Closed loop control and optimization in centrifugal partition chromatography

Centrifugal partition chromatography (CPC) is a powerful liquid-liquid separation technique, but its productivity is often limited by fixed operating conditions, stationary-phase loss, and idle times between consecutive separations. In this work, a proof of concept for an integrated closed-loop CPC operation is presented. The approach combines online phase-retention control via redosing with model-based operating-point optimization and online evaluation of chromatographic resolution. Prediction models for retention time and peak width were used to calculate productivity-oriented operating points, while experimentally determined resolutions were fed back to compensate plant-model mismatch through a setpoint correction. Once stable baseline separation was reached, model-assisted stacked injections were initiated to reduce idle time between separation sequences. The concept was validated in both CPC operating modes using nicotinic acid/coumarin in descending mode and coumarin/vanillin in ascending mode. In descending mode, stacked operation increased solvent efficiency by 86% and productivity by 115% compared with sequential operation. In ascending mode, solvent efficiency increased by 88%, while productivity increased by 153%. These results demonstrate that closed-loop operating-point optimization combined with stacked injections can substantially improve CPC productivity and solvent efficiency while maintaining the required separation performance.