Modelling of an innovative membrane crystallizer for the production of magnesium hydroxide from waste brine

Background The discharge of industrial waste brines into natural water bodies has gained large interest in recent years, both for its possible environmental impact, but also for the high potential of raw materials recovery contained in brines, pushing towards a circular economy approach. Among such raw materials, magnesium is often abundant and has been defined as aCritical Raw Material by EU [1]. Within this framework, a Crystallizer with Ion Exchange Membrane (CrIEM) has been proposed as an innovative process to recover magnesium from waste brines exploiting low-cost alkaline reactants. In the present work, a novel mathematical model of the CrIEM process is proposed providing a useful tool for its design in different working conditions (Batch and feed & bleed configurations). Materials and Methods The CrIEM (Figure 1) consists of an Anion Exchange Membrane interposed between a brine and an alkaline compartment, which allows the selective passage of OH-into the brine compartment, promoting the precipitation of Mg(OH)2. Figure 1.The CrIEM’s pilot set-up in a feed and bleed configuration. The basics of the modelling tool developed lay on the Donnan Dialysis theory. One dimensional (1D) differential mass balance equations were solved adopting a numerical method. An original algorithm was also developed to describe the time-dependent phenomena occurring in the CrIEM and in two feed tanks. Batch and feed & bleed configurations were investigated considering: (i) the variation of the alkaline and brine concentration in their tanks over time and (ii) the spatial 1D variation of the main parameters inside the CrIEM. Original experimental data and literature information [2] were used for model validation. Results and Conclusions Good agreements between model predictions and experimental data were found for both configurations, proving the reliability of the proposed model for the design of the CrIEM. Magnesium conversion was well predicted exhibiting discrepancies < 5%. Moreover, ions concentrations in the brine channel, hydroxyl transmembrane fluxes and pH trends were well predicted (max discrepancies of 4%, 17% and 5%, respectively). Acknowledgements This work was funded by the ZERO BRINE project (ZERO BRINE – Industrial Desalination – Resource Recovery – Circular Economy) - Horizon 2020 programme, Project Number: 730390: www.zerobrine.eu.