Electrodeless reverse electrodialysis

1 Introduction

Electrodialysis (ED) and reverse electrodialysis (RED) are closely related techniques that employ essentially identical stack architectures (Nijmeijer and Metz, 2010). In ED, ion transport is driven by an externally applied electric field and is widely used for desalination. In RED, ion transport is driven by concentration gradients, with ions diffusing from high- to low-salinity solutions. The resulting ionic flux is converted into electrical current at the electrodes, enabling the RED stack to operate as a power generator. The RED principle is illustrated in Figure 1a, where compartments are separated by alternating cation exchange membranes (CEMs) and anion exchange membranes (AEMs).

The relationship between power density and current density in a RED stack is shown in Figure 1b. A galvanostat was used instead of an ohmic load, allowing operation beyond the conventional RED range (0 to −80 A m^-2) (Veerman et al., 2009). Positive current densities (J > 0 A m^-2) correspond to ED operation, whereas strongly negative current densities (J < −80 A m^-2) correspond to assisted RED (ARED). ARED can be applied for pre-desalination of seawater when a lower-salinity stream, such as wastewater, is available water (Vanoppen et al., 2018; Scialdone, 2024; Pellegrino et al., 2024; Campisi et al., 2023; Ma et al., 2020; Veerman et al., 2009).

Three characteristic operating points can be distinguished in Figure 1b. At point A (short-circuit condition), ion transport is lower than in the ARED regime, but no external power supply is required. At point B, the power output is maximal; RED systems are therefore typically operated near this condition. At point C (open-circuit condition), ion transport ideally vanishes. However, due to incomplete membrane permselectivity, a small leakage current persists.

Electrodes provide the interface for conversion between ionic and electronic currents and are central to RED operation. In conventional systems, Faradaic reactions occur at the electrodes, often involving gas evolution (e.g., H₂, O₂, Cl₂) (Veerman et al., 2010). When reversible redox couples such as Fe²⁺/Fe³⁺ are employed, no net chemical conversion occurs, as the electrode potentials cancel. Alternative approaches include battery-type electrodes that actively participate in redox reactions (Kim et al., 2016; Jiang et al., 2020), and capacitive electrodes, which avoid Faradaic reactions but require periodic reversal of the feed streams to restore the electrode state (Vermaas et al., 2013). Recently, capacitive flow slurries have been introduced, eliminating the need for periodic current reversal (Simões et al., 2023).

Despite their functionality, electrodes introduce several drawbacks, including additional electrical resistance, parasitic power consumption, and the need for auxiliary systems such as electrode rinsing. Furthermore, gas evolution may pose safety risks (toxicity and explosion hazards), and periodic current reversal may be required (Vermaas et al., 2013).

Electrodeless ED and RED systems eliminate these limitations by operating as closed ionic circuits in which charge transport is mediated solely by ions. These systems are typically configured as cyclic stacks of alternating AEMs and CEMs. As shown in Figure 1b, two RED stacks can be connected via bridging compartments. Supplying the terminal compartments with the highest-salinity solution minimizes the overall system resistance. In some configurations, the cycle is interrupted by a bipolar membrane, a hydraulic shutter, or even a biological interface.

RED is driven by the salinity gradient between concentrated and dilute solutions, which may originate from natural or engineered sources. Typical examples include seawater–river water systems and combinations of desalination brine with seawater. Alternatively, closed-loop systems can be employed in which the salinity gradient is continuously regenerated using waste heat (Bevacqua et al., 2017; Papapetrou et al., 2019; Tamburini et al., 2017).

2 Existing applications

In a previous publication, we provided an overview of all known applications of RED (Veerman, 2025b). In the present work, we take a closer look at the electrodeless variants, focusing on the main principles.

2.1 RED–ED coupled desalination

Mannecke can be regarded as the founder of RED technology. In his 1952 paper Membranakkumulator (Manecke, 1952), he described a device based on ED and RED that could store electrical energy. A few years later, Pattle (1954) pointed out that the RED principle could be used to tap an enormous energy source by controlled mixing of river water and seawater (Pattle, 1954; Pattle, 1955). In 1958, Murphy proposed Osmionic Demineralization, using a salinity gradient to further desalinate a dilute solution (Murphy, 1958). In essence, this represents a direct coupling between a RED and an ED device. In 1967, Murphy and Matthews published a paper on the thermodynamic foundations of ion transport, with particular focus on osmotic demineralization (Murphy and Matthews, 1967). Unaware of Murphy’s work, Veerman independently described a similar system in 1994 (Veerman, 1994). The RED–ED coupling principle was later redeveloped by Chen et al., although they employed electrodes in their design (Chen et al., 2015).

Figure 2a illustrates the basic configuration of the RED–ED concept. Four compartments (A, B, C, D) are separated by ion-exchange membranes: a CEM between A–B and C–D, and an AEM between B–C and D–A. We constructed and tested such a system experimentally (Veerman, 1994). The motivation arose from elevated salinity in the Rhine River due to discharges from potash mines in Alsace, France (RIWA-Rijn, 2026), which hampered cultivation of salt-sensitive crops in the Westland region near the Rhine delta.

Experiments were performed with NaCl concentrations of 30 g/L in compartment A and 1 g/L in compartments B, C, and D. The D–A–B segment functioned as the RED generator powering the ED unit (B–C–D). A rough estimate of equilibrium concentrations assumes ideal membranes and unit activity coefficients. Let x g/L of salt be transferred from compartments A and C to B and D. The equilibrium condition, where RED and ED potentials are equal, can be written as:where R is the gas constant, T the temperature, and F the Faraday constant. Solving yields x = 0.88 g/L, giving a final concentration in compartment C of 0.12 g/L.

R T F ln ⁡ 30 − x 1 + x = R T F ln ⁡ 1 + x 1 − x (1)

Due to the relatively large driving force, substantial desalination can be achieved. For seawater desalination using brine–with a lesser driving force due to the lower concentration ratio - Murphy proposed increasing the number of RED cell pairs relative to the ED section. Figure 2b illustrates this configuration. The stack is fed alternately with high-concentration brine (H) and lower-concentration seawater (L). At selected positions, a desalination compartment contains seawater instead of brine; these locations correspond to desalination (product) compartments, indicated by P.

2.2 Transdermal drug delivery

Pharmaceutical drugs can enter the body via tablets, suppositories, or injections. An alternative route is transdermal administration, in which the drug reaches the bloodstream through the skin. Medicated patches allow this without medical supervision, but conventional methods are slow. Kwon et al. (2018) developed a method based on RED (Figure 2c) for ionized drugs, positioned at one end of the RED stack depending on their charge. Notably, in this configuration, the RED stack can operate without electrodes, as the electrical circuit is closed and charge transport is facilitated by ionic species (both cations and anions). The authors tested lidocaine. HCl (a cationic anesthetic) and sodium risedronate (an anionic anti-osteoporosis drug) in vitro on mouse skin. Transport rates were several times faster than diffusion-based topical administration without an electric field. The complete RED stack was miniaturized to the size of a standard patch.

2.3 Iontronic systems

Information transfer in biological systems relies on the transport of ions through nerves. In contrast, modern electrical and electronic devices operate with electrons and holes as charge carriers. Iontronics is an emerging field analogous to electronics, in which ions, rather than electrons, act as charge carriers (Liu et al., 2025). In biological organisms, ionic transport in nerves enables information transfer and control. Historically, electronic sensors interfacing with living tissue required metallic electrodes, but fully iontronic systems offer improved biocompatibility. Powering these systems has been a challenge. In 2017, Han et al. demonstrated the first iontronic logic gate powered by RED, creating a fully metal-free system (Han et al., 2017). Figure 2d illustrates their experiment. A logical OR gate was embedded in a microchip, with inputs connected to a RED generator via capillary tubes. Input interruption—representing a logical “0”—was achieved by pinching the tubes (T1 and T2). Fluorescein in the output indicated current flow: yellow when current passed, colorless otherwise.

2.4 Donnan dialysis processes

Although not strictly RED, Donnan dialysis (DDP) is closely related as it uses a salinity-gradient-induced electric field to drive ions against their concentration gradient. A DDP cell consists of a donor compartment and a sweep compartment, separated by a CEM for cations or an AEM for anions. Figure 2e depicts a DDP system for water softening (Gueccia et al., 2020). The sweep phase contains 0.5 M NaCl, sufficient to exchange Mg²⁺ (0.5 mM) and Ca²⁺ (1.5 mM) for Na⁺. Applications include removal of undesirable ions, recovery of valuable ions, or both (Asante-Sackey et al., 2021).

3 New concepts

We introduce three novel concepts, which, to the best of our knowledge, have not been previously reported.

3.1 ARED driven by RED

In a previous study, we proposed a freshwater supply system for a coastal city (Veerman, 2025b), in which seawater desalination is performed using reverse osmosis (RO). It was assumed that the city produces a wastewater volume equal to its freshwater consumption, with a similarly low salinity. Although this wastewater could, in principle, be recycled to the RO unit, several considerations argue against this approach. First, the system would be vulnerable to membrane defects, which could allow contaminated water to enter the drinking water supply and pose a public health risk. Second, the reuse of wastewater for potable purposes may be restricted by cultural or religious practices.

Figure 3a shows a configuration in which wastewater is used to extract salt from the RO feed while another part generates energy through two reverse electrodialysis (RED) modules. In RED-1, energy is recovered from the controlled mixing of a fraction of the wastewater with incoming seawater, resulting in partial desalination and improved RO performance. In RED-2, energy is extracted from another fraction of the wastewater combined with the RO concentrate. RED-1 can be operated in assisted RED (ARED) mode to further enhance desalination efficiency. Although this mode requires additional energy input, the associated reduction in RO energy demand more than compensates for this (Vanoppen et al., 2018). The specifications of the two RED stacks are provided in Table 1 (RED-1) and in Table 2 (RED-2).

TABLE 1. Specifications of the pre-desalination stack using ARED (RED-1).

TABLE 2. Specifications of the power generating RED stack (RED-2).

The system is designed for a drinking water production of 3,000 m³ day^-1, sufficient to supply a city of 10,000 inhabitants with 300 L capita^-1 day^-1. For clarity, the required treatment steps for the different streams are not included in the schematic. The RO unit is assumed to operate at a recovery of 50%, producing permeate with a salinity of 0.5 g L^-1. All solutions are assumed to be pure NaCl.

A new concept is the direct, electrodeless coupling of RED-1 and RED-2, as shown in Figure 3b. In this configuration, electrodes and associated electrical equipment are no longer required. To ensure proper matching between the two RED units, detailed knowledge of the specifications of both stacks is essential. In the present numerical example, the average concentrations of the feed streams for RED-1 are (35 + 25)/2 = 30 g L^-1 and (35 + 0.5)/2 = 17.75 g L^-1. For RED-2, these values are (49.5 + 35)/2 = 42.25 g L^-1 and (35 + 0.5)/2 = 17.75 g L^-1.

The potential across an ion-exchange membrane separating a high-concentration solution (H) and a low-concentration solution (L) is given by:where E is the electromotive force (EMF), α is the membrane permselectivity, z is the ionic valency (z = 1 for Na⁺ and Cl⁻), R is the gas constant (8.31 J mol^-1 K⁻¹), T is the absolute temperature, F is the Faraday constant (96,485 C mol^-1), a is the activity, γ is the activity coefficient, and C is the concentration in the high (H) and low (L) concentration compartments (Veerman et al., 2016).

E = α . R T z F ln a H a L = R T z F ln γ H γ L C H C L (2)

Based on Equation 2, the EMF of RED-2 is higher than that of RED-1. Since γ_L is identical for both stacks and γ_H is comparable, the ratio C_H/C_L is the determining factor. If both stacks contain the same number of cell pairs, the open-circuit voltage of RED-2 will therefore exceed that of RED-1. In addition, the higher salt concentrations in RED-2 result in a lower internal resistance. Under these conditions, RED-2 can effectively drive the assisted RED (ARED) operation of RED-1. For optimal performance in practice, however, the number of cell pairs in each stack should be adjusted to achieve proper electrical matching.

3.2 Acid–base production using a bipolar membrane

A bipolar membrane (BM) can be regarded as a sandwich consisting of a cation exchange membrane (CEM) and an anion exchange membrane (AEM). Figure 4 shows a BM incorporated in an experimental setup that is electrically driven by a galvanostat. The compartments A and F are part of the electrode rinse circuit and contain a solution of FeCl₂ and FeCl₃ in a NaCl bulk electrolyte. The compartments B, C, and E are supplied with NaCl solutions. Compartment D is extremely narrow and essentially acts as a junction. There are three modes in which the BM can operate (Pärnamäe et al., 2021).

3.2.1 Forward bias in NaCl

An electric current is passed through the stack from left to right, which is referred to as forward bias. In compartment D, water dissociates into H⁺ and OH⁻ ions. The OH⁻ ions migrate through the AEM part of the BM toward compartment C, while the H⁺ ions move through the CEM part toward compartment E. The water in D is replenished by diffusion through the membranes from compartments C and E. The net result is the formation of NaOH in compartment C and HCl in compartment E. A special application is to use the generated acid for lower the pH in a Fe²⁺/Fe³⁺ electrode rinse solution in a system as in Figures 1a, 4 (Hamelers et al., 2007).

3.2.2 Reverse bias in NaCl

We assume that compartments B, C, and E are still fed with NaCl solutions. The polarity of the electrodes is now reversed, so the current would be expected to flow from right to left. In this case there should be a migration of Na⁺ from E to D and of Cl⁻ from C to D. However, compartment D quickly becomes saturated with NaCl, leaving no space for additional ions. As a result, the current is effectively blocked, and the BM behaves like a semiconductor diode (Malti et al., 2015).

3.2.3 Reverse bias with HCl and NaOH

If compartment C is filled with NaOH and compartment E with HCl, the situation differs from the case where these compartments contain NaCl. H⁺ and OH⁻ ions now recombine in compartment D to form water. This water can leave D again by diffusion. In this mode of operation there is electrical power generated (Pärnamäe et al., 2020). If the acid and base were previously produced by forward bias in NaCl, this effectively creates an acid–base battery with an energy storage capacity roughly ten times higher than that of a salinity gradient (fresh–salt) battery (van Egmond et al., 2018). The key technical challenge is to ensure sufficiently fast water diffusion to prevent delamination of the BM (Al-Dhubhani et al., 2023).

The acid–base production system can also be powered by an electrode-free reverse electrodialysis system (Figure 3c). This principle can drive most BM-based applications, including ammonia production, SO₂ capture, and CO₂ capture (Zaffora et al., 2020; Pärnamäe et al., 2020; Cassaro et al., 2023).

3.3 An alternating current RED generator

Figure 3d illustrates an electrodeless RED generator. A ring-shaped RED stack generates an ionic loop current in the primary winding of a transformer. A mechanical shutter periodically interrupts the current, allowing transfer of electrical power to the secondary winding. Adding a second circular RED stack generating current in the opposite direction can produce a square-wave electric field, advantageous over interrupted DC. There is a rule of the thumb indicating that the weight of a transformer is inversely proportional with the frequency (Wadsworth, 2026). Therefore, efficient operation requires a shutter frequency of at least 50–60 Hz. Replacing the mechanical shutter with an iontronic interrupter remains challenging. Emerging research on membrane transistors (Cheng et al., 2022; Kaihovirta et al., 2009) may provide solutions, but for a power transformer a large active membrane transistor area is required. Closed-loop systems as used in heat-to-power systems may allow the addition of chemical aids, whereas natural feed water systems must operate without additives.

4 Conclusion and outlook

Originally, the primary application of reverse electrodialysis (RED) was energy generation. However, the field has rapidly evolved, and a broad spectrum of applications has emerged, including desalination, drug delivery, hydrogen and oxygen production, chemical synthesis and decomposition, energy storage, and heat engines. RED is now explored across both macro- and microscale domains, most often in systems that rely on electrodes. At the microscale, applications such as transdermal drug delivery and body-coupled iontronics highlight the unique advantages of electrode-free operation, enabling safer, more compact, and more integrable systems. At the macroscale, promising directions include RED–electrodialysis (RED–ED) coupling for desalination and RED–assisted RED (RED–ARED) configurations for pre-desalination in combination with reverse osmosis (RO), offering new pathways toward energy-efficient water treatment.

Looking ahead, electrodeless RED systems may provide a versatile platform for powering industrial processes involving bipolar membranes, using feed waters of natural or industrial origin. In parallel, closed-loop concepts in which ionic streams are continuously regenerated using low-grade industrial waste heat could enable sustainable and circular process designs.

Beyond these applications, the electrodeless RED generator introduced here points toward a new class of iontronic systems. It may serve as a conceptual stepping stone toward the development of ionic transistors and other controllable ion-based switches, opening perspectives for fully ionic circuits that operate without electronic components. Such systems could ultimately bridge the gap between electrochemical engineering, membrane science, and emerging iontronic technologies.

Pursuing these directions will require addressing fundamental challenges in materials, system integration, and process control. At the same time, history suggests that progress in this field will not be linear: advances are likely to be accompanied by unexpected insights and discoveries, which may redefine both the scope and impact of RED-based technologies.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Enver Güler, Atılım University, Türkiye

Reviewed by: Aydın Cihanoğlu, Ege University, Türkiye

Footnotes

AEM, anion exchange membrane; BM, bipolar membrane; CEM, cation exchange membrane; DDP, Donnan dialysis process; ED, electrodialysis; H, high concentration feedwater; L, low concentration feedwater; RED, reverse electrodialysis; RO, reverse osmosis.