Cyano‐Triggered Strong Anion‐π Interactions: Unlocking Anion–Cation Adsorption Bifunction

ABSTRACT

Due to the inherent differences in anion and cation adsorption mechanisms, designing a single Faradaic material that functions as an efficient symmetric capacitive deionization (CDI) electrode for desalination poses a significant challenge. Herein, by employing a simple n‐type azabenzene compound, hexaazatrinaphthalene (HATN), as a template, we introduce strong electron‐withdrawing cyano (C≡N) groups to drastically reduce its surface electron density, constructing an electron‐deficient system with a positive surface potential and a highly positive permanent quadrupole moment (Q _zz), denoted as HCNAP. This unique electronic structure triggers anion‐π interactions and consequently enables anion adsorption. Meanwhile, the C═N and C≡N groups on the HCNAP skeleton maintain the function of cation adsorption. As a proof of concept, the symmetric CDI device fabricated with HCNAP exhibits outstanding desalination performance in a 500 mg L⁻¹ NaCl solution, achieving a salt adsorption capacity of 53.88 mg g⁻¹ and a removal rate of 10.1 mg g⁻¹ min⁻¹. Theoretical calculations and experimental results clearly reveal the adsorption mechanism of Na⁺ and Cl⁻. Besides, HCNAP presents favorable adsorption toward three additional cations and anions. This innovative strategy establishes a new paradigm for constructing bifunctional Faradaic electrodes for highly efficient desalination.

The introduction of electron‐withdrawing cyano groups yields an electron‐deficient aromatic system that synergistically realizes anion capture via anion‐π interactions while providing additional coordination sites for cations, thereby achieving efficient cation‐anion co‐adsorption within a single organic material for symmetric capacitive deionization (CDI) desalination.

1. Introduction

The rapid development of industry and agriculture, coupled with population growth, has led to a severe freshwater shortage, making brackish water desalination a promising solution [1, 2]. Capacitive deionization (CDI) is an emerging water treatment technology that has the advantages of low energy consumption, a green environment, and energy recyclability, providing a resource‐friendly method for ion removal from brackish water [3, 4]. Within the CDI technology framework, symmetric configurations utilizing identical electrode materials offer distinct merits, including simplified manufacturing processes, inherent electrode compatibility, and favorable charge balance [5]. These characteristics render them particularly desirable for applications demanding high operational stability and economic viability. Currently, research on electrode materials for symmetric CDI remains predominantly focused on carbon based materials [6]. However, their reliance on the electric double‐layer capacitance (EDLC) mechanism fundamentally limits the salt adsorption capacity (8–20 mg g⁻¹) and desalination kinetics, thereby hindering their practical application [7, 8, 9]. To overcome these performance bottlenecks, Faradaic electrode materials, such as CoNi‐LDH@HMCSs, CS@MnO₂‐2 have recently been explored for symmetric CDI [10, 11]. By enabling ion storage through reversible redox reactions, these materials can significantly enhance ion storage capacity and accelerate ion removal kinetics, representing a highly effective approach to boost the performance and sustainability of CDI [8]. However, designing a single Faradaic material capable of efficient desalination as a symmetric CDI electrode remains challenging, primarily due to the inherent differences in anion and cation adsorption mechanisms.

Redox‐active organic materials have recently generated significant attention as competitive candidates for CDI [12, 13]. In contrast to inorganic Faradaic electrodes, organic materials exhibit superior structural design flexibility and enhanced biocompatibility. Through rational molecular design, redox‐active sites can be incorporated into their framework, significantly improving ion storage capacity, ion removal kinetics, and cycling stability [14, 15]. Moreover, these materials exhibit unique structural adaptability during ion capture through a functional group mediated coordination mechanism, which effectively alleviates lattice expansion and structural strain [16]. In general, these redox‐active organic materials can be broadly classified into n‐type and p‐type [17]. In n‐type materials, the reduction of active sites generates negative charges that bind with cations in the electrolyte, enabling cation adsorption [18]. Conversely, in p‐type materials, the oxidation process produces positive charges that interact with anions, thereby facilitating anion removal [19]. Both n‐type (as cathode) and p‐type (as anode) organic materials have demonstrated excellent desalination performance when configured in hybrid CDI with activated carbon (AC). However, only a few studies have reported their use as bifunctional electrodes capable of adsorbing both cations and anions [5, 6, 10, 20, 21]. Despite this scarcity, such findings highlight the great potential of organic active materials for broader applications in CDI. Taking n‐type electrode materials as an example, their electron‐rich nature offers a distinct advantage in capturing cations, rendering them highly suitable for this purpose [22]. Nevertheless, the inherent electrostatic repulsion between anions in the electrolyte and the electron‐rich aromatic π‐systems makes n‐type organic materials generally ineffective for anion capture [23]. Therefore, overcoming this intrinsic repulsion to achieve concurrent adsorption of both cations and anions presents a significant challenge.

Studies have observed the presence of anion‐π interactions between anions and electron‐deficient aromatic compounds. First experimentally confirmed in 1993 [24], these attractive noncovalent interactions have been widely applied in supramolecular chemistry, molecular recognition, and polymer science, while their applications in battery systems have also attracted increasing attention in recent years [25, 26, 27, 28]. A key determinant of anion–π interaction strength is the permanent quadrupole moment (Q _zz) value of the electron‐deficient aromatic ring, where an increased positive Q _zz value results in enhanced anion‐π interactions [29, 30]. Given the presence of anions in electrolytes, anion‐π interactions hold promise for facilitating anion adsorption [23, 31]. However, to the best of our knowledge, no studies have yet reported the application of anion‐π interactions in organic materials for CDI.

Inspired by the aforementioned studies, we focused our attention on a structurally simple n‐type azabenzene compound, hexazatrinaphthalene (HATN) [32, 33, 34]. This molecule possesses six C═N active sites, making it suitable as a cathode material for cation adsorption. However, the inherent repulsive interaction between its electron‐rich aromatic π‐system and anions severely limits its anion capture capability. Herein, we engineer the molecular structure of HATN by embedding six strong electron‐withdrawing cyano (C≡N) groups, constructing an electron‐deficient system named HCNAP. Detailed theoretical calculations and characterization analyses confirm that this modification reverses both the surface potential and Q _zz value from their initial negative states to positive ones, thereby substantially reducing the surface electron density, which in turn triggers favorable anion‐π interactions. As a result, HCNAP exhibits efficient Cl⁻ adsorption, a dramatic reversal from the Cl⁻‐repelling nature of its precursor HATN, as demonstrated in Figure 1. Besides, its capability to also capture Br⁻, NO₃ ⁻ and SO₄ ²⁻ collectively demonstrates that constructing anion‐π interactions represents a viable design principle for boosting anion uptake in organic material systems. Meanwhile, the cyano groups in HCNAP can enhance the adsorption ability of Na⁺ through additional coordination. Moreover, the lone pair electrons from nitrogen atoms in cyano groups act as hydrogen bond acceptors, promoting a stable hydrogen‐bonding network in HCNAP that concurrently enhances both structural stability and electrical conductivity. These structural advantages endow the prepared HCNAP electrode with both high specific capacitance (331.8 F g⁻¹ at 1 A g⁻¹) and ultralong cycling stability (89.9% capacitance retention over 10 000 cycles at 5 A g⁻¹). Strikingly, the symmetrical HCNAP//HCNAP CDI device demonstrates a top‐tier salt adsorption capacity of 53.88 mg g⁻¹ at 1.2 V in a 500 mg L⁻¹ NaCl solution, surpassing not only its asymmetric counterparts AC//HCNAP and HCNAP//AC but also symmetric AC//AC and HATN//HATN, as well as the vast majority of reported symmetrical electrodes, while maintaining excellent desalination performance over 30 cycles. This work realizes a single organic material capable of simultaneously capturing both anions and cations in CDI, which provides a novel symmetrical CDI device and establishes a pioneering strategy for designing dual‐functional organic electrode materials.

FIGURE 1.

Illustration of C≡N group‐induced electron modulation and the triggered anion‐π interactions for dual‐ion adsorption.

2. Results and Discussion

In the presence of acetic acid as the catalyst, monomers trquinoyl and 4,5‐diaminophthalonitrile were reacted under simple reflux conditions to synthesize HCNAP, which incorporates both strong electron‐withdrawing cyano groups and pyrazine redox‐active sites. Meanwhile, the basic azine compound, hexaazatrinaphthalene (HATN), was also synthesized using a similar method for a comparison, employing monomers trquinoyl and phenylenediamine as the reactants. The molecular structures of HCNAP and HATN were confirmed by nuclear magnetic resonance (NMR) spectroscopy (solid‐state ¹³C and ¹H NMR) and mass spectrometry (MS), with detailed discussions illustrated in Figures S1–S5, verifying the successful synthesis of the target products. Furthermore, as shown in Figure 2a, the Fourier transform infrared (FT‐IR) spectrum further reveals the presence of imine (C═N) bonds in HCNAP via a characteristic adsorption peak around 1528 cm⁻¹ [35]. Moreover, compared with the HATN, a new stretching vibration band ascribed to the C≡N bond appears at 2240 cm⁻¹ in HCNAP, indicating the successful incorporation of cyano groups into the framework of HCNAP [36]. Additionally, the presence of C═N and C≡N bonds within HCNAP is further supported by Raman spectroscopy (Figure S6). The chemical composition and valent state of the as‐obtained products were examined by x‐ray photoelectron spectroscopy (XPS). The XPS survey spectrum of HCNAP (Figure S7) exhibits distinct signals corresponding to the C and N elements. In the high‐resolution C 1s spectrum (Figure 2b), four clearly fitting peaks of HCNAP are observed, which are attributed to the C═C (284.0 eV), C–C (284.8 eV), C═N (286.1 eV), and C≡N (288.5 eV) bonds, respectively [37]. Compared to the HATN, the peak of C═N bond in HCNAP shifts toward higher binding energy, indicating a reduced electron density at C═N bond due to the introduction of six C≡N groups, which leads to the formation of an electron‐deficient aromatic system. Scanning electron microscopy (SEM) (Figure S8a) and transmission electron microscopy (TEM) (Figure S8c) images demonstrate that the HCNAP possesses nanosheet morphology. Moreover, the corresponding element mapping (Figure S8d) confirms the uniform distribution of C and N elements in HCNAP.

FIGURE 2.

Structural characterizations. (a) FT‐IR spectra of HCNAP and HATN. (b) C 1s XPS spectra of HCNAP and HATN. (c) PXRD pattern and Pawley refinement based on the HCNAP P62m space group. (d) Structural simulation diagram of HCNAP. (e) and (f) HRTEM images of HCNAP. (g) N₂ adsorption/desorption isotherms and the corresponding pore size distribution of HCNAP.

The crystal structure of HCNAP was determined by powder x‐ray diffraction (PXRD) analysis. As shown in Figure 2c, the PXRD pattern displays two intense diffraction peaks at 18.11° and 27.83°, which are attributed to (131) and (200) crystal planes, respectively, suggesting high crystallinity. The crystal model of HCNAP was also simulated, with the final parameters presented in Table S1. Based on the Pawley refinement, the AB flip stacking mode agrees well with the experimental PXRD data, as demonstrated by the good agreement factors of R _wp = 2.84% and R _p = 1.71%. The calculated lattice parameters are a = 6.42 Å, b = 26.32 Å, c = 15.21 Å, α = 89.96°, β = 90.11°, and γ = 90.05° in a P62m group space. The structural model indicates an interlayer spacing of 0.32 nm in HCNAP (Figure 2d–i), corresponding to π–π stacking interactions as evidenced by the (200) diffraction peak. This interlayer distance was independently verified by high‐resolution transmission electron microscopy (HRTEM) (Figure 2e). The C≡N groups in HCNAP can form intermolecular hydrogen bonds with C–H groups, creating a hydrogen bond network that promotes rapid charge transfer [38, 39, 40]. As displayed in Figure 2d‐ii, the constructed HCNAP model exhibits a hydrogen bond with a bond length of 2.49 Å, consistent with typical values reported in previous literature [40]. Furthermore, the model reveals that the hydrogen bond network possesses polygonal pores with a diameter of 6.055 Å (Figure S9 and Figure 2d‐iii), which aligns with the result obtained from HRTEM measurement (Figure 2f). The surface area and pore size distribution were studied by N₂ adsorption–desorption measurement. As displayed in Figure 2g, HCNAP has a specific surface area of 106.29 m² g⁻¹. Moreover, pore size analysis further confirms the existence of a microporous structure with a pore diameter of approximately 6 Å. These combined results collectively demonstrate the successful fabrication of HCNAP, which features a π–π stacking motif integrated with an intermolecular hydrogen bond network.

Zeta potential measurements in Figure 3a reveal a positive value of 45.7 mV for HCNAP, in contrast to a negative surface charge of −39.2 mV for HATN. This inversion indicates that the introduction of C≡N groups modulates the surface electron density, creating an electron‐deficient aromatic ring in HCNAP. Consequently, this configuration activates the anion‐π interactions, enabling the adsorption of anions from solution. In general, the strength of anion‐π interactions can be evaluated by Q _zz [29, 30]. As shown by theoretical calculations (Figure 3b), HATN exhibits a negative Q _zz value of −21.95, suggesting repulsive interactions with anions and thus poor anion capture capability. Interestingly, the introduction of six C≡N groups endows HCNAP with a large positive Q _zz value of 81.12, indicating strong anion‐π interactions and thus facilitating efficient anion capture. Detailed theoretical calculations were further conducted to study the charge distribution of HCNAP and HATN. As displayed in Figure 3b, the calculated electrostatic potential (ESP) diagrams display their corresponding charge distribution. In the ESP map, blue regions indicate negative ESP with high electron density, while orange shows positive zones with low electron density. In HCNAP, the C═N and C≡N regions exhibit distinct negative ESP characteristics and are identified as the active sites for cation adsorption. By contrast, in HATN, only the C═N groups display a negative ESP region. This difference significantly improves the cation‐capture capacity of HCNAP. Notably, apart from the active sites C═N and C≡N exhibiting blue electron‐rich regions, the other domains present electron‐deficient characteristics. Specifically, compared with HATN, the introduction of the C≡N groups imparts a markedly more electron‐deficient character to the aromatic ring in HCNAP, which is consistent with the zeta potential trends. These results collectively confirm that HCNAP possesses both promising anion‐adsorption potential and superior cation‐adsorption capability relative to HATN.

FIGURE 3.

Structural properties: (a) Zeta potentials. (b) Calculated Q _zz values and electrostatic potential of HCNAP and HATN. LOL‐π color‐filled contour image of (c) HCNAP and (d) HATN. (e) Plots of RDG versus sign(λ ₂)ρ for HCNAP. (f) Calculated LUMO‐HOMO energy levels of HCNAP and HATN. (g) TGA curves of HCNAP and HATN under N₂ flow.

The theoretical calculation of the localized orbital locator (LOL) is an effective method to reveal the delocalization of π electrons within molecules. Normally, an LOL value above 0.5 implies the occurrence of electron delocalization [41]. Figure 3c clearly illustrates the electron delocalization in HCNAP, with LOL values above 0.5, verifying efficient π‐electron delocalization. Furthermore, in contrast to HATN (Figure 3d), the C≡N groups in HCNAP establish spatial conjugation with the aromatic ring, leading to a more extended π‐electron delocalization. This pronounced electron delocalization in HCNAP improves its electrical conductivity, which consequently benefits the reaction kinetics [42, 43]. To go further, the reduced density gradient (RDG) diagram in Figure 3e reveals the distinct green peaks in the scope of −0.02 to 0.00 a.u (λ ₂)ρ [44], providing additional evidence for the existence of π–π interactions between adjacent planes of the HCNAP. The energy levels of the lowest unoccupied molecular orbital (LUMO) and the highest occupied molecular orbital (HOMO) of HCNAP and HATN were also analyzed using theoretical calculations. As shown in Figure 3f, HCNAP exhibits a lower HOMO energy (−7.95 eV) than HATN (−6.40 eV), indicating a stronger electron affinity. The energy gap (E _g) between the HOMO and LUMO is often used as an indicator of the electrical conductivity of a molecule, and a narrower gap means that easier electron transition from the HOMO orbital to the LUMO orbital [45]. The E_g value of HCNAP is determined to be 3.57 eV, while the value of HATN is 3.77 eV, suggesting that the introduction of a cyano group helps to reduce the E _g, which is conducive to the rapid electron transfer. In addition, the electronic conductivity of HCNAP and HATN was measured at room temperature using the four‐probe method, yielding values of 9.43 × 10⁻³ S/cm and 7.19 × 10⁻³ S/cm, respectively. It is evident that HCNAP possesses higher electrical conductivity than HATN, which is consistent with the predictions from the above calculations. Collectively, these results confirm that HCNAP exhibits favorable electronic conductivity, which is beneficial for the overall electrochemical reaction kinetics.

The aqueous stability of HCNAP was also comprehensively evaluated. Specifically, the HCNAP electrode was immersed in 10 mL of 1 M NaCl, 1 M KCl,1 M CaCl₂, and 1 M MgCl₂ electrolytes, respectively, and the color changes of the electrolytes were observed after 2 weeks. As shown in Figure S10, the four different electrolytes remain colorless and transparent, suggesting the insolubility of HCNAP in aqueous media. Subsequent ultraviolet‐visible (UV‐vis) spectral analysis (Figure S11a,b) reveals that no characteristic absorption peaks of HCNAP are detected after soaking in these electrolytes for 2 weeks, demonstrating its exceptional chemical stability. Further characterization of post‐immersion samples by XRD and FT‐IR (Figure S11c,d) spectroscopy reveals that both the crystalline structure and functional groups are well preserved, with no detectable structural changes. Besides, the thermal stability of HCNAP was also evaluated by thermogravimetry (TGA). As displayed in Figure 3g, HCNAP shows a thermal decomposition temperature of 600°C, with only 7.57% mass loss due to water evaporation prior to decomposition. In contrast, HATN decomposes at 400°C, implying that the introduction of cyano groups contributes to the thermal stability of HCNAP, which is mainly attributed to the formation of a robust hydrogen bond network. Based on the above conclusions, the unique combination of strong anion‐π interactions, multiple redox‐active sites, good electrical conductivity, and structural stability renders HCNAP a highly promising electrode material for CDI.

To validate this, the electrochemical performance of HCNAP was first evaluated in 1 M NaCl solution using a three‐electrode system. For comparison, the electrochemical performance of HATN was also investigated. As shown in Figure 4a, HCNAP exhibits superior electrochemical performance compared to HATN, evidenced by more numerous redox peaks (detailed discussions are in the Supporting Information), higher peak currents, and a larger integrated CV area. These features collectively demonstrate that the introduced cyano groups significantly enhances the charge storage capacity. By integrating the enclosed area of the CV curve, the specific capacitance of HATN is calculated to be 52.5 F g⁻¹. In contrast, HCNAP delivers a remarkably higher value of 331.8 F g⁻¹, directly demonstrating its superior charge storage performance, which is further verified by galvanostatic charge‐discharge (GCD) analysis. As displayed in Figure 4b, the charge‐discharge time of the HCNAP electrode is markedly longer than that of HATN. Moreover, across various current densities, HCNAP consistently exhibits prolonged charge‐discharge durations compared to HATN (Figure S12), further corroborating its exceptional capacitance performance, which is consistent with the observations from the CV curves. Additionally, HCNAP shows excellent electrochemical reversibility and rate performance, with well‐preserved CV profiles (Figure S13) and 50.8% capacitance retention even at a 50‐fold higher scan rate (Figure 4c). Furthermore, the slopes (b‐value) of the redox peak current versus scan rate for HCNAP range between 0.696 and 0.938 (Figure S14 and Figure 4d), indicating that its capacitance is governed by a mixed mechanism involving both diffusion‐controlled and surface‐controlled processes. The relative percentage of each contribution is quantified in Figure 4e and Figures S15–S17. Within the scan rate range of 1–10 mV s⁻¹, HCNAP exhibits greater capacitive contribution compared to HATN, providing evidence for its reduced ion diffusion limitations and enhanced ion kinetics.

FIGURE 4.

Electrochemical performances of HATN and HCNAP in 1 M NaCl solution: (a) CV curves at a scan rate of 1 mV s⁻¹. (b) GCD curves at a density of 1 A g⁻¹. (c) Specific capacitance of HCNAP at various scan rates. (d) b‐values obtained from redox peaks of HCNAP. (e) Capacitive contributions of HCNAP at various scan rates. (f) Nyquist plots of the EIS spectra. (g) Ex situ Nyquist plots. (h) Bode plots of HCNAP at different charging and discharging states.

Electrochemical impedance spectroscopy (EIS) was further conducted to reveal the reaction kinetics of HCNAP and HATN electrodes. Figure 4f presents the Nyquist plot under the open‐circuit voltage. HCNAP displays a near‐vertical slope in the low‐frequency region, indicating near‐ideal capacitive behavior and fast ion diffusion kinetics. In contrast, HATN shows a smaller slope of ∼45°, suggesting more sluggish diffusion. This pronounced difference validates that the introduced cyano groups play a positive role in enhancing ion storage kinetics. On this basis, the kinetic behavior of HCNAP during the charge‐discharge process was also thoroughly studied via EIS. As shown in Figure 4g, HCNAP consistently exhibits a low charge transfer resistance (R _ct < 1 Ω) across various potentials, confirming rapid and stable charge transfer kinetics during the electrochemical charge‐discharge process. A Bode diagram was utilized to elucidate the redox kinetics of the HCNAP electrode during the charge‐discharge process. Typically, ideal EDLC materials exhibit a phase angle close to ‐90° in Bode plots, whereas pseudocapacitive materials often show a deviation from this theoretical value due to involvement of Faradaic reactions [46]. As shown in Figure S18, the phase angle of the HCNAP electrode consistently deviates from −90° across various applied potentials, further confirming that its charge storage behavior is predominantly governed by a pseudocapacitive mechanism. The reaction kinetics were further probed by calculating the characteristic relaxation time (τ ₀) from the voltage‐dependent Bode plot in Figure 4h, using the formula τ ₀ = 1/f ₀ [47, 48], (f ₀ is the frequency at −45°, marked by the white dashed line, ranging from 0.011 to 5.20 Hz). As shown in Table S2, the calculated results reveal a minimum τ ₀ of only 0.19 s within the potential range where pronounced redox reactions occur. These kinetic analyses confirm the fast redox reaction kinetics of HCNAP, which can be attributed to its good electronic conductivity arising from the tailored molecular structure. Stability is also one of the important indexes to evaluate the electrochemical performance of electrode materials. Therefore, the cyclic stability of HCNAP was estimated by a long‐term GCD test, as shown in Figure S19. At a current density of 5 A g⁻¹, HCNAP exhibits exceptional long‐term cycling durability, maintaining a capacitance retention of 89.9% even after 10 000 cycles (the data obtained after 50 cycles as the initial cycle), which indicates its ultra‐high stability. Notably, HCNAP undergoes an electrochemical activation process during the stability test. The samples after 200 and 1000 GCD cycles were collected and characterized by XRD, TEM, XPS, FT‐IR, BET, and water contact angle measurements (Figure S20). The results demonstrate that the phase, morphology, and composition of HCNAP remain unchanged during the cycle. Such an activation phenomenon arises primarily from the synergistic optimization of improved hydrophilicity and increased specific surface area of HCNAP driven by the electric field. These factors collectively promote the exposure of more active sites, thereby leading to a performance enhancement period during the stability test.

After evaluating the capacitive performance and stability of HCNAP in a three‐electrode system, a symmetrical HCNAP//HCNAP CDI device was assembled using HCNAP as both cathode and anode, and its CDI performance was investigated in a 500 mg L⁻¹ NaCl solution at 1.2 V under a two‐electrode system. For comparison, the desalination performance of symmetric HATN//HATN and commercial active carbon (AC) AC//AC CDI devices was also evaluated. The standard curves of various salt solutions are presented in Figures S21–S23. A schematic diagram of a symmetrical HCNAP//HCNAP device is shown in Figure 5a. Following the application of an electric field to the symmetrical HCNAP//HCNAP device, the anions and cations move rapidly toward the anode and cathode, respectively, causing a rapid drop in solution concentration, which is evidenced by a marked decrease in the time‐dependent conductivity curve (Figure S24). Moreover, the desalination process is highly reversible, as indicated by the rapid restoration of the initial conductivity once the electric field is removed. Conversely, the conductivity of the HATN//HATN and AC//AC devices changes very slowly over time (Figure S24), which shows that their desalination performance is significantly inferior to that of the HCNAP//HCNAP device. Based on the data obtained from the conductivity meter, the desalination performance of HCNAP//HCNAP, HATN//HATN, and AC//AC devices was calculated. As shown in Figure 5b–i, the symmetric HCNAP//HCNAP device achieves a salt adsorption capacity (SAC) of 53.88 mg g⁻¹ within 3600 s, significantly outperforming HATN//HATN (8.98 mg g⁻¹) and AC//AC (10.78 mg g⁻¹) devices. In addition, ion chromatography analysis confirms that the adsorption capacities of HCNAP//HCNAP device for Na⁺ and Cl⁻ are 47.42 and 16.53 mg g⁻¹, respectively, which are much higher than those of HATN//HATN (16.65 mg g⁻¹ for Na⁺ and 2.82 mg g⁻¹ for Cl⁻) and AC//AC (8.60 mg g⁻¹ for Na⁺ and 7.44 mg g⁻¹ for Cl⁻) devices (Table S3). The above results confirm that the excellent desalination performance of HCNAP//HCNAP stems from the synergistic adsorption contributions of anions and cations, rather than the effect of a single ion. Moreover, it exhibits a distinct advantage as a symmetrical electrode for high‐efficiency desalination, overcoming the constraints of HATN, which functions exclusively for cation adsorption, and avoiding the inherent limitations of commercial AC, whose capacitance is solely dependent on EDLC.

FIGURE 5.

Desalination performance of the assembled CDI devices: (a) Schematic illustration of CDI working mechanism of symmetrical HCNAP//HCNAP device. (b) SAC performance diagram at 500 mg L⁻¹ NaCl solution and a voltage of 1.2 V. (c) SAC performance of HCNAP//AC device under different electrolytes. (d) Comparison of the desalination performance of the seven assembled CDI devices. (e) Comparison of desalination performance of symmetric HCNAP//HCNAP device with various reported electrode materials. (f) Kim–Yoon CDI Ragone diagram. (g) The calculated energy consumption and charge efficiencies of HCNAP//HCNAP device in contrast with AC//HCNAP and HCNAP//AC devices. (h) SAC of the HCNAP//HCNAP device at different applied voltages. (i) SAC of HCNAP//HCNAP device under different electrolytes. (j) Long‐term cycling performance of HCNAP//HCAP device at a constant voltage of 1.2 V.

To verify that the HCNAP has excellent anion adsorption and enhanced cation adsorption capability compared to its counterpart HATN, hybrid CDI devices were assembled and studied in detail. Specifically, HCANP//AC and HATN//AC devices were constructed to highlight anion adsorption, while AC//HCANP and AC//HATN were designed to probe the enhanced cation adsorption. As displayed in Figure S25, the HCNAP//AC device shows a sharp conductivity decrease over time, contrasting with the minimal change in the HATN//AC device, indicative of more effective desalination. Based on the conductivity measurements, the SAC of the HCANP//AC device is four times that of HATN//AC device (18.52 vs. 5.39 mg g⁻¹) (Figure 5b‐ii). Given the identical cathode, this enhancement effect can be attributed to the stronger adsorption capacity of the HCNAP anode for Cl⁻. This conclusion is corroborated by ion chromatography data, which indicate a much higher Cl⁻ adsorption capacity in HCANP//AC device compared to HATN//AC device, while the Na⁺ adsorption capacities of both devices remain comparable (Table S3). To investigate whether the anion adsorption capability of HCNAP is universally applicable, the desalination performance of the HCNAP//AC device was further evaluated using three other different electrolyte solutions (NaBr, NaNO₃, and Na₂SO₄). As shown in Figure 5c and Figure S27, its SAC for these three solutions is 30.15, 23.04, and 17.83 mg g⁻¹, respectively, which are comparable to the performance in NaCl. Moreover, the molar SAC reveals the affinity order of HCNAP for the anions as Cl⁻ > Br⁻ > NO₃ ⁻ > SO₄ ²⁻, a trend dictated by their hydration energies and ionic radius. The low adsorption of SO₄ ²⁻ is attributed to its highest hydration energy [49]. For Cl⁻, Br⁻, and NO₃ ⁻, which share comparable hydration energies [50], the variation in ionic radius emerges as the critical factor. The progressively smaller radius (Cl⁻ < Br⁻ < NO₃ ⁻) [51, 52] enhance charge density and thus the electrostatic attraction to HCNAP, ultimately driving the observed adsorption trend. These results confirm the universal applicability of HCANP for anion adsorption.

On the other hand, Figure 5b‐iii shows that the desalination performance of the AC//HCNAP device is significantly enhanced compared with that of the AC//HATN device, with a more pronounced decrease in its conductivity (Figure S26), thus yielding a higher SAC (39.57 vs. 17.96 mg g⁻¹). With the anode conditions kept identical, the improved desalination performance stems from the enhanced Na⁺ adsorption by HCNAP cathode, a finding also supported by ion chromatography data indicating higher Na⁺ adsorption in AC//HCNAP relative to AC//HATN device, with both devices exhibiting similar Cl⁻ adsorption capacities (Table S3). Figure 5d summarizes the aforementioned SAC values. The data collectively demonstrate that the introduction of electron‐withdrawing C≡N groups not only unlocks anion adsorption capability but also enhances cation adsorption. More importantly, as visualized in the comparative plot of Figure 5e and supported by the extensive literature survey summarized (over 100 publications) in Figure S28 and Table S4, HCNAP//HCNAP device holds the best desalination performance among all symmetric CDI devices based on organic materials. Moreover, under identical testing conditions (1.2 V, 500 mg L⁻¹ NaCl), its performance ranks among the top‐tier of all reported symmetric devices (inorganic and organic electrodes), highlighting its exceptional performance. A Ragone plot, which reflects the average desalination performance of electrode materials, features a top‐right region representing higher capacity and faster rates. As shown in Figure 5f, all data points for HCNAP//HCNAP are located in this top‐right corner among the seven symmetric/hybrid devices, indicating it consistently achieves the highest SAC and salt adsorption rate (SAR) throughout the entire process, with a peak SAR of 10.1 mg g⁻¹ min⁻¹. The achieved desalination rate represents a significant advance over previously reported values in the literature, as displayed in Figure S29. These findings not only confirm the dual‐function capability of HCNAP as a symmetric electrode for adsorbing both anions and cations but also emphasize its exceptional overall desalination performance.

It is noteworthy that a key insight from Figure 5d–f is the superior desalination performance of the symmetric HCNAP//HCNAP device compared to the asymmetric HCANP//AC and AC//HCNAP counterparts. Despite the fact that HCNAP possesses dual functionality for adsorbing both Na⁺ and Cl⁻ ions and AC acts as a multifunctional electrode partner, this finding underscores the critical importance of symmetrical electrode matching for achieving optimal performance. Beyond performance, further evaluation of the charging efficiency (Λ) and energy consumption is essential to reveal the potential of symmetric devices for low‐energy‐consumption water treatment applications. Figure 5g shows that the symmetric HCNAP//HCANP device achieves the highest charging efficiency and the lowest energy consumption, outperforming its hybrid counterparts (AC//HCNAP, HCNAP//AC), which further validates the superiority of the HCNAP//HCNAP symmetric configuration. This finding is expected to underscore the importance of symmetric design in future research on CDI.

To go further, the desalination performance of the HCNAP//HCNAP device shows a distinct voltage dependence within the applied voltage range of 1.0 to 1.4 V. As the voltage increased from 1.0 to 1.4 V, both the rate and extent of the decrease in solution conductivity are significantly enhanced (Figure S30), accompanied by a synchronous improvement in the SAC from 34.12 to 67.35 mg g⁻¹ (Figure 5h). This phenomenon is primarily attributed to the intensified external electric field, which provides a greater driving force for ion migration, accelerating ion transport to the electrode surface and, in turn, enhancing their adsorption at the active sites. Increasing the operating voltage improves performance but comes at the cost of higher energy consumption (Figure S31). In a comprehensive assessment, 1.2 V was identified as the most energy‐efficient condition.

To comprehensively evaluate the desalination performance of the symmetric HCNAP//HCNAP device in brackish water environments, the tests were conducted not only in NaCl but also extended to KCl, CaCl₂ and MgCl₂. As shown in Figure S32, the desalination performance of HCNAP//HCNAP device in KCl and NaCl is comparable and remarkable, with SAC of 55.75 and 53.88 mg g⁻¹, respectively. In CaCl₂ and MgCl₂, its SAC is relatively low, measuring 41.98 and 39.67 mg g⁻¹, respectively. Further analysis of the molar SAC shows that the HCNAP//HCNAP device exhibits highly similar performance in 1 M NaCl (0.92 mmol g⁻¹) and KCl (0.75 mmol g⁻¹), with both reaching comparable levels. However, the molar SAC in CaCl₂ (0.38 mmol g⁻¹) and MgCl₂ (0.42 mmol g⁻¹) is significantly reduced, reaching only approximately half that of the observed in the former two solutions (Figure 5i). This series of results reveals two critical information: (i) the symmetric HCNAP//HCNAP device shows significant advantages and broad application prospects in the treatment of low‐salinity brackish water. It can effectively remove common cations such as K⁺, Na⁺, Mg²⁺ and Ca²⁺ and the most common anion (Cl⁻), providing strong support for brackish water desalination; (ii) the charge carried by ions is likely a key factor determining their maximum electro‐sorption capacity. Since both K⁺ and Na⁺ carry a single unit charge, their desalination behavior exhibits certain similarities. In contrast, Ca²⁺/Mg²⁺, which carry two units of charge, demonstrate significantly lower desalination capacities compared to the former. These findings provide deeper experimental insights for the development of HCNAP//HCANP CDI device in practical brackish water desalination. Cyclic stability is also critical for the practical application of CDI. Therefore, the HCNAP//HCNAP device was tested over 30 cycles at a constant voltage of 1.2 V. As can be seen from Figure 5j, the SAC retention of the device can still maintain 84.4% after 30 cycles, showing its good desalination stability. Besides, the post‐reaction sample displays no structural changes in its XRD pattern, and no characteristic peaks of organic molecules are detected in the recovered solution (Figures S33 and S34). Together, the above results demonstrate that the HCNAP, featuring strong anion‐π interactions and rich redox‐active sites, exhibits universal and stabile dual‐functional adsorption characteristics in a CDI device.

The structure information of HCNAP during the charge‐discharge process was also studied by XRD, XPS, Raman, and FT‐IR spectra. As shown in Figure 6a, the XRD pattern of HCNAP shows that with the increase of charging voltage, the (002) peak of HCNAP shifts progressively toward a smaller angle, indicating a gradual expansion of its π–π interlayer spacing. Given the positive surface potential of HCNAP, the possibility of interlayer spacing variation induced by Na⁺ adsorption can be ruled out. Furthermore, the hydrated radius of Cl⁻ is 0.330 nm (ionic radius: 0.167 nm) [53, 54], which renders it theoretically unable to penetrate the interlayer structure. Therefore, the change in interlayer spacing during the charging process can be attributed to interlayer expansion caused by Cl⁻ adsorption at the edge of the highly electron‐deficient aromatic layer of HCNAP. In the subsequent discharging process, the π–π layer spacing returns to its original state, revealing that HCNAP exhibits excellent reversibility during Cl⁻ absorption. Besides, the Cl 2p XPS spectrum (Figure 6b) shows a prominent Cl⁻ characteristic peak at full charge, which disappears completely after full discharge. This provides compelling evidence that HCNAP enables efficient capture of Cl⁻ in brackish water through anion‐π interactions and delivers outstanding reversible performance. Furthermore, the Raman spectra (Figure 6c) reveal that the characteristic peaks of C═N and C≡N gradually decrease during the discharge process but return to the initial state upon charging, demonstrating the highly reversible Na⁺ intercalation and deintercalation process in HCNAP. This reversible behavior is further corroborated by FT‐IR, N 1s and Na 1s XPS analyses, as detailed in Figures S35–S37.

FIGURE 6.

Mechanism of ion adsorption by HCNAP: (a) XRD. (b) Cl 2p XPS spectra. (c) Raman. Calculated adsorption behavior of (d) Na⁺ and (e) Cl⁻ on HCNAP. (f) Schematic of the ion adsorption mechanism for HCNAP.

DFT calculations were employed to investigate the energy evolution during Na⁺ and Cl⁻ adsorption processes on HCNAP. As shown in Figure 6d, theoretically, HCNAP is capable of holding nine Na⁺ ions; specifically, the inner azabenzene structure contains six C═N active sites that can adsorb six Na⁺ ions, while the remaining three Na⁺ ions are captured by the introduced cyano group [55]. The coordination process between Na⁺ with active sites occurs in three energetically favorable steps: in the first stage, 3Na⁺ is preferentially chelated by two nitrogen atoms in the inner layer to form [N–Na–N]⁺ intermediates, causing an energy reduction to −6.23 × 10⁴ eV. In the subsequent stage, 3Na⁺ continues to be adsorbed by nitrogen atoms in the inner layer to form 6 [N–Na]⁺ intermediates, which lowers the energy further to −7.55 × 10⁴ eV. The final stage involves the chelation of the other 3Na⁺ with two nitrogen atoms in the outer layer –C≡N, bringing the total energy reduction to −8.89 × 10⁴ eV. The energy of the ion coordination model shows a downward trend, indicating the stable adsorption of ions at the active sites. Notably, although HCNAP possesses a layered structure, its interlayer distance is merely 0.32 nm, which prevents Cl⁻ from entering the interlayer space. Furthermore, in highly electron‐deficient aromatic systems, anions tend to accumulate preferentially on the outer periphery of the aromatic rings [56]. Accordingly, Cl⁻ adsorption on HCNAP occurs at the edges of the aromatic rings, which is also supported by theoretical calculation. As shown in Figure 6e, 3Cl⁻ ions are adsorbed on the outer side of cyano‐containing aromatic rings, and the energy after adsorption is reduced to −8.66 × 10⁴ eV, indicating that the adsorption process is thermodynamically favorable. Based on these results, the schematic diagram of the adsorption mechanism of Na⁺ and Cl⁻ by HCNAP is shown in Figure 6f. Overall, the experimental results and theoretical calculations above confirm that HCNAP effectively utilizes the C═N and C≡N active sites to remove Na⁺ from brackish water. More importantly, the incorporation of strong electron‐withdrawing cyano groups into HCNAP creates a highly electron‐deficient structure, which switches surface electron density to active anion‐π interactions, thereby achieving efficient removal of Cl⁻.

Furthermore, in addition to introducing cyano groups into HATN, ‐F and ‐COOH groups with strong electron‐withdrawing ability were also designed and incorporated into HATN, denoted as FAP and COOAP, respectively (Figures S38–S40a). DFT calculations reveal that both FAP and COOAP exhibit positive Q _zz (3.29 for FAP and 30.67 for COOAP, Figure S40b), whereas the value of the original HATA is −21.95. Their positive Q _zz facilitates anion‐π interactions and thus enhances anion adsorption. As shown in Figure S40c, the desalination capacities of the symmetric CDI devices with FAP//FAP and COOAP//COOAP configurations reach 23.35 and 43.10 mg g⁻¹, respectively, both of which were higher than that of the HATN//HATN symmetric CDI device (8.98 mg g⁻¹), demonstrating enhanced ion adsorption capability. Moreover, the hybrid CDI devices with FAP//AC and COOAP//AC (FAP and COOAP as anodes) configurations deliver desalination capacities of 13.47 and 16.16 mg g⁻¹ (Figure S40d), respectively. These values are higher than that of HATN//AC (5.39 mg g⁻¹), revealing that the incorporation of electron‐withdrawing ‐F and ‐COOH groups can effectively promote the Cl⁻ adsorption. Furthermore, the desalination capacities of the AC//FAP and AC//COOAP (FAP and COOAP as cathodes) hybrid CDI devices are 14.37 and 26.04 mg g⁻¹ (Figure S40e), respectively. Specifically, the AC//COOAP device shows a higher desalination capacity than AC//HATN (17.96 mg g⁻¹), while the AC//FAP device exhibits a lower capacity. This phenomenon can be attributed to the fact that the introduction of F groups in HATN can partially reduce the electron cloud density on the C═N functional groups, thereby weakening their adsorption ability toward Na⁺ and resulting in a lower desalination capacity than AC//HATN. In contrast, the –COOH and –CN, as additional n‐type functional groups, can serve as redox‐active sites for the extra adsorption of Na⁺, thus contributing to a higher Na⁺ adsorption capacity. In a word, the desalination performance follows the order HCNAP (–CN) > COOAP (–COOH) > FAP (–F). This result arises from the combined effects of the electron‐withdrawing ability of the functional groups, the overall molecular conjugation effect, and the redox activity of these groups. Collectively, these observations reaffirm that constructing an electron‐deficient aromatic framework within n‐type organic materials enables the dual adsorption of both anions and cations.

3. Conclusion

In summary, we demonstrate that electrode material HCNAP, synthesized through an electronic tailoring strategy involving the incorporation of strongly electron‐withdrawing cyano groups into azabenzene compounds, exhibits a remarkable desalination capability via a symmetric CDI device. Combined theoretical calculations and experimental characterizations reveal that cyano groups reverse the surface potential and Q _zz value of aromatic layers in HCNAP, shifting them from initially negative to positive states. This reversal significantly reduces its surface electron density, thereby inducing favorable anion‐π interactions that effectively capture anions from brackish water. Simultaneously, the C═N and C≡N groups in HCNAP serve as highly efficient adsorption sites for Na⁺, significantly enhancing Na⁺ capture efficiency. Consequently, the symmetrical HCNAP//HCNAP CDI device achieves a SAC of 53.88 mg g⁻¹ in 500 mg L⁻¹ NaCl solution at 1.2 V, which considerably outperforms the AC//AC device and five other comparable devices. This performance positions it among the most advanced symmetric CDI devices reported to date. This breakthrough provides a novel strategy for designing high‐performance organic materials to realize efficient desalination and opens up new avenues for advanced symmetrical CDI systems.

Author Contributions

Li Dong: data curation, methodology, writing‐original draft. Shuang Li: conceptualization, methodology, writing‐original draft and review & editing, supervision. Jiayi Liu: data curation. Ya‐Pan Wu: visualization. Meidi Wang: validation. Xue‐Qian Wu: formal analysis. Guangtong Hai: software. Meng Lu: formal analysis. Ya‐Qian Lan: conceptualization, supervision. Dong‐Sheng Li: conceptualization, supervision.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: anie72775‐sup‐0001‐SuppMat.docx

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.