Introduction

Ensuring access to clean water and sustainable energy remains a fundamental global challenge, particularly in decentralized or resource-limited settings. These two demands are often intertwined: water purification processes require significant energy input, while many energy production systems rely heavily on water resources. This interdependence calls for diverse technological efforts aimed at improving water and energy systems either independently or in tandem, depending on application needs.

On energy production, electrochemical hydrogen production has emerged as a leading technology for clean energy generation1,2. In particular, ion-exchange membranes play a central role in electrolyzer systems by enabling selective ion transport between electrodes while physically separating the anodic and cathodic reactions. This architecture offers efficient hydrogen and oxygen generation in separate chambers, thereby improving overall conversion efficiency and gas purity3,4. Beyond their role in energy systems, ion-exchange membranes are also widely used in water treatment applications5. A representing example is electrodialysis (ED), which produces freshwater utilizing the alternative stacking of both cation exchange membrane (CEM) and anion exchange membrane (AEM) with DC electric field across the stack6,7. Thus, there have been reports on ED’s ability to produce hydrogen gas and freshwater simultaneously8,9. However, protons favorable for hydrogen production are mostly trapped between the laminated space composed of CEM and AEM layers in ED, limiting continuous proton transport across the membrane. On the other hand, ion concentration polarization (ICP) purification allows protons to move freely through a stacked membrane structure composed solely of CEMs, thereby creating a favorable ionic environment that can support hydrogen production at the cathode10. Therefore, ICP purification enables a one-step process for combining purification and electrolysis.

Building upon the nanoelectrokinetic transport of ions through ion exchange membranes, this study aims to create an integrated system capable of achieving simultaneous water purification and hydrogen generation in a one-step process as shown in Fig. 1. In a nanoelectrokinetic phenomenon known as ICP through unipolar ion exchange membranes and their surrounding fluids, by applying a voltage across CEM, an ion depletion zone (IDZ) is formed on the anodic side and an ion enrichment zone (IEZ) on the cathodic side11. In the IDZ, not only salt ions but also larger impurities can be actively repelled or redirected, enabling ICP purification to remove a wide range of contaminants—including heavy metals, suspended solids, bio-contaminants, and even oils—in a single step, whereas ED is largely limited to salt ion removal10,12,13. Importantly, ICP-based desalination systems have been experimentally demonstrated to achieve higher salt removal ratios than ED at comparable or even lower energy consumption levels. This advantage stems from the unique ion transport dynamics within the depletion zone, which enhances separation efficiency without requiring alternating AEM/CEM stacks12,13. Furthermore, as previously described, the continuously supplied protons are reduced in the IEZ to generate hydrogen gas. A detailed comparison between ICP and ED is provided in the Supplementary Information. Leveraging this principle, our approach extracts purified water from the IDZ, as successfully demonstrated in our previous seawater desalination study10, while simultaneously reducing protons to hydrogen gas at the cathodic electrode. This integrated process enables multifunctional treatment capabilities that go beyond those of conventional ED or reverse osmosis (RO) systems.

Fig. 1: Schematic concept of simultaneous freshwater and hydrogen production via ion concentration polarization.

a Schematic illustration of the system for simultaneous hydrogen gas and freshwater production using ICP. In the upper anodic channel, saltwater splits into brine (green region) and purified water (sky blue region) streams by the formation and expansion of IDZ near CEM. In the lower cathodic channel, protons transported through CEM create a proton-rich stream (acidic red region), which is reduced to hydrogen gas at the cathode, eventually resulting in a proton-depleted stream (basic purple region). b Magnified view of near CEM region from (a), illustrating the competitive ion transport processes. Na+ ions, moving from the anodic channel to the cathodic channel, primarily contribute to water purification, while H+ ions play a crucial role in simultaneous hydrogen production.

Another key advantage of employing nanoelectrokinetic ICP water purification is the inherent scalability and compact design, making the technology well-suited for small-scale or portable applications. Traditional RO systems are optimized for centralized plants that treat thousands of tons of water per day and face limitations in miniaturization for decentralized use. In contrast, the ICP module presented in this study can be designed to produce only a few liters of purified water per day, and the overall throughput can be linearly increased by parallelizing multiple modules without the need to redesign the system architecture. Thanks to this scalable, lightweight form factor and the additional benefit of simultaneous hydrogen production, ICP becomes especially attractive for deployment in size- and infrastructure-constrained environments such as wearable healthcare sector14, disaster zones15,16, military operations17,18, and even space missions19,20. The hydrogen gas generated at the cathode can be collected and reconverted into electricity, for instance using a fuel cell, thus improving the total energy efficiency of the purification process. Rather than competing with large-scale desalination infrastructure to solve the energy-water nexus at global scale, the proposed ICP approach fills a critical gap by providing compact, energy-efficient “micro-utilities” that deliver both clean water and power in resource-limited settings.

Despite these advantages, a major hurdle to develop the simultaneous production system is that the movement of Na+ through ion-exchange membranes is favorable for water purification system, while the H+ transport is preferable for hydrogen production system. In practice, however, ion transport within ion-exchange membranes involves a competitive process between Na+ and H+ in both systems, necessitating a thorough analysis and a refined operational strategy to achieve optimal ion selectivity and system efficiency. Therefore, a deep understanding and precise control of the nanoelectrokinetic processes through ion exchange membrane play a pivotal role in both water purification and electrolytic hydrogen production.

In this work, we demonstrate the simultaneous production of freshwater and hydrogen gas using branched microfluidic channels with ion exchange membrane. Specifically, by incorporating microelectrodes within the channels, we qualitatively visualize simultaneous hydrogen gas bubble generation at the reduction electrode, while water purification occurs at the anodic side which is quantitatively confirmed by ion chromatography. The gas bubble formed at the cathodic side was evidenced as hydrogen by the color change of pH indicator of the solution collected at the cathodic reservoir. For the quantitative analysis and direct visualization of hydrogen gas production, we build a scaled-up device capable of producing sufficient hydrogen gas for gas chromatography (GC) analysis. Furthermore, we observed that the transport of Na+ and H+ varied in a competitive manner depending on the applied current, with H+ transport becoming more dominant as the current increased. Through this nanoelectrokinetic study of ion transport through the membrane, we propose a simple and straightforward strategy to regulate the competition between the Na+ and H+ ions, thereby enhancing proton transport across the membrane for efficient and stable hydrogen production with simultaneous water purification.

Results and Discussion

The concept for simultaneous hydrogen gas and freshwater production proposed in this study is described in Fig. 1a. The ICP desalination technology, which we intend to use as a water purification method, can effectively remove salt ions and suspended particles from saltwater10. By applying an electric field across a cation-exchange membrane, an ion depletion zone is formed adjacent to the membrane, which repels substances larger than ions, causing most of contaminants to reroute to brine channel due to the electroneutrality10,21. In the meantime, cations including Na+ and H+ preferentially transport through the membrane due to the cation-selectivity of the membrane22,23,24. Finally, one could collect the purified water as shown in Fig. 1a. During this water purification process at the upper channel, one would concurrently expect to have a proton- or sodium-rich stream flowing toward GND reservoir. By installing a microelectrode in the middle of the channel, hydrogen gas is generated on the surface of the microelectrode, and the proton-rich stream would turn to be the proton-depleted stream, which will be verified by analyzing pH changes in lower channel.

As outlined in the introduction section, unlike conventional ED systems, ICP purification enables the continuous supply of cations all the way down from the anode to the cathode because of utilizing only CEM. However, to achieve a simultaneous process, as illustrated in Fig. 1b, it is necessary to regulate the competitive transport between Na+ and H+ ions through CEM. Enhanced transport of Na+ ions benefits the purification process, while increased transport of H+ ions favors hydrogen gas generation. We first focus on the demonstration of simultaneous production of hydrogen and freshwater using a microfluidic device in Section “Demonstration of simultaneous hydrogen gas generation and freshwater collection in a microfluidic platform”. Subsequently, the same process is demonstrated with a scaled-up device in Section “Demonstration of simultaneous hydrogen gas generation and freshwater collection in a mesoscale platform”. Finally, a strategy to control aforementioned competitive transport will be discussed in Section “Competitive transport of Na+ and H+ ions through the ion-exchange membrane Analysis by transported Na+ ions”.

Demonstration of simultaneous hydrogen gas generation and freshwater collection in a microfluidic platform

The microfluidic device used for this demonstration is shown in Fig. 2a. See experimental method section for the fabrication method. The internal structure of each channel within this microfluidic device is shown in the magnified image (Fig. 2b). Note that this image consists of a fluorescent image (upper) and a brightfield image (lower), depicting the concurrent production of purified water and hydrogen gas generation, respectively. The two channels are connected by a Nafion membrane as a representative CEM, and two kinds of electrolyte solution were injected into upper and lower channels from the inlet on the left side with the flow rate of 150 nL/min using a syringe pump. For the demonstration of water purification, NaCl solution was injected into the upper channel, as a representative saltwater feed. Although any electrolyte can be used in the lower channel, referred to as the ‘hydrogen channel,’ LiCl was selected as the electrolyte because it contains a cation different from Na+, allowing for more accurate measurement of Na+ ion migration from the upper channel to the lower channel. Note that Li+ ions were used solely for analytical purposes in lab-scale experiments, while arbitrary supporting electrolytes (e.g., Na+ , K+ salts or others) could be used in practical applications. The upper channel is branched at the vicinity of the Nafion membrane to separate the purified water and remaining brine. By adding a small amount of fluorescent dye to the NaCl saltwater and observing the solution flow in dark field, a clear IDZ boundary was visualized, confirming effective water purification. Since the fluorescent dye is also charged, it is repelled from the IDZ and flows out with the brine. Therefore, in the anodic side channel, the fluorescent-colored brine solution exits through the ‘brine channel’ outside the IDZ, while the purified water, appearing black, flows through the ‘purified water channel’ located closer to the CEM. Simultaneously, a microelectrode was integrated into the hydrogen channel to clearly visualize gas generation. The generated gas bubbles are carried out through the outlet along with the rinsing water. Supplementary Video 1 shows the simultaneous generation of freshwater in the upper channel and gas bubble formation at the microelectrode in the lower channel, carried out with the flow.

Fig. 2: Microfluidic demonstration of concurrent hydrogen generation and freshwater purification.

a An image of the microfluidic device demonstrating simultaneous hydrogen gas and freshwater production. The upper channel processes the saltwater feed introduced from the inlet on the left-hand side, which is split into brine and purified water channels near the Nafion membrane. The lower hydrogen channel receives rinsing water from its inlet and carries cations transported through CEM. In this microfluidic device, the Nafion membrane is formed by coating a Nafion solution onto a glass substrate, and it connects the upper and lower PDMS microchannels. The current is applied via an anode ( + V) and a cathode (GND) positioned at the reservoirs of the brine and hydrogen channels, respectively. b Magnified views of the microfluidic device showing IDZ formation in the upper anodic channel with fluorescent image and hydrogen gas generation at the microelectrode in the lower cathodic channel with brightfield image. See Supplementary Video 1. c Concentration of Na+ in brine, purified water, and hydrogen channels at various feed saltwater concentrations (10–500 mM NaCl). Data are presented as the mean ± standard deviation from measurements on 5 different devices. The inset depicts the molar ratio of Na+ in each channel. d Visualization of pH changes with pH indicators added to both channels: the brine channel exhibited a basic environment due to OH− generation, while the hydrogen channel turned acidic due to H+ transport through Nafion. e Final pH changes in each reservoir with pH indicators, showing a basic environment in the reservoir of the hydrogen channel (purple), a neutral pH in the reservoir of the purified water channel (green), and an acidic pH in the reservoir of the brine channel (yellow). The inset depicts the pH measurements in the reservoir of the hydrogen channel via a pH meter during current application, starting with 10 μL of 10 mM KCl in the reservoir without pH indicator. f Schematic of the electrochemical reactions contributing to the observed pH changes, including hydrogen production at the cathode and potential oxidation reactions at the anode, such as chlorine gas evolution or water oxidation.

To verify the efficiency of water purification in this micro/nanofluidic device, we measured the concentration of Na+ ions in the solution flowing through each channel. As shown in Fig. 2c, the comparison of the concentrations in each stream of the brine, purified water, and hydrogen channels demonstrates that water purification occurs on this platform as we expected from the fluorescent image in Fig. 2b, even when the Na+ concentration of the saltwater feed varies from low (10 mM, 20 mM and 40 mM) to high (200 mM and 500 mM, representing brackish water and seawater, respectively). Notably, across all feed concentrations, the Na+ concentration remains lowest in the purified water channel. For feed concentration of 10–40 mM and 200 mM, the Na+ concentration in the purified water decreased to approximately 10% of the initial feed concentration, indicating effective purification (e.g., 10 mM to 1.01 ± 0.54 mM, 20 mM to 1.79 ± 0.77 mM, 40 mM to 4.63 ± 3.21 mM, and 200 mM to 25.78 ± 13.09 mM). However, for the 500 mM feed, the concentration decreased to above 20% of the initial feed concentration (e.g., 500 mM to 121.93 ± 54.43 mM), primarily due to the diminished effectiveness by instability of ICP at higher salt concentration25. As the saltwater concentration increases to higher levels, such as 500 mM, the thickness of the electric double layer on the inner walls of nanoporous structures such as Nafion can become thinner, potentially reducing ion selectivity of the membrane and weakening the ICP effect, leading to unstable IDZ formation. Furthermore, this figure shows that the Na+ concentration in the hydrogen channel is higher than that in the brine channel, indicating a stronger migration of Na+ ions from the inlet toward the hydrogen channel than toward the brine channel. This trend is further corroborated by the molar ratio of Na+ ions in each channel, as shown in the inset of Fig. 2c. This result suggests that the movement of Na+ ions into the hydrogen channel plays a significant role in the water purification process, as the removal of Na+ ions is essential for effective purification.

While it appears that the significant migration of Na+ ions into the hydrogen channel contributes substantially to the water purification process as shown in Fig. 2c, it should be emphasized that H+ ions also migrate through CEM. This phenomenon was observed in Fig. 2d. For this experiment, pH indicators were added to both feed solutions, which consisted of 10 mM NaCl in the upper channel and 10 mM LiCl in the lower channel, and the color changes of pH indicators were monitored. The pH indicator color in the hydrogen channel changed to red (i.e. acidic), while that in the brine channel shifted to blue (i.e., basic). The red color change in the hydrogen channel can be attributed to the addition of H+ ions from the upper channel. Due to the extremely amplified electric field inside IDZ26,27, especially at the interface between Nafion and saltwater, water dissociation vigorously occurs so that one has a number of generated H+ and OH− 28,29,30,31,32,33,34. Then, H+ ions can pass through CEM so that pH indicator turns red at the interface between Nafion and rinsing water, while OH− ions are rejected from IDZ so that pH indicator turns blue at “certain distance” from the Nation. This certain distance is the thickness of IDZ and one can extract fluid within it, leading to water purification process. This experiment demonstrated that the flow of H+ ions through Nafion can be observed even from the neutral NaCl saltwater. These H+ ions are advantageous for promoting simultaneous hydrogen gas production within the system. Figure 2c, d demonstrated that the migration of Na+ ions advantageous for water purification and the transport of H+ ions beneficial for hydrogen production occur concurrently through CEM. Therefore, to enable the simultaneous production of hydrogen and freshwater using a single membrane, a comprehensive study of the competitive transport between Na+ and H+ ions is required, which will be discussed in more detail in Section “Competitive transport of Na+ and H+ ions through the ion-exchange membrane Analysis by transported Na+ ions”.

Following this, we observed the color changes of pH indicator in pipette tips inserted at the outlet of the microchannel to verify the eventual fate of H+ ions. In Fig. 2d, the pH indicator near CEM of the hydrogen channel turned red inside the microsystem, indicating an acidic environment due to the hydrogen ions passing through the CEM. However, as shown in Fig. 2e, the pH indicator at the reservoir of the hydrogen channel exhibited a deep purple color, signifying high alkalinity. This observation provides indirect evidence that the gas bubbles produced in the microsystem were hydrogen gas, since hydrogen gas can be generated either by the reduction of hydrogen ions (H+ + e− → 1/2H2, E0 = 0.0 VSHE) or by the reduction of water (H2O + e− → 1/2H2 + OH−, E0 = −0.828 VSHE) at the cathode, both of which result in the pH increase of the solution due to the consumption of H+ ions or the generation of OH− ions, as illustrated in Fig. 2(f). While the other cations such as Na+ and Li+ are present near cathode in our system, the reduction of hydrogen ions or water is more favorable than the reduction of other species, such as Na+ and Li+ (Na+ + e− → Na, E0 = −2.71 VSHE; Li+ + e− → Li, E0 = −3.04 VSHE)35. Thus, we would expect all applied current to be converted to hydrogen gas, assuming a Faradaic Efficiency (FE) of nearly 100% with the Pt electrode. To neutralize pH at cathodic reservoir, a proportional amount of H+ ions would need to be supplied to the cathode side via CEM. However, as observed in Fig. 2c, d, both Na+ ions and H+ ions contribute to the charge transport through CEM. Consequently, when using saltwater (i.e., not pure water or acidic water) as the feed, the total charge passing through CEM cannot be attributed solely to H+ ion transport due to competitive ion transport. This competition results in a deficit of hydrogen ions at the cathode, leading to a significant increase in pH. Thus, the color change of the pH indicator at the cathode indirectly confirms both the production of hydrogen gas and the occurrence of competitive ion transport through CEM. Moreover, Fig. 2e corroborates that water purification is occurring concurrently with the hydrogen production process, consistent with the results shown in Fig. 2c. At the pipette tips located at the end of the brine channel and the water purified channel, the brine reservoir exhibited acidity, while the purified water reservoir showed neutral pH. In an aqueous NaCl solution, possible oxidation reactions at the anode include chlorine gas evolution (Cl− → 1/2Cl2 + e−, E0 = 1.358 VSHE) and water oxidation (1/2H2O → 1/4O2 + H+ + e−, E0 = 1.229 VSHE)35, both of which can acidify the solution (Fig. 2f). In addition, other chlorine-related oxidation reactions such as hypochlorous acid (HClO) formation (Cl− + H2O → HClO + H+ + 2e−, E0 = 1.482 VSHE) may also occur, further contributing to the local acidification due to proton generation35. This explains the acidic nature of the brine reservoir. In contrast, the reservoir of the purified water channel displayed a neutral pH (green color), as it contained purified water with most ionic species removed during the purification process within the microsystem.

Demonstration of simultaneous hydrogen gas generation and freshwater collection in a mesoscale platform

In the microfluidic platform demonstration described above, the indirect evidence of hydrogen gas generation was obtained through the color change of pH indicators, but the volume of the generated bubbles was too small for quantitative hydrogen production analysis. Therefore, we scaled up the microfluidic platform and fabricated a mesoscale device using 3D printing to provide the direct evidence of the simultaneous production of hydrogen and purified water. Notably, this mesoscale approach overcomes the limitations of microfluidic desalination systems such as low throughput and high pressure drop by enabling scalable design via 3D printing while preserving key nanoelectrokinetic behaviors. Multiple modules can be integrated in parallel to increase capacity while maintaining a compact, modular form factor. The modular and expandable units are well-suited for decentralized applications, including disaster relief, military operations, and space missions. A recent study by our group demonstrated this concept in a wearable artificial kidney platform, highlighting the potential of ICP-based systems in extreme environments14.

To realize higher throughput in hydrogen and freshwater production, however, hydrodynamic and electrokinetic instabilities must be carefully addressed. Simply increasing the width of microchannels can induce hydrodynamic instabilities, leading to IDZ boundary layer collapse and local mixing36,37,38, which decreases the quality of freshwater. In previous works, we studied the suppression of hydrodynamic instabilities using microstructures39,40,41,42,43,44,45,46,47,48. Based on this, we designed a device that maintains the hydrodynamic domain of a microfluidic platform, while the design can significantly enhance the throughput of freshwater and hydrogen production. The designed mesoscale device was fabricated using a 3D printer, as shown in Figs. 3a–c. Detailed fabrication process is described in Experimental methods. The purification chamber has an internal volume of 11 mm × 10 mm × 7 mm, and the hydrogen chamber had an internal volume of 11 mm × 10 mm × 4 mm, representing a scaled-up version of the microfluidic platform and serving as a mesoscale module in this study. The purification chamber and hydrogen chamber were separated by a Nafion sheet (contact area of 11 mm × 10 mm). To prevent mixing between purified water formed near the Nafion membrane and brine pushed outside the IDZ, fin-type structures were designed to guide the two streams separately. Argon-purged saltwater and rinsing water were introduced at the inlets of the device using syringe pumps at the flow rate of 0.2 mL/min. In the meantime, redox reactions occurred at the electrodes inserted in each chamber. This allows the production of freshwater and acidic by-products in the purification chamber. Then, the freshwater and the by-product split toward purified stream and brine stream, respectively. On the other hand, hydrogen gas was produced in hydrogen chamber and flow out naturally with waste water to collect the gas. The amount of hydrogen gas was enough to collect for GC analysis in this mesoscale device. See Supplementary Video 2 for the hydrogen gas production. Therefore, we can achieve stable IDZ separation, improved hydrogen gas collection, and quantify the quality of the produced freshwater and the amount of hydrogen generated using this mesoscale module.

Fig. 3: Scaled-up mesoscale device enabling quantitative analysis of hydrogen and freshwater production.

a Schematic diagram of the 3D mesoscale device, scaled-up from the geometry of the microfluidic device, employing the same strategy for simultaneous hydrogen gas and freshwater production. The device comprises a purification chamber and a hydrogen chamber, separated by a Nafion sheet. Fin-type structures within the purification chamber stabilize the IDZ and guide the separation of purified water and brine streams. b Cross-sectional diagram of the 3D mesoscale device. The device operates via same mechanism as the microfluidic device shown in Fig. 2b but is extended along the z-axis to form a scaled-up structure. c Actual image of the 3D-printed mesoscale device, showing the separation of streams and gas production. Saltwater and rinsing water are introduced into the purification and hydrogen chambers, respectively, and hydrogen gas is produced at the cathode in the hydrogen chamber. d Quantitative results for hydrogen gas production and Na+ ion concentration in purified water at 20 mM feed concentration under various applied currents (4, 10, 15 and 20 mA). Hydrogen production increased linearly with the current, while Na+ concentration in purified water plateaued, indicating diminishing returns in purification performance. Data are presented as the mean ± standard deviation from measurements on 4 different devices. e Ratio of Na+ ions transported to each stream at 20 mM feed concentration. Higher currents increased the fraction of Na+ ions in the hydrogen chamber, while decreasing the fractions in the purified water and brine streams. f Quantitative results for hydrogen gas production and Na+ ion concentration in purified water at 200 mM feed concentration under various applied currents (40, 100, 150 and 200 mA). Hydrogen production and Na+ concentration have similar trends of (c). Data are presented as the mean ± standard deviation from measurements on 4 different devices. g Ratio of Na+ ions transported to each stream at 200 mM feed concentration. Higher currents increased the fraction of Na+ ions in the hydrogen chamber, while decreasing the fractions in the purified water and brine streams.

Figure 3d, e show the quantitative evaluation of hydrogen and freshwater production at the representative low concentration of 20 mM, which was a previously tested concentration in the microfluidic platform for comparison. The experiment was conducted under constant current conditions by applying an electric field across the purification chamber and hydrogen chamber, each equipped with Pt wire electrodes. As shown in Fig. 3d, hydrogen production increased proportionally with the current (e.g., 1.65 ± 0.03 mL/h at 4 mA, 4.69 ± 0.11 mL/h at 10 mA, 7.43 ± 0.15 mL/h at 15 mA, and 9.91 ± 0.37 mL/h at 20 mA). The Faradaic Efficiency (FE), which represents the fraction of charges converted into hydrogen gas compared to the total charge, was approximately 98% at 4 mA and reached up to 100% at 10 mA, 15 mA and 20 mA, attributed to the use of Pt wire optimized for hydrogen generation. For water purification, when the feed concentration was 20 mM, the quality of freshwater improved with increasing current (e.g., 11.04 ± 1.04 mM at 4 mA, 6.16 ± 0.94 mM at 10 mA, 4.54 ± 0.74 mM at 15 mA, and 3.37 ± 0.83 mM at 20 mA), as shown in Fig. 3d, while the rate of improvement slightly saturated at higher currents. To track Na+ transportation, the ratio of moles of Na+ ions migrating to each stream was plotted in Fig. 3e as a function of applied current, showing that the fraction in the hydrogen chamber increased with higher current. Contrarily, the fraction in brine and purified water keeps decreasing. These analyses suggest that while Na+ ions are transported more through CEM as the current increases, they eventually reach a saturation point (indicating that purification performance saturates). Concurrently, however, hydrogen gas production continues to increase linearly with the applied current.

The same conclusion was also obtained at a high concentration representative value of 200 mM as shown in Fig. 3f, g. Given that the over-limiting current regime for ICP is expected to shift to higher voltage and current ranges at higher feed concentrations25, constant currents of 40, 100, 150, and 200 mA—ten times higher than those used for the 20 mM tests—were applied. As a result, hydrogen production increased proportionally with the applied current (e.g., 20.87 ± 1.83 mL/h at 40 mA, 53.51 ± 2.61 mL/h at 100 mA, 83.97 ± 0.83 mL/h at 150 mA, and 111.50 ± 3.73 mL/h at 200 mA), as shown in Fig. 3f. For water purification, with a 200 mM feed, the quality of freshwater improved with increasing current (e.g., 125.79 ± 5.40 mM at 40 mA, 59.18 ± 10.79 mM at 100 mA, 52.04 ± 4.51 mM at 150 mA, and 46.30 ± 6.68 mM at 200 mA). The slightly lower purification levels observed with the 200 mM feed compared to the 20 mM feed are interpreted as a weakening of ICP due to the thinner electric double layer within the nanostructure of the Nafion membrane at higher concentrations. However, similar to the 20 mM case, we observed that as the current increased, more Na+ ions migrated to the hydrogen chamber in Fig. 3g, but the rate of decrease in freshwater concentration relative to the feed concentration diminished at higher currents (Fig. 3f). The latter observation is attributed to the competitive transport of Na+ and H+ ions through the ion-exchange membrane, which will be discussed in the next section.

Next, we discuss potential applicability of hydrogen produced concurrently with freshwater in our system. As observed above, although the quality of freshwater improves with increasing current, the degree of improvement reaches a saturation point. This indicates that achieving higher purification quality requires significantly greater energy consumption, which has been a primary limitation in most of desalination systems. However, by utilizing the hydrogen produced simultaneously with ICP purification, we can improve the energy efficiency of ICP purification systems through the following approach, inspired by hybrid type electric vehicle. Producing 1 liter of freshwater from 200 mM brackish water reduced to a concentration of 50 mM at 200 mA requires 397.71 Wh, which is a substantial energy demand but is able to produce 1.59 g of hydrogen at the same time using our device. Converting this hydrogen back into electricity using a commercial fuel cell with a 60% energy conversion efficiency yields 31.80 Wh of electrical energy. When this energy is recycled to power the ICP purification system, the energy efficiency improves by 8.00% compared to conventional ICP purification systems that do not recycle hydrogen. For detailed calculations on energy efficiency, refer to the Supplementary Information. In the desalination research community, while the energy consumption of our system may appear higher than that of conventional ED (20–200 Wh/L at optimized condition) or RO systems (3.5–5 Wh/L, achievable only at large-scale plants), it should be noted that our system performs functions beyond ionic desalination including simultaneous hydrogen production and non-ionic contaminant rejection, which cannot be fully captured by a direct Wh/L comparison alone. Importantly, by optimizing the device through strategies such as enhancing surface conduction45,46,47 or utilizing recirculation flow[41](https://www.nature.com/articles/s43246-025-01001-z#ref-CR41 “Lee, H. et al. Overlimiting Current in Nonun