Energy consumption and constant current operation in membrane capacitive deionization

Optimization of salt adsorption rate in membrane capacitive deionization

Currently, 2 billion people live in water scarce regions and 4 billion people experience water scarcity for at least 1 month per year [1]. Besides implications to the broad masses, the U.S. Marine Corps also face enormous challenges of generating potable water during squad level field missions. Reverse osmosis is the most reliable and economical process for desalinating water, but it requires piping that can tolerate large pressures and is not conducive for portable missions. Membrane capacitive deionization (MCDI) is a modular and electrified water desalination platform that does not generate significant acoustic, thermal, or electromagnetic signatures nor does it require high pressure piping. Flow-by MCDI, which is commercialized, uses electrical energy to remove ions from water. The system architecture feature two porous electrodes covered by ion-exchange membranes – the latter material is used to prevent co-ion adsorption and greater Coulombic efficiency than the variant that does not utilize ion-exchange membranes.Our previous research improved the energy efficiency of MCDI by using IEMs with lower area-specific resistance (ASR) values and porous ionic conductors in the spacer channel that augmented solution conductivity and curtailed ohmic losses. Reducing these resistances enabled the MCDI to operate at 700 mV lower cell voltage at 2 mA cm-2 current density. This is important for shrinking the size of the MCDI unit and lowering the overall capital costs.This talk presents our recent work examining ionic charge transport resistance at patterned and non-patterned ion-exchange membrane interfaces in MCDI. Membrane surface patterning increased the interfacial area between the membrane and aqueous stream promoting greater salt removal fluxes. Soft-lithography methods were used to micropattern poly(phenylene alkylene) cation exchange membranes and anion exchange membranes with systematically varied topographies (e.g., cylindrical well and line structures that vary from 9 to 20 mm). We can reduce the MCDI cell voltage by 1 V when utilizing micropatterned ion-exchange membranes and ionomer infiltrated electrodes (operating at 2 mA cm-2 with a 2000 ppm NaCl feed). This translated to a 2.4x greater energy normalized adsorbed salt (ENAS) value - a metric used to assess the energy consumption for MCDI. Our future work examines electrode materials deposited at pattern membrane interfaces to reduce interfacial transport resistances from the membrane to the electrode. This strategy has been shown to be effective in reducing charge-transfer resistances in both proton exchange membrane (PEM) and AEM fuel cells and reducing the water dissociation kinetics resistance in bipolar membranes [2-5]. References Somini Sengupta, Weiyi Cai, A Quarter of Humanity Faces Looming Water Crises, in The New York Times. 2019: United States of America.Kole, S., et al., Bipolar membrane polarization behavior with systematically varied interfacial areas in the junction region. Journal of Materials Chemistry A, 2021. 9(4): p. 2223-2238.Breitwieser, M., et al., Tailoring the Membrane-Electrode Interface in PEM Fuel Cells: A Review and Perspective on Novel Engineering Approaches. Advanced Energy Materials, 2018. 8(4).Hee-Tak Kim, T.V.R., Ho-Jin Kweon, Microstructured Membrane Electrode Assembly for direct methanol fuel cell. Journal of The Electrochemical Society, 2007. 154(10).Tomizawa, M., et al., Heterogeneous pore-scale model analysis of micro-patterned PEMFC cathodes. Journal of Power Sources, 2023. 556. Figure 1

US Marine missions require technologies that produce potable water from various water sources that include both seawater and ground water. Reverse osmosis (RO) is the most wide-spread and economical process for desalinating water from seawater. However, it is only cost-competitive and practical when deployed in large, centralized production facilities and is not conducive for marine squad units. Membrane capacitive deionization (MCDI) is an alternative water desalination platform that is enticing to military missions as it does not require high pressure piping and it does not generate significant acoustic, thermal, or electromagnetic signals. MCDI removes ions from the liquid solution using electrical energy, and the commercialized variant, flow-by MCDI, feature two porous electrodes covered by ion-exchange membranes(1). During deionization, the positively biased electrode has an anion exchange membrane (AEM) in front of it while the negatively biased electrode contains a cation exchange membrane (CEM) in front of it.Our previous research improved the energy efficiency of MCDI by using ion-exchange membranes with lower area-specific resistance (ASR) values(2) and porous ionic conductors in the spacer channel that augment solution conductivity and curtail ohmic losses(3). Reducing these resistances enabled the MCDI to operate at higher current density – which is important for shrinking the size of the MCDI unit and lowering the overall capital costs.This talk highlights our recent work examining ionic charge transport resistance at the membrane-electrode interface in MCDI. We studied the impact of this resistance on MCDI performance metrics (salt removal and energy normalized adsorbed salt (ENAS)) by systematically changing the interfacial area of the ion-exchange membrane-electrode interface via patterning of the ion-exchange membrane surfaces using soft-lithography. This strategy has been shown to be effective in reducing charge-transfer resistances in both proton exchange membrane (PEM)(4) and AEM(5) fuel cells and reducing the water dissociation kinetics resistance in bipolar membranes(6). Micropatterned AEMs and CEMs based upon poly(arylene ether) and all-carbon poly(arylene) backbones(7) were fabricated with periodic line and cylinder topographies and lateral feature sizes that vary from 2 to 20 μm. The patterned membranes were coated with graphitic carbon electrodes. It is posited that smaller lateral feature sizes, which gives rise to larger interfacial area values, reduces interfacial resistance values in MCDI and larger ENAS values. Keywords: Desalination, Membrane capacitive deionization, Ion exchange membranes, Soft lithography, Energy normalized adsorbed salt, Membrane electrode assembly References S. Porada, L. Zhang and J. E. Dykstra, Desalination, 488, 114383 (2020).V. M. Palakkal, J. E. Rubio, Y. J. Lin and C. G. Arges, ACS Sustainable Chemistry & Engineering, 6, 13778 (2018).V. M. Palakkal, M. L. Jordan, D. Bhattacharya, Y. J. Lin and C. G. Arges, Journal of The Electrochemical Society, 168, 033503 (2021).Y. Jeon, D. J. Kim, J. K. Koh, Y. Ji, J. H. Kim and Y.-G. Shul, Scientific Reports, 5, 16394 (2015).S. Jang, M. Her, S. Kim, J.-H. Jang, J. E. Chae, J. Choi, M. Choi, S. M. Kim, H.-J. Kim, Y.-H. Cho, Y.-E. Sung and S. J. Yoo, ACS Applied Materials & Interfaces, 11, 34805 (2019).S. Kole, G. Venugopalan, D. Bhattacharya, L. Zhang, J. Cheng, B. Pivovar and C. G. Arges, Journal of Materials Chemistry A, 9, 2223 (2021).W.-H. Lee, Y. S. Kim and C. Bae, ACS Macro Letters, 4, 814 (2015).

Membrane capacitive deionization (MCDI) is an energy efficient, cost effective desalination technique for brackish water to produce water for drinking and processes found in semiconductor and pharmaceutical manufacturing, energy production, and other water related industries.[1-3] The operating principle in MCDI is a potential induced electro-sorption of salt ions that migrate across the ion-exchange membrane layers to the porous carbon electrodes. Hence, the application of electric work reduces the salt content from the feed water. Saturation of the carbon electrodes then leads to discharge of the salt from the carbon electrodes, leading to recovered energy and an increase in salt concentration of the emanating flow stream from the MCDI cell. The discharge process regenerates the electrodes so salt removal can take place again. From a materials aspect of MCDI, most research has focused on carbon-based electrodes with the aim to improve device performance.[4-6] Conversely, little innovation has been made in alternative ion-exchange membrane materials for MCDI. Most reports of MCDI leverage commercially available membranes for electrodialysis.[1,3,7] In this work, MCDI performance was correlated to ion-exchange membrane thickness. The experimental design evaluated examined: i.) commercially available ion-exchange membranes, ii.) ion-exchange layered electrodes of different thicknesses with electrode samples being prepared by drop casting followed by spray painting of ionomer layer on top of the electrode; and iii.) drop casting a mixture of electron conducting carbon with dissolved ionomer. A home-built, single-cell MCDI module was characterized with the different ion-exchange materials using a 275 ppm salt feed. MCDI performance was determined by quantifying salt removal and energy efficiency and the individual resistance contributions in the system through electrochemical impedance spectroscopy.

The energy-water nexus, a 21st century challenge, represents a symbiotic relationship as it takes a lot of energy to purify water and a significant amount of water is needed to produce energy. Prudent manage of these precious resources is paramount for a sustainable planet. Today, numerous industrial processes, such as thermal electric power plants, consume 40% of fresh water withdrawals within the United States.1 Industry demand is often met by treating brackish water streams, which are about 5,000 ppm in total dissolved salts (TDS), and the current annual energy consumption in the U.S. amounts to 2.7x1011 kWh annually. Based on thermodynamic analysis constrained to a 50% minimal water recovery, today’s state-of-the-art for water desalination of brackish water streams is 20% efficient. Improving the efficiency from 20% to 30% could amount to a savings of $4 billion dollars worth of electricity annually in the U.S. and a reduction of energy needs by 20 GW. One technology capable of providing low energy water desalination of brackish water streams with over 30% thermodynamic efficiency is membrane capacitive deionization (MCDI). MCDI, an electrochemical technology, offers the enticing prospect of low energy water desalination as it can recover energy (demonstrated up to 83%)2 when discharging the saturated electrodes as a supercapacitor. Most research today has focused on porous carbon electrodes to increase MCDI capacity. However, material innovation with regard to the ion-exchange membranes used in MCDI has been neglected. Today’s MCDI currently uses electrodialysis (ED) membranes – not originally designed for MCDI. The ion-exchange membranes for ED tend to be thick as they serve as a separator in addition to an ion-conducting electrolyte in electrodialysis. In MCDI, the ion-exchange membranes do not serve as separator and they are employed to prevent co-ion adsorption to the porous carbon electrodes. This talk will demonstrate how thinner ion-exchange membranes with higher ionic conductivity substantially reduces the energy footprint for MCDI - approximately 35 kT per ion removed with ED membranes to 15 kT per ion removed with new ion-exchange membranes. Additionally, the new ion-exchange membranes displayed a significantly higher Coulombic efficiency. The talk will showcase how ex-situ properties of ion-exchange membranes, such as ionic conductivity, thickness, and permselectivity, are instrumental for lowering the energy footprint of MCDI. Note: Figure represents single-cell MCDI operation under constant current with newly prepared ion-exchange membranes. Green line corresponds to cell voltage, while the purple line corresponds to the salt effluent stream.

Comparison of constant voltage (CV) and constant current (CC) operation in the membrane capacitive deionisation process

Energy consumption analysis of constant voltage and constant current operations in capacitive deionization

Theory of membrane capacitive deionization including the effect of the electrode pore space

Energy consumption in capacitive deionization – Constant current versus constant voltage operation

Membrane capacitive deionization (MCDI) is a unique electrochemical separations platform that allows for energy recovery during electrode regeneration. Similar to other electrochemical separation technologies producing deionized water (e.g. electrodialysis), ohmic resistances in the spacer channel significantly hampers the performance and energy efficiency of the process. This work devised a series of ionomer coated nylon mesh nets to address spacer channel resistances in MCDI. Under constant current operation, the ionomer coated nylon meshes displayed a 300 mV lower cell voltage rise during deionization while sustaining the same deionization rate. Furthermore, energy recovery was improved by 1.4x to 5.5x depending on the saline feed concentration. The lower cell voltage rise during deionization combined with the greater energy recovery with ionomer coated meshes resulted in energy normalized adsorbed salt (ENAS) values that were 2x to 3x greater. Addressing the spacer channel resistances in MCDI allowed for 8% to 19% increase in current density without the cell voltage exceeding 1.6 V—the upper bound set for mitigating parasitic reactions. Operating at higher current density leads to smaller MCDI units for a given deionization requirement and has implications for reducing the capital costs of the MCDI unit.

Membrane capacitive deionization

Membrane capacitive deionization (MCDI) can be typically operated with constant voltage (CV) or constant current (CC) mode in the charging stage. While a series of previous studies have compared both charging modes to identify the better operating mode, neither their performance evaluation protocols were consistent, nor did their conclusions unanimously converge. This study presents a new framework to evaluate and compare MCDI performance, considering the kinetic efficiency, the energetic efficiency, and the intrinsic trade-off between the two. A key prerequisite for making rational comparison of performance between MCDI operations is that the operations being compared should all result in the same target adsorption. With this key prerequisite and the new evaluation framework based on the trade-off curve between kinetic and energetic efficiencies, our experimental assessment and theoretical analysis suggest that whether CC or CV charging is more efficient is strongly dependent on the target adsorption and, to a less extent, on the kinetic rate of charging. However, the advantage in energy or kinetic efficiency of one charging mode over that of the other is relatively small in all cases. Our study also reveals that, for a given MCDI system, there exist regimes of target adsorptions and kinetic rates that can only be achieved by either CC or CV charging, or even regimes that can be achieved by neither charging mode. In summary, this study revises our current understanding regarding the comparison of the two typical charging modes in MCDI, and introduces a new framework for comparing the performance of different MCDI and CDI operations.

Capacitive deionization (CDI) is an electrochemical technology that utilizes porous carbon electrodes to remove ionic substances. Ions from the solution are adsorbed during the charging stage and the adsorbed ions are desorbed during the discharging process, repeating the process to operate the system. To enhance ion removal capacity and prolong life of systems compared to CDI, membrane capacitive deionization (MCDI) has been developed. MCDI includes both anion and cation exchange membranes to induce the co-ion expulsion effects during the discharging stage, thereby exhibiting superior ion adsorption capacity than CDI. However, the utilization of ion exchange membranes causes a new issue of membrane fouling. Natural organic matter (NOM) serves as a common precursor to membrane fouling in nature environment. Especially, NOM could lead to the formation of fouling layers that disrupt transport phenomena across membrane systems, including ion exchange membranes used in MCDI. Numerous modeling studies have been conducted to investigate the effect of fouling on MCDI, aiming to predict fouling and evaluate the performance of MCDI in advance. Although a mathematical model was developed to predict MCDI performance under natural organic matter conditions, it had limitations of achieving high accuracy in natural environments. While a modified mathematical model showed high accuracy in predicting adsorption time reduction and capacity, it demonstrated low accuracy at points where current was applied and varied. This highlights the limitations of the modified mathematical model, as it is inadequate to consider the actual operating environment of MCDI. Therefore, this study adopted a robust machine learning model to address the limitations of the mathematical models. Electrochemical profiles and feed water conditions were used as input parameter of the model, enabling the prediction of both salt adsorption capacity and the growth of fouling layer on ion exchange membranes during operation. Consequently, this approach could improve the prediction accuracy of MCDI performance while providing detailed insights into fouling layer development, facilitating efficient operation and management of the MCDI system.

Chapter 8 - Principles and Theoretical Models of CDI: Experimental Approaches

Membrane capacitive deionization (MCDI) is an emerging water desalination platform that is compact, electrified, and does not require high‐pressure piping. Herein, highly conductive poly(phenylene alkylene) ion‐exchange membranes (IEMs) are micropatterned with different surface geometries for MCDI. The micropatterned membranes increase the interfacial area with the liquid stream leading to a 700 mV reduction in cell voltage when operating at constant current (2 mA cm−2; 2000 ppm NaCl feed) while improving the energy normalized adsorbed salt (ENAS) value by 1.4 times. Combining the micropatterned poly(phenylene alkylene) IEMs with poly(phenylene alkylene) ionomer‐filled electrodes reduces the cell voltage by 1000 mV and improves the ENAS values by 2.3 times relative to the base case. This reduction in cell voltage allows for higher current density operation (i.e., 3–4 mA cm−2) . The reduction in cell voltage is ascribed to the ameliorating ohmic resistances related to ion transport at the membrane‐process stream interface and in the carbon cloth electrode. Finally, porous ionic conductors are implemented into the spacer channel with flat and micropatterned IEM configurations and ionomer infiltrated electrodes. For the configuration with flat IEMs, the porous ionic conductor improves ENAS values across the current density regime (2–4 mA cm−2), while for micropatterned IEMs it gets improved only at 4 mA cm−2.