Predicting ion selectivity in water purification by capacitive deionization: electric double layer models

Selective adsorption in ion exchange membranes: The effect of solution ion composition on ion partitioning

Electrodialysis is a water desalination technology that enables selective separation of ions, making it a promising solution for sustainable water reuse. The selectivity of the process is mainly determined by the properties of ion exchange membranes that can vary depending on the composition of ions in water, such as water uptake and charge density. In this work, we studied selective adsorption of Na+ and K+ ions in various ion exchange membranes considering the effect of solution ion composition on membrane water volume fraction. For that purpose, we conducted membrane adsorption experiments using solutions with Na+ and K+ ions with different ion compositions including Li+, Ca2+ or Mg2+ ions at different concentrations (0.001 - 0.25 M). The experiments showed that with the total ion concentration and the amount of divalent ions in solution, the membrane water volume fraction decreases while the selective adsorption of the smaller (hydrated) K+ ions over the Na+ ions in the membrane increases. We developed a theoretical framework based on Boublik-Mansoori-Carnahan-Starling-Leland (BMCSL) theory to describe the effect of membrane water volume fraction on selective adsorption of the ions by including volumetric effects, such as size exclusion. The developed framework was used to describe ion partitioning results of the membrane adsorption experiments. In addition, the effect of solution ion composition on selective ion removal during electrodialysis operation was evaluated using experimental data and theoretical calculations. The results of this study show that considering volumetric effects can improve the ion partitioning description in ion exchange membranes for solutions with various ion compositions.

Hydrolyzed Hydrated Titanium Oxide on Laser-Induced Graphene as CDI Electrodes for U(VI) Adsorption

Recently, laser-induced graphene (LIG), which has been successfully applied in CDI technology (directly without a complex preparation process), has gained considerable attention. However, the raw LIG electrode with a limited number of active sites exhibits low adsorption efficiency. Therefore, the search for a suitable and effective method to modify LIG to improve its electroadsorption performance is significant. Herein, a very simple titration hydrolysis method is adopted to modify LIG, resulting in a layer of hydrated titanium oxide (HTO) being synthesized on the surface of LIG. The LIG/HTO composites possess a good adsorption property since covering the surface of LIG with a layer of HTO can greatly improve the adsorption capacity of LIG. Moreover, with the addition of HTO, not only the proton transfer ability of LIG has been enhanced but also considerable specific capacitance has been enlarged. As a result, LIG/HTO composite as CDI electrode displays a maximum theoretical adsorption capacity of 1780.89 mg/g at 1.2 V, and the capacitance of LIG/HTO composite material is 4.74 times higher than LIG. During the electroadsorption process, Ti4+ is reduced to Ti3+ under external voltage, and O2- is produced through oxidation. Meanwhile, part of the U (VI) is hydrolyzed into UO3·2H2O under the action of -OH, and some combine with O2- to produce UO4·4H2O.

Automation of membrane capacitive deionization process using reinforcement learning

Capacitive deionization (CDI) is an alternative desalination technology that uses electrochemical ion separation. Although several attempts have been made to maximize the energy efficiency and productivity of CDI with conventional control methods, it is difficult to optimize the CDI processes because of the complex correlation between the operational conditions and the composition of feed water. To address these challenges, we applied deep reinforcement learning (DRL) to automatically control the membrane capacitive deionization (MCDI) process, which is one of the representative CDI processes, to accomplish high energy efficiency while desalinating water. In the DRL model, the numerical model is combined as the environment that provides states according to the actions. The feed water conditions, that is, the input state of the DRL, were assumed to have a random salt concentration and constant foulant concentration. The model was constructed to minimize energy consumption and maximize desalted water volume per cycle. After training of 1,000 episodes, the DRL model achieved a 22.07% reduction in specific energy consumption (from 0.054 to 0.042 kWh m^-3) and 11.60% increase in water desalted water volume per cycle (from 1.96×10^-5 to 2.19×10^-5 m³), achieving the desired degree of desalination, compared to the first episode. This improved performance was because the trained model selected the optimized operating conditions of current, voltage, and the number and intensity of flushing. Furthermore, it was possible to train the model depending on demand by modifying the reward function of the DRL model. The fundamental principle described in this study for applying the DRL model in MCDI operations can be the cornerstone of a fully automated water desalination process.