Prediction of collector flotation performance based on machine learning and quantum chemistry: A case of sulfide minerals

Flotation, as a pivotal separation technology in the 21st century, facilitates the large-scale utilization of mineral resources. The development of high-performance surfactants, particularly collectors, is crucial for enhancing flotation efficiency. This study introduces a novel machine learning (ML) model designed to evaluate and predict the recoveries of sulfide minerals (chalcopyrite, galena, pyrite, and sphalerite) under various flotation conditions, including pulp pH, flotation time, and collector concentration. Quantum chemistry (QC) computations were employed to characterize the features of 116 collectors (e.g., electrostatic properties, atomic charges, molecular orbitals) and the sulfide minerals (e.g., surface charges, band gap, adsorption energies). These features, along with flotation conditions from the literature, served as input, while experimental recoveries of the four minerals were the output. The model was refined using 10 randomly selected collectors, achieving a mean absolute error (MAE) of 10.0%. The optimized ML model demonstrated high accuracy, successfully predicting the flotation performance of 23 new collectors with an MAE of 5.2%. This QC-ML approach offers a powerful tool for the high-throughput screening and rational design of flotation reagents, significantly advancing the field of mineral processing.