Efficient extraction of hydrogen fluoride using hollow fiber membrane contactors with the aid of active-learning

This study presents a sustainable approach to extracting hydrogen fluoride (HF) from wastewater using hollow fiber membrane contactors. HF, a widely used yet hazardous chemical, requires efficient separation techniques due to its environmental and health concerns. Our research compared two operational modes, vacuum mode and liquid-liquid extraction mode, revealing the latter as more efficient for HF separation. Notably, this study introduces a novel aspect by employing a data-driven decision-making method, Bayesian optimization (BO) for process optimization of the liquid-liquid extraction mode, aiming to maximize HF removal efficiency at low experimental costs. Subsequent validation through a 96-h experimental run confirms the suitability of the optimized conditions for industrial applications. This study not only demonstrates an efficient HF separation process using hollow fiber membrane contactors but also establishes a new standard for complex industrial process optimization.