from secondary resources.
• To provide real-time recommendations for optimizing leaching conditions and enhancing recovery rates.

Main Methods:

• An explainable AI model was trained on 572 experimental datasets from the Web of Science database.
• The system predicts leaching efficiency and provides explanations for key influencing parameters.
• The AI suggests condition adjustments to improve REE recovery.

Main Results:

• Silica concentration was identified as the most critical factor affecting REE leaching efficiency.
• REE classification (light vs. heavy) was the second most influential parameter.
• Acid strength (pH), aluminum content, and temperature showed moderate impacts on leaching performance (R²=0.81).

Conclusions:

• Explainable AI effectively bridges the gap between empirical data and process innovation in REE extraction.
• The developed AI framework enhances decision-making and process efficiency for complex extraction systems.
• This methodology has broad applicability beyond REEs to other resource-intensive industries.