DATA: Diafiltration apparatus for high-throughput analysis

Abstract Improved characterization techniques, which address knowledge gaps related to the interfacial processes that govern solute–solute selectivity and the performance of membranes in complex multi-component feed streams, are necessary to advance membrane processes. In this study, guided by the tools of data science, a diafiltration apparatus is developed to inform material and process design by rapidly characterizing membrane performance over a broad range of feed solution compositions. The apparatus doses a fixed-concentration diafiltrate solution into a stirred cell to achieve a predetermined change in the retentate concentration. Here, using an 80 mM potassium chloride (KCl) diafiltrate solution, it was shown that membrane performance, within a 5 mM to 80 mM KCl phase space, could be probed five times more quickly with one diafiltration experiment (8 h) than with an experimental campaign using traditional filtration processes (47 h). Additionally, the synergy between data analytics and instrumentation led to the incorporation of an inline conductivity probe that monitored the real-time retentate concentration. This additional information provided key insights to distinguish between the mechanisms that govern membrane separations (e.g., discriminating between adsorption or rejection based separations) and allowed for the membrane transport coefficients to be determined accurately. Ultimately, incorporating the appropriate governing phenomena identified a single set of self consistent transport parameters for commercial NF90 membranes.