Skip to content
Publicly Available Published by De Gruyter July 1, 2021

Principles of reverse electrodialysis and development of integrated-based system for power generation and water treatment: a review

Nur Hidayati Othman ORCID logo EMAIL logo , Nalan Kabay and Enver Guler


Reverse electrodialysis (RED) is among the evolving membrane-based processes available for energy harvesting by mixing water with different salinities. The chemical potential difference causes the movement of cations and anions in opposite directions that can then be transformed into the electrical current at the electrodes by redox reactions. Although several works have shown the possibilities of achieving high power densities through the RED system, the transformation to the industrial-scale stacks remains a challenge particularly in understanding the correlation between ion-exchange membranes (IEMs) and the operating conditions. This work provides an overview of the RED system including its development and modifications of IEM utilized in the RED system. The effects of modified membranes particularly on the psychochemical properties of the membranes and the effects of numerous operating variables are discussed. The prospects of combining the RED system with other technologies such as reverse osmosis, electrodialysis, membrane distillation, heat engine, microbial fuel cell), and flow battery have been summarized based on open-loop and closed-loop configurations. This review attempts to explain the development and prospect of RED technology for salinity gradient power production and further elucidate the integrated RED system as a promising way to harvest energy while reducing the impact of liquid waste disposal on the environment.

1 Introduction

Global energy expenditure is predicted to increase at an exponential rate due to the rapid growth of populations. Therefore, carbon dioxide (CO2) emissions, global warming, and air pollution issues are also expected to increase. The future energy of the world should be clean and sustainable as it significantly impacts our climate, health, and economy. By developing alternative and renewable energy resources, issues related to increased demand for clean energy and associated environmental concerns might be resolved. Salinity gradient power (SGP) or identified as ‘blue energy’ can generate electrical or mechanical energy by converting the chemical potential when two solutions with different salinities are mixed (Jia et al. 2014). As it does not release any toxic gas emissions, it is seen as sustainable and clean energy. It has a massive prospect to harvest energy and produce electricity simply by mixing seawater and river water at global estuaries, thus increasing the world’s renewable energy supply. It can be associated with the Earth’s complex water cycle where first water evaporates from water bodies due to solar radiation. These low salt content solutions are then transported in clouds form and finally precipitate to the ground. Figure 1 illustrates the notion of SGP as a potential renewable energy source when seawater/brine, brackish water, or river water are mixed (Tufa et al. 2018). Although it is currently being explored for the generation of energy from river water/seawater, there are vast opportunities to use concentrated brine due to its higher power efficiencies and alleviation of undesirable impact to the environment related to brine dumping (Chung et al. 2017; Ramon et al. 2011). Other potential brine sources include saline wastewater from domestic sources, industrial processes, anthropogenic sources, and thermolytic solutions (e.g. ammonium bicarbonate) (Bevacqua et al. 2017; Cusick et al. 2012).

Figure 1: 
Salinity gradient power (SGP) from the natural water cycle. Redrawn from Tufa et al. (2018). Copyrighted from Elsevier Ltd.
Figure 1:

Salinity gradient power (SGP) from the natural water cycle. Redrawn from Tufa et al. (2018). Copyrighted from Elsevier Ltd.

SGP concept has been recognized since the 1950s after it was proposed by Pattle (1954). Since then, a growing number of research and patents have been published (Avci et al. 2018; Helfer and Lemckert 2015; Jones and Finley 2003; Wick 1978; Yip et al. 2016). Another significant benefit of SGP compared to wind and solar power is that it can be exploited continuously for 24 h per day throughout the whole year without depending on seasonal climate (Güler and Nijmeijer 2018). However, the total of SGP that can be extracted might be different based on various factors, such as temperature, flow rate, salinity level, fouling behavior, and other aspects. Other than these technical and operational factors, ecological and legal constraints also play an important role (Alvarez-Silva et al. 2016).

Several ways have been to exploit SGP such as by using a hydrostatic generator, vapor compression, capacitive system and membrane-based pressured-retarded osmosis (PRO), and reverse electrodialysis (RED) technologies (Jones and Finley 2003). Among those, membrane-based processes have emerged as a promising route based on pilot-scale studies (D’Angelo et al. 2017; Gómez-Coma et al. 2020; Tedesco et al. 2016a). PRO utilizes semipermeable membranes while RED uses ion-exchange membranes (IEMs). Both membranes have different operation principles. PRO uses osmotic pressure as a driving force for water to permeate across salt-rejecting membranes into highly concentrated draw solution compared to ion flux that passes through a charged membrane in the case of the RED system (Figure 2). Typically, a RED stack composed of repetitive units, known as cell pairs ranges from few (lab-scale), some tens (bench-/pilot-scale) to hundreds and thousands (commercial stacks), which are sandwiched between two end plates containing electrode (Gurreri et al. 2020). As more membranes are piled in a stack, higher electrical power can be generated by the stack. These SGP harnessing techniques have been extensively reviewed by Cipolina and Micale (2016), Mei and Tang (2018), Post et al. (2007), Tian et al. (2020), Wick (1978), Yip and Elimelech (2014), and Yip et al. (2016).

Figure 2: 
Comparison of (a) pressure retarded osmosis (PRO) and (b) reverse electrodialysis (RED) with one cell pair. Reproduced with permission (Yip and Elimelech 2014). Copyrighted from Elsevier Ltd.
Figure 2:

Comparison of (a) pressure retarded osmosis (PRO) and (b) reverse electrodialysis (RED) with one cell pair. Reproduced with permission (Yip and Elimelech 2014). Copyrighted from Elsevier Ltd.

Over several years ago, the scope of RED has been swiftly extended due to the increased demand for sustainable energy. However, the technological progress of RED is hampered mainly by (1) the unattainability stable, high performance and low-cost membrane materials, and (2) the need for a broader and deeper investigation of the influence of operating parameters on RED performances in order to achieve optimal design and operation of RED units. A high-performance IEM must be chemically stable and possess high power density, high permselectivity, and low resistance. The membrane materials used should be cheaper and have a longer lifetime to ensure lower operational and design costs. Besides membrane development, there is also a need for broader and deeper investigation on the influence of operating parameters toward RED performances. This aspect is important to achieve the optimal design and operation of RED units. Up until today, most of the investigation on the RED performances was carried out using artificial feedwater. However, as real or natural feedwater sources contain multivalent ions and various types of organic compounds, this significantly introduces challenges toward the readiness level of RED in real applications (Chon et al. 2020). Therefore, it is essential to perform RED experiments using natural feedwater to understand the process under real conditions.

This review begins with an introduction to the basic principles of the RED system, major progress on IEMs used for RED from the past decades up to now, followed by the effects of operational factors toward the performance of the RED system. Recently, integrated/hybrid-based systems have been proposed to overcome the inherent limitations of RED. These innovative systems work synergistically to optimize RED performances while reducing operating and capital costs. Therefore, a dedicated section on the integration of RED with other systems is presented to highlight the future outlook of this technology.

2 Principles and mechanism of RED

Figure 3 shows a simple schematic diagram of a RED system, which consists of IEMs selective for either cations or anions and a compartment for feed solutions. Multiple cells consisting of anion-exchange membranes (AEMs) and cation-exchange membranes (CEMs) are arranged in an alternated pattern to provide contact between high-salinity solution in the high-concentration compartment (HCC) and low-salinity solution in the low-concentration compartment (LCC). This arrangement creates an electrical current or voltage difference known as the Donnan potential (Najmiyah et al. 2018), which is the driving force for the process. The number of cells can be amplified to increase the system voltage (Veerman and Vermaas 2016). The ion transport phenomenon in RED is mainly depending on the transport of the ions in the membranes and solutions. This comprises electro-migrative ions to and toward the surface and in IEMs and the convective flux through spacer channels along IEMs. AEMs contain fixed positive (+) charges that permit anions to be transported toward the anode, while CEMs contain negative (−) charges which permit cations to pass through to the cathode. As the RED stack is connected to an external load, a flux of ions will pass through the membranes. Detail mathematical modeling on the ion transport phenomenon in RED has been carried out and it was observed that the transport phenomenon in RED becomes more complex when the feed streams contain multivalent ions. A number of approaches have been adopted such as Nernst–Planck transport (Tedesco et al. 2016b), irreversible thermodynamics formalism (Nikonenko et al. 2002), Stefan–Maxwell theory (Tedesco et al. 2017), and semi-empirical model (Tedesco et al. 2015b) in order to evaluate different contributions to the resultant ionic flow and to present a basis to understand limiting factors as in the case of polarization phenomena, water passage, and co-ions transport.

Figure 3: 
Schematic diagram of the RED system.
Figure 3:

Schematic diagram of the RED system.

Electrodes are connected to both ends of cells to mitigate electron transfer and transform the accumulated ionic flow to an electrical current through redox reaction (Moreno et al. 2018b). The efficiency of the redox reaction is affected by the type of electrodes and electrolytes utilized in the system. The electrolyte recirculation leads to no net chemical reaction, where the generation of electricity is caused by electronic current and potential difference over the electrodes (Veerman et al. 2010a).

Recently, several works have been conducted to investigate the sustainable electrode system that is low in cost, safe, and environmentally friendly by selecting appropriate electrodes and electrode rinse systems. Veerman et al. (2010a) compared performances of various types of electrode systems looking into several aspects such as electrochemical reactions, energy consumption, transport through the membranes of the electrode system, and reversibility. Up until today, the most common electrodes used are titanium mesh electrodes coated with Ru–Ir mixed metal oxides or less expensive carbon materials. These electrodes can be utilized as an anode as well as a cathode, thus allowing current reversal. As for electrolyte solutions (electrode rinse solution), iron-based redox couples such as FeCl2/FeCl3, hexacyanoferrate(III)/hexacyanoferrate(II), and Fe(III)–EDTA/Fe(II)–EDTA in a NaCl–HCl supporting electrolyte, and hexacyanoferrate(III)/hexacyanoferrate(II) are typically used. The redox couple Fe2+/Fe3+ is preferable as it is stable at a lower pH condition (acidic) and in the presence of inert gas, such as nitrogen, which minimiszes potential contamination of concentrated and dilutes solutions flowing into the stack (Scialdone et al. 2013).

Gibbs free energy of mixing equation can be to calculate the theoretical total of energy produced after dilute and concentrated solution streams are mixed (Daniilidis et al. 2014; Nijmeijer and Metz 2010):

(1) Δ G mix = i [ E i , c + E i , d E i , b ]

(2) Δ G mix = i [ c i , c V c . R . T . ln ( x i , c ) + c i , d d . R . T . ln ( x i , d ) c i , b V b . R . T . ln ( x i , b ) ]

where E is the free energy (J), c is the species concentration (mol m−3), x is the mole fraction, V is the volume (m3), R is the ideal gas constant (8.314 J mol−1 K−1), and T is the temperature (K). The subscripts of c, d, and b are concentrated, dilute, and brackish water, respectively. The generated energy (W, J) is the voltage produced (U, V) at a specific current density (I, A) and interval time (Δt, s). It is calculated using Equation (3) (Nijmeijer and Metz 2010):

(3) W = t 0 t end U . I . Δ t

where t 0 and t end are the time of current start and stop, respectively. Energy efficiency is then calculated as the ratio between the actual energy W and the ideal Gibbs energy.

Two important keys for measuring the RED system performance for power generation are (i) power density, i.e. the power produced per unit membrane area (W m−2) and (ii) energy efficiency where the proportion of energy available is converted into electrical energy (Yip and Elimelech 2014). Other parameters include open-circuit voltage (OCV, V) which indicates the highest voltage available from the system when no current is permitted, which can be calculated using the Nernst equation. IEMs are only permeable to the oppositely charged species (counter-ions) while retaining the similarly charged ions (co-ions) and water (Fan and Yip 2019). Therefore, the OCV is multiplied with the apparent membrane permselectivity (α) to rectify the nonideal behavior of the IEMs. α is defined as the ratio of membrane voltage to the theoretical voltage (Długołecki et al. 2008; Kingsbury et al. 2017; Vermaas et al. 2012):

(4) OCV = α × N m . R . T z . F ln C H γ H C L γ L

where N m is the total number of membranes, Z is the valence of the ionic species, F is the Faraday constant (96,485 C mol−1), C is the salt concentration (mol L−1), and γ is the activity coefficient of the salt. The subscripts H and L refer to HCC and LCC, respectively. Potential differences over each pair of membranes need to be totaled to obtain the system’s whole potential.

OCV and stack resistance (R stack) are typically obtained from fitted data (at I = 0 A) and the slope of the I–V curve, respectively. Gross power density (P d ) and current density (A d ) can be then determined and fitted as a parabola (Avci et al. 2020b). Due to the ohmic behavior of RED, the maximum power density can then be calculated from OCV and internal resistance of the RED system using Equation (5):

(5) P d , max = OC V 2 4 . R i . N m

where R i is the total electrical area resistance per stack (Ω cm2). The sum of three different sections, i.e. cell resistance, change of concentration in the bulk solution, and the boundary layer gives total resistance of the system, as shown in Equation (6):

(6) R i = R ohmic + R Δ C + R B L

where R ohmic is the ohmic area resistance per cell (Ω cm2), R ΔC and R BL are the area of resistance when concentration change in bulk solution and boundary layer, respectively. The values of R i and R ohmic can be obtained from the experiment using voltage produced, electrical current, and Ohm’s law. Meanwhile, the nonohmic resistances (R ΔC and R BL ) are the reduced electromotive forces (emfs) when the salinity gradient over the membrane decreases (Daniilidis et al. 2014).

The energy productivity is restricted by ohmic loss in the RED stack, unexploited existing energy. It might be due to the passage of co-ions and water across the IEMs in the case of membrane imperfection (perfect membrane only allows opposite charge ions to cross over). The Coulombic efficiency is used to determine the quantity of charge transport used in electricity production, which shows the loss caused by the charge transport of co-ions (Equation (7)):

(7) η C E = I Δ c . Q . F

where Δc is the concentration difference between inflow and outflow of the feed water (mol m−3) and Q is the volumetric flow rate of the feed stack (m3 s−1).

3 IEMs for RED

IEMs are the most vital element in the RED stack, which are currently one of the factors constraining RED performances for SGP. IEMs typically consist of a hydrophobic substrate that has been immobilized with fixed functional groups and movable counter-ions. It is used to split up a concentrated solution from a dilute solution by selectively allowing the passing of oppositely charged ions (counter-ions) while blocking similarly charged ions (co-ions). In general, the transfer of corresponding ions occurs due to polar solvents, such as water, leading to the dissociation of functionalized groups attached to the IEMs (Kamcev et al. 2017). IEMs are broadly classified into CEM, AEM, amphoteric IEMs, bipolar membranes (BPMs), and mosaic IEMs (Figure 4). These are based on the charge and distribution of fixed ionic groups (Luo et al. 2018). CEMs and AEMs are known as charge selective monopolar IEMs, and by combining a CEM and an AEM, a BPM is formed. The BPM offers the possibility of mixing multiple functionalities into IEMs such as antifouling, water dissociation, and the separation of monovalent and divalent ions (Yao et al. 2020). Amphoteric IEM possesses weak acidic (negative charge) groups and weak basic (positive charge) groups, which are randomly dispersed in the membrane matrix. It permits both cations and anions simply by controlling the pH and has potential as an antifouling material that prevents the adsorption of organic molecules and biological macromolecules on the surface (Matsumoto et al. 2003). In comparison, the mosaic membrane contains anion and cation exchange groups arranged in parallel to one another. This arrangement allows both anions and cations to flow through the membrane through the individual channels (Xu 2005). The mosaic membrane is suitable for separating salts from water-soluble organic substances, treating waste streams from various industries as it possesses negative salt rejection and osmotic pressure (Linder and Kedem 2001). However, the preparation of mosaic membrane is complex and not suitable for commercial production. Besides, it is difficult to fabricate a thin selective layer of the mosaic membrane while controlling domain size and ensuring no interfacial leaks between the domains (Wang et al. 2016). This review focuses on monopolar CEMs and AEMs as they are among the most common membrane type used in RED due to their easy manufacturing, low price, and high stability and lifetime. However, as the RED technology is shifting from an artificial feed solution lab-scale system into a large-scale system using natural feed water or wastewater, the use of BPM particularly in suppressing inorganic scaling on membranes will also be discussed.

Figure 4: 
Various IEM types: anion exchange membranes (AEMs), cation exchange membranes (CEMs), bipolar IEMs, amphoteric IEMs, and mosaic IEMs.
Figure 4:

Various IEM types: anion exchange membranes (AEMs), cation exchange membranes (CEMs), bipolar IEMs, amphoteric IEMs, and mosaic IEMs.

CEMs possess functional groups of negative charges namely sulfonic acid (–SO3 ), phosphonic acid (–PO3H), and carboxylic acid (–COO) that selectively allow the transport of cations but refuse anions (Ran et al. 2017). Assorted polymer materials including polyethersulfone (PES) (Klaysom et al. 2011b; Zhao et al. 2018), polyetherketone (Güler et al. 2013), polybenzimidazole (Pal et al. 2016), polyphenylene (Yoon et al. 2019), polyimide (Shukla et al. 2016), and polyvinylidene fluoride (PVDF) (Reig et al. 2015) have been explored as CEMs backbones. The discrepancy in the topology and morphology of polymeric ionomers such as block CEM, dense functionalized CEM, and chain-type CEMs has been proven to alter the performance of CEMs substantially.

AEMs have positively charged functional groups anchored onto the polymer backbones such as quaternary ammonium (–NH3 +), primary amine (–RNH3 +), secondary amine (R2NH+), tertiary amine (–R3N+), imidazole, and guanidinium cationic groups which selectively permit anions transport while omitting cations (Ran et al. 2015). The type, quantity, and distribution of the ion-exchange groups can significantly affect the membrane physicochemical properties and performance (Fernandez-Gonzalez et al. 2017). Anion conductivity, chemical stability, and dimensional stability have been the major issues faced by AEMs and significant research efforts have been dedicated, which can be categorized into two main groups: (i) synthesis of novel anion-conducting groups (Hagesteijn et al. 2018; Lee et al. 2019; Ran et al. 2017) and (ii) design of a specific polymer architecture (Fan and Yip 2019; Lee et al. 2017; Hong and Chen 2014; Wu et al. 2020).

IEM contains high concentrations of fixed charges, and the permselectivity of an IEM is due to Donnan equilibrium where the IEM matrix is assumed as a homogeneously distributed fixed charge solution. An electrical potential (Donnan potential) at the membrane–solution boundary is accountable for the rejection of co-ions from the membrane matrix (Donnan effect) (Hong and Kim 2016). As shown in Figure 5 (Domenech et al. 2012), the Donnan exclusion effect occurs when the charge of co-ions is similar to that of the polymer functional groups, thus repels the ions of the same charge (Kamcev et al. 2017). Therefore, the gradient of ion concentration and Donnan effects driving forces in opposite directions allows the balance of ion penetration inside the matrix.

Figure 5: 
Donnan exclusion effect. Redrawn from Domenech et al. (2012).
Figure 5:

Donnan exclusion effect. Redrawn from Domenech et al. (2012).

3.1 Characteristics of IEMs

The development of high-performance RED membranes remains a challenge that delays RED technology utilization for energy production. Therefore, it is essential to understand the properties of the required membrane to maximize the overall efficiency of the RED system. Among the important IEMs characteristics that need to be considered for power generation are electrical resistance, ion exchange capacity, permeability, permselectivity, and water content (Ortiz-Imedio et al. 2019). Although these characteristics can be easily controlled during the membrane manufacturing process, it is difficult to study their effects independently as they correlate with each other. The electrical resistance of IEMs can be modified by changing the membrane charge density and the fixed charge groups, which strongly affect the ion selectivity through interactions with different counter-ions. Typically, IEMs with higher conductivities and charge densities will have a greater ion transport that can be obtained by enhancing the ion-exchange capacity (IEC) of IEMs (Fontananova et al. 2017). Nevertheless, higher IEC increases the swelling effects on the IEM and reduces the permselectivity. Membrane permeability affects the transfer of different ions in the membrane phase. By having membranes with higher porosity and larger pore size distribution, less effective Donnan exclusion can occur, where co-ions enter the membrane phase like counter-ions. As a result, IEM with lower porosity, smaller pore size distribution, and higher fixed charge density (CDfix) is required to achieve a higher transport number for counter-ions and better permselectivity (Klaysom et al. 2011a).

3.1.1 Membrane homogeneity

IEM can be categorized into homogeneous or heterogeneous membranes, based on the distribution of ion-charged groups or ion-conducting functional groups in a continuous membrane matrix. In a homogenous membrane, fixed charged groups are uniformly dispersed in the membrane matrix. The most common method used in the preparation of homogenous membranes is polymerization and polycondensation of functional monomers (e.g. phenylosulfonic acid with formaldehyde) (Kariduraganavar et al. 2006; Xu 2005). In comparison, ion-exchange groups in heterogeneous membranes are commonly contained in small domains distributed throughout the membrane matrix. Heterogeneous membranes can be easily fabricated through compressing and melting of ion-exchange resin with polymer granules However, this process can lead to unevenly structured and clustered distribution of ion-exchange groups in the membrane (Zabolotskii et al. 2005). Thus, the charge density is fairly low. The membrane possesses higher electrical resistance owing to a longer mobile ion route due to the uncharged domains in the heterogeneous morphology (Hosseini et al. 2014). However, these limitations do not hinder the utilization of the heterogeneous membrane for RED as the production method and cost of heterogeneous membranes are much simpler and lower compared to those of homogeneous membranes (Merino-Garcia et al. 2020).

3.1.2 Water uptake, swelling degree, and swelling ratio

Water content can significantly affect membrane dimensional stability and ionic transport properties. Water uptake (WU) also known as swelling degree (SD) is defined as the amount of water that has been held up by the membrane and measures the change of the mass of membranes when exposed to water (Güler et al. 2013):

(8) WU = W wet W dry W dry × 100 %

where W wet and W dry are the weights of wet and dry IEM in g, respectively. The swelling ratio (SR) is related to the changes in dimension or measure of the linear expansion of the membranes when exposed to water. It can be quantified experimentally by measuring the length difference between wet and dry membranes, as shown in Equation (9) (Hagesteijn et al. 2018):

(9) SR = l wet l dry l dry × 100 %

where l wet and l dry are the lengths of wet and dry IEM, respectively.

Although higher WU might lead to better conductivity, it also infers a weak mechanical structure and frequently leads to poor permselectivity (Hong et al. 2015a). The WU or water absorption highly depends on membrane material, ionic functional groups, cross-linking degrees, and surrounding solution conditions (Nagarale et al. 2006). The WU and SD are directly correlated with the structural development of the fixed charge group (i.e. IEC) (Villafaña-López et al. 2019). If the WU of the membrane is too high, the ionic functional groups in the polymer matrix can be interrupted and the fixed charge group density declines (Geise et al. 2014). Therefore, swelling might occurs significantly as more ionic functional groups are present, leading to lower resistances.

3.1.3 Electrochemical properties

IEC signifies the amount of fixed charge groups available in the membrane matrix and is defined as milliequivalents (meq) of charged groups per gram of dry membrane. Several methods can be used to measure IEC, namely titration, ion chromatography, spectrophotometric determination of NO3 ion concentrations, and ion-selective pH determination of H+/OH ions in solution (Karas et al. 2014). For the acid/base titration method, various procedures are reported depending on acid/base strengths used and soaking times. In general, the number of counter-ions (i.e. cations for CEM and anions for AEM) can be calculated after turning CEMs and AEMs into H+- and Cl-saturated form, respectively. The IEC value is determined using the following equation:

(10) I E C = V t i t r a n t C t i t r a n t W d r y × 100 %

where V titrant is the volume of titrant, W dry is the dry weight of the membrane in g, and C titrant is the strength or concentration of titrant used to determine IEC.

However, human errors can occur during the titration mainly in determining the changes of color at the termination point that significantly affects the IEC value. To improve the titration method’s accuracy, a potentiometric technique that measures the potential of an electrode sensitive to one specific ion available in the solution can be used. Ion-selective electrode potential is determined against a suitable reference electrode, connected to the solution by an appropriate salt bridge. Besides that, spectroscopic techniques using Fourier transform infrared or nuclear magnetic resonance can be used to provide information on the functional groups that can dissociate in the polymer membrane structure. Karas et al. (2014) found that UV–visible (UV–vis) spectroscopy produces a more accurate IEC value compared with theoretical IEC determined from elemental analysis among the methods available. However, the theoretical value obtained does not provide real IEC, such as the number of functional groups in the polymer structure accessible to ion exchange, which is typically different from the total content.

Membrane charge density is another parameter used to characterize a charged membrane where the fixed charge group is typically attached to the polymer backbone. It indicates the concentration of fixed charge groups per unit volume of water (meq L−1) (Długołecki et al. 2008). This value is highly dependent on the IEC and the SD. As the membrane swells, the space between the ion-exchange groups increases, lowering the CDfix. The fixed negative charges in CEM are in electrical equilibrium with the mobile cations while the opposite relationship occurs in AEM. The movement of counter ions across the membrane depends on the CDfix and the electrolyte concentrations in contact with the membrane. As a result, these can highly affect the permselectivity and electrical resistance of the membrane. Avci (2020b) compared the performance of Nafion 117 and Nafion 115 with CMX (Neosepta) and Fuji-CEM 80050 (Fujifilm) CEM for NaCl and NaCl+ in high salinity conditions. Nafion 117 and Fuji-CEM had the highest and lowest maximum power density, respectively. This result was due to the effects of fixed charged density where high CDfix in Nafion 117 (8.0 mol L−1) assisted retain the exclusion capacity while low CDfix of Fuji-CEM (3.2 mol L−1) made it to be susceptible to high salinity conditions.

3.1.4 Permselectivity

Permselectivity indicates the charge selectivity of the IEM and its capability to move counter-ions and exclude co-ions selectively. The ideal permselectivity for IEM should be 1, which indicates that the co-ions are entirely barred from transferring across the membrane matrix (Mei and Tang 2018). In reality, the membrane permselectivity is typically less than 1 as the transport of co-ions is unavoidable. It is calculated based on the ratio of electrical potential difference measured across a membrane sample (Emeasure) to the theoretical value of an ideal membrane (Etheoretical) under a given concentration gradient as per Equation (11).

(11) Permselectivity = E measure E theoretical × 100 %

3.1.5 Membrane resistance

Membrane resistance is referred to as the deterrent of the polymer matrix to ionic current transportation, which signifies an important impact on a RED stack’s internal resistance. The electrical resistance of IEM is directly correlated to the energy utilization in electrodialysis (ED) processes or maximum power output in RED (Urano et al. 1986). Greater membrane resistance reduces the voltage and hence lowers the available power output (Benneker et al. 2018). The value is highly affected by operating temperature, where a higher temperature substantially increases the ion mobility, thus reducing the resistance value. The resistance value can be obtained from the IEC and mobility of ions within the membrane matrix. Although the specific membrane resistance is reported in Ω cm, the unit commonly used in the literature is Ω cm2 as the area resistance of the membrane.

The movement of ions in solutions leads to the production of electrical currents and therefore, it can be measured similarly as in electrical circuits. Ohm’s law describes the relationship between current, imposed electrical field, and the resistance to current flow in the system. It is widely used to determine the membrane resistance through the slope of the I–V curves (Equation (11)).

(11) V = I R

where V is the voltage (V), I is current (A), and R is the resistance (Ω) to current flow. However, Ohm’s law is not suitable for fouling characterization as it could not identify microscopic and structural changes of fouling because electrical responses of fouling phenomena to a direct current (DC) signal provide information only on ionic movement (Park et al. 2006).

Electrical impedance spectroscopy (EIS) can be used to determine the impedance (ideal resistance in Ohm’s law) over a wide range of frequencies in which alternate current (AC) potential (or current) is applied to an electrochemical cell, and the current (or potential) response through the system is measured. It can quantitatively measure the electrical resistance in bulk and interfacial regions of solid and liquid electrolyte materials, including membranes (Fontananova 2016) and serves as a nondestructive characterization of membranes particularly in investigating the concentration polarization and fouling phenomena of IEM (Fontananova et al. 2012). The information on charge storage properties of the fouling layers themselves allows better insight on ongoing fouling attractions between foulants and membrane.

3.1.6 Membrane durability

Chemical stability indicates the robustness of membranes in various acidic or alkaline solutions. Usually, the chemical and thermal stabilities of CEMs are better than those of AEMs due to the quaternary ammonium groups in AEMs. In RED applications, the chemical stability of IEMs is not critical compared to in ED applications as the feeding solutions (river water and saltwater) are usually nearly neutral, and the pH of the solution does not change significantly during the operation (Tufa et al. 2018). The thermal stability required for IEMs in RED is not as high as other applications, such as fuel cells. It is typically operated at room temperature with a 30 K seasonal variant (Hong et al. 2015a). This stability is highly reliant on the degree of cross-linking, thermal properties of inert polymers, and supporting fabric used such as polyvinyl chloride and polyethylene.

Higher mechanical strength is essential for maintaining good resilience during the flow of feed. Besides the osmotic pressure caused by a concentration gradient, hydraulic pressure over the membrane can also be formed. The mechanical strength of the IEMs for RED is not as important as in other applications such as pressure-retarded osmosis, where it needs to withstand tremendous hydraulic pressure (Han et al. 2014). Commonly, cross-linked membranes have better mechanical strength, but their membrane resistances tend to increase too. Thus, for RED application where lower membrane resistance is more important, membrane resistance properties should not be compromised to the improved mechanical strength.

Membrane resistance and permselectivity can be impacted by IEC and SD. As fixed charges can encourage swelling, higher IEC will lead to a high SD membrane (Zhu et al. 2016). While high IEC can increase membrane permselectivity, a high SD could also have an adverse effect on permselectivity. A trade-off between permselectivity and resistance is needed to attain the optimum power performance of the RED system. For RED stack application in harvesting SGP, the loss of permselectivity might be negligible as the power density increases significantly when the resistance is lowered, owing to the internal resistance role in the production of power (Geise et al. 2013; Hong et al. 2015a). In general, the membrane conditions for RED vary slightly compared to many membrane systems because IEMs typically operate in a neutral condition (pH ∼7) where Na+ and Cl ions can highly affect the ion-transport process. The IEM for RED should be thin with high permselectivity and reasonable mechanical properties. The cheaper polymeric material is considered a good candidate for membrane fabrication to ease the production of commercial IEMs.

3.1.7 Antifouling membrane

RED is anticipated to be less sensitive to fouling as it only permits ions to pass through the membranes while retaining water and other compounds. Nevertheless, fouling is still a major concern as it can trigger several operational issues such as an increase of feed channel pressure drop and an increase in electrical resistance. These conditions can dramatically reduce the effective net power density output, hindering the long-term application of RED for sustainable energy generation. Several fouling types have been observed in RED, i.e. colloidal fouling, organic fouling, biofouling, multivalent ions, and scaling. As compared to CEM, AEM tends to be more affected by fouling due to unfavorable interactions between the positively charged fixed groups in AEM and negatively charged foulant materials present in natural streams. Several techniques such as polymerization, dip coating, and layer-by-layer have been adopted to fabricate monovalent permselective AEMs with antifouling properties (Zhang et al. 2017b; Zhao et al. 2016). By modifying the AEMs surface through the addition of a charged layer, it was found to enhance the ion-exchange capacities, antifouling and antibiofouling properties, and rejection of divalent anions (Kotoka et al. 2020). As there is a trade-off between the stability of the added layer and the monovalent anion permselectivity, it is crucial to have better knowledge and understanding of the IEM fouling phenomenon as this is the key to resolve the problems and driving membrane technology ahead.

3.2 Advances in IEMs for RED

For IEMs to be utilized for separation purposes, the membranes should have high strength and a long lifespan irrespective of electrical resistance and thickness. Nevertheless, for RED application, IEMs should possess low electrical resistance and high preferential permselectivity as it can affect the energy efficiency of the system. As a result, the quest for IEMs development with excellent properties has been stressed by many researchers. The IEMs’ properties largely depend on two factors: (1) membrane material that greatly affects chemical, mechanical, and thermal stabilities and (2) ion-exchange groups’ type, concentration, and distribution that establish the electrochemical characteristics of IEMs. Much effort has been documented on the fabrication of engineered-design membranes for RED applications. The preparation techniques and materials used are highly dependent on their role of either cation-/anion-exchange or monovalent-selective membrane and required structural modifications (Gao et al. 2018; Pawlowski et al. 2019; Tufa et al. 2020). This section highlights the IEMs preparation and modifications for RED applications to present a current insight into RED membranes development.

Güler et al. (2014a) prepared membranes with selectivity for monovalent ion using in situ UV-irradiation techniques of the reactive polymeric coating. 2-Acrylamido-2-methylpropanesulfonic acid was used as the polyanion and sulfonic group, and N,Nʹ-methylenebisacrylamide (MBA) was used as the cross-linking agent. The bulk membrane structure was preserved, and the antifouling properties and the monovalent ion selectivity improved significantly. When the coating layer thickness was more than 110 µm, no significant changes in gross power densities were observed. As a result, a thinner membrane was proposed for future membrane development. Guler et al. (2012) also proposed the use of safer and green polyepichlorohydrin (PECH), which eliminates the need for chloromethylation reaction for AEM preparation. Then, positively charged ion-exchange groups and instantaneous cross-linking of the polymer were carried out using amine 1,4-diazabicyclo[2.2.2]octane. Polyacrylonitrile (PAN) was then added to circumvent the fragility of AEM in water caused by the high SD. Sufficient mechanical strength was observed even at a high SD due to the addition of PAN. However, the permselectivity decreased, and no substantial resistance changes were observed when PECH to PAN ratio (mPECH/mPAN) was higher. This result might be due to the high amount of chloromethyl groups in the PECH active polymer. Better permselectivity and lower membrane resistance were observed when a blend ratio (mPECH/mPAN) of 0.333 and an excess diamine ratio of 4.2 were used. Power density up to 1.27 W m−2, beyond the power output of commercial AMX membranes, was observed due to the faster ions migration when a thinner membrane was used.

Composite membranes are prepared by incorporating the inorganic nanomaterials into the organic polymer matrices, and it is a widely used method to enhance the overall membrane performances. Developing nanocomposite IEMs for RED applications permits additional ion-exchangeable functional groups into the membrane matrix to improve its electrochemical properties. Numerous works have focused on the combination of various inorganic nanoparticles and organic materials.

Susanto et al. (2020) utilized 0.5–2 wt% fly ash to improve the performance of polyvinyl chloride (PVC) membranes for RED. A preliminary study shows that the composite PVC membrane possesses significant improvement in terms of SD (83.78%), IEC (0.163 meq/g), and conductivity (8.7 × 10−2 S cm−1). The increased IEC and conductivity were due to the adsorption ability of fly ash, which promoted interactions between ions and membrane surface and the conductivity of minerals in the fly ash.

Hong and Chen (2014) prepared an organic–inorganic nanocomposite CEM by combining functionalized iron(III) oxide (Fe2O3–SO4 2−) as an inorganic filler with sulfonated poly(2,6-dimethyl-1,4-phenylene oxide) (SPPO) polymer matrix to promote the energy generation. The sulfonated Fe2O3 introduced ion-exchangeable functional groups which can enhance IEC and ion conductivity. The ion-exchange resin was homogenized throughout the polymer matrix to ensure that the fixed charged groups were evenly distributed and to avoid distinct regions of uncharged polymer (Figure 6i). An optimized amount of 0.5–0.7 wt% Fe2O3–SO4 2− increased the permselectivity up to 87.7% and area resistance of 0.87 Ω cm2. The highest power density was observed for 0.7 wt% of Fe2O3–SO4 2− nanocomposite membrane (up 1.3 W m−2). The value was comparatively better than the commercial CSO (Selemion, Japan) membranes (Figure 6ii).

Figure 6: 
(i) SEM micrographs of cross-sectional surfaces: (a) pristine sPPO membrane, (b) composite sPPO–0.2 wt%, (c) composite sPPO–0.5 wt%, (d) composite sPPO–0.7 wt%, (e) composite sPPO–1.0 wt%, and (f) composite sPPO–2.0 wt% and (ii) performance potential ratio (P
2/R) (solid bars) versus maximum gross power density (shaded bars). Reproduced with permission (Hong and Chen 2014).
Figure 6:

(i) SEM micrographs of cross-sectional surfaces: (a) pristine sPPO membrane, (b) composite sPPO–0.2 wt%, (c) composite sPPO–0.5 wt%, (d) composite sPPO–0.7 wt%, (e) composite sPPO–1.0 wt%, and (f) composite sPPO–2.0 wt% and (ii) performance potential ratio (P 2/R) (solid bars) versus maximum gross power density (shaded bars). Reproduced with permission (Hong and Chen 2014).

Hong and Chen (2015) then fabricated custom porous nanocomposite Fe2O3–SO4 2− CEMs by varying the thicknesses, aging time, and inorganic nanoparticle loadings through the two-step phase inversion process. This method permits the membrane morphology control, particularly on the porosity of the polymeric membranes, leading to lower resistance and higher electrical current in the RED system. The thinnest, porous nanocomposite membrane (30 μm) containing 0.7 wt% of Fe2O3–SO4 2− possessed the best performance with a permselectivity of 85.6% and resistance of 0.82 Ω cm2, achieving gross power density up to 1.4 W m2. A series of nanocomposite CEMs containing sulfonated polymer, poly(2,6-dimethyl-1,4-phenylene oxide), SPPO, and sulfonated silica (SiO2–SO3H) were prepared (Hong et al. 2015b). The permselectivity increased with the loading of sulfonated silica up to 0.5 wt%, then started to decrease, perhaps due to the deficit of available functional groups when more sulfonated silica was added. Low IEC and CDfix were observed, reducing selective ion transport. In addition, more swelling occurred when the content of ionic groups presented in the matrix was higher, thus lower resistance was observed. When 70 nm of SiO2 was used, the resistance was 10% lower compared to 15 nm of SiO2. This observation might be due to the increase of ion-accessible functionalized groups surface at the polymer and filler boundary region when the size of the nanoparticle increased, therefore increasing the ability to exchange ions. A lower degree of agglomeration for bigger particles created a bigger interface void area and poor polymer–filler interface, which provides extremely open-pore structures and porous channels for better ions contact (Hong et al. 2015b).

The nanocomposite CEM was also fabricated by blending SPPO with oxidized multiwalled carbon nanotubes (O–MWCNTs) (Tong et al. 2016). Significant improvement in the antifouling properties and power produced was observed when 0.3–0.5 wt% of O–MWCNTs were used compared to pristine SPPO CEMs. This result indicates the potential of O–MWCNTs to enhance the properties of IEMs for applications in electrochemical systems such as RED. As the ionic resistance of an IEM displays a great dependence on the concentration of the external solution and hydrodynamic environment, it is vital to apprehend the IEM process so that the ionic resistance can be precisely modeled.

Zhang et al. (2017a) established a new model by considering membrane properties that can influence bulk resistance and hydrodynamic conditions, simultaneously influencing the nonohmic behavior of membrane resistance. The developed model highlighted the dependency of external solution’s concentration and explicitly established a correlation between the assessed membrane resistance and current density. The DC and AC of membranes obtained from modeling results were checked with experimental data measured at various solution concentrations and current densities. The model precisely predicted the behaviors of sulfonated poly(2,6-dimethyl-1,4-phenylene oxide), Fumasep-FKS, and Fumasep-FAS membranes.

A pore-filling membrane comprises a porous and thin substrate for mechanical strength, and the pores of the substrate are filled with an electrolyte polymer for ion conductivity enhancement. Kim et al. (2015) prepared one AEM and two CEM pore filling membranes and compared their performances with commercial membranes. AEM1 consisted of a polymer cross-linked using N,Nʹ-bis(acryloyl)piperazine and (vinylbenzyl)trimethylammonium chloride (1:11.01 mol mol−1) in a porous polyolefin substrate, and KIER–CEM1 contained a polymer cross-linked using N,N′-ethylenebis(acrylamide) and vinylsulfonic acid (1:8.83 mol mol−1). Both membranes were used to enhance the anion and cation movement. KIER–CEM2, which consisted of a cross-linked polymer with N,N′-ethylenebis(acrylamide), and acrylamido-2-methyl-1-propanesulfonic acid (1:3.25 mol mol−1) was then used as a shield membrane to avoid the crossover of redox couples and water back-diffusion toward the electrode sides for cation transport. To enhance the power generation in the RED system, the shielding CEM was designed to have better permselectivity instead of fast ion transport. The prepared membranes possessed better tensile strength up to 130 MPa. The CDfix of IEMs strongly affected its permselectivity and resistance. Interestingly, in this work, although the prepared membranes have similar CDfix and permselectivities, their membrane resistances were significantly lower than those of the commercial IEMs. This observation shows that membrane thickness significantly affects membrane resistance rather than CDfix.

Radiation grafting is a common technique used for the development of IEMs for electrochemical applications such as fuel cells. The polymer film is radiated with γ rays or electron beams to introduce radicals and/or peroxide groups into the film. This step will then induce copolymerization of vinyl monomers, forming graft copolymer membranes. Safronova et al. (2016) investigated the use of radiation chemical grafting polymerization of styrene with divinylbenzene on a polyethylene film to prepare CEM. Various amounts of cross-linking agent (up to 3.5%) and degrees of polystyrene grafting (23–32%) were used, and the performances of fabricated membranes were compared with Nafion 117 and Neosepta CMX membranes. The properties of these membranes strongly depended on the amount of the cross-linking agent. Tensile strength increased with increasing cross-linker amount (15–20 MPa). The conductivity of membranes containing cross-linking agents increased the WU and reduced their conductivity at high relative humidity. The diffusion permeability of the cross-linked membranes was low, while a higher cation transport number was observed. The theoretical power obtained was 10% higher than that of the Neosepta CMX membrane.

Willson et al. (2019) prepared a series of radiation-grafted CEM (RG-CEM) via a high-dose-rate electron-beam peroxidation. Two types of precursor films, PVDF and poly(ethylene-co-tetrafluoroethylene) were utilized for grafting purposes. Without co-monomer cross-linking, low permselectivities were observed (≤80%) although the RG-CEMs were conductive to Na+ cations. The type of cross-links highly affected the performance of RF-CEMs, where higher permselectivities but decreased Na+ cation conductivity were observed when inflexible divinylbenzene cross-links were used. The use of a more flexible cross-linking monomer such as bis(vinylphenyl)ethane increased the permselectivity up to 92% without decreasing the Na+ ion conductivity.

Other than incorporating functional materials into membrane matrix and chemical grafting, the design strategy of creating profiled membrane, which is known as microstructured membranes, corrugated membranes, or patterned membranes with a topography design that allows the channeling of feed solutions have been proposed (Pawlowski et al. 2019). Through this design, the need for nonconductive spacers can be eliminated as the membrane can act as spacers too. This design induces hydrodynamic changes and leads to a better ion transport rate (Pawlowski et al. 2016). The use of a single ion-conducting body also avoids the possibility of ion transport obstruction. Hot-pressing (Vermaas et al. 2011a) and steel mold (Güler et al. 2014b) techniques have been proposed to fabricate this membrane type. Two commercials heterogeneous Ralex CMH and AMH membranes have been fabricated using the hot-pressing (or hot embossing) method. Higher power densities obtained were compared to membranes that contain spacers due to lower resistance. Different steel mold casting shapes have been utilized as it is hard to preserve the microstructure form when the membrane film is removed from the mold after the hot-pressing. Ridges, waves, and pillars are the three types of structures that have been investigated in a RED stack by Güler et al. (2014b). An improvement in power density of up to nearly 40% was observed for pillar-structured membranes used in the RED system compared to the typical flat membrane-spacers due to better structure geometry, lower deterrent, and better water distribution flow on the membrane surface.

A wave-patterned IEM using nonconductive materials supported by thin (16-μm-thick) pore-filling membranes was prepared by Choi et al. (2020) to develop stable electrolyte channels on wave lines with mirror images. The resistance of AEMs and CEMs increased due to nonconductive material without affecting the permselectivity. The net power density of the patterned membrane was observed to enhance at a higher flow rate. Through this membrane design, the net energy efficiency can be further boosted by optimizing the channel’s design or adding ionic conductivity properties to the patterned structure.

Gurreri et al. (2013) used computational fluid dynamics tools to forecast the performance of profiled membranes with various pillar shapes. Lower pumping power was expected when profiled membranes were used compared to the flat membranes with a spacer. Nevertheless, stagnant zones responsible for the concentration polarization phenomenon were also observed near the pillar corrugations. The performances of the RED system were highly dependent on the membrane’s pattern profile and thickness. Commonly, a thinner membrane is preferred for minimum membrane resistance in ion transport. However, the membrane should be robust enough for the flow channel without spacers and to circumvent the possibility of membrane deformation. The massive possibilities of creating new profile geometries, particularly with the aid of three-dimensional printing, provide an exciting opportunity for the improvement of RED (Pawlowski et al. 2019).

As the establishment of ion channels is vital in developing superior performance IEMs, the orientation of ion channels in the desired direction was proposed by Lee et al. (2015a) to enhance the ion conductivity of a membrane. The sulfonic groups attached to poly (2,6-dimethyl-1,4-phenylene oxide) (SPPO) were arrayed on a glass plate according to the applied electric field direction while drying the polymer solution. The membrane’s ion conductivity increased up to 12 times without deprivation in ion selectivity, highlighting the advantages of using an electrical arrangement of nanoscale ion channels to enhance the ion conductivity of an IEM at a given IEC. The use of DC electric field for the formation of low-resistance membranes was proposed by Lee et al. (2015b) for effective CEM alignment where the membrane was first fabricated under a pulsed electric field (AC mode). Lower membrane electrical resistance (MER) was observed for the AC membrane (0.86 Ω cm2) compared to those for the DC membrane (2.13 Ω cm2) and pristine membrane (4.30 Ω cm2). The highest power densities of 1.34 W m−2 and 1.14 W m−2 were observed for AC and DC membranes, respectively. The values are higher compared to that of the commercial CEM (1.07 W m−2).

Until today, the performance of RED in practical application is constrained by the presence of multivalent ions, in which IEMs with high selectivity for monovalent/multivalent ions are vital. However, many research works indicated that there is a trade-off effect between permselectivity and resistance. As a result, the development of IEM with low resistance and high permselectivities is significant. Table 1 summarizes recent studies on IEM development for RED via various techniques that aim to enhance power density and minimize fouling to ensure possible practical application of RED for power generation.

Table 1:

Comparison of recent progress IEM development for RED (2019–2020).

Membranes/conditions Properties t (μm); IEC (meq g dry−1); WU/SD (%); PS (%); σ Na+ (mS cm−1); R i (Ω cm2); P d (W m−2); OCV (V) Significant findings References
Modified commercial membranes
Polypyrrole-chitosan (PPyCS) coated Fujifilm CEMT1; Fuji CEMT1-PPyCS-0.05 tested in 0–30% MgCl2 in NaCl t (122 ± 1); IEC (2.1); PS (47.4); R i (2.41–8.3); P d (1.5–0.6); OCV (0.24–0.17) Chemical polymerization of PPyCS on CEMs led to rigid and tight structure, which restricted the transport of Mg2+. A high amount of pyrrole could reduce the SD and IEC. Tufa et al. (2020)
Nafion-0.2 sulfonated single-wall carbon nanotube (SWCNT)/Fumatech FAA-3 AEM tested in 35 g L−1/0.35 g L−1 NaCl t (40); WU (25); P d (0.0794 at a flow rate of 75 mL min−1); OCV (0.115) Salt deposition can be restricted by modifying the Nafion membrane with micron length SWCNT. Sulfonation ion at the end side of SWCNT could interact with hydronium ion and provide an efficient channel and continuous network to restrict the deposition of salts. The impact of ion transport became more apparent at the minimum flow rates that might indicate the significant drop of electromotive forces. The use of a porous and thicker spacer (500 μm) restricted the fouling and increased the power density at a higher flow rate compared to a dense 300 μm spacer. Shah et al. (2020)
Home-made membranes
Electrospun PVC CEM membranes using sodium dodecyl sulfate (SDS) t (70 ± 1); SD (1.7 ± 0.1); IEC (1.7 ± 0.1) Electrospun-membranes had thinner and homogenous structures. Excellent pH resistance and apparent cross-linking. However, adding SDS affected the mechanical integrity of the membranes. Jaime-Ferrer et al. (2020)
Sulfonated polyethersulfone (sPES) CEM dense (sPES-D) and asymmetric porous (sPES-P) structure prepared through solvent evaporation or immersion precipitation phase inversion method. Tested in brackish water (0.1 M NaCl)/hypersaline brine (4 M NaCl) sPES-P: t (83 ± 6); IEC (1.15 ± 0.02); SD (67.2 ± 0.5); PS (84), P d (3.65); OCV (0.291)

sPES-D: t (63 ± 6); IEC (1.19 ± 0.04); SD (28 ± 0.4); PS (95); P d (3.92); OCV (0.314)
The use of 5 M NaCl electrolyte for immersion precipitation coagulation bath helped form a self-standing membrane due to electrostatic interaction between fixed charged groups and electrolyte solutions. Asymmetric sPES-P had very low resistance, especially for high ionic gradients and low permselectivity, while SPES-D had low resistance compared to commercial membranes but displayed high permselectivity. (Avci et al. 2020a)
Radiation grafted using a high-dose rate electron-beam peroxidation CEM

RG-CEM (no cross-linker)

RG ETFE-divinylbenzene (DVB)

ETFE-bis(vinylphenyl)ethane (BVPE)
t (91 ± 2); IEC (1.88 ± 0.02); WU (42 ± 1); PS (80); σ Na+ (15.0 ± 4.6); R i (2.41–8.3)

t (71 ± 2); IEC (1.92 ± 0.01); WU (15 ± 1); PS (97); σ Na+ (6.4 ± 1.9); R i (2.16)

t (91 ± 2); IEC (3.07 ± 0.02); WU (41 ± 1); PS (92); σ Na+ (15.7 ± 1.4); R i (0.58)
Without the cross-linker, RG-CEMs were conductive (to Na+ cations), but the permselectivities were ≤ 80%. In contrast, with (inflexible) divinylbenzene cross-links, higher permselectivities but undesirable decreases in Na+ cation conductivity were observed. The use of flexible BVPE cross-linker led to high permselectivity without a significant conductivity decrease. BVPE enhanced IEC without increasing water uptakes. Willson et al. (2019)
Anionic monomeric electrolytes [(2-acrylamido-2-methyl-1-propanesulfonic acid (AMPS) and 2-acrylamido-2-methyl-1-propanesulfonic acid sodium salt solution (AMPS–Na)] were used to fabricate pore-filling cation-exchange membranes (PCEM) using roll to roll (R2R) equipment. Tested using 0.5 M NaCl and 0.017 M NaCl T (16); IEC (1.80 ± 0.05); WU (49.5 ± 2.8); PS (95.7 ± 0.03); R i (0.42 ± 0.06); P d (1.95 at 50 mL min−1) Simple, energy-efficient, and environmentally friendly industrial-scale R2R process using an aqueous photo-curable solution. Repetitive impregnation was used to re-filling CEM with a photo-cured electrolyte polymer in the pores of the substrate. The PCEMs exhibited a lower resistance compared with commercial CEMs, thus, exhibited better RED performances. Yang et al. (2019a)
Pore-filling monomeric electrolytes (3-acrylamidopropyl)-trimethylammonium chloride) PAEM fabricated using roll to roll (R2R) equipment. Tested using 0.5 M NaCl and 0.017 M NaCl T (39); IEC (1.84); WU (56.58); PS (94.6); R i (0.661); P d (1.76 at 50 mL min−1); OCV (0.72) Green AEM with low swelling and high selectivity was fabricated using pore-filling methods with porous substrates and electrolytes where only water was used as a solvent. The R2R comprises six processes: pretreatment, impregnation, lamination, photo-, delamination, and polishing. High permselectivities of PAEM were due to the hydrophobic porous substrate. Fast ion transfer through PAEM was due to low electrical resistance and low boundary-layer resistance in the RED stack, leading to higher power density. Yang et al. (2019b)
Sulfonated poly(2,6-dimethyl-1,4-phenylene oxide), PPO CEM

PPO-0.5 wt% Fe2O3 (PPO-NP1)

PPO-0.7 wt% Fe2O3 (PPO-NP2)

Poly(diallyldimethylammonium chloride) AEM

PDDA -PVA2 (blend ratio 1.0)

PDDA -PVA3 (blend ration 1.5)

CEM/AEM 4 cell stack


t (100); IEC (1); SD (20); PS (84.4); R i (1.09)

t (100); IEC (1.4); SD (26); PS (87.7); R i (0.97)

t (55); IEC (1.35); SD (115); PS (55); R i (1)

t (55); IEC (1.50); SD (148); PS (59.5); R i (0.71)

PS (69.7); R i (8.36); P d (0.48)

PS (73.6); R i (6.72); P d (0.59)
Gross power density is affected by membrane’s resistance area, film thickness, and SD. The permselectivity is influenced by IEC and charge density but showed minor importance to power density in RED. The combination of PPO-NP2/PDDA-PVA3 possesses relatively low permselectivity (73.6%), low total area resistance but still produces the highest power density. Hong et al. (2019)
Wave-patterned IEM membranes supported by 16 µm thin pore-filling membranes (anionic or cationic electrolyte filled into a porous polyethylene substrate) fabricated using a dispenser-type photo-printer t AEM (130 ± 2); t CEM (121 ± 3); PS AEM (94.9 ± 1.5); PS CEM (95.19 ± 0.6); R i AEM (1.41 ± 0.04); R i CEM (1.21 ± 0.04); P d (1.39 at 200 mL min−1); OCV (1.47) Wave-lines of mirror-image patterns aimed to provide stable electrolyte channels when assembled into the RED stack and increased the net power density. Pore-filling membranes had lower ohmic resistances and higher gross power density due to fast ion penetration through the narrow IEMs. Non-conductive photocurable ink was used for patterning to avoid the formation of electrostatic junctions between EMs and CEMs after RED stacking. The single-sided patterned membrane stack exhibited superior net power density (0.9 W m−2) particularly at high flow rates, which indicates better mixing in the patterned membranes than in the flat membranes. Choi et al. (2020)
Nanoporous carbon membranes via thermal cross-linking of polycyclic aromatic hydrocarbons core–rim structured monomers tested in NaCl 0.5/0.01 M t (0.002 ± 0.0005); P d (67) The current RED membranes have limited conductivity which can be improved by reducing the membrane thickness to the atomic level. Molecularly thin carbon membranes with a tight size distribution of nanometer pores could be formed from the core–rim structured monomers. The thermally stable core formed the membrane’s frame, and the thermally unstable rim decomposed and cross-linked the cores. The ultra-permeable to cations properties made it excellent for RED. Liu et al. (2020)
Poly(ethylene)-reinforced AEM based on cross-linked quaternary-aminated polystyrene and quaternary-aminated poly(phenylene oxide); PErC(5)QPS-QPPO t (51); IEC (1.2); R i (0.69–1.67); P d (1.82 at a flow rate of 100 mL min−1); OCV (1.09) To develop low-cost IEMs with high efficiency, PS and PPO were used in the backbone of ionomers and cross-linked with a diamine-based cross-linker. The ionomers were impregnated into a PE matrix support by atmospheric plasma treatment to increase the compatibility and maximize the mechanical properties. Lee et al. (2019)
Ultra-thin polyepichlorohydrin (PECH)-based AEM supported onto nanoporous alumina using spin coating technique; UTFCS5 with 5 μm thickness t polymer (5); t ceramic (40); PS (88); R i (1.2); P d (36.6 at a flow rate of 5 mL min−1); OCV (1.09) The ultra-thin ion-exchange films on the ceramic supporter (UTFCS) with various thickness of polymer layer and porosity of ceramic support were prepared. The UTFCS possessed low electrical resistance, excellent mechanical strength, and higher PD than commercial AEM (up to 24%). Suitable for RED and other fuel cells. Jung et al. (2019)
Chloromethylation and quaternization of polystyrene within a graft copolymer films based on UV-oxidized polymethylpentene AEM with different grafting degrees (GD) and cross-linking degrees (CD) t (51); IEC (1.1–2.9); PS (91–95); R i (0.4–3.0); P d (0.8–0.9) Graft polymerization made it possible to fabricate thinner membranes with similar compositions and no secondary porosity. The properties of the membranes can be changed by varying the cross-linking degree, polystyrene’s content in pristine copolymer film, and chloromethylation reaction conditions. An excellent combination of permselectivity and specific ionic conductivity might reduce the membrane ohmic losses in the RED stack. Golubenko et al. (2020)

4 Effects of operating conditions on the performance of RED

4.1 Flow velocity

Flow velocity is described as the average fluid velocity inside a single spacer-filled channel. The flow rate of the solution directly affects the ion transport rate. In general, higher flow velocity reduces the diffusion boundary layer thickness as the solution mixing rate increases. This condition increases dilute concentration on the membrane surface (Długołecki et al. 2010a; Tanaka 2012) and reduces electrical resistance near the boundary layer, especially at higher current density (Kim et al. 2012). High flow rate also leads to maximum power density as the difference in concentration between fresh water and brine is not affected by the movement of the salt through the membrane, subsequently causes a high potential difference across the membrane (Veerman et al. 2010b).

In contrast, higher feed flow velocities also lead to a minimal residence time of ions in the RED stack. Thus, they probably will not have sufficient time to pass flow across the membrane and will leave the stack before being removed from the dilute stream to the concentrated stream (Karimi and Ghassemi 2016a). Several works were carried out to investigate the dominant factors and establish whether an increased flow rate has a positive, negative, or no impact on ion removal rate (Gurreri et al. 2014; Tedesco et al. 2015a). Many of them produced inconsistent results, mainly due to the difference in types of membrane used, stack cell design, and other operational parameters. As a result, it is important to develop standard protocols for evaluating membrane performances.

Fu et al. (2020) investigated the effects of velocity from 0.24 to 1.19 cm s−1 using eight pairs of repeating units. At first, the increased flow velocity rapidly increased OCV (from 1.25 to 1.39 V) and maximized power density (from 0.357 to 0.389 W m−2). Once the velocity value reached 0.71 cm s−1, OCV increment started to slow down, and the maximum power density decreased to 0.365 W m−2 at 1.19 cm s−1. Although the flow velocity might improve hydrodynamic distribution in the RED stack and directly reduce the concentration polarization in the boundary layers adjacent to the membranes, if it is too fast, it could reduce the residence time of ions in the stack. Therefore, the migration of ions from HCC to LCC would be impaired, increasing the internal resistance.

Recently, Ortiz-Martínez et al. (2020) investigated the effects of Reynolds number and LC concentration toward gross and net power at 0.5 M NaCl (average seawater salinity) for the HCC. A vigorous mathematical model was investigated using Aspen Custom Modeler software to examine various RED optimization scenarios and upscaling scenarios. It was observed that when 0.02 M NaCl of LC solution was employed with ReHC = 3.4 and ReLC = 7, the highest net power density was attained. The modeling results were verified in a laboratory-scale plant. Mehdizadeh et al. (2019) indicated that the flow rate and stack types could greatly affect the power output of the RED system. As the flow rate of the feed solution was increased from 2 to 6 L min−1, the power became more reliant on temperature. The stack resistance was mainly reliant on the membrane physicochemical properties and solution resistances, while the concentration polarization effects decreased upon increasing the flow rate. Kang et al. (2017) investigated the effects of flow rate using a microbial fuel cell (MFC) with a RED stack to produce electricity from salinity gradient and organic wastewater. They showed that by increasing the feed flow rate, maximum power and optimum current densities increased by 17.7 and 16.2%, respectively. Zhu et al. (2015a) observed a compromise between increased diffusion boundary layer resistance and decreased solution resistance of the LC channels. The power was not consistently reduced, even at a lower flow rate. Figure 7 shows that OCV, maximum power density, and maximum current decreased with LC flow rates, indicating that the RED performances were highly affected by the LC flow rate (Zhu et al. 2015a). This finding is similar to the work by Vermaas et al. (2014c), where the solution resistance in LC channels was observed to affect the RED stack’s internal resistance significantly. The solution conductivity in the LC channels was higher because of more ions transportation from the HC into the LC channels at lower flow rates. This condition aided the reduction of the RED stack’s ohmic resistance (Vermaas et al. 2011b). As a result, comparable or improved performance was attained at lower flow rates.

Figure 7: 
(a) Open circuit voltage, (b) maximum power density, (c) maximum current when the flow rates of LC decreased from 50 to 2 mL min−1 while HC solution flow rate was fixed at 50 mL min−1. Reproduced with permission (Zhu et al. 2015). Copyrighted from Elsevier Ltd.
Figure 7:

(a) Open circuit voltage, (b) maximum power density, (c) maximum current when the flow rates of LC decreased from 50 to 2 mL min−1 while HC solution flow rate was fixed at 50 mL min−1. Reproduced with permission (Zhu et al. 2015). Copyrighted from Elsevier Ltd.

The flow rate also has different effects depending on the type of ions. Walker (2010) observed that higher feed velocity had positive effects on removing sodium and sulfate but trivial effects on chloride and calcium removals, which might explain the differences in the initial concentrations. According to Karimi and Ghassemi (2016b), the discrepancy of the flow velocity effects might be due to the experimental method. The first approach used a continuous desalination process with no recirculation of dilute and concentrated streams. The second approach used a batch desalination process in which the dilute and concentrated streams are recycled. Most studies conducted using a nonrecycled batch process reported the benefit of having high superficial velocity toward the total removal of ions, particularly at the beginning of the process. However, this is actually due to the effect of the number of solution passages through the electrodialyzer rather than superficial velocity if the effects of duration are eliminated. As the superficial velocity of the feed solution in a batch system increases, the boundary layer thickness decreases, which then increases the number of ion transfers and the number of passes where the solution is subjected through the system. Therefore, the continuous experiments that consistently report negative effects from higher flow rates are more reliable for investigating the effect of the superficial velocity.

Although a higher flow rate can increase the rate of energy production, it also has adverse effects on hydrodynamic power. The hydrodynamic power losses are usually maximal at high flow rates (Veerman et al. 2010b). In contrast, energy efficiency is expected to be maximal at lower flow rates (long residence times) as the equilibrium concentrations and minimal hydrodynamic power losses can be achieved. Nevertheless, efficiency can decrease if the flow rate is too low because of co-ion transport and osmosis (Veerman et al. 2009). As a result, an optimal flow rate should be chosen based on the requirement of energy production rate and economic feasibility. In practice, a higher feed flow rate is not recommended for the RED system due to higher hydraulic capital and operating costs.

4.2 Temperature

The temperature has a major impact on the transport properties of membrane and solution. Higher feed temperature increases the ion separation rate due to the ion mobility and electrical resistance of the solution. Equation (10) shows the Nernst–Einstein equation, which correlates the operating conditions with ion mobility (Karimi and Ghassemi 2016b; Tanaka 2004):

(10b) μ i = | z i | F D i R T

where μ i is ion mobility, z i is the electrical charge of the ion, F is the Faraday constant, D i is the ion diffusion coefficient, R is the gas constant, and T is temperature. In general, the diffusion coefficient changes inversely with viscosity and linearly with temperature. Higher feed temperature leads to increased ion separation rate due to increased diffusivity (Equation (11)) (Benneker et al. 2018):

(11c) D i ( T ) = D i ( T 0 ) × T × μ ( T 0 ) T 0 × μ ( T )

where D i (T) and D i (T 0) are diffusion coefficients at temperature T and reference temperature T 0, and μ(T) and μ(T 0) are solution dynamic viscosities at temperature T and reference temperature T 0, respectively.

A comprehensive model was proposed by Tedesco et al. (2015a) using seawater and concentrated brine as feed solutions. Higher feed temperature led to a favorable impact on solution conductivity and ion mobility, reducing the overall resistance of the RED stack. The model predicted a 45% increase in the maximum power density as the feed temperature was increased from 20 to 40 °C. A similar trend was observed by several researchers (Długołȩcki et al. 2009; Fontananova et al. 2014). This observation indicates the suitability of the RED system where the salt solutions are available at temperatures higher than 20 °C. The effects of temperature on the electrical resistance of AEMs and CEMs and their interfaces (electrical double layer and diffusion boundary layer) were investigated by Fontananova et al. (2014). The resistance of ions to be transported through the membrane decreased with the temperature due to increasing ion mobility, particularly for CEMs. Brauns (2009) examined the effect of temperature on the gross output of SGP–RED cell power by determining the equivalent conductance at various temperatures. Higher SGP–RED dilute feed temperature increased the conductivity of the dilute solution and increased the cell power output. The simulation model also demonstrated the role of specific parameters such as feed solution temperature, membrane thickness, and thickness of the dilute solution compartment in producing high SGP–RED output. Mei and Tang (2017) studied the influence of feed temperature in HCC and LCC and observed that the power generated was in the following order: heated HCC and LCC, heated LCC only, heated HCC only, and no heating. The internal resistance decreased from 3.61 to 2.67 Ω and the generation of energy boosted up to 60% when both compartments’ temperature increased from 25 to 60 °C. As higher temperature commonly improves ion transfer rate and reduces the energy consumption of RED, a warmer feed stream might lead to better performance of RED and decrease the overall operating cost. However, the operation of the RED membrane should be below 40 °C (extreme temperatures) as it might affect the stability and durability of IEMs.

4.3 Feed properties

The feed stream properties to the RED stack vary greatly in terms of ion concentration and ion types. The ion removal rate and the total ion removal are highly dependent on ionic concentration and composition. These factors directly affect the ion diffusivity and the separation rate through the interactions between counter-ions and fixed charge groups in the IEM matrix (Gómez-Coma et al. 2019). Seawater and river water are the most abundant feed water stream for RED. On average, the salinity of seawater in the world is approximately 35 g L−1. According to Pinti (2011), major concentration of seawater ions are sodium (Na+, 0.48616), magnesium (Mg2+, 0.05475), calcium (Ca2+, 0.0105), potassium (K+, 0.01058), chloride (Cl, 18.95 g L−1), sulfate (SO4 2−, 0.02927), bicarbonate (HCO3 , 0.00183), bromide (Br, 0.00087), carbonate (CO3 2−, 0.00027), boric acid [B(OH)3, 0.00033], tetrahydroxyborate [B(OH)4 , 0.00010], and fluoride (F, 0.00007). However, river water composition is typically different from seawater, and it varies significantly based on continents.

While the composition of multivalent ions is relatively small compared to monovalent ions in natural feedwater, the negative influence on the power generated by RED is substantial. The transport of ions becomes complex when a mixture of monovalent and multivalent ions is present at both sides of IEM where ions are exchanged to obtain chemical potential equilibrium (Post et al. 2009). Owing to the uphill transport of these ions against their concentration gradient, increased resistance of IEMs and decreased power density are typically observed (Avci et al. 2018). The monovalent ion potential decreases to enhance the multivalent ion emf. As the transport is electroneutral, the divalent ion from the dilute solution is transported to the concentrated solution in exchange for two monovalent ions in the opposite direction (Vermaas et al. 2014a). The uphill transport of divalent ions causes a loss in salinity difference of the monovalent species, without no net charge transport. This situation leads to an irreversible loss in the available energy from monovalent ions and lowers the available emf.

The high concentration of multivalent ions in the highly concentrated compartment could also lead to salt precipitation on the IEM. Besides that, the multivalent ion entering the IEM could be transported across the membrane and forms an attachment to a single fixed charge group or multiple fixed charge groups in the IEM (Pintossi et al. 2020). As the electrical resistance for multivalent ions transport is bigger than that for monovalent ions, a higher ohmic drop is observed. In the case of multivalent ion attachment to a single fixed charge group, IEC and permselectivity will decrease as the multivalent ion reverses the charge of the fixed group (Lacey 1980). If the multivalent ions are attached to multiple fixed charge groups, it can neutralize two fixed charged groups in the IEM forming a higher membrane resistance over time due to the decreased free volume available for transport of other ions (Badessa and Shaposhnik 2016).

Recently, many works have been carried out in understanding the effects of multivalent ions on RED performance and approaches to alleviate this issue. Most of the work concentrates primarily on multivalent cations (Moreno et al. 2018a; Rijnaarts et al. 2017) or mixtures of cations and anions. It presumes that the negative effect of multivalent anions is inadequate owing to the smaller hydrated radii of multivalent anions.

Ion concentration has been positively associated with ion removal rate and total ion removal. Higher NaCl concentration increases the rate of ion removal, and total ion removal increases as more ions are affected by the applied voltage (Banasiak et al. 2007; Karimi and Ghassemi 2016a). According to Długołecki et al. (2010b), the membrane resistance depends on concentration under different current conditions and increases with decreasing concentration. Daniilidis et al. (2014) showed decreased energy productivity and permselectivity when highly concentrated salt concentrations higher gradients were used. On the other hand, power density increased when higher salinity gradients were utilized. The effect of concentration bulk solution, spacers, and boundary layer resistances are more pronounced for less concentrated and lower gradients, whereas membrane resistance is highly affected by high ion concentrations. However, this effect might be limited to lower concentrations, as several researchers observed no significant changes at higher concentrations (Mohammadi et al. 2004). Zhu et al. (2015b) observed that the power generated by the RED stack enlarged from 0.6 to 3.6 M when higher NaCl concentrations in the HC solution were used. The power then remained similar due to the capacity limitation of IEMs.

Lee et al. (2013) revealed that the removal of monovalent ions is less than that of divalent ions when both types of ions are available in the feed solution. As the flow rates were increased, the efficiencies decreased linearly due to a shorter contact time between cations and membrane surfaces (Figure 7a). Kabay et al. (2006) also investigated the effect of feed composition on monovalent and divalent ions removal at room temperature and constant flow using different binary mixtures. At a lower voltage, it was easier to remove monovalent cations compared to divalent anions due to the strong attraction of monovalent cations by divalent anions. However, this effect disappeared at higher voltages.

The use of different hydrated ionic radii and charge properties could also affect the RED performances when compared with similar salinity of NaCl solution. Fontananova et al. (2017) observed higher membrane resistance for a solution containing Mg2+ or Ca2+ compared to a solution of Na+ ions due to bigger hydrated radii of Mg2+ (0.428 nm) or Ca2+ (0.412 nm) compared to Na+ (0.358 nm) and slow transport of Mg2+ or Ca2+. Using a similar ionic valence state system, the internal resistance of a NaCl–KCl mixture was higher than that of pure NaCl. The hydrated ionic radii of Na+ and K+ are 0.358 and 0.331 nm, respectively, while the diffusion coefficients in water for Na+ and K+ are 1.334 × 10−9 and 1.957 × 10−9 m2 s−1, respectively (Tufa et al. 2014; Guo et al. 2018). Lower R CEM was observed for K+ due to its smaller hydrated ionic radius size and higher diffusion coefficient. However, the conductivity of HCC and LCC solutions decreased under the same salinity due to the bigger relative atomic mass of K compared to that of Na (Guo et al. 2018). The resistances of high salinity (R H ) and low salinity (R L ) solutions would then increase, especially for the R L as the electrical resistance has significant effects on the LCC (Mei and Tang 2017). Besides that, if SO4 2− is present, an increase of resistance could occur as SO4 2− has lower relative mobility and lower diffusion coefficient than Cl due to the increase in hydrated ionic radius (Fontananova et al. 2017). This condition will then lead to a stronger combination with AEMs (Wu et al. 2009).

Gao et al. (2018) examined the impact of co-ions (K+, Mg2+, Ca2+, and SO4 2−) in NaCl solution and found that the removal efficiency decreases with flow rate (Figure 8a) and the presence of all the ions resulted in lower OCV (Figure 8b). The order of OCV in different systems was NaCl > NaCl–KCl > NaCl–Na2SO4 > NaCl–MgCl2 > NaCl–CaCl2, which is consistent with the Nernst equation (Equation (1)). From the equation, OCV is inversely related to the ionic valence state (Z) and positively related to the membrane permselectivity (α). Based on the membrane permselectivity equation (Equation (2)), the reduction of membrane permselectivity also led to lower OCV values of NaCl–MgCl2 and NaCl–CaCl2 systems. A monovalent NaCl system was observed to have higher OCV compared to NaCl–KCl system. Multivalent cations had a more undesirable influence on OCV due to the higher ion charges. Multivalent cations in the feed solution also merged with sulfonic fixed ionic groups in CEM as electrostatic forces formed between them. As a result, membrane IEC decreased and a low transport rate of multivalent cations across the membrane from HCC to LCC was obtained (Martí-Calatayud et al. 2014; Fontananova et al. 2017).

Figure 8: 
(a) Removal efficiencies as a function of flow rates in synthetic solutions (hardness concentration: 50 mg/L as CaCO3) (Lee et al. 2013). (b) OCV for different aqueous solutions at different temperatures. Reproduced with permission (Guo et al. 2018). Copyrighted from Elsevier Ltd.
Figure 8:

(a) Removal efficiencies as a function of flow rates in synthetic solutions (hardness concentration: 50 mg/L as CaCO3) (Lee et al. 2013). (b) OCV for different aqueous solutions at different temperatures. Reproduced with permission (Guo et al. 2018). Copyrighted from Elsevier Ltd.

Vermaas et al. (2014a) indicated that while the content of multivalent ions (such as Mg2+ and SO4 2−) was lower compared to the content of monovalent ions (such as Na+ and Cl), their impact on power generation is significant. A mixture of 10% of MgSO4 and 90% of NaCl was fed into the RED system and the power density with different types of IEMs decreased between 29 and 50%. In addition, the cation type (Mg2+ or Na+) only significantly affected the voltage of CEMs, while the type of anion (SO4 2− or Cl) only influenced the voltage of AEMs. Interestingly, contrasting findings were observed when the feed solution contained both monovalent and divalent ions, in which divalent ions had a higher removal rate (Klaysom et al. 2011a). This finding might be due to the significant electrostatic correlations between monovalent and divalent ions with opposite charges. As a result, electrostatic interactions hindered the movement of monovalent ions and divalent ions were removed at a higher rate. The monovalent cations are attracted to the divalent anions, and vice versa for monovalent anions. This attraction anchors the monovalent ions, consequently hampering their removal rate (Karimi and Ghassemi 2016a). Based on several works, this effect is commonly noticeable at lower voltages and disappears when a higher voltage is applied (Karimi and Ghassemi 2016b; Klaysom et al. 2011a).

Tufa et al. (2014) observed more than 50% decrease in power density when artificial solutions were used instead of pure NaCl solutions, although the salt concentration was similar. The effect of Mg2+ ions toward the power density was also investigated by Avci et al. (2016), and the power density of MgCl2 (0.06 W m−2) was much lesser when NaCl (1.06 W m−2) was used. The membrane resistance in pure MgCl2 solution increased due to the interaction of Mg2+ with the CEM membrane. The impact of MgCl2 toward commercial Nafion and Fuji-CEM membranes performances was also investigated, and it was found that the gross power density decreased from 1.38 to 1.08 W m−2 and 1.24 to 0.824 W m−2, respectively (Avci et al. 2020b).

Moreno et al. (2018a) found that the type of IEM is crucial, especially when various ion types are available in the feed. The RED performances of different membrane types were investigated, and multivalent ion-permeable membranes were found to be more favorable compared to monovalent permeable membranes. Similarly, Rijnaats et al. (2017) investigated the impacts of IEM types. They found that monovalent-selective Neosepta CMS blocked the transport of divalent cations, which mitigated decreases in stack voltage. At the same time, new multivalent-permeable Fuji T1 was capable of transferring divalent cations with no significant rise in resistance.

In general, several strategies can be used to alleviate the impact of multivalent ions in the RED stack. Depending on the feedwater properties, pretreatment of the feed solutions could be the best option when highly concentrated brines or natural feed water is used. In terms of the membrane, the use of monovalent ion-selective membranes signifies a promising route to minimize both uphill transport and membrane resistance due to low-cost materials and simple membrane fabrication techniques. However, the membrane resistance might increase due to the lower ion mobility of divalent ions inside these membranes. Several techniques were proposed to fabricate new monovalent selective CEMs and AEM that can exclude magnesium and other large multivalent cations from the membrane through thin layer coating, cross-linking, and layer-by-layer method. However, investigation on the long-term operation of these types of membranes should be carried out to look into the possibility of the layer’s damage due to solid particles in the feed streams. The more open structure of IEM was also proposed to allow the free movement of both sodium (mono) and magnesium (divalent) ions through the membrane. Such membranes were found to have low resistance values and reasonable OCV, even at high magnesium concentrations. The high power densities obtained make this type of open structure IEM suitable for the long-term RED application (Moreno et al. 2018a).

5 Fouling issue

As RED only allows ion to pass through the membranes, while the membrane restrains water and additional colloids, the tendency of membrane clogging or fouling occurs in RED are less pronounced in contrast to other processes such as pressure retarded osmosis (Ramon et al. 2011). Nevertheless, when working with real water conditions, fouling of IEMs and spacers can lead to several operating issues and, consequently, reduce RED stack performances (Post et al. 2007). Various fouling forms might happen in RED systems, i.e. organic fouling, biofouling, inorganic fouling (scaling, uphill transport of multivalent cations like Ca2+ and Mg2+, and membrane poisoning), adhesion, and deposition of colloids. The effects of each fouling are different based on the ion-exchange groups of membranes (Warsinger et al. 2018).

In general, many studies found that AEMs are mostly sensitive toward organic fouling and colloidal fouling, while CEMs are found to be predominantly affected by scaling. Colloidal fouling/scaling will cause the feedwater channels to clog and built-up of pressure drop in the feed compartment, disrupting the flow distribution. As organic fouling in real feed waters can lead to AEMs fouling, it is vital to understand and regulate this type of fouling when real feed waters are used to increase the power output of the RED system. The organic fouling is generally contained with negatively charged large molecules attracted by the surface charge of AEMs (Vaselbehagh et al. 2017). Besides, the real feedwaters/natural water bodies generally contain a series of Ca2+, Mg2+, and SO4 2− which can increase the resistance of IEMs and can consequently reduce the power output. Chon et al. (2020) investigated the fouling characteristics of dissolved organic matter (DOM) such as humic- and protein-like substances in the freshwater and seawater compartments and found that fouling formation of the IEMs was caused by partial deposition of DOM within the membrane pores during its passage from the freshwater compartment to the seawater compartment. In addition, the changes of color in AEMs that have been in contact with freshwater were visible than the ones with seawater, which indicates that the deposition of hydrophobic DOM onto AEM could increase the pressure drop of the freshwater compartment, decreasing the power density.

Ratkje et al. (1986) compared the effects of feedwater types (seawater and artificial river water) toward the power density attained from a RED stack with spacers. After 180 days, the power density decreased by 67% because of the biological fouling. Thus, Post et al. (2009) proposed several strategies to overcome biological fouling and pressure drop by reversing the direction of flow and swapping feed waters. The biofouling process could be slowed down by interchanging the feed water and periodically reversing the flow direction. Cifuentes-Araya et al. (2011) also found that the fouling phenomenon can be reduced by applying short electrical pulses frequently. Vermaas et al. (2011a) also proposed the utilization of profiled membranes that combines membrane and spacer functionality to overcome the tendency of fouling at spacers.

The utilization of chemical cleaning agents has also been proposed to overcome the fouling issue, although the efficiency is not as high as other methods. Besides, it can shorten the membrane lifetime, and therefore, physical cleaning such as in two-phase flow cleaning is more preferred. Air is commonly used in physical cleaning due to its simplicity in handling and storing. Vermaas et al. (2014b) utilized water/air sparging as an antifouling approach to maintain minimal pressure drop over 67 days of operation. Commercial thin woven net-spacers were used to separate membranes and to create flow channels. However, this approach is not appropriate for two-phase flow cleaning as the gas flow can break the net-spacer structure, simultaneously obstructing the compartment of the feed flow. Air bubbles in the net-spacer-filled channels with low liquid flow velocities have been widely employed in nanofiltration and reverse osmosis (RO) membranes. However, one of the major downsides is the formation of undesired stagnant bubbles due to the air coming into the system. Consequently, this led to the reduction of active membrane surface area for water permeation and ion diffusion in RED and finally increased the stack ohmic resistance and lowered the gross power density output and the sum of energy produced by the RED system.

Owing to the high solubility of CO2 in water (27 g L−1 at 1 atm at 25 °C) in water compared to air (0.023 g L−1 at 1 atm at 25 °C), it is considered to be much more effective to be used for two-phase flow cleaning (Moreno et al. 2018a). CO2 was dissolved in water at a pressure higher than the working pressure to allow the supersaturation of water in CO2 before entering the RO membrane module (Ngene et al. 2010). The depressurization of CO2 will create spontaneous bubbles, particularly at the spacer filaments due to local pressure changes. After the CO2-saturated water leaves the spacer-filled channel, it is free into the atmosphere for minimal natural system changes. Moreover, this can lead to cleaning effects as pH in the system will decrease as CO2 in water can form a carbonated solution. The use of two-phase flow CO2/saturated water for RED fouling control has been widely investigated. It was observed that the flow creates hydrodynamic instabilities, interrupts concentration polarization allowing the peeling of formed cake layers, simultaneously removing biofouling from the membrane surfaces or the net-spacers (Moreno et al. 2017).

The impact of fouling can be evaluated using energy-dispersive X-ray spectroscopy. Currently, the RED fouling investigations occur at two stages, i.e. stack and single membrane stage, which imitates the compromise during the fouling investigations. In general, the characterization of the RED system can be limited due to the inherent system complexity or a simplified RED system that might not reflect the hydrodynamic and physical environments in the real RED systems (Pintossi et al. 2019). Thus, a great understanding and better control of process parameters are important for translating laboratory work experiments into large industrial scale stacks, which is essential for commercializing the RED technology (Moreno et al. 2018b). Recent work by Pintossi et al. (2019) where EIS was used as nondestructive fouling monitoring is considered a significant step in utilizing natural salinity gradient for RED applications. This is because it provides detail on how foulants affected the ohmic resistance of the stack, the nonohmic resistance of the AEM, and the nonohmic resistance of the CEM on different time scales. Thus, it guides how to control this fouling by minimizing the cleaning resources while maximizing the cleaning effects.

The use of BPM as a shielding membrane at the cathode to minimize inorganic scaling was investigated recently (Han et al. 2019). This design allowed water splitting in the bilayer structure of the BPM that could restrict the ions diffusing from the cathode and the feed solution, thereby ensuring low stack’s internal resistance, and establishing a highly stable electrode system. A study using artificial feed solutions with 100 cell pairs found that dense precipitates were formed on one side of the AEM and CEM, used as the shielding membranes. However, there was little deposition on both sides of the BPMs. The difference in depositions indicates that the water splitting at the bilayer structure of the BPM blocked the crossover of hydroxide ions and multivalent ions, due to electroneutrality. This fouling minimization strategy is different from pre-/post-treatment, which requires additional electric power or facility as it is a concurrent process with the RED operation.

To summarize, the fouling can be managed using several methods depending on the operational strengths. The first stage is prevention, which can be carried out by designing membranes with smooth surfaces or applying spacers or profiled membranes with an open structure and fewer obstacles. The second stage is pretreatment which includes filtration of feedwater and settling. Next is cleaning in place (CIP), in which fouling detection techniques such as measuring the nonohmic resistance number and two-dimensional fluorescence spectroscopy can be used. Several CIP methods can be used, such as feedwater reversal, air sparging, and brine treatment. Although the treated-fouled membrane’s power density was observed to increase after the treatment, it commonly could not regain its original performance. Other fouling management techniques include cleaning the membrane using mechanical or chemical treatment after demounting the RED system, or if the membrane’s performance decreases significantly, the membranes should be replaced.

6 Integrated RED membrane systems

The conventional RED process aims to extract SGP still faces several economic issues toward scaling-up and commercialization, where low power densities and high costs of commercial membranes are observed. To overcome these challenges, hybrid or integrated RED processes have received great attention in recent years. Several integrated systems have been proposed to intensify process versatility, particularly energy conversion, desalination technology, and water treatment. RED can be coupled with desalination technologies to enhance power density while minimizing brine discharge impacts from desalination technology and can be used for direct and indirect water treatment. In addition, RED can be coupled with a heat engine (HE) to transform waste heat into electricity. For practical application, the usage and storage of electricity produced by RED in situ are vital, and by combining RED with a fuel cell or battery technologies, the electricity produced by RED can be transformed to hydrogen or other energy forms and stored in a battery, respectively.

The integrated systems are typically designed based on two main factors, i.e. (i) a simple retrofit in the existing desalination plants or (ii) as part of novel integrated systems to be ad-hoc designed and built. The following section discusses the synergistic effects of the integrated RED system that can facilitate the RED technology, based on the availability of salinity gradients. There are two types of configurations available for RED integrated systems, i.e. open-loop and closed-loop configurations. In an open-loop configuration, the source of salinity gradients is largely available. The source can be based on natural streams (e.g. seawater and river water) or streams deriving from human activity. These sources typically require extensive pretreatment and fouling-control measures besides having environmental impact due to the withdrawal of natural water from ecosystems. In contrast, a closed-loop configuration is more suitable when the access to salinity gradients is limited, and thus, artificial solutions need to be utilized. It tends to have a less fouling issue and possibly generate much higher power densities. The main parameter to signify the energy performance in an open-loop system is the amount of energy produced based on the available salinity known as energy yield. In contrast, for a closed-loop system, salinity gradient conversion efficiency plays a significant role (Giacalone et al. 2018).

6.1 Open-loop configuration

The traditional open-loop integrated RED system exploits natural salinity gradients that are readily available to produce electricity. The sources can be river water, seawater, brine, or brackish water. Several challenges have been observed in the development of open-loop RED configuration, such as location constraints and the requirement of the pretreatment system, which can increase the system’s overall cost.

6.1.1 RED–RO integrated system

In general, hybrid RED–RO unit is proposed to harvest electricity from the salinity gradient between a highly concentrated seawater solution and a low salinity secondary effluent. Assisted RED (ARED) has been proposed to overcome this initial resistance and decrease the RED investment cost without the need for additional infrastructure (Vanoppen et al. 2018). In ARED, a small potential difference is applied in the natural salinity gradient direction, increasing the ionic transport rate and rapidly decreasing the initial diluate resistance. This decreasing resistance is shown to outweigh any negative effects caused by, for example, concentration polarization, resulting in a process that is more efficient than theoretically expected. As this effect is mainly important at low diluate concentrations (up to 0.1 M), ARED is proposed as the first step in an economical and energy-efficient (A) RED–RO hybrid process.

Vanoppen et al. (2019) investigated the effects of integrating RED to RO for seawater desalination to produce sustainable energy through the RED process while decreasing the energy demand due to the partial desalination entering the RO unit. Secondary-treated wastewater was utilized as the low salinity solution, and the effects of the pretreatment process were evaluated. A pretreatment using a sand filter and 100 µm filtration is required to enhance the process efficiency where the presence of bacteria and a biofilm was minimal after the pretreatment.

Li et al. (2013) evaluated several integrated RED–RO process schemes by comparing their specific energy consumption and discharge brine concentration. In the RED–RO mode system (Figure 9a), high salinity seawater first passes the RED unit where at this point, the energy of the seawater will be extracted. Then, the treated seawater with low salinity will be fed into the RO unit. This condition decreases the pump workload and reduces the concentration of brine discharge. Consequently, the same recovery rate during the RO process can be achieved as the feed solution possesses relatively low osmotic pressure. The other scheme investigated was RO–RED (Figure 9b), where RED was used as a post-treatment unit after RO. A higher emf electrical circuit can be obtained due to highly saline RO discharge. This situation can aid the brine management load as salts in the concentrated seawater are partly removed through the RED process. A biologically treated secondary effluent with a higher concentration than typical river water was used as a low salinity solution in the RED system to improve the system efficiency. A more complex integrated system, i.e. RED–RO–RED (Figure 9c) that combines the advantages of the two basic modes was then proposed. However, this design requires a high capital cost. Overall, the integrated system offers a reduction in energy consumption due to the RED system and better discharge brine management leading to the possibility of a zero-discharge system. Nevertheless, these integrated conceptual schemes need to be tested experimentally in the future to validate their potential.

Figure 9: 
Schematic diagrams of the basic RED–RO hybrid processes: (a) RED–RO mode; (b) RO–RED mode, and (c) RED–RO–RED mode. Reproduced with permission (Li et al. 2013). Copyrighted from Elsevier Ltd.
Figure 9:

Schematic diagrams of the basic RED–RO hybrid processes: (a) RED–RO mode; (b) RO–RED mode, and (c) RED–RO–RED mode. Reproduced with permission (Li et al. 2013). Copyrighted from Elsevier Ltd.

Choi et al. (2019) also proposed a hybrid RO–membrane capacitive deionization (MCDI)–RED as a seawater desalination system to overcome the power loss during the electricity to turbine power conversion. MCDI and RED were utilized to replace brackish water reverse osmosis and produce energy by exploiting the RO brine and MCDI brine (Figure 10). The typical two-pass RO system can be replaced by the integrated RO–MCDI–RED as it was observed to display better performance. An increase in voltage would lead to better MCDI desalination performances (lower TDS) as it highly depends on the potential difference between a pair of porous carbon electrodes (Figure 9b). Jande and Kim (2014) investigated the co-integration system of RED with CDI as a way to minimize the environmental impact due to the high salinity of CDI brine. Through this system, pure water and electricity could be simultaneously produced. The 15,000 ppm feed concentration of CDI stack produced ultrapure water and two feed streams for the RED stack, i.e. low salinity (∼17.4 mol NaCl m−3) and high salinity (∼512.8 mol NaCl m−3). The simulated total power generation was 0.57 W m−2 using infinitely divided electrodes. With RED integration, better energy efficiency can be obtained, making it potential as the next-generation water treatment system. However, a detailed study should be carried out as capital costs remain as the main concern.

Figure 10: 
The performance of integrated RO–MCDI–RED system: (a) Permeate water quality and (b) Energy consumption of the overall desalination system. Reproduced with permission (Choi et al. 2019). Copyrighted from Elsevier Ltd.
Figure 10:

The performance of integrated RO–MCDI–RED system: (a) Permeate water quality and (b) Energy consumption of the overall desalination system. Reproduced with permission (Choi et al. 2019). Copyrighted from Elsevier Ltd.

6.1.2 RED–ED integrated system

Figure 11 shows an integrated reverse electrodialysis–electrodialysis (RED–ED) system developed by Wang et al. (2017) to treat wastewater containing phenol. In this system, phenol wastewater was used as the high-salinity solution instead of seawater to recover greater osmotic energy. The integrated RED–ED system utilized less electricity (i.e. 17.65 versus 25 kWh m−3) while gaining higher SGP (i.e. 0.35 versus 0 kWh m−3) at a predesalination rate of 27.44% in RED (Wang et al. 2017). Besides lowered-energy desalination systems, other benefits include the utilization of salinity-based energy and the recovery of high-value resources.

Figure 11: 
Schematic diagram of the hybrid RED/ED system for simultaneous osmotic energy recovery and complete desalination of high-salinity wastewater. Reproduced with permission (Wang et al. 2017). Copyrighted from Elsevier Ltd.
Figure 11:

Schematic diagram of the hybrid RED/ED system for simultaneous osmotic energy recovery and complete desalination of high-salinity wastewater. Reproduced with permission (Wang et al. 2017). Copyrighted from Elsevier Ltd.

6.1.3 Reverse osmosis–direct contact membrane distillation–reverse electrodialysis integrated system

Tufa et al. (2019) performed energy–exergy and economic evaluation of integrated reverse osmosis–direct contact membrane distillation–reverse electrodialysis (RO–DCMD–RED) system for energy and pure water generation (Figure 12). Several scenarios were considered on the existence of the RED unit, where it was used to recover and convert the electrochemical and thermal potentials of hot hypersaline brine of MD into electricity. The data were theoretically calculated and supported by lab-scale testing. The exergetic efficiency can be reached up to 49% when the MD feed temperature and brine concentrations were 60 °C and 5 M NaCl, respectively, and RED power density was 2.2 W m−2. This condition can instantaneously reduce electrical energy consumption and specific energy consumption up to 23 and 16.6%, respectively.

Figure 12: 
Diagrammatic sketch of the proposed MD–RED hybrid electricity generation system to harvest low-grade thermal energy. Reproduced with permission (Tufa et al. 2019). Copyrighted from Elsevier Ltd.
Figure 12:

Diagrammatic sketch of the proposed MD–RED hybrid electricity generation system to harvest low-grade thermal energy. Reproduced with permission (Tufa et al. 2019). Copyrighted from Elsevier Ltd.

6.2 Closed-loop

One of the most important advantages of closed-loop systems compared to open systems is the possibility to select ad-hoc salt solutions to achieve high efficiencies. Therefore, the properties of the salt solutions are essential to assess the performance of the energy generation and solution regeneration processes. The closed-loop characteristics enable the system to have merits such as no repetitive pretreatment cost and removal of a locational constraint than open-loop RED with sea and river water.

6.2.1 Reverse electrodialysis–membrane distillation–heat engine integrated system

Long et al. (2017) proposed an integrated membrane distillation (MD) and RED system to recover low-temperature thermal energy (Figure 13). The thermal separation unit consists of an MD and a regenerator functions to absorb thermal energy from the heat source while producing highly saline NaCl solution. In the electricity generation unit, RED functions to convert Gibbs free energy of mixing from high salinity solution into electricity. The simulation results indicate that the optimal relative permeate/feed solution flow rate is required in the RED–MD system to yield maximum electrical efficiency. An efficiency of up to 1.15% can be observed when the system was operated using 5 mol kg−1 of NaCl solution between 20 and 60 °C, which specifies its potential application for low-grade heat harvesting.

Figure 13: 
Scheme of the integrated membrane desalination process that includes the recovery of energy from brine by RED. Reproduced with permission (Long et al. 2017). Copyrighted from Elsevier Ltd.
Figure 13:

Scheme of the integrated membrane desalination process that includes the recovery of energy from brine by RED. Reproduced with permission (Long et al. 2017). Copyrighted from Elsevier Ltd.

6.2.2 RED–HE integrated system

Tamburi et al. (2017) proposed a co-integration of RED with a HE to transform low-grade/waste heat (<100 °C) into electric power using a simplified mathematical model (Figure 14). The RED–HE works when the concentration gradient of two saline solutions pushes the ions to pass through the membranes. The performance is controlled by ions mobility, membrane permselectivity, and selective ion transport across the membranes, forming a net ionic current through the stack. The ionic current is then transformed into electric current through redox reactions occurring at the electrodes positioned at both ends of the stack. In their work, various salts were investigated as solutes in the feed stream, and the optimum conditions were then evaluated. Several REDHE scenarios were used, and higher power densities were obtained using lithium-based salt (i.e. LiBr and LiCl). The efficiency was observed to improve up to 15% for the closed-loop system, matching the exergetic efficiency of ∼85%. This result indicates that the REDHE system is one of the promising technologies for the low-grade heat industry.

Figure 14: 
Simplified scheme of the REDHE process. Reproduced with permission (Tamburini et al. 2017). Copyrighted from Elsevier Ltd.
Figure 14:

Simplified scheme of the REDHE process. Reproduced with permission (Tamburini et al. 2017). Copyrighted from Elsevier Ltd.

Bevacqua et al. (2017) investigated the use of ammonium hydrogen-carbonate (NH4HCO3) aqueous solutions as fluid in REDHE. The system consists of a RED unit and a regeneration unit (stripping and absorption). The RED unit functions to generate electrical power from various salt solutions at the regeneration unit and form a closed-loop cycle by reinstating the initial salinity gradient. The integrated process was modeled using two routes: (i) a lumped parameter model for the RED validated with experimental data and (ii) a model established using a process simulator to evaluate the thermal duty of the stripping column. A sensitivity analysis was used to investigate the effects of solution concentrations and velocities. At optimal conditions, 9 W m−2 of cell pair of power density was forecast for the RED while the whole cycle maximum exergetic efficiency was ∼22%. Giacalone et al. (2020) recently reported on the testing of the first REDHE prototype unit in the world (Figure 15). The RED system was utilized to convert the salinity gradient of two thermolytic salts solutions into electricity, and the regeneration unit used low-temperature heat to reinstate the conditions of the feed streams. The degradation of salts into gaseous ammonia and CO2 allowed the regeneration process to occur. The initial salinity gradient was installed by removing the exhaust gas in dilute solution using vapor stripping followed by reabsorption into the exhausted concentrate solution. The highest exergy efficiency was up to 1.1% when 0.05–1.9 M of ammonium bicarbonate solutions were used, and the system possessed good stability during 55 h of operation.

Figure 15: 
Simplified P&I diagram of the thermolytic RED–HE. Main components: (1) RED unit; (2) stripping column; (3) reboiler; (4) auxiliary circuit; (5) venturi ejectors; (6) barometric condenser; (7) thermal integration heat exchanger; (S) sampling points. Reproduced with permission (Giacalone et al. 2020). Copyrighted from Elsevier Ltd.
Figure 15:

Simplified P&I diagram of the thermolytic RED–HE. Main components: (1) RED unit; (2) stripping column; (3) reboiler; (4) auxiliary circuit; (5) venturi ejectors; (6) barometric condenser; (7) thermal integration heat exchanger; (S) sampling points. Reproduced with permission (Giacalone et al. 2020). Copyrighted from Elsevier Ltd.

6.2.3 Microbial fuel cell–reverse electrodialysis

MFC is a unit that directly transforms chemical energy from organic matter into electrical energy by microorganisms. In MFC, the bioanodes consist of microorganisms supported on conductive carriers that oxidize organic pollutants and release electrons. The electrons are accumulated and transported to the cathode where O2 (an electron acceptor) generates H2O. In an integrated MFC–RED or simply known as MRC, the low voltage RED stack between the bioanode and cathode can accelerate the movement of electrons from the anode to the cathode to enhance the output power. The bioanode degrades organic pollutants and exports electrons (Nam et al. 2012), and at the cathode, H+ or O2 is used as an electron acceptor to generate H2, H2O, and H2O2. Besides producing H2, a reduction of CO2 to valuable carbonaceous energy can also be observed (Li et al. 2018).

Jwa et al. (2020) proposed the use of microbial electrolysis cells (MECs) to improve multivalent cation removal rates for the pretreatment system of seawater (Figure 16). The MEC is comprised two electrodes and a hydrogen collection system in which organic matter can be oxidized by exoelectrogens at the anode to produce electrons and protons, while hydrogen is produced by reducing protons at the cathode using an applied voltage. The hydroxyl ions increase the pH of the catholyte, leading to the precipitation of Ca2+ and Mg2+ as CaCO3 and Mg(OH)2, respectively. It was found that through MEC–RED system, the seawater was softened by removing 84 ± 5% of Ca2+, and 99.5 ± 4% of Mg2+ and high-purity H2 was produced at 2.00 ± 0.09 m3 m3d−1. The treated seawater (47 mS cm−1) was then fed into the RED stack as an HC solution. The maximum power density was observed to be 26% higher compared to the use of untreated seawater (53.7 mS cm−1) as the high pH of treated seawater reduced the MERs and further increased the power generation.

Figure 16: 
Schematic diagram of MEC–RED system. Reproduced with permission (Jwa et al. 2020). Copyrighted from Elsevier Ltd.
Figure 16:

Schematic diagram of MEC–RED system. Reproduced with permission (Jwa et al. 2020). Copyrighted from Elsevier Ltd.

6.2.4 Reverse electrodialysis–flow battery

One of the main challenges faced by RED is efficient energy storage to ensure a sustainable electric supply. By flow battery (FB) systems, the energy storage problem can be overcome, allowing the decentralized use of renewable energy. There are two types of integrated RED–FB system proposed, i.e. RED with a concentration flow battery (RED–CFB) and RED with an acid–base flow battery (RED–ABFB).

RED–CFB consists of an anode chamber, a series of CEMs and AEMs, and a cathode chamber. When solutions with different salinities flow through the channels separated by CEMs and AEMs, a voltage is generated across each membrane due to the ion flux caused by the differences in salt concentrations. Reversible redox reactions using [Fe(CN)6]4−/[Fe(CN)6]3− and Fe2+/Fe3+ that are recycled between the electrodes were investigated to reduce the electrode over-potentials and thus improve the power output. In the charging process, Fe(CN)6 4− is to Fe(CN)6 3− at the anode, and 2,6-dihydroxyanthraquinone (2,6-DHAQ) is reduced to 2,6-reDHAQ at the cathode (Zhu et al. 2017). These two charged electrolyte solutions are then pumped to the CFB for discharging. Therefore, the chemical energy of the salinity gradient is converted to another type of chemical energy. In the discharge process, Fe(CN)6 3− is reduced to Fe(CN)6 4− at the cathode, and 2,6-reDHAQ is oxidized to 2,6-DHAQ at the anode of the CFB. The electrons are transferred by connecting the external circuit with a collector on the electrodes. Therefore, the chemical energy is converted to electric energy.

A RED–ABFB system is composed of three compartments separated by two monopolar membranes and one bipolar membrane (BPM) (Xia et al. 2020). Each repeating cell unit consists of acid (HCl) and base (NaOH) chambers at both sides of the BPM. A salt (NaCl) compartment is separated from the acid or base chambers by an AEM and CEM, respectively, to ensure electroneutrality. As the high electric potential difference is applied between the stacks’ electrodes during charging, the water dissociates at the interface of the BPM into H+ and OH. H+ ions that migrate from the NaCl solution across the AEM combine with Cl ions to form HCl, while OH ions which migrate from the NaCl solution across the CEM combine with Na+ ions to form NaOH. Consequently, the electric energy is converted to chemical energy with a pH gradient in the charging process. In the discharge process, H+ and OH ions migrate from the acid and base solutions into the BPM interface, where they are neutralized to water. An equivalent number of Na+ and Cl ions to preserve electroneutrality. RED-ABFB was observed to provide a stable charge/discharge in cyclic operation (Kim et al. 2016). Figure 17 compares the RED–CFB and RED–ABFB systems.

Figure 17: 
Schematic diagram of (a) RED–CFB system. Reproduced with permission (Zhu et al. 2017), copyrighted from Wiley-VCH, and (b) RED–ABFB. Reproduced from Xia et al. (2020). The image is published by MDPI as in open access mode.
Figure 17:

Schematic diagram of (a) RED–CFB system. Reproduced with permission (Zhu et al. 2017), copyrighted from Wiley-VCH, and (b) RED–ABFB. Reproduced from Xia et al. (2020). The image is published by MDPI as in open access mode.

7 Conclusions

RED converts the free energy of two streams with different salinities directly into electrical energy. Owing to its clean, renewable, and sustainable characteristics, research on this technology has increased substantially in the last decade and is considered future energy generation technology. The development of IEM, optimization of the RED system and integrated RED system for various applications are crucial for sustainable SGP generation. Over the last decade, many IEMs have been developed, but many works concentrate on using single ions feed composition. Multivalent ions have a strong impact on the membrane characteristics, particularly in reducing membrane permselectivity and increasing membrane resistance. This scenario can significantly decrease the obtainable power density. With the advancement of nanomaterials and membrane preparation techniques, the performances of IEMs can be tailored easily to fulfill the specific applications, particularly feed type, and composition, when natural feedwaters are used. Various routes can be used to improve the physicochemical structure, particularly thickness and selectivity of IEM, such as adding functional nanomaterials, controlling sulfonation reaction for CEM, controlling amination for CEM, grafting, and chemical cross-linking agent.

Besides the intrinsic properties of IEMs, the selection of suitable operating conditions such as flow velocity, feed temperature, and feed properties are critically important to achieving the desired RED performances. A deeper understanding of the fouling issue and mathematical modeling of the RED system is vital for implementing RED in the real environment. The threat of membranes fouling and plugging can be significantly raised when natural feedwater is used. Many methods have been adopted to suppress the fouling phenomenon in RED such as pretreatment (using sand-filter and precipitation) and post-treatment (chemical and physical cleaning). However, the key challenge in RED is to adopt proper and cost-effective fouling management. Therefore, long-term experiments should be carried out to investigate the optimal trade-off between membrane performances and the fouling phenomenon.

Recently, a significant increase in the development of a sustainable membrane system through the integration of various processes has been observed, particularly for RED processes. Besides being cost- and energy-efficient, this synergized RED system offers numerous functions, such as cleaner water production, energy extraction from wastewater and water desalination, and transformation of low-grade waste heat, opening several exciting perspectives, particularly in implementing the zero liquid discharge concept, which is in line with the sustainable goal development of clean water and clean energy. Higher power density and water recovery rate with minimal brine discharge can be achieved by utilizing RED with desalination technologies. In addition, the electricity produced by the integrated RED system can also be transformed through hydrogen conversion or to an energy storage battery to overcome the issue of low current power densities and high cost of commercial membranes in RED. Integrating different systems with RED depends on the source of feedwater and whether the design is part of retrofitting existing plants or in a newly built plant. Studies on its energy efficiency and capital cost need to be evaluated for practical application as optimization of size and capacity of the system provides efficient renewable energy at the lowest investment.

Corresponding author: Nur Hidayati Othman, School of Chemical Engineering, College of Engineering, Universiti Teknologi MARA, Shah Alam, 40450 Selangor Darul Ehsan, Malaysia, E-mail:

Funding source: Scientific and Technological Research Council of Turkey (TÜBİTAK)

Award Identifier / Grant number: TÜBİTAK-NCBR Project no. 117M023

Funding source: Universiti Teknologi MARA 10.13039/501100004625

Award Identifier / Grant number: 600-RMC/SRC/5/3 (021/2020)

  1. Research funding: The authors would like to acknowledge the Scientific and Technological Research Council of Turkey (TÜBİTAK) by 2221- Fellowship for Visiting Scientists and Scientists on Sabbatical Leave (short-term), and the financial support for the initiation of RED studies through a research project (TÜBİTAK-NCBR Project no. 117M023). This work was also supported by Universiti Teknologi MARA [grant no: 600-RMC/SRC/5/3 (021/2020)].

  2. Competing interests: The authors declare that they have no conflicts of interest regarding this article.



Alternating current


Anion-conducting groups


Anion exchange membrane


Brackish water reverse osmosis


Cation exchange membrane


Capacitive deionization


Computational fluid dynamics


Direct current


Direct contact membrane distillation




Forward osmosis


High concentration compartment


Heat engine


Ion exchange membrane


Low salinity solution


Low concentration compartment


Membrane electrode assembly


Membrane distillation


Membrane capacitive deionization


Osmotic heat engine


Oxygen evolution reaction


Pressure retarded osmosis


Reverse electrodialysis


Reverse electrodialysis heat engine


Reverse osmosis


Radiation-grafted CEM


Salinity gradient power


Swelling degree


Swelling ratio


Vacuum membrane distillation


Water uptake


Seawater desalination units

Nomenclature and symbols


2-Acrylamido-2-methylpropanesulfonic acid




Chloride ion


Calcium(II) chloride


Carboxylic acid








Iron(III) oxide

SO4 2−

Sulfonated ion

Sulfate ion


Potassium chloride




Magnesium ion


Magnesium chloride


Sodium ion


Sodium chloride


Sodium sulfate

NH3 +

Quaternary ammonium cation

NRH2 +

Primary amine


Secondary amine

NR3 +

Tertiary amine


Oxidized multiwalled carbon nanotubes














Phosphonic acid




Polyvinylidene fluoride

SO4 2−

Sulfonated ion

Sulfate ion


Sulfonated silica


Sulfonic acid


Sulfonated polyethersulfone


Sulfonated polyethersulfone-dense structure


Sulfonated polyethersulfone-porous structure


Sulfonated poly(2,6-dimethyl-1,4-phenylene oxide)


Vinyl benzyl trimethylammonium chloride


Species concentration (mol m−3)


Salt concentration (M/mol L−1)


Concentration of titrant used to determine IEC (mol m−3)


Concentration difference (mol m−3)

CD fix

Fixed charge

c p

Specific heat of solution (kJ kg−1 K−1)

D i

Ion diffusion coefficient


Free energy


Measured electrical potential difference


Theoretical electrical potential difference


Theoretical electromotive force (V)

E x,in

Exergy input (kJ h−1)

E x,out

Exergy output (kJ h−1)


Electric voltage (V)

E t-

Amount of electricity generated


Faraday constant (96485 C mol−1)

Δ G mix

Gibbs energy


Specific enthalpy (kJ kg−1)


Current (A)



m wet

Weight of wet IEM

m dry

Weight of dry IEM

N m

Total number of membranes


Open-circuit voltage (V)

P d

Power density (W m−2)


Volumetric feed flow rate (m3 s−1)


Universal gas constant (8.314 J mol−1 K−1)

R i

Total electrical area resistance per stack (Ω cm2).

R ohmic

Ohmic area resistance per cell (Ω cm2)


Area of resistance caused by the concentration change (Ω cm2)


Area resistance due concentration changes in the boundary layer (Ω cm2)


Temperature (K)

T 0

Reference temperature


time interval


Voltage (V)


Volume (m3)



W dry

Weight of the dry membrane (g)


Molar fraction


Valence of the ionic species

z i

Ion electrical charge


Activity coefficient of the salt species

η C E

Coulombic efficiency


Solution dynamic viscosity at temperature T

μ(T 0)

Solution dynamic viscosity at reference temperature

μ i

Ion mobility


Alvarez-Silva, O.A., Osorio, A.F., and Winter, C. (2016). Practical global salinity gradient energy potential. Renew. Sustain. Energy Rev. 60: 1387–1395. in Google Scholar

Avci, A.H., Sarkar, P., Tufa, R.A., Messana, D., Argurio, P., Fontananova, E., Di Profio, G., and Curcio, E. (2016). Effect of Mg2+ ions on energy generation by Reverse Electrodialysis. J. Membr. Sci. 520: 499–506. in Google Scholar

Avci, A.H., Tufa, R.A., Fontananova, E., Di Profio, G., and Curcio, E. (2018). Reverse Electrodialysis for energy production from natural river water and seawater. Energy 165: 512–521. in Google Scholar

Avci, A.H., Rijnaarts, T., Fontananova, E., Di Profio, G., Vankelecom, I.F.V., De Vos, W.M., and Curcio, E. (2020a). Sulfonated polyethersulfone based cation exchange membranes for reverse electrodialysis under high salinity gradients. J. Membr. Sci. 595: 117585. in Google Scholar

Avci, A.H., Messana, D.A., Santoro, S., Tufa, R.A., Curcio, E., Di Profio, G., and Fontananova, E. (2020b). Energy harvesting from brines by reverse electrodialysis using nafion membranes. Membranes 10: 1–16. in Google Scholar PubMed PubMed Central

Badessa, T. and Shaposhnik, V. (2016). The electrodialysis of electrolyte solutions of multi-charged cations. J. Membr. Sci. 498: 86–93. in Google Scholar

Banasiak, L.J., Kruttschnitt, T.W., and Schäfer, A.I. (2007). Desalination using electrodialysis as a function of voltage and salt concentration. Desalination 205: 38–46. in Google Scholar

Benneker, A.M., Rijnaarts, T., Lammertink, R.G.H., and Wood, J.A. (2018). Effect of temperature gradients in (reverse) electrodialysis in the Ohmic regime. J. Membr. Sci. 548: 421–428. in Google Scholar

Bevacqua, M., Tamburini, A., Papapetrou, M., Cipollina, A., Micale, G., and Piacentino, A. (2017). Reverse electrodialysis with NH4HCO3-water systems for heat-to-power conversion. Energy 137: 1293–1307. in Google Scholar

Brauns, E. (2009). Salinity gradient power by reverse electrodialysis: effect of model parameters on electrical power output. Desalination 237: 378–391. in Google Scholar

Choi, J., Kim, W.S., Kim, H.K., Yang, S.C., and Jeong, N.J. (2020). Ultra-thin pore-filling membranes with mirror-image wave patterns for improved power density and reduced pressure drops in stacks of reverse electrodialysis. J. Membr. Sci. 620: 118885.10.1016/j.memsci.2020.118885Search in Google Scholar

Choi, J., Oh, Y., Chae, S., and Hong, S. (2019). Membrane capacitive deionization-reverse electrodialysis hybrid system for improving energy efficiency of reverse osmosis seawater desalination. Desalination 462: 19–28. in Google Scholar

Chon, K., Jeong, N., Rho, H., Nam, J.Y., Jwa, E., and Cho, J. (2020). Fouling characteristics of dissolved organic matter in fresh water and seawater compartments of reverse electrodialysis under natural water conditions. Desalination 496: 114478. in Google Scholar

Chung, H.W., Nayar, K.G., Swaminathan, J., Chehayeb, K.M., and Lienhard V, J.H. (2017). Thermodynamic analysis of brine management methods: zero-discharge desalination and salinity-gradient power production. Desalination 404: 291–303.10.1016/j.desal.2016.11.022Search in Google Scholar

Cifuentes-Araya, N., Pourcelly, G., and Bazinet, L. (2011). Impact of pulsed electric field on electrodialysis process performance and membrane fouling during consecutive demineralization of a model salt solution containing a high magnesium/calcium ratio. J. Colloid Interface Sci. 361: 79–89. in Google Scholar PubMed

Cipollina, A. and Micale, G. (2016). Sustainable energy from salinity gradients. Elsevier.Search in Google Scholar

Cusick, R.D., Kim, Y., and Logan, B.E. (2012). Energy capture from thermolytic solutions in microbial reverse- electrodialysis cells. Science 335: 1474–1477. in Google Scholar PubMed

D’Angelo, A., Tedesco, M., Cipollina, A., Galia, A., Micale, G., and Scialdone, O. (2017). Reverse electrodialysis performed at pilot plant scale: evaluation of redox processes and simultaneous generation of electric energy and treatment of wastewater. Water Res. 125: 123–131.10.1016/j.watres.2017.08.008Search in Google Scholar PubMed

Daniilidis, A., Vermaas, D.A., Herber, R., and Nijmeijer, K. (2014). Experimentally obtainable energy from mixing river water, seawater or brines with reverse electrodialysis. Renew. Energy 64: 123–131. in Google Scholar

Długołecki, P., Nymeijer, K., Metz, S., and Wessling, M. (2008). Current status of ion exchange membranes for power generation from salinity gradients. J. Membr. Sci. 319: 214–222. in Google Scholar

Długołȩcki, P., Gambier, A., Nijmeijer, K., and Wessling, M. (2009). Practical potential of reverse electrodialysis as process for sustainable energy generation. Environ. Sci. Technol. 43: 6888–6894. in Google Scholar PubMed

Długołecki, P., Anet, B., Metz, S.J., Nijmeijer, K., and Wessling, M. (2010a). Transport limitations in ion exchange membranes at low salt concentrations. J. Membr. Sci. 346: 163–171.10.1016/j.memsci.2009.09.033Search in Google Scholar

Długołecki, P., Ogonowski, P., Metz, S.J., Saakes, M., Nijmeijer, K., and Wessling, M. (2010b). On the resistances of membrane, diffusion boundary layer and double layer in ion exchange membrane transport. J. Membr. Sci. 349: 369–379.10.1016/j.memsci.2009.11.069Search in Google Scholar

Domenech, B., Bastos-Arrieta, J., Alonso, A., Macanas, J., Munoz, M., and N, D. (2012). Bifunctional polymer-metal nanocomposite ion exchange materials. In: Kilislioglu, Ayben (Ed.), Ion exchange technologies. London: IntechOpen, pp. 35–72.10.5772/51579Search in Google Scholar

Fan, H. and Yip, N.Y. (2019). Elucidating conductivity-permselectivity tradeoffs in electrodialysis and reverse electrodialysis by structure-property analysis of ion-exchange membranes. J. Membr. Sci. 573: 668–681. in Google Scholar

Fernandez-Gonzalez, C., Dominguez-Ramos, A., Ibañez, R., Chen, Y., and Irabien, A. (2017). Valorization of desalination brines by electrodialysis with bipolar membranes using nanocomposite anion exchange membranes. Desalination 406: 16–24. in Google Scholar

Fontananova, E., Cucunato, V., Curcio, E., Trotta, F., Biasizzo, M., Drioli, E., and Barbieri, G. (2012). Influence of the preparation conditions on the properties of polymeric and hybrid cation exchange membranes. Electrochim. Acta 66: 164–172. in Google Scholar

Fontananova, E. (2016). Impedance spectroscopy, membrane characterization by. In: Drioli, E. and Giorno, L. (Eds.), Encyclopedia of membranes. Berlin, Heidelberg: Springer Berlin Heidelberg, pp. 1025–1027. in Google Scholar

Fontananova, E., Messana, D., Tufa, R.A., Nicotera, I., Kosma, V., Curcio, E., van Baak, W., Drioli, E., and Di Profio, G. (2017). Effect of solution concentration and composition on the electrochemical properties of ion exchange membranes for energy conversion. J. Power Sources 340: 282–293. in Google Scholar

Fontananova, E., Zhang, W., Nicotera, I., Simari, C., van Baak, W., Di Profio, G., Curcio, E., and Drioli, E. (2014). Probing membrane and interface properties in concentrated electrolyte solutions. J. Membr. Sci. 459: 177–189.10.1016/j.memsci.2014.01.057Search in Google Scholar

Fu, L.J., Ji, Z.Y., Tumba, K., Yang, F.J., Liu, J., Zhao, Y.Y., and Yuan, J.S. (2020). Performance parameters analysis of reverse electrodialysis process: sensitive to the repeating unit pairs, inflow velocity and feed concentration. Int. J. Energy Res. 44: 7093–7102. in Google Scholar

Gao, H., Zhang, B., Tong, X., and Chen, Y. (2018). Monovalent-anion selective and antifouling polyelectrolytes multilayer anion exchange membrane for reverse electrodialysis. J. Membr. Sci. 567: 68–75. in Google Scholar

Geise, G.M., Hickner, M.A., and Logan, B.E. (2013). Ionic resistance and permselectivity tradeoffs in anion exchange membranes. ACS Appl. Mater. Interfaces 5: 10294–10301. in Google Scholar PubMed

Geise, G.M., Paul, D.R., and Freeman, B.D. (2014). Fundamental water and salt transport properties of polymeric materials. Prog. Polym. Sci. 39: 1–42. in Google Scholar

Giacalone, F., Catrini, P., Tamburini, A., Cipollina, A., Piacentino, A., and Micale, G. (2018). Exergy analysis of reverse electrodialysis. Energy Convers. Manag. 164: 588–602. in Google Scholar

Giacalone, F., Vassallo, F., Scargiali, F., Tamburini, A., Cipollina, A., and Micale, G. (2020). The first operating thermolytic reverse electrodialysis heat engine. J. Membr. Sci. 595: 117522. in Google Scholar

Golubenko, D.V., Van der Bruggen, B., and Yaroslavtsev, A.B. (2020). Novel anion exchange membrane with low ionic resistance based on chloromethylated/quaternized-grafted polystyrene for energy efficient electromembrane processes. J. Appl. Polym. Sci. 137: 48656. in Google Scholar

Gómez-Coma, L., Ortiz-Martínez, V.M., Carmona, J., Palacio, L., Prádanos, P., Fallanza, M., Ortiz, A., Ibañez, R., and Ortiz, I. (2019). Modeling the influence of divalent ions on membrane resistance and electric power in reverse electrodialysis. J. Membr. Sci. 592. in Google Scholar

Gómez-Coma, L., Ortiz-Martínez, V.M., Fallanza, M., Ortiz, A., Ibañez, R., and Ortiz, I. (2020). Blue energy for sustainable water reclamation in WWTPs. J. Water Process Eng. 33: 101020. in Google Scholar

Guler, E., Zhang, Y., Saakes, M., and Nijmeijer, K. (2012). Tailor-made anion-exchange membranes for salinity gradient power generation using reverse electrodialysis. ChemSusChem 5: 2262–2270. in Google Scholar PubMed

Güler, E., Elizen, R., Vermaas, D.A., Saakes, M., and Nijmeijer, K. (2013). Performance-determining membrane properties in reverse electrodialysis. J. Membr. Sci. 446: 266–276. in Google Scholar

Güler, E., van Baak, W., Saakes, M., and Nijmeijer, K. (2014a). Monovalent-ion-selective membranes for reverse electrodialysis. J. Membr. Sci. 455: 254–270. in Google Scholar

Güler, E., Elizen, R., Saakes, M., and Nijmeijer, K. (2014b). Micro-structured membranes for electricity generation by reverse electrodialysis. J. Membr. Sci. 458: 136–148. in Google Scholar

Güler, E. and Nijmeijer, K. (2018). Reverse electrodialysis for salinity gradient power generation: challenges and future perspectives. J. Membr. Sci. Res. 4: 108–110.Search in Google Scholar

Guo, Z.Y., Ji, Z.Y., Zhang, Y.G., Yang, F.J., Liu, J., Zhao, Y.Y., and Yuan, J.S. (2018). Effect of ions (K+, Mg2+, Ca2+ and SO42−) and temperature on energy generation performance of reverse electrodialysis stack. Electrochim. Acta 290: 282–290. in Google Scholar

Gurreri, L., Tamburini, A., Cipollina, A., Micale, G., and Ciofalo, M. (2013). CFD simulation of mass transfer phenomena in spacer filled channels for reverse electrodialysis applications. Chem. Eng. Trans. 32: 1879.Search in Google Scholar

Gurreri, L., Tamburini, A., Cipollina, A., Micale, G., and Ciofalo, M. (2014). CFD prediction of concentration polarization phenomena in spacer-filled channels for reverse electrodialysis. J. Membr. Sci. 468: 133–148. in Google Scholar

Gurreri, L., Tamburini, A., Cipollina, A., and Micale, G. (2020). Electrodialysis applications in wastewater treatment for environmental protection and resources recovery: a systematic review on progress and perspectives. Membranes 10: 1–93. in Google Scholar PubMed PubMed Central

Hagesteijn, K.F.L., Jiang, S., and Ladewig, B.P. (2018). A review of the synthesis and characterization of anion exchange membranes. J. Mater. Sci. 53: 11131–11150. in Google Scholar

Han, G., Zhang, S., Li, X., and Chung, T.S. (2014). Progress in pressure retarded osmosis (PRO) membranes for osmotic power generation. Prog. Polym. Sci. 51: 1–27.10.1016/j.progpolymsci.2015.04.005Search in Google Scholar

Han, J.H., Jeong, N., Kim, C.S., Hwang, K.S., Kim, H., Nam, J.Y., Jwa, E., Yang, S.C., and Choi, J. (2019). Reverse electrodialysis (RED) using a bipolar membrane to suppress inorganic fouling around the cathode. Water Res. 166: 115078. in Google Scholar PubMed

Helfer, F. and Lemckert, C. (2015). The power of salinity gradients: an Australian example. Renew. Sustain. Energy Rev. 50: 1–16. in Google Scholar

Hong, J.G. and Chen, Y. (2014). Nanocomposite reverse electrodialysis (RED) ion-exchange membranes for salinity gradient power generation. J. Membr. Sci. 460: 139–147. in Google Scholar

Hong, J.G., and Chen, Y. (2015). Evaluation of electrochemical properties and reverse electrodialysis performance for porous cation exchange membranes with sulfate-functionalized iron oxide. J. Membr. Sci. 473: 210–217, doi: in Google Scholar

Hong, J.G., Zhang, B., Glabman, S., Uzal, N., Dou, X., Zhang, H., Wei, X., and Chen, Y. (2015a). Potential ion exchange membranes and system performance in reverse electrodialysis for power generation: a review. J. Membr. Sci. 486: 71–88. in Google Scholar

Hong, J.G., Glabman, S., and Chen, Y. (2015b). Effect of inorganic filler size on electrochemical performance of nanocomposite cation exchange membranes for salinity gradient power generation. J. Membr. Sci. 482: 33–41. in Google Scholar

Hong, J.G. and Kim, J.J. (2016). Salinity gradient energy: current membrane development and challenges for reverse electrodialysis system. New Renew. Energy 12: 53. in Google Scholar

Hong, J.G., Park, T.W., and Dhadake, Y. (2019). Property evaluation of custom-made ion exchange membranes for electrochemical performance in reverse electrodialysis application. J. Electroanal. Chem. 850: 113437. in Google Scholar

Hosseini, S.M., Askari, M., Koranian, P., Madaeni, S.S., and Moghadassi, A.R. (2014). Fabrication and electrochemical characterization of PVC based electrodialysis heterogeneous ion exchange membranes filled with Fe3O4 nanoparticles. J. Ind. Eng. Chem. 20: 2510–2520. in Google Scholar

Jaime-Ferrer, J.S., Mosqueda-Quintero, M., Suárez-Toriello, V.A., Anderson, S.M., González Vargas, O.A., and Villafanã-López, L. (2020). Heterogeneous PVC cation-exchange membrane synthesis by electrospinning for reverse electrodialysis. Int. J. Chem. React. Eng. 18. in Google Scholar

Jande, Y.A.C. and Kim, W.S. (2014). Integrating reverse electrodialysis with constant current operating capacitive deionization. J. Environ. Manag. 146: 463–469. in Google Scholar PubMed

Jia, Z., Wang, B., Song, S., and Fan, Y. (2014). Blue energy: current technologies for sustainable power generation from water salinity gradient. Renew. Sustain. Energy Rev. 31: 91–100. in Google Scholar

Jones, A.T., and Finley, W. (2003). Oceans 2003. Celebrating the past ... teaming toward the future. In: Recent developments in salinity gradient power, Vol. 4. IEEE, pp. 2284–2287.10.1109/OCEANS.2003.178265Search in Google Scholar

Jung, D.H., Han, E.D., Kim, B.H., and Seo, Y.H. (2019). Ultra-thin ion exchange film on the ceramic supporter for output power improvement of reverse electrodialysis. Sci. Rep. 9: 17440. in Google Scholar PubMed PubMed Central

Jwa, E., Yun, Y.M., Kim, H., Jeong, N., Hwang, K.S., Yang, S.C., and Nam, J.Y. (2020). Energy-efficient seawater softening and power generation using a microbial electrolysis cell-reverse electrodialysis hybrid system. Chem. Eng. J. 391: 123480. in Google Scholar

Kabay, N., Kahveci, H., Ipek, Ö., and Yüksel, M. (2006). Separation of monovalent and divalent ions from ternary mixtures by electrodialysis. Desalination 198: 74–83. in Google Scholar

Kamcev, J., Paul, D.R., and Freeman, B.D. (2017). Effect of fixed charge group concentration on equilibrium ion sorption in ion exchange membranes. J. Mater. Chem. A 5: 4638–4650. in Google Scholar

Kang, H., Kim, E., and Jung, S.P. (2017). Influence of flowrates to a reverse electro-dialysis (RED) stack on performance and electrochemistry of a microbial reverse electrodialysis cell (MRC). Int. J. Hydrogen Energy 42: 27685–27692. in Google Scholar

Karas, F., Hnát, J., Paidar, M., Schauer, J., and Bouzek, K. (2014). Determination of the ion-exchange capacity of anion-selective membranes. Int. J. Hydrogen Energy 39: 5054–5062. in Google Scholar

Kariduraganavar, M.Y., Nagarale, R.K., Kittur, A.A., and Kulkarni, S.S. (2006). Ion-exchange membranes: preparative methods for electrodialysis and fuel cell applications. Desalination 197: 225–246. in Google Scholar

Karimi, L. and Ghassemi, A. (2016a). Effects of operating conditions on ion removal from brackish water using a pilot-scale electrodialysis reversal system. Desalination Water Treat. 57: 8657–8669. in Google Scholar

Karimi, L. and Ghassemi, A. (2016b). How operational parameters and membrane characteristics affect the performance of electrodialysis reversal desalination systems: the state of the art. J. Membr. Sci. Res. 2: 111–117.Search in Google Scholar

Kim, H.K., Lee, M.S., Lee, S.Y., Choi, Y.W., Jeong, N.J., and Kim, C.S. (2015). High power density of reverse electrodialysis with pore-filling ion exchange membranes and a high-open-area spacer. J. Mater. Chem. A 3: 16302–16306. in Google Scholar

Kim, J.H., Lee, J.H., Maurya, S., Shin, S.H., Lee, J.Y., Chang, I.S., and Moon, S.H. (2016). Proof-of-concept experiments of an acid-base junction flow battery by reverse bipolar electrodialysis for an energy conversion system. Electrochem. Commun. 72: 157–161. in Google Scholar

Kim, Y., Walker, W.S., and Lawler, D.F. (2012). Competitive separation of di- vs. mono-valent cations in electrodialysis: effects of the boundary layer properties. Water Res. 46: 2042–2056. in Google Scholar

Kingsbury, R.S., Liu, F., Zhu, S., Boggs, C., Armstrong, M.D., Call, D.F., and Coronell, O. (2017). Impact of natural organic matter and inorganic solutes on energy recovery from five real salinity gradients using reverse electrodialysis. J. Membr. Sci. 541: 621–632. in Google Scholar

Klaysom, C., Moon, S.H., Ladewig, B.P., Lu, G.Q.M., and Wang, L. (2011a). Preparation of porous ion-exchange membranes (IEMs) and their characterizations. J. Membr. Sci. 371: 37–44. in Google Scholar

Klaysom, C., Ladewig, B.P., Lu, G.Q.M., and Wang, L. (2011b). Preparation and characterization of sulfonated polyethersulfone for cation-exchange membranes. J. Membr. Sci. 368: 48–53. in Google Scholar

Kotoka, F., Merino-Garcia, I., and Velizarov, S. (2020). Surface modifications of anion exchange membranes for an improved reverse electrodialysis process performance: a review. Membranes 10: 1–22. in Google Scholar

Lacey, R.E. (1980). Energy by reverse electrodialysis. Ocean Eng. 7: 1–47. in Google Scholar

Lee, H.J., Song, J.H., and Moon, S.H. (2013). Comparison of electrodialysis reversal (EDR) and electrodeionization reversal (EDIR) for water softening. Desalination 314: 43–449. in Google Scholar

Lee, J.Y., Lee, J.H., Ryu, S., Yun, S.H., and Moon, S.H. (2015a). Electrically aligned ion channels in cation exchange membranes and their polarized conductivity. J. Membr. Sci. 478: 19–24. in Google Scholar

Lee, J.Y., Kim, J.H., Lee, J.H., Kim, S., and Moon, S.H. (2015b). Morphologically aligned cation-exchange membranes by a pulsed electric field for reverse electrodialysis. Environ. Sci. Technol. 49: 8872–8877. in Google Scholar PubMed

Lee, K.H., Cho, D.H., Kim, Y.M., Moon, S.J., Kim, J.F., and Lee, Y.M. (2017). Isomeric influences of naphthalene based sulfonated poly(arylene ether sulfone) membranes for energy generation using reverse electrodialysis and polymer electrolyte membrane fuel cell. J. Membr. Sci. 535: 35–44. in Google Scholar

Lee, Y.J., Cha, M.S., Oh, S.G., So, S., Kim, T.H., Ryoo, W.S., Hong, Y.T., and Lee, J.Y. (2019). Reinforced anion exchange membrane based on thermal cross-linking method with outstanding cell performance for reverse electrodialysis. RSC Adv. 9: 27500–27509. in Google Scholar

Li, W., Krantz, W.B., Cornelissen, E.R., Post, J.W., Verliefde, A.R.D., and Tang, C.Y. (2013). A novel hybrid process of reverse electrodialysis and reverse osmosis for low energy seawater desalination and brine management. Appl. Energy 104: 592–602. in Google Scholar

Li, X., Angelidaki, I., and Zhang, Y. (2018). Salinity-gradient energy driven microbial electrosynthesis of value-added chemicals from CO2 reduction. Water Res. 142: 396–404. in Google Scholar

Linder, C. and Kedem, O. (2001). Asymmetric ion exchange mosaic membranes with unique selectivity. J. Membr. Sci. 181: 39–56. in Google Scholar

Liu, X., He, M., Calvani, D., Qi, H., Gupta, K.B.S.S., de Groot, H.J.M., Sevink, G.J.A., Buda, F., Kaiser, U., and Schneider, G.F. (2020). Power generation by reverse electrodialysis in a single-layer nanoporous membrane made from core–rim polycyclic aromatic hydrocarbons. Nat. Nanotechnol. 15: 307–312. in Google Scholar

Long, R., Li, B., Liu, Z., and Liu, W. (2017). Hybrid membrane distillation-reverse electrodialysis electricity generation system to harvest low-grade thermal energy. J. Membr. Sci. 525: 107–115. in Google Scholar

Luo, T., Abdu, S., and Wessling, M. (2018). Selectivity of ion exchange membranes: a review. J. Membr. Sci. 555: 429–454. in Google Scholar

Martí-Calatayud, M.C., Buzzi, D.C., García-Gabaldón, M., Bernardes, A.M., Tenório, J.A.S., and Pérez-Herranz, V. (2014). Ion transport through homogeneous and heterogeneous ion-exchange membranes in single salt and multicomponent electrolyte solutions. J. Membr. Sci. 466: 45–57. in Google Scholar

Matsumoto, H., Koyama, Y., and Tanioka, A. (2003). Interaction of proteins with weak amphoteric charged membrane surfaces: effect of pH. J. Colloid Interface Sci. 264: 82–88. in Google Scholar

Mehdizadeh, S., Yasukawa, M., Abo, T., Kuno, M., Noguchi, Y., and Higa, M. (2019). The effect of feed solution temperature on the power output performance of a pilot-scale reverse electrodialysis (RED) system with different intermediate distance. Membranes 9: 73.10.3390/membranes9060073Search in Google Scholar PubMed PubMed Central

Mei, Y. and Tang, C.Y. (2017). Co-locating reverse electrodialysis with reverse osmosis desalination: synergies and implications. J. Membr. Sci. 539: 305–312. in Google Scholar

Mei, Y. and Tang, C.Y. (2018). Recent developments and future perspectives of reverse electrodialysis technology: a review. Desalination 425: 156–174. in Google Scholar

Merino-Garcia, I., Kotoka, F., Portugal, C.A.M., Crespo, J.G., and Velizarov, S. (2020). Characterization of poly(acrylic) acid-modified heterogenous anion exchange membranes with improved monovalent permselectivity for red. Membranes 10. in Google Scholar PubMed PubMed Central

Mohammadi, T., Razmi, A., and Sadrzadeh, M. (2004). Effect of operating parameters on Pb2+ separation from wastewater using electrodialysis. Desalination 167: 379–385. in Google Scholar

Moreno, J., de Hart, N., Saakes, M., and Nijmeijer, K. (2017). CO2 saturated water as two-phase flow for fouling control in reverse electrodialysis. Water Res. 125: 23–31. in Google Scholar PubMed

Moreno, J., Díez, V., Saakes, M., and Nijmeijer, K. (2018a). Mitigation of the effects of multivalent ion transport in reverse electrodialysis. J. Membr. Sci. 550: 155–162. in Google Scholar

Moreno, J., Grasman, S., Van Engelen, R., and Nijmeijer, K. (2018b). Upscaling reverse electrodialysis. Environ. Sci. Technol. 52: 10856–10863. in Google Scholar PubMed PubMed Central

Nagarale, R.K., Gohil, G.S., and Shahi, V.K. (2006). Recent developments on ion-exchange membranes and electro-membrane processes. Adv. Colloid Interface Sci. 119: 97–130. in Google Scholar PubMed

Najmiyah, T.W., Aziyah, L., Hendrawan, Y., Maharani, D.M., Hawa, L.C., and Wibisono, Y. (2018). Low cost carbon electrodes to produce salinity gradient energy using reverse electrodialysis membranes: effect of feed flow velocities and addition of Mg2+. MATEC Web of Conferences, Vol. 197: 09006. in Google Scholar

Nam, J.Y., Cusick, R.D., Kim, Y., and Logan, B.E. (2012). Hydrogen generation in microbial reverse-electrodialysis electrolysis cells using a heat-regenerated salt solution. Environ. Sci. Technol. 46: 5240–5246. in Google Scholar PubMed

Ngene, I.S., Lammertink, R.G.H., Kemperman, A.J.B., Van De Ven, W.J.C., Wessels, L.P., Wessling, M., and Van Der Meer, W.G.J. (2010). CO2 nucleation in membrane spacer channels remove biofilms and fouling deposits. Ind. Eng. Chem. Res. 49: 10034–10039. in Google Scholar

Nijmeijer, K. and Metz, S. (2010). Chapter 5: salinity gradient energy. Sustain. Sci. Eng. 2: 95–139. in Google Scholar

Nikonenko, V., Zabolotsky, V., Larchet, C., Auclair, B., and Pourcelly, G. (2002). Mathematical description of ion transport in membrane systems. Desalination 147: 369–374. in Google Scholar

Ortiz-Imedio, R., Gomez-Coma, L., Fallanza, M., Ortiz, A., Ibañez, R., and Ortiz, I. (2019). Comparative performance of salinity gradient power-reverse electrodialysis under different operating conditions. Desalination 457: 8–21. in Google Scholar

Ortiz-Martínez, V.M., Gómez-Coma, L., Tristán, C., Pérez, G., Fallanza, M., Ortiz, A., Ibañez, R., and Ortiz, I. (2020). A comprehensive study on the effects of operation variables on reverse electrodialysis performance. Desalination 482: 114389. in Google Scholar

Pal, S., Dixit, S., and Lal, S. (2016). Synthesis and characterization of cation exchange PVA-g-PAA/PBI sulfone membrane for the electrolysis of sodium chloride. J. Membr. Sci. Res. 2: 147–154.Search in Google Scholar

Park, J.S., Choi, J.H., Yeon, K.H., and Moon, S.H. (2006). An approach to fouling characterization of an ion-exchange membrane using current-voltage relation and electrical impedance spectroscopy. J. Colloid Interface Sci. 294: 129–138. in Google Scholar PubMed

Pattle, R.E. (1954). Production of electric power by mixing fresh and salt water in the hydroelectric pile [19]. Nature 174: 660. in Google Scholar

Pawlowski, S., Geraldes, V., Crespo, J.G., and Velizarov, S. (2016). Computational fluid dynamics (CFD) assisted analysis of profiled membranes performance in reverse electrodialysis. J. Membr. Sci. 502: 179–190. in Google Scholar

Pawlowski, S., Crespo, J.G., and Velizarov, S. (2019). Profiled ion exchange membranes: a comprehensible review. Int. J. Mol. Sci. 20. in Google Scholar PubMed PubMed Central

Pinti, D.L. (2011). Ocean, chemical evolution of. In: Gargaud, M., et al.. (Eds.), Encyclopedia of astrobiology. Berlin, Heidelberg: Springer Berlin Heidelberg, pp. 1164–1167. in Google Scholar

Pintossi, D., Saakes, M., Borneman, Z., and Nijmeijer, K. (2019). Electrochemical impedance spectroscopy of a reverse electrodialysis stack: a new approach to monitoring fouling and cleaning. J. Power Sources 444. in Google Scholar

Pintossi, D., Chen, C.-L., Saakes, M., Nijmeijer, K., and Borneman, Z. (2020). Influence of sulfate on anion exchange membranes in reverse electrodialysis. npj Clean Water 3: 29. in Google Scholar

Post, J.W., Veerman, J., Hamelers, H.V.M., Euverink, G.J.W., Metz, S.J., Nymeijer, K., and Buisman, C.J.N. (2007). Salinity-gradient power: evaluation of pressure-retarded osmosis and reverse electrodialysis. J. Membr. Sci. 288: 218–230. in Google Scholar

Post, J.W., Hamelers, H.V.M., and Buisman, C.J.N. (2009). Influence of multivalent ions on power production from mixing salt and fresh water with a reverse electrodialysis system. J. Membr. Sci. 330: 65–72. in Google Scholar

Ramon, G.Z., Feinberg, B.J., and Hoek, E.M.V. (2011). Membrane-based production of salinity-gradient power. Energy Environ. Sci. 4: 4423–4434. in Google Scholar

Ran, J., Wu, L., Ru, Y., Hu, M., Din, L., and Xu, T. (2015). Anion exchange membranes (AEMs) based on poly(2,6-dimethyl-1,4-phenylene oxide) (PPO) and its derivatives. Polym. Chem. 6: 5809–5826. in Google Scholar

Ran, J., Wu, L., He, Y., Yang, Z., Wang, Y., Jiang, C., Ge, L., Bakangura, E., and Xu, T. (2017). Ion exchange membranes: new developments and applications. J. Membr. Sci. 522: 267–291. in Google Scholar

Ratkje, S.K., Holt, T., and Fiksdal, L. (1986). Industrial membrane processes, AIChE symposium series, Vol. 248. American Institute of Chemical Engineers (AIChE), New York, pp. 39–44.Search in Google Scholar

Reig, M., Farrokhzad, H., Van der Bruggen, B., Gibert, O., and Cortina, J.L. (2015). Synthesis of a monovalent selective cation exchange membrane to concentrate reverse osmosis brines by electrodialysis. Desalination 375: 1–9. in Google Scholar

Rijnaarts, T., Huerta, E., Van Baak, W., and Nijmeijer, K. (2017). Effect of divalent cations on RED performance and cation exchange membrane selection to enhance power densities. Environ. Sci. Technol. 51: 13028–13035. in Google Scholar PubMed PubMed Central

Safronova, E.Y., Golubenko, D.V., Shevlyakova, N.V., D’yakova, M.G., Tverskoi, V.A., Dammak, L., Grande, D., and Yaroslavtsev, A.B. (2016). New cation-exchange membranes based on cross-linked sulfonated polystyrene and polyethylene for power generation systems. J. Membr. Sci. 515: 196–203. in Google Scholar

Scialdone, O., Albanese, A., D’Angelo, A., Galia, A., and Guarisco, C. (2013). Investigation of electrode material - redox couple systems for reverse electrodialysis processes. Part II: experiments in a stack with 10-50 cell pairs. J. Electroanal. Chem. 704: 1–9. in Google Scholar

Shah, S.A., Choi, S.Y., Cho, S., Shahbabaei, M., Singh, R., and Kim, D. (2020). Modified single-wall carbon nanotube for reducing fouling in perfluorinated membrane-based reverse electrodialysis. Int. J. Hydrogen Energy 45: 30703–30719. in Google Scholar

Shukla, G., Pandey, R.P., and Shahi, V.K. (2016). Temperature resistant phosphorylated graphene oxide-sulphonated polyimide composite cation exchange membrane for water desalination with improved performance. J. Membr. Sci. 520: 972–982. in Google Scholar

Susanto, H., Fahmi, Y., Hutami, A.T., and Hadi, Y.T. (2020). Effects of fly ash loading on the characteristics of PVC-based cation exchange membranes for reverse electrodialysis. Int. J. Technol. 11: 544–553. in Google Scholar

Tamburini, A., Tedesco, M., Cipollina, A., Micale, G., Ciofalo, M., Papapetrou, M., Van Baak, W., and Piacentino, A. (2017). Reverse electrodialysis heat engine for sustainable power production. Appl. Energy 206: 1334–1353. in Google Scholar

Tanaka, Y. (2004). Concentration polarization in ion-exchange membrane electrodialysis: the events arising in an unforced flowing solution in a desalting cell. J. Membr. Sci. 244: 1–16. in Google Scholar

Tanaka, Y. (2012). Mass transport in a boundary layer and in an ion exchange membrane: mechanism of concentration polarization and water dissociation. Russ. J. Electrochem. 48: 665–681. in Google Scholar

Tedesco, M., Brauns, E., Cipollina, A., Micale, G., Modica, P., Russo, G., and Helsen, J. (2015a). Reverse electrodialysis with saline waters and concentrated brines: a laboratory investigation towards technology scale-up. J. Membr. Sci. 492: 9–20. in Google Scholar

Tedesco, M., Cipollina, A., Tamburini, A., Bogle, I.D., and Micale, G. (2015b). A simulation tool for analysis and design of reverse electrodialysis using concentrated brines. Chem. Eng. Res. Des. 93: 441–456. in Google Scholar

Tedesco, M., Scalici, C., Vaccari, D., Cipollina, A., Tamburini, A., and Micale, G. (2016a). Performance of the first reverse electrodialysis pilot plant for power production from saline waters and concentrated brines. J. Membr. Sci. 500: 33–45. in Google Scholar

Tedesco, M., Hamelers, H.V.M., and Biesheuvel, P.M. (2016b). Nernst-Planck transport theory for (reverse) electrodialysis: I. Effect of co-ion transport through the membranes. J. Membr. Sci. 510: 370–381. in Google Scholar

Tedesco, M., Hamelers, H.V.M., and Biesheuvel, P.M. (2017). Nernst-Planck transport theory for (reverse) electrodialysis: II. Effect of water transport through ion-exchange membranes. J. Membr. Sci. 531: 172–182. in Google Scholar

Tian, H., Wang, Y., Pei, Y., and Crittenden, J.C. (2020). Unique applications and improvements of reverse electrodialysis: a review and outlook. Appl. Energy 262: 114482. in Google Scholar

Tong, X., Zhang, B., and Chen, Y. (2016). Fouling resistant nanocomposite cation exchange membrane with enhanced power generation for reverse electrodialysis. J. Membr. Sci. 516: 162–171. in Google Scholar

Tufa, R.A., Curcio, E., Van Baak, W., Veerman, J., Grasman, S., Fontananova, E., and Di Profio, G. (2014). Potential of brackish water and brine for energy generation by salinity gradient power-reverse electrodialysis (SGP-RE). RSC Adv. 4: 42617–42623. in Google Scholar

Tufa, R.A., Pawlowski, S., Veerman, J., Bouzek, K., Fontananova, E., di Profio, G., Velizarov, S., Goulão Crespo, J., Nijmeijer, K., and Curcio, E. (2018). Progress and prospects in reverse electrodialysis for salinity gradient energy conversion and storage. Appl. Energy 225: 290–331. in Google Scholar

Tufa, R.A., Noviello, Y., Di Profio, G., Macedonio, F., Ali, A., Drioli, E., Fontananova, E., Bouzek, K., and Curcio, E. (2019). Integrated membrane distillation-reverse electrodialysis system for energy-efficient seawater desalination. Appl. Energy 253: 113551. in Google Scholar

Tufa, R.A., Piallat, T., Hnát, J., Fontananova, E., Paidar, M., Chanda, D., Curcio, E., di Profio, G., and Bouzek, K. (2020). Salinity gradient power reverse electrodialysis: cation exchange membrane design based on polypyrrole-chitosan composites for enhanced monovalent selectivity. Chem. Eng. J. 380: 122461. in Google Scholar

Urano, K., Masaki, Y., and Kawabata, M. (1986). Electric resistances of ion-exchange membranes in dilute solutions. Desalination 58: 171–176. in Google Scholar

Vanoppen, M., Criel, E., Walpot, G., Vermaas, D.A., and Verliefde, A. (2018). Assisted reverse electrodialysis—principles, mechanisms, and potential. npj Clean Water 1: 9. in Google Scholar

Vanoppen, M., van Vooren, T., Gutierrez, L., Roman, M., Croué, L.J.P., Verbeken, K., Philips, J., and Verliefde, A.R.D. (2019). Secondary treated domestic wastewater in reverse electrodialysis: what is the best pre-treatment?. Separ. Purif. Technol. 218: 25–42. in Google Scholar

Vaselbehagh, M., Karkhanechi, H., Takagi, R., and Matsuyama, H. (2017). Biofouling phenomena on anion exchange membranes under the reverse electrodialysis process. J. Membr. Sci. 530: 232–239. in Google Scholar

Veerman, J., de Jong, R.M., Saakes, M., Metz, S.J., and Harmsen, G.J. (2009). Reverse electrodialysis: comparison of six commercial membrane pairs on the thermodynamic efficiency and power density. J. Membr. Sci. 343: 7–15. in Google Scholar

Veerman, J., Saakes, M., Metz, S.J., and Harmsen, G.J. (2010a). Reverse electrodialysis: evaluation of suitable electrode systems. J. Appl. Electrochem. 40: 1461–1474. in Google Scholar

Veerman, J., Saakes, M., Metz, S.J., and Harmsen, G.J. (2010b). Electrical power from sea and river water by reverse electrodialysis: a first step from the laboratory to a real power plant. Environ. Sci. Technol. 44: 9207–9212. in Google Scholar PubMed

Veerman, J., and Vermaas, D.A. (2016). Reverse electrodialysis: fundamentals. In: Sustainable energy from salinity gradients. Woodhead Publishing, Sawston, UK, pp. 77–133. in Google Scholar

Vermaas, D.A., Guler, E., Saakes, M., and Nijmeijer, K. (2012). Theoretical power density from salinity gradients using reverse electrodialysis. Energy Procedia 20: 170–184. in Google Scholar

Vermaas, D.A., Veerman, J., Saakes, M., and Nijmeijer, K. (2014a). Influence of multivalent ions on renewable energy generation in reverse electrodialysis. Energy Environ. Sci. 7: 1434–1445. in Google Scholar

Vermaas, D.A., Kunteng, D., Veerman, J., Saakes, M., and Nijmeijer, K. (2014b). Periodic feedwater reversal and air sparging as antifouling strategies in reverse electrodialysis. Environ. Sci. Technol. 48: 3065–3073. in Google Scholar PubMed

Vermaas, D.A., Saakes, M., and Nijmeijer, K. (2011a). Power generation using profiled membranes in reverse electrodialysis. J. Membr. Sci. 385–386: 234–242. in Google Scholar

Vermaas, D.A., Saakes, M., and Nijmeijer, K. (2011b). Doubled power density from salinity gradients at reduced intermembrane distance. Environ. Sci. Technol. 45: 7089–7095. in Google Scholar

Vermaas, D.A., Saakes, M., and Nijmeijer, K. (2014c). Enhanced mixing in the diffusive boundary layer for energy generation in reverse electrodialysis. J. Membr. Sci. 453: 312–319. in Google Scholar

Villafaña-López, L., Reyes-Valadez, D.M., González-Vargas, O.A., Suárez-Toriello, V.A., and Jaime-Ferrer, J.S. (2019). Custom-made ion exchange membranes at laboratory scale for reverse electrodialysis. Membranes 9: 145. in Google Scholar

Walker, W.S. (2010). Improving recovery in reverse osmosis desalination of inland brackish groundwaters via electrodialysis. University of Texas Austin.Search in Google Scholar

Wang, J., Zhang, Y., Zhu, J., Hou, J., Liu, J., and Van der Bruggen, B. (2016). Zwitterionic functionalized layered double hydroxides nanosheets for a novel charged mosaic membrane with high salt permeability. J. Membr. Sci. 510: 27–37. in Google Scholar

Wang, Q., Gao, X., Zhang, Y., He, Z., Ji, Z., Wang, X., and Gao, C. (2017). Hybrid RED/ED system: simultaneous osmotic energy recovery and desalination of high-salinity wastewater. Desalination 405: 59–67. in Google Scholar

Warsinger, D.M., Chakraborty, S., Tow, E.W., Plumlee, M.H., Bellona, C., Loutatidou, S., Karimi, L., Mikelonis, A.M., Achilli, A., Ghassemi, A., et al. (2018). A review of polymeric membranes and processes for potable water reuse. Prog. Polym. Sci. 81: 209–237. in Google Scholar

Wick, G.L. (1978). Power from salinity gradients. Energy 3: 95–100. in Google Scholar

Willson, T.R., Hamerton, I., Varcoe, J.R., and Bance-Soualhi, R. (2019). Radiation-grafted cation-exchange membranes: an initial ex situ feasibility study into their potential use in reverse electrodialysis. Sustain. Energy Fuels 3: 1682–1692. in Google Scholar

Wu, Q., Wang, C., Wang, R., Chen, C., Gao, J., Dai, J., Liu, D., Lin, Z., and Hu, L. (2020). Salinity‐gradient power generation with ionized wood membranes. Adv. Energy Mater. 10: 1902590. in Google Scholar

Wu, Z.S., Ren, W., Gao, L., Liu, B., Jiang, C., and Cheng, H.M. (2009). Synthesis of high-quality graphene with a pre-determined number of layers. Carbon 47: 493–499. in Google Scholar

Xia, J., Eigenberger, G., Strathmann, H., and Nieken, U. (2020). Acid-base flow battery, based on reverse electrodialysis with bi-polar membranes: stack experiments. Processes 8: 99. in Google Scholar

Xu, T. (2005). Ion exchange membranes: state of their development and perspective. J. Membr. Sci. 263: 1–29. in Google Scholar

Yang, S.C., Choi, Y.W., Choi, J., Jeong, N., Kim, H., Nam, J.Y., and Jeong, H. (2019a). R2R Fabrication of pore-filling cation-exchange membranes via one-time impregnation and their application in reverse electrodialysis. ACS Sustain. Chem. Eng. 7: 12200–12213.10.1021/acssuschemeng.9b01450Search in Google Scholar

Yang, S.C., Choi, Y.W., Choi, J., Jeong, N., Kim, H., Jeong, H., Byeon, S.Y., Yoon, H., and Kim, Y.H. (2019b). Green fabrication of pore-filling anion exchange membranes using R2R processing. J. Membr. Sci. 584: 181–190. in Google Scholar

Yao, L., Qiu, Y., Zhao, Y., Tang, C., and Shen, J. (2020). A continuous mode operation of bipolar membrane electrodialysis (BMED) for the production of high-pure choline hydroxide from choline chloride. Separ. Purif. Technol. 233: 116054. in Google Scholar

Yip, N.Y. and Elimelech, M. (2014). Comparison of energy efficiency and power density in pressure retarded osmosis and reverse electrodialysis. Environ. Sci. Technol. 48: 11002–11012. in Google Scholar PubMed

Yip, N.Y., Brogioli, D., Hamelers, H.V.M., and Nijmeijer, K. (2016). Salinity gradients for sustainable energy: primer, progress, and prospects. Environ. Sci. Technol. 50: 12072–12094. in Google Scholar PubMed

Yoon, H., Jo, K., Kim, K.J., and Yoon, J. (2019). Effects of characteristics of cation exchange membrane on desalination performance of membrane capacitive deionization. Desalination 458: 116–121. in Google Scholar

Zabolotskii, V.I., Loza, S.A., and Sharafan, M.V. (2005). Physicochemical properties of profiled heterogeneous ion-exchange membranes. Russ. J. Electrochem. 41: 1053–1060. in Google Scholar

Zhang, B., Hong, J.G., Xie, S., Xia, S., and Chen, Y. (2017a). An integrative modeling and experimental study on the ionic resistance of ion-exchange membranes. J. Membr. Sci. 524: 362–369. in Google Scholar

Zhang, H., Ding, R., Zhang, Y., Shi, B., Wang, J., and Liu, J. (2017b). Stably coating loose and electronegative thin layer on anion exchange membrane for efficient and selective monovalent anion transfer. Desalination 410: 55–65. in Google Scholar

Zhao, J., Guo, L., and Wang, J. (2018). Synthesis of cation exchange membranes based on sulfonated polyether sulfone with different sulfonation degrees. J. Membr. Sci. 563: 957–968. in Google Scholar

Zhao, Z., Cao, H., Shi, S., Li, Y., and Yao, L. (2016). Characterization of anion exchange membrane modified by electrodeposition of polyelectrolyte containing different functional groups. Desalination 386: 58–66. in Google Scholar

Zhu, L., Zimudzi, T.J., Li, N., Pan, J., Lin, B., and Hickner, M.A. (2016). Crosslinking of comb-shaped polymer anion exchange membranes via thiol-ene click chemistry. Polym. Chem. 7: 2464–2475. in Google Scholar

Zhu, X., He, W., and Logan, B.E. (2015a). Reducing pumping energy by using different flow rates of high and low concentration solutions in reverse electrodialysis cells. J. Membr. Sci. 486: 215–221.10.1016/j.memsci.2015.03.035Search in Google Scholar

Zhu, X., He, W., and Logan, B.E. (2015b). Influence of solution concentration and salt types on the performance of reverse electrodialysis cells. J. Membr. Sci. 494: 154–160.10.1016/j.memsci.2015.07.053Search in Google Scholar

Zhu, X., Kim, T., Rahimi, M., Gorski, C.A., and Logan, B.E. (2017). Integrating reverse-electrodialysis stacks with flow batteries for improved energy recovery from salinity gradients and energy storage. ChemSusChem 10: 797–803. in Google Scholar PubMed

Received: 2020-09-04
Accepted: 2021-04-14
Published Online: 2021-07-01
Published in Print: 2022-11-25

© 2021 Walter de Gruyter GmbH, Berlin/Boston

Downloaded on 30.1.2023 from
Scroll Up Arrow