Translating Selective Lithium Leaching from LFP to Industrial Black Mass: Experimental Validation and Statistical Quantification

In lithium iron phosphate (LFP) battery recycling, lithium represents the principal element of economic value, making its selective recovery a key process objective. Selective leaching conditions are often optimized using idealized materials at laboratory scale; however, these conditions may not translate directly to industrial black masses, which contain copper, aluminum, graphite, and electrolyte residues in addition to the LFP material. This study evaluates the transferability of a sulfuric acid–hydrogen peroxide leaching system developed at laboratory scale to an industrial LFP stream obtained after pyrolysis and mechanical processing. The effects of acid and oxidant concentration, temperature, time, and liquid-to-solid ratio on the dissolution of Li, Fe, P, Cu, and Al were quantified using replicated experimental design. The combination of industrial feed variability and near-complete lithium extraction resulted in nonlinear and noisy datasets, for which the generalized linear model with logit link (GLM-logit) model achieved R2 = 0.89, approaching the experimental noise ceiling. Conventional regression and machine-learning models showed poor predictive performance under these conditions. The GLM-logit approach enabled identification of significant process parameters and construction of safe-operation maps indicating regions where > 95% Li extraction with < 5% Fe and P coleaching occur. Preliminary purification of the leachate enabled effective removal of Fe, Cu, Al albeit with some Li loss. Overall the integrated experimental–statistical approach offers a robust route to identify key leaching parameters and define safe operating regions.