Selective Removal and Recovery of Cu²⁺ From Complex Water via Asymmetric Electrochemical Separation System

ABSTRACT

Industrial wastewater pollution and resource shortage are major issues affecting environmental sustainability today. Copper, a critical metal, is also a prevalent heavy metal contaminant in aquatic environment. Selective extraction/recovery of low‐concentration Cu²⁺ from wastewater remains a hurdle. Herein, we develop an asymmetric electrochemical separation system to selectively capture Cu²⁺ from complicated water with the composite of covalent organic frameworks (COFs) and hollow mesoporous carbon spheres (HMCS) serving as cathode. By taking advantage of their synergistic effects, the HMCS@COF with suitable component proportion enables good electrical conductivity, fast ion movement, and full exposure of Cu²⁺ capture sites, achieving 97% Cu²⁺ removal and a Cu²⁺/Na⁺ selectivity of 10.54 at 1.2 V in 20 mg L⁻¹ Cu²⁺ and 100 mg L⁻¹ Na⁺ with additional involvement of minor electrodeposition route. Even in mixed solutions containing various metal ions (Na⁺, K⁺, Ca²⁺, Mg²⁺, Zn²⁺, Ni²⁺, Fe³⁺), it maintains high removal efficiency and selectivity for Cu²⁺. Moreover, satisfactory regenerability and stability are displayed over continuous operational cycles. The remarkable preferential removal of Cu²⁺ is also demonstrated by theoretical calculations. This work provides a new route for the selective removal of Cu²⁺ and offers the potential for copper recovery from wastewater containing multiple metals.

An asymmetric electrochemical separation system with HMCS@COF composite serving as the cathode was developed to tackle the challenge of selective extraction/recovery of low‐concentration Cu²⁺ from complicated wastewater containing competitive metal ions. HMCS@COF‐1 with suitable component mass proportion could achieve excellent Cu²⁺ removal efficiency and removal selectivity vs. multiple metal ions, along with satisfactory regenerability and stability over continuous operational cycles.

1. Introduction

Industrial development has led to the discharge of a large amount of copper‐containing wastewater, for example, from energy storage devices, copper‐plating factories, dye production, and so on [1, 2]. As a typical metal ion, Cu²⁺ is crucial for biochemical processes and is widely present in many plants and animals, participating in respiration or photosynthesis [3]. However, excessive Cu²⁺ intake by humans can lead to severe headaches, hair loss, hypoglycemia, rapid heartbeat, nausea, kidney and liver damage, as well as other potential psychological issues [4]. On the other hand, in terms of metal production and consumption, copper is the third largest metal in this industry, following after iron and aluminum [5]. According to the current copper reserves, copper depletion may occur within 50 years [6]. Therefore, effectively removing excessive Cu²⁺ from wastewater while extracting copper to achieve resource recovery is an important approach to address the current issues of copper‐related heavy metal pollution and resource shortage. Currently, most of techniques reported for treating Cu²⁺‐containing wastewater are chemical precipitation [7], ion exchange [8, 9], membrane separation [10, 11], and adsorption [12, 13]. Nevertheless, these techniques often suffer from disadvantages like secondary pollution, low effectiveness, high operating costs, etc [14, 15]. In particular, it is a large challenge to achieve selective removal and recovery of Cu²⁺ from actual wastewater because of the co‐existence of multiple metal ions. To tackle these challenges, great research efforts need to be devoted to developing novel and effective techniques.

Electrodeposition is a widely adopted technique for removal/recovery of Cu²⁺. For instance, Li et al. [16] proposed an electrochemically activated limestone system that can cause copper, cadmium, and zinc metal ions in acidic mine drainage (AMD) to migrate to the cathode under the action of an electric field, where they deposit or precipitate in the high pH environment formed by the hydrogen evolution reaction. However, it faces several significant limitations and drawbacks. Generally, electrodeposition is only suitable for treating wastewater with high concentrations of Cu²⁺. When dealing with low concentrations of Cu²⁺ (e.g., ≤100 mg/L), it becomes ineffective and suffers from high energy consumption [17]. Electrosorption has emerged as a promising technology for the selective removal of heavy metal ion pollutants, offering advantages such as environmental friendliness, convenient operation, high energy efficiency, and low cost [18, 19, 20, 21]. In a typical electrosorption process, two porous electrodes are arranged in parallel, with aqueous solution flowing between them. When a certain voltage is applied across the electrodes, charged ions in the solution would migrate toward the electrode of opposite charge under the electric field. These ions are ultimately retained on or in the electrodes, leading to a reduction in the total ion concentration of the solution. When the voltage is reversed or an appropriate desorption reagent is used under short‐circuit condition, the retained ions could be released back into the solution, regenerating the electrodes [22, 23, 24].

As one of the most important components of the electrosorption system, the electrodes play a crucial part in the electrosorption performance. In recent years, carbon‐based materials, such as activated carbon (AC) [25], carbon fibers [26], graphene [27], and carbon aerogels [28], have been used as electrode materials for electrosorption. However, they usually suffer from poor selectivity and weak removal capacity. Faradaic materials can capture ions through special redox reactions, thereby realizing greater selectivity and higher removal capacity. For instance, Wang et al. [29]. reported WO₃/C electrodes to selectively capture and extract U(VI) in the capacitive deionization (CDI) system, achieving a high removal capacity of 449.9 mg g⁻¹ at 1.2 V. The ZnS‐modified carbon felt electrodes prepared by Jin et al. [30]. exhibited a relatively high Cu²⁺ removal capacity (27.4 mg g⁻¹) and maintained saturation adsorption capacities of 99.1% and 85.8% in the presence of 150 mg L⁻¹ CrCl₃ and 40 mg L⁻¹ natural organic matter (humic acid), respectively. Nevertheless, despite such advances, it's worth noting that electrodes made from metal‐containing materials may experience metal leaching during operation, causing secondary pollution. Consequently, it is essential to develop alternative electrodes that are both environmentally friendly and effective for the selective removal/recovery of Cu²⁺ from wastewater.

Covalent organic frameworks (COFs) have attracted considerable attention in recent years. COFs are an emerging type of porous, crystalline polymer formed by covalent bonds between molecular components, possessing several advantageous characteristics such as long‐range order, uniform porosity, stable functional frameworks, and a high degree of freedom in structural design and performance tuning [31, 32, 33]. Additionally, COFs do not contain metal elements, which prevents metal leaching and secondary pollution during wastewater treatment processes. Furthermore, the lone‐pair electrons on functional groups (e.g., N and O atoms) within COF structures exhibit a tendency toward metal ion binding [34], making them potentially capable of selectively removing Cu²⁺ from aqueous media through coordination interactions. Nonetheless, of particular note is that COFs have drawbacks of poor electrical conductivity and strong agglomeration, restricting their application in electrosorption processes.

Inspired from the previous reports on the combination of carbon substrates with functional frameworks (e.g., MXene‐ and graphene‐based systems) for selective metal removal [35, 36], the integration of COFs with an appropriate conductive substrate is expected to be able to enhance the electrochemical separation of Cu²⁺. This improvement is supposed to stem from the combined advantages of enhanced conductivity, better dispersion of the COFs, and synergistic effects from the substrate material. Three‐dimensional (3D) hollow mesoporous carbon sphere (HMCS) represents a promising substrate candidate for hosting COF‐based cathodes aimed at Cu²⁺ capture. This potential stems from its advantageous properties: [37, 38, 39, 40, 41, 42] good electrical conductivity and mesoporous structure promote ion transport; high surface area and spherical geometry support COFs’ growth while preventing COFs’ aggregation. Moreover, when endowed with oxygen‐containing groups, HMCS itself may exhibit a degree of selectivity for Cu²⁺ removal. Consequently, HMCS is a promising substrate for COFs, and their integration into a composite cathode is expected to yield synergistic effects for the selective removal of Cu²⁺ from water. To date, as far as we know, such exploration has been not implemented and the underlying mechanism keeps unclear. Conducting this research will contribute to filling in the knowledge gaps and promoting the advancement of relevant technologies.

2. Results and Discussion

2.1. Property Characterization

Figure 1 shows the synthetic process of HMCS@COF, and Figure 2a shows the HMCS@COF composites of three different HMCS:COF component mass proportions. The morphology and structure of the as‐synthesized samples were examined by scanning and transmission electron microscopy (SEM/TEM). Pure COFs (Figure 2b) form crystalline nanorods and display strip‐like structure (around 200–400 nm long, 10–20 nm wide; aspect ratio 15–25) with smooth surfaces and specific features suggesting a tendency for aggregation. By contrast, the as‐synthesized HMCS are mono‐dispersed spheres (approximately 500 nm in diameter) and exhibit outstanding stability showing no signs of agglomeration, along with uniform mesoporosity (Figure 2c and Figure S3). Due to the presence of the spherical HMCS framework, the HMCS@COF composites retain spherical morphology (Figure 2d–g). With increasing COF coating quantity, the sphere diameter grows proportionally from approximately 550 nm (HMCS@COF‐0.5) to 600 nm (HMCS@COF‐1) and 650 nm (HMCS@COF‐2), beneath which the underlying HMCS sphere remains faintly discernible. TEM imaging provides further detailed structural insight. The hollow spherical HMCS features a carbon shell of approximately 50 nm thickness permeated by mesopores (Figure 2h). For the HMCS@COF‐1 composite, the shell thickness increases to approximately 90 nm (Figure 2i), and comparison with pure HMCS demonstrates the presence of a COF coating grown uniformly and vertically on the carbon sphere surface. Elemental mapping reveals the positional distribution of C, N, and O elements for the HMCS@COF‐1 composite, with intensified signals around the periphery also confirming the successful surface COFs’ coating (Figure 2j). By employing an optimized HMCS:COF mass proportion such as 1:1, the synergistic configuration achieves two key objectives: it retains the excellent conductivity and mass‐transfer capabilities of the HMCS framework, while simultaneously improving COFs’ dispersion to reduce agglomeration. This leads to greater exposure of active sites and a boost in coordination‐based Cu²⁺ capture.

FIGURE 1.

Schematic of the synthetic process of HMCS@COF.

FIGURE 2.

(a) Schematic of the HMCS@COF composites with different HMCS:COF mass proportions. SEM graphs of (b) COF, (c) HMCS, (d,f) HMCS@COF‐1, (e) HMCS@COF‐0.5 and (g) HMCS@COF‐2. TEM graphs of (h) HMCS, (i) HMCS@COF‐1, and (j) elemental mapping of HMCS@COF‐1.

The specific surface areas and pore size distributions of the COF, HMCS, and HMCS@COF composite were determined from nitrogen adsorption‐desorption isotherm with the results presented in Figure S3a–e and Table S1. The HMCS@COF‐1 composite possesses a significantly larger BET surface area (245.28 m² g⁻¹) compared to the pure COF (72.44 m² g⁻¹). This enhancement is mainly attributed to the introduction of HMCS component, whose well‐designed mesoporous walls contribute to a high surface area. The pore size distribution results indicate that HMCS@COF‐1 composite retains the pore characteristics of HMCS to a large degree, and a majority of the pores of HMCS@COF‐1 composite are mesoporous, mainly ranging within 2–8 nm.

The FTIR spectra of HMCS, COF, and the HMCS@COF composite also provide comparative structural insights (Figure S3f). While all samples show expected bands for adsorbed water (─OH stretch ∼3400 cm⁻¹) and framework C═C bonds (∼1610 cm⁻¹) [43, 44, 45, 46], the composite spectrum clearly inherits specific signatures from the COF coating. These include the stretching vibrations of the aromatic backbone (C═C/N─H) at ∼1523 cm⁻¹ and a C─O stretch at ∼1010 cm⁻¹, which are distinctive of COF and absent in HMCS. Moreover, for vibrational modes present in both components, such as the tertiary alcohol ─OH stretch (∼1382 cm⁻¹) and a different C‐O stretch (∼1235 cm⁻¹) [47, 48, 49], the peak intensities in HMCS@COF closely match those of pure COF and are significantly enhanced relative to the weak features in HMCS. This effective inheritance of COF‐derived spectral features directly corroborates its successful coating on the HMCS substrate.

XPS survey spectra (Figure S4) verify the presence of C, N, and O in HMCS@COF, which is in agreement with the mapping results (Figure 2j). Quantitatively, the elemental composition shows a median N/O level for HMCS@COF relative to its monomers (Table S2), suggesting the successful hybridization between COF and HMCS. Further evidence comes from detailed peak fitting. The C 1s spectrum of HMCS@COF‐1 shows 5 peaks assigned to C─C (284.8 eV), C─O (286.5 eV), C═O (288.2 eV), O─C═O (289.9 eV), and π–π ^* (291.8 eV) (Figure S4b) [50, 51, 52, 53]. Notably, a marked increase in the C─O signal relative to HMCS was observed (Figure S5a), suggesting the possible formation of C─O─C covalent linkages between the two components. The O 1s spectrum of HMCS@COF‐1 is fitted to C═O (530.7 eV), C─O─C (532.4 eV), C─O─NO^* ₂ (534.2 eV), and adsorbed H₂O (536.5 eV) [54, 55, 56]. The N 1s curve of HMCS@COF‐1, displaying signatures of pyridinic N (398.6 eV), N‐C₃ (399.8 eV), NH_x (401.1 eV), ─N(CH₃)₃ (402.8 eV), and ─NO₂ (405.0 eV), closely resembles that of pure COF (Figure S6) and differs from HMCS, confirming that the nitrogen species of HMCS@COF‐1 primarily originate from the COF component.

2.2. Selective Removal and Recovery of Cu²⁺ by HMCS@COF Electrode

To assess the advantages of the HMCS@COF composite over common AC and its individual components, and to identify the optimum component mass proportion for Cu²⁺ selective removal, a series of comparative tests were performed. A double‐electrode set‐up was adopted via pairing an activated carbon (AC) anode with various cathodes (AC, HMCS, COF, or HMCS@COF composites with component mass proportions of 0.5:1, 1:1, and 2:1) operated at 1.2 V in a solution of 20 mg/L Cu²⁺ and 100 mg/L Na⁺. First, the Visual Minteq software was used to simulate and reveal the changes in copper species and their corresponding concentrations vs. pH values. As shown in Figure S7, at pH smaller than 5, the main copper species is soluble Cu²⁺. When the pH exceeds 6, the concentration of Cu²⁺ drops significantly and the precipitation of Cu ions may occur. In this study, the changes in pH during the electrosorption process were measured for the various cathodes. It was found from Figure S8 that the pH kept smaller than 6, ruling out the likelihood of Cu²⁺ precipitation.

The Cu²⁺/Na⁺ selectivity (namely Cu²⁺ removal efficiency/Na⁺ removal efficiency) results (Figure 3a,b) clearly establish the superiority of the HMCS@COF‐1 composite cathode since it achieves the largest selectivity for Cu²⁺ removal relative to Na⁺ among all the tested cathodes. As anticipated, the AC cathode shows the lowest selectivity (stabilizing around 2). The HMCS cathode performs slightly better, reaching a Cu²⁺/Na⁺ selectivity of 4.81 at 240 min, likely due to the existence of its oxygen‐containing functional groups [57]. The COF cathode attains a higher selectivity for the removal of Cu²⁺ relative to Na⁺, reaching 6.03 at 240 min. However, its Cu²⁺ removal efficiency at 240 min is only 52.5%, and because of the inherent drawbacks of pure COF such as massive agglomeration and poor wettability/electrical conductivity (Figures S9 and S10), its performance remains far from satisfactory. The HMCS@COF‐1 composite cathode overcame these limitations, delivering a peak Cu²⁺/Na⁺ selectivity of 11.11 at 180 min and sustaining 10.54 at 240 min, which corresponds to 97% Cu²⁺ removal efficiency and 9.2% Na⁺ removal efficiency. In addition, the HMCS@COF‐1 composite cathode can achieve a removal capacity of 30.31 mg g⁻¹ for Cu²⁺ in a mixed solution of 20 mg L⁻¹ Cu²⁺ and 100 mg L⁻¹ Na⁺ (Figure 3c), and the concentration of Cu²⁺ can be reduced to 0.6 mg L⁻¹ (Figure S11), which was far below the health‐based upper limit for copper (i.e., 2 mg L⁻¹) in the World Health Organization drinking water quality guidelines. This is probably because the hybrid design at a 1:1 ratio optimally combines components: insufficient COF in HMCS@COF‐0.5 yielded low selectivity, while excessive COF in HMCS@COF‐2 led to suboptimal performance compared to the optimized ratio.

FIGURE 3.

Comparison of various cathodes regarding (a) Cu²⁺ removal efficiency and Na⁺ removal efficiency, (b) Cu²⁺/Na⁺ selectivity, and (c) removal capacity in the treatment of 20 mg L⁻¹ Cu²⁺ and 100 mg L⁻¹ Na⁺. Comparison of (d) ion removal efficiency and (e) removal selectivity vs. Cu²⁺ in the treatment of mixed solution of 20 mg L⁻¹ Cu²⁺, 100 mg L⁻¹ Na⁺, 100 mg L⁻¹ Mg²⁺, 100 mg L⁻¹ K⁺ and 100 mg L⁻¹ Ca²⁺ for the HMCS@COF‐1 cathode. Comparison of (f) ion removal efficiency and (g) removal selectivity vs. Cu²⁺ in the treatment of mixed solution of 20 mg L⁻¹ Cu²⁺, 20 mg L⁻¹ Ni²⁺, 20 mg L⁻¹ Zn²⁺ and 20 mg L⁻¹ Fe³⁺ for the HMCS@COF‐1 cathode. (h) Fluctuations in Cu²⁺ removal efficiency and Cu²⁺/Na⁺ selectivity of the HMCS@COF‐1 cathode during fifty consecutive cycling tests.

In fact, the composition of Cu²⁺‐containing wastewater is complex [58]. To verify the selective removal ability of HMCS@COF‐1 cathode for Cu²⁺ in complex aqueous environments, we prepared two different multi‐component solutions for the experiment. One was a coexisting salt ion system (20 mg L⁻¹ Cu²⁺, 100 mg L⁻¹ Na⁺, 100 mg L⁻¹ Mg²⁺, 100 mg L⁻¹ K⁺, and 100 mg L⁻¹ Ca²⁺), and the other was a coexisting heavy metal ion system (20 mg L⁻¹ Cu²⁺, 20 mg L⁻¹ Ni²⁺, 20 mg L⁻¹ Zn²⁺, 20 mg L⁻¹ Fe³⁺). K⁺, Ca²⁺, Mg²⁺, Ni²⁺, Zn²⁺ and Fe³⁺ came from the addition of KNO₃, Ca(NO₃)₂, Mg(NO₃)₂, Ni(NO₃)₂, Zn(NO₃)₂, Fe(NO₃)₃, respectively. The relevant experimental results are shown in Figure 3d–g. In the coexisting salt ion system, the HMCS@COF‐1 cathode could reduce the Cu²⁺ concentration to 0.75 mg L⁻¹ (Figure S13), and its removal efficiency and removal capacity for Cu²⁺ (Figure S14) reached 96.25% and 30.08 mg g⁻¹, respectively, both of which were much higher than those for Na⁺ (7.9% and 12.34 mg g⁻¹), Mg²⁺ (8.75% and 13.67 mg g⁻¹), K⁺ (9.5% and 14.84 mg g⁻¹) and Ca²⁺ (10.8% and 16.88 mg g⁻¹). Moreover, the HMCS@COF‐1 cathode exhibited particularly outstanding selectivity for Cu²⁺ relative to the above four salt ions. The Cu²⁺/Na⁺ selectivity gradually increased with electrosorption time, reaching a maximum of 12.30 at 180 min and remaining at 12.18 at the end of the 240 min electrosorption process. The Cu²⁺/Mg²⁺ selectivity also generally increased with electrosorption time, reaching 11.00 at 240 min. The Cu²⁺/K⁺ selectivity fluctuated over time, peaking at 10.80 at 120 min and remaining at 10.13 at the conclusion of electrosorption. The Cu²⁺/Ca²⁺ selectivity reached its maximum of 9.04 at 180 min and finally stabilized at 8.91. In the coexisting heavy metal ion system, the concentration of Cu²⁺ could be reduced to 0.9 mg L⁻¹ by the HMCS@COF‐1 cathode (Figure S15), with a Cu²⁺ removal efficiency of 91%, while the removal efficiencies of Ni²⁺, Zn²⁺ and Fe³⁺ were all below 24%. Similarly, in terms of removal capacity, the removal capacity for Cu²⁺ was 29.84 mg g⁻¹ significantly higher than that for Ni²⁺ (6.72 mg g⁻¹), Zn²⁺ (6.80 mg g⁻¹), and Fe³⁺ (7.34 mg g⁻¹) (Figure S16). It was observed that the removal selectivity Cu²⁺ vs. Ni²⁺ showed a gradual decline over time, attaining 4.44 by the end of the electrosorption. In contrast, the removal selectivity of Cu²⁺ relative to Zn²⁺ gradually increased most of the time, reaching a maximum of 4.71 at 180 min and ending at 4.39. The removal selectivity of Cu²⁺ relative to Fe³⁺ fluctuated around 4.20 and remained relatively stable during the electrosorption process.

The potential recovery of Cu²⁺ during the desorption stage as well as the regeneration capability and the cyclic performance of the electrodes were also investigated. By analyzing the Cu²⁺ desorption concentrations at different steps of the desorption process (reverse voltage and acid immersion) for various cathodes (Figure S17), it can be inferred that, for the electrosorption induced Cu²⁺ removal (electrodopistion additionally played a minor role in Cu²⁺ removal and we will discuss later in the mechanism section), the COF‐based electrode removed Cu²⁺ primarily through coordination interaction, whereas the AC electrode and HMCS electrode removed Cu²⁺ primarily through electrostatic interaction. Specifically, when using the HMCS@COF‐1 cathode to treat a feed solution of 20 mg L⁻¹ Cu²⁺ and 100 mg L⁻¹ Na⁺ (the corresponding copper purity is 16.7%), the desorption concentration of Cu²⁺ and Na⁺ during the reversing voltage step was 5.3 and 7.5 mg L⁻¹, respectively. During the acid immersion step, the desorption concentration of Cu²⁺ and Na⁺ was 13.9 and 1.7 mg L⁻¹, respectively, corresponding to a high copper purity of about 90%, which could be potentially recovered and reused for other beneficial purposes. The energy consumption per gram of Cu recovered was calculated to be 2.23 × 10⁻³ kWh g⁻¹. During fifty successive cycling tests, it can be known from Figure S18 that all the captured Cu²⁺ during electrosorption could be almost desorbed after the reversing voltage and acid immersion steps. Moreover, as shown in Figure 3h, a high Cu²⁺ removal efficiency of at least 95% was maintained for a majority of the fifty cycles, with an average Cu²⁺/Na⁺ selectivity of 9.92. The good regeneration and recyclability of HMCS@COF‐1 cathode paired with AC anode over fifty successive cycles suggest its great potential in treatment of Cu²⁺‐containing wastewater, even in the presence of high levels of competing metal ions such as Na⁺.

The above tests have indicated the efficacy and advantages of HMCS@COF‐1 cathode toward selective removal and recovery of Cu²⁺ from complex water. For the sake of understanding the impact of important factors such as applied voltage, initial solution pH, and initial Cu²⁺ concentration, we carried out the following experiments (Figure 4 and Figures S19–S31). As shown in Figure 4a,b, as the applied voltage increased from 0.6 to 1.5 V, the Cu²⁺ removal efficiency after 240‐min treatment gradually increased. It was noteworthy that the Cu²⁺/Na⁺ selectivity at 1.5 V was much lower than that at other voltages, mainly due to the significant increase in Na⁺ removal efficiency. Overall, the Cu²⁺/Na⁺ selectivity was greater than 8.90 at voltages between 0.6 and 1.2 V, reaching a maximum of 10.54 at 1.2 V. Figure 4c showed that when the initial solution pH was 2, the Cu²⁺ removal efficiency after 240‐min treatment was significantly lower than that at initial pH values of 3, 4, and 5. The reason may be that under the condition of low pH, the solution contains more H⁺, and due to the competitive effect, a large number of H⁺ would occupy the site where Cu²⁺ binds to the HMCS@COF‐1 electrode, leading to the reduction of Cu²⁺ electrosorption [58, 59, 60]. Additionally, after 240‐min treatment, the Cu²⁺/Na⁺ selectivity was found to remain around 9.30 for initial solution pH values ranging from 2 to 5, reaching a maximum of 10.57 at an initial pH of 4 (Figure 4d). Regarding the initial Cu²⁺ concentration, the results revealed that although the initial Cu²⁺ concentrations varied, the Cu²⁺ concentration (Figure S28) and Cu²⁺ removal efficiency (Figure 4e) after 240‐min treatment were not significantly affected. The final Cu²⁺ concentration was less than 1 mg L⁻¹, and the Cu²⁺ removal efficiency was higher than 95%. Moreover, as the initial Cu²⁺ concentration increased, the Cu²⁺/Na⁺ selectivity (Figure 4f,g) and Cu²⁺ removal capacity (Figure 4h) gradually increased. When the initial Cu²⁺ concentration was 100 mg L⁻¹, although the Cu²⁺ removal efficiency decreased (Figure S31), the Cu²⁺ removal capacity reached as high as 148.58 mg g⁻¹, and the Cu²⁺/Na⁺ selectivity could reach a maximum of 19.12, remaining 16.12 after 240‐min treatment. Furthermore, the Cu²⁺ removal capacity of the HMCS@COF‐1 cathode in the pure Cu²⁺ solution was also tested. As shown in Figure 4i, under initial Cu²⁺ concentrations of 20, 50 and 100 mg L⁻¹, the HMCS@COF‐1 electrode exhibited Cu²⁺ removal capacity of 30.47, 77.11 and 150.16 mg g⁻¹, respectively, following 240‐min treatment at 1.2 V.

FIGURE 4.

(a–f) Changes of Cu²⁺ removal efficiency, Na⁺ removal efficiency and Cu²⁺/Na⁺ selectivity over time under different voltages, initial pH conditions and initial Cu²⁺ concentrations when the HMCS@COF‐1 cathode was used to treat a feed solution containing Cu²⁺ and 100 mg L⁻¹ Na⁺. (g) Changes of Cu²⁺ and Na⁺ concentrations and the Cu²⁺/Na⁺ selectivity over time when the HMCS@COF‐1 cathode was used to treat a feed solution containing 100 mg L⁻¹ Cu²⁺ and 100 mg L⁻¹ Na⁺. (h) Changes of Cu²⁺ and Na⁺ removal capacity at different initial Cu²⁺ concentrations when the HMCS@COF‐1 cathode was used to treat a feed solution containing Cu²⁺ and 100 mg L⁻¹ Na⁺. (i) Changes of Cu²⁺ removal capacity over time when the HMCS@COF‐1 cathode was used to treat pure Cu²⁺ solutions with different initial Cu²⁺ concentrations.

2.3. Mechanism Exploration

The total XPS spectrum of HMCS@COF‐1 cathode at the end of charging (Figure 5a) shows peaks around 930 and 950 eV, confirming the presence of Cu, which indicated that copper ions were successfully retained onto the HMCS@COF‐1 cathode. This is also in agreement with the EDS elemental mapping results presented in Figure S32. Apart from electrosorption, Cu²⁺ ions may be also removed via electroreduction generating metal copper (Cu) and cuprous oxide (Cu₂O) at the cathode through Reactions 1 and 2. Therefore, we further probed into the mechanisms for Cu²⁺ removal. From the XRD characterization of HMCS@COF‐1 cathode after charging (Figure S33), it can be known that Cu⁰ and Cu₂O were indeed produced, substantiating the electrodeposition of Cu²⁺. To determine the contribution ratio of electrodeposition and electrosorption, we further analyzed the Cu 2p XPS spectrum. As shown in Figure 5e, peaks at 934.8 and 954.8 eV correspond to Cu²⁺, while peaks at 932.7 and 952.7 eV were also observed, corresponding to Cu⁺/Cu⁰. More specifically, the reduced forms of Cu (i.e., Cu⁰ and Cu₂O) account for about 16.5%, while Cu²⁺ account for about 83.5%. These results suggest that the removal of Cu²⁺ is related to the synergistic effect of electrosorption and electrodeposition, with electrosorption playing a major role while electrodeposition playing a minor role. The cyclic voltammetry (CV) curves for the various cathodes (AC, HMCS, COF, HMCS@COF composites with different component mass proportions) plotted in Figure 5g also substantiate this. All the cathodes including AC exhibit redox peaks, indicating the occurrence of Cu²⁺ electrodeposition, while the redox peaks of COF and HMCS@COF composites are apparently more prominent than those of AC and HMCS, which indicates the additional occurrence of COF‐mediated Faradaic Cu²⁺ coordination capture.

$$\mathrm{C}{\mathrm{u}}^{2+}+2{\mathrm{e}}^{-}\to \mathrm{Cu}$$ (1)

$$2\mathrm{C}{\mathrm{u}}^{2+}+{\mathrm{H}}_2\mathrm{O}+2{\mathrm{e}}^{-}\to \mathrm{C}{\mathrm{u}}_2\mathrm{O}+2{\mathrm{H}}^{+}$$ (2)

FIGURE 5.

(a) Total XPS survey and high‐resolution XPS spectra for HMCS@COF‐1 cathode: (b) C 1s, (c) N 1s and (d) O 1s, (e) Cu 2p after charging. (f) Schematic diagram of the electrochemical operation of the AC/HMCS@COF‐1 system. (g) CV curves for the various cathodes at a sweep rate of 50 mV s⁻¹ in 100 mg L⁻¹ Cu(NO₃)₂ and 1 mol L⁻¹ NaCl solution.

As for Cu²⁺ removal induced by the major contributor electrosorption, HMCS and COF are anticipated to work synergistically in two principles, i.e., electrostatic interactions and coordination interactions: HMCS offers good electrical conductivity and mesoporous structure enabling quick ion movement, and HMCS's high surface area and spherical geometry support COFs’ growth while preventing COFs’ aggregation. COF possesses conjugated aromatic skeleton as well as a number of N and O atoms which could act as Lewis soft base sites and prefer to bind with soft acids like Cu²⁺ through orbital hybridization rather than binding with hard acids like Na⁺, thereby resulting in the selective removal of Cu²⁺ from aqueous solution. Figure S4 and Figure 5b–e displayed the detailed XPS characterization of C 1s, N 1s and O 1s for HMCS@COF‐1 cathode prior to and after charging. Regarding the O 1s spectrum, the peaks of C═O and C─O─C shifted from 530.7 and 532.4 to 530.5 and 532.2 eV, respectively. Similarly, regarding the N 1s spectrum, the peaks of C═N─C and N─C₃ shifted from 398.6 and 399.8 to 398.4 and 399.6 eV, respectively. Such phenomena could be ascribed to Cu²⁺‐induced electronic perturbations of nitrogen‐ and oxygen‐containing functional groups [61, 62]. The formation of O─Cu (529.8 eV) and N─Cu (400.6 eV) bonds also suggests that carbonyl oxygen and amino nitrogen coordinated with Cu²⁺ to form stable complexes [63, 64, 65, 66]. To enable a better understanding, the coordination interaction is shown below:

$$-\mathrm{C}=\mathrm{O}+-\mathrm{N}{\mathrm{H}}_2+\mathrm{C}{\mathrm{u}}^{2+}\to -\mathrm{C}=\mathrm{O}\cdots \mathrm{C}{\mathrm{u}}^{2+}\cdots \mathrm{N}{\mathrm{H}}_2-$$ (3)

After the electrosorption process is completed, the captured Cu can be effectively released into the solution via voltage reversal and acid immersion, as already demonstrated by the desorption data in Figure 3h, Figures S17, S18 and S29. On the one hand, the generated Cu⁰ and Cu₂O will lose electrons and transform into Cu²⁺ under the action of voltage reversal and acid immersion, as shown in Reactions 4 and 5. On the other hand, via acid immersion, an acidic environment is created, making the adsorption of competing protons (H⁺) thermodynamically favorable and coordinated Cu²⁺ displaced from specific binding sites via the ion exchange mechanism as shown in Reaction 6. Simultaneously, the protonation of amine nitrogen occurred, saturating the lone pairs of the nitrogen atoms and thereby breaking the Cu–N bond. As a result, the HMCS@COF‐1 cathode was regenerated and regained the ability to capture Cu²⁺ in the following cycles. The entire electrosorption and desorption process is schematically shown in Figure 5f.

$$\mathrm{Cu}-2{\mathrm{e}}^{-}\to \mathrm{C}{\mathrm{u}}^{2+}$$ (4)

$$\mathrm{C}{\mathrm{u}}_2\mathrm{O}+2{\mathrm{H}}^{+}-2{\mathrm{e}}^{-}\to 2\mathrm{C}{\mathrm{u}}^{2+}+{\mathrm{H}}_2\mathrm{O}$$ (5)

$$-\mathrm{C}=\mathrm{O}\cdots \mathrm{C}{\mathrm{u}}^{2+}\cdots \mathrm{N}{\mathrm{H}}_2-+2{\mathrm{H}}^{+}\to -\mathrm{C}-\mathrm{O}{\mathrm{H}}^{+}+\mathrm{C}{\mathrm{u}}^{2+}+-{\mathrm{NH}}_3^{+}$$ (6)

To elucidate the aforementioned preferential adsorption removal of Cu²⁺ by the MCHS@COF‐1 cathode in multi‐ion systems (e.g., Na⁺, Mg²⁺) as displayed in Figure 3d,e, DFT calculations were further performed. The resulting adsorption configurations and adsorption energies (ΔE) are depicted in Figure 6. In comparison with Na⁺ and Mg²⁺, the larger absolute ΔE (1.71 eV) for Cu²⁺ confirms the MCHS@COF‐1 cathode's higher affinity for this cation, aligning with experimental observations. The model further shows that cations primarily occupy sites in vacancies between N and O atoms. Originating from their unique electronic structures, the cations exhibit differing adsorption: Cu²⁺, with its vacant d‐orbital, engages in stronger coordination with N/O atoms, forming robust Cu‐N/O bonds as verified by modeling optimization and XPS analysis. To sum up, DFT modeling corroborates the Cu²⁺ selectivity mechanism.

FIGURE 6.

Adsorption energy (ΔE) and adsorption models for HMCS@COF‐1 cathode interacting with various metal ions (top view).

3. Conclusion

Aiming at tackling the challenge of selective extraction/recovery of low concentrations of Cu²⁺ from complicated wastewater containing competitive metal ions, in this work, we developed an asymmetric electrochemical separation system with the HMCS@COF composite serving as the cathode. It was found that both electrosorption and electrodeposition were involved in the removal of Cu²⁺, with electrosorption playing a major role while electrodeposition playing a minor role. As for the major contributor electrosorption route, HMCS and COF worked synergistically in two principles, i.e., electrostatic interactions and coordination interactions: HMCS offers good electrical conductivity and mesoporous structure enabling quick ion movement, and HMCS's high surface area and spherical geometry support COFs’ growth while preventing COFs’ aggregation. COF possesses conjugated aromatic skeleton as well as a number of N and O atoms which could act as Lewis soft base sites and prefer to bind with soft acids like Cu²⁺, thereby resulting in the selective removal of Cu²⁺ from aqueous solution. Results revealed that HMCS@COF‐1 with suitable component mass proportion could achieve excellent Cu²⁺ removal efficiency and removal selectivity vs. multiple metal ions such as Na⁺, K⁺, Ca²⁺, Mg²⁺, Zn²⁺, Ni²⁺ and Fe³⁺, along with satisfactory regenerability and stability over at least 50 cycles. This study demonstrates the potential of HMCS@COF‐1 for efficient and selective copper capture and recovery, providing a sustainable strategy for environmental remediation and resource recycling.

4. Experimental Section

4.1. Preparation of HMCS

Hollow mesoporous carbon spheres (HMCS) were synthesized by first stirring a mixture of 10 mL of deionized water, 70 mL of ethanol and 3 mL ammonia solution (25 wt.%) for 15 min. After adding 3.46 mL of tetrapropoxysilane and stirring for another 15 min, 0.4 g of resorcinol and 0.56 mL of formaldehyde solution (37 wt.%) were introduced, and the reaction was allowed to proceed under stirring for 24 h. The product was collected by repeated centrifugal washing (5 times, 6000 rpm) using ethanol and deionized water. After vacuum drying at 60°C for 12 h and subsequent calcination in N₂ environment at 700°C for 5 h with a heating rate of 5°C min⁻¹, the resulting powder was etched with a 20 wt.% NaOH solution at 80°C for 2 days. The final material was acquired after washing with deionized water until the pH of the washed solution reaches neutral.

4.2. Construction of HMCS@COF

A mixture of 0.75 mmol melamine (Tt), 0.75 mmol 2,4,6‐triformyltriphenyltriol (Tp) and 100 mg HMCS in 50 mL dimethyl sulfoxide (DMSO) undergoes ultrasonic dispersion for 30 min, which was then subjected to a solvothermal reaction in a sealed vessel at 120°C for 12 h. The resulting solid is isolated by filtration, washed with acetone, and vacuum‐dried at 60°C to obtain the composite, with the product designated as HMCS@COF‐1. The detailed washing step using acetone was carried out as follows: centrifuge was employed at 7000 rpm/min running for 5 min each time, for a total of six washes. Acetone was added as the initial solvent for washing, and the solution was orange‐red in color. The washing was completed by centrifugation until the aqueous solution became clear. HMCS@COF composites of other HMCS:COF component mass proportions were also synthesized for performance optimization. When the mass of HMCS added was 200 and 50 mg, the yielded products are named to be HMCS@COF‐0.5 and HMCS@COF‐2, respectively. For comparison, pure COF was also synthesized under identical conditions but without HMCS addition, as shown in Figure S1.

4.3. Experimental Procedures

The selective electrochemical removal of Cu²⁺ by the prepared materials was assessed using a custom asymmetric hybrid capacitive deionization (HCDI) cell (see Figure S2). In this setup, the anode was made from activated carbon (AC), while the cathode was made from either HMCS, COF, or the HMCS@COF composite. A 1 mm‐thick silicone gasket, cut to match the electrode dimensions, created a spacer where 300‐mesh nylon mesh was placed enabling fluid flow and avoiding direct contact of the anode and cathode. The entire assembly was secured between acrylic plates using screws around the perimeter. The flow of the aqueous solution was controlled using the peristaltic pump and the tests are operated under batch mode.

The preparation method for HMCS@COF or activated carbon (AC) electrode was as follows: First, the active material (HMCS@COF or AC), carbon black, and polyvinylidene fluoride were blended at an 8:1:1 mass ratio and thoroughly ground. Then a small quantity of N‐methyl‐2‐pyrrolidone was introduced to this mixture, followed by 1 h of ultrasonication to produce a homogeneous slurry (the ultrasonic power was 300 W). After uniformly coating the slurry onto the graphite sheet to get a coverage of 4 cm × 4 cm size, the coated electrode was vacuum‐dried at 65°C to fully evaporate the N‐methyl‐2‐pyrrolidone solvent, resulting in a final product containing an active material loading of 16 mg.

Prior to the testing, the HCDI device was rinsed with deionized water and emptied. Subsequently, 50 mL of a mixed Cu(NO₃)₂/NaNO₃ solution was fed into the HCDI device and circularly flows at 16 mL/min. Upon filling the device (a matter of seconds), a potential difference was applied immediately. A series of tests were conducted to evaluate performance under varying conditions: applied potential difference (0.6–1.5 V), Cu²⁺ concentration (10–50 mg L⁻¹), initial solution pH (2–5, adjusted with 0.1 m NaOH/HNO₃), and in the presence of multiple competing cations. Desorption phase commences after 240 min of electrosorption. The procedure involved reversing the potential difference until a conductivity decrease was noticed, immediately followed by sample collection. Short‐circuit desorption is then carried out for 1 h by circulating 50 mL of 50 mm HNO₃ as the desorption reagent, driven by a peristaltic pump. A final sample was collected at the end of the process, and the metal ion concentrations are determined by inductively coupled plasma optical emission spectrometry. After the first cycle of desorption was completed, ultrapure water was circulated for washing until the conductivity of the effluent reached below 5 µS/cm prior to the start of the second cycle test. Long‐term performance stability was assessed via fifty repeated adsorption‐desorption cycles.

4.4. Theoretical Investigation

Apart from the experimental investigation, the theoretical investigation, namely density functional theory (DFT) calculation, was also conducted to validate the observed phenomena. Vienna Ab initio simulation package (VASP) software was employed. The detailed information about the DFT calculations is presented in the Supplementary Information.

Funding

National Natural Science Foundation of China (22276048), the Natural Science Foundation of Chongqing (CSTB2025NSCQ‐GPX0438), the Science and Technology Innovation Program of Hunan Province (2025RC1030), the Natural Science Foundation of Hunan Province (2024JJ6146, 2023JJ40234, 2023JJ40230), and the Scientific Research Fund of Hunan Provincial Education Department (23B0601, 24B0568).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: smll73989‐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.