A Review on Recent Progress in Metal–Organic Frameworks for Capacitive Deionization

Abstract

The escalating challenge of supplying freshwater stems from an alarming rise in water consumption driven by population growth and increased economic activity. With worsening freshwater scarcity, water pollution, and 80% of wastewater worldwide containing high TDS, heavy metals, and inorganic and organic contaminants, the need to obtain freshwater from nonfreshwater sources is essential. Capacitive deionization (CDI) is an emerging, promising electrochemical technique that offers robustness, cost‐effectiveness, and high efficiency. CDI is a cost‐ and energy‐efficient alternative for applications such as desalination and the elimination of heavy or radioactive metals. This review discusses the general concepts of CDI, including different cell designs, adsorption and desorption cycles, and evaluation parameters, to provide a foundation. Later, we describe the potential of metal–organic frameworks (MOFs) as electrode materials for CDI, the recently reported modifications and developments in MOF‐based electrode materials, their CDI performance, and the underlying mechanisms for desalination and ion capture. This review concludes with future perspectives for further developments in the field of capacitive deionization.

Capacitive deionization (CDI) is an energy‐ and cost‐efficient alternative for obtaining fresh water from nonfresh water sources and for eliminating heavy or radioactive metals. This review highlights the fundamentals of CDI and focuses on metal–organic frameworks as an emerging, promising electrode material for desalination and ion removal, concluding with a future perspective.

1. Introduction

Water scarcity is a global challenge, and the need to develop efficient, environmentally friendly water treatment techniques for fresh water has become urgent. As global scarcity worsens, clean and fresh water is essential for industrial operations, daily life, and maintaining ecological balance and economic prosperity [1]. The agricultural sector owns the largest share of freshwater consumption of about 50%–60%, followed by industries, which consume 5%–10%, and 10%–20% for other urban activities [2, 3, 4]. Many of these operations produce polluted water, contaminating available freshwater. Water consumption is also rising due to population growth and increased economic activity, placing stress on freshwater resources. Therefore, the world faces a challenge in supplying fresh water due to contamination caused by rapid industrialization, population growth, and climate change [5]. Approximately 80% of the world's wastewater, which contains high levels of TDS, heavy metal ions, and organic and inorganic contaminants, is discharged into the environment without proper water purification treatment. 97% of the earth's water resources are saline water, and only 3% consists of freshwater [6, 7, 8, 9]. Therefore, it's necessary to obtain freshwater from nonfreshwater sources that can be utilized [10]. Desalination of seawater and brackish groundwater can provide a steady water supply without harming the natural freshwater ecosystem [11, 12, 13].

Heavy metals are naturally occurring elements, but due to human activities, their concentrations in the environment increase to levels that pose a risk to humans upon exposure through breathing, ingestion, and skin contact [14, 15]. Based on the origin of heavy metals, they can be in the form of organic compounds as complexes of heavy metal or in the form of inorganic free ions [16]. Heavy metals tend to accumulate in plants, animals, and humans. Long‐term exposure to these toxic metals can cause human health issues. They tend to remain in the environment and bioaccumulate over time, as the natural degradation of heavy metals is slow and difficult. Therefore, adopting cleaner technologies and increasing awareness of the harmful effects of heavy metals are necessary, with profound management implications [17, 18]. Key heavy metals found in the environment include iron, lead, nickel, cadmium, chromium, strontium, barium, selenium, mercury, manganese, and arsenic [19, 20]. Metals like Fe, Zn, Co, Mn, and Cu are necessary to carry out crucial biochemical processes occurring in the human body, but accumulation of these metals at toxic levels due to prolonged exposure can disrupt human health [21].

Industries such as the steel and metal processing industry, electronic circuit production, and fine chemical synthesis industry generate industrial wastewater containing harmful and toxic heavy metals that are released into the environment without proper treatment [22]. Chemical precipitation, chemical reduction, ion exchange, ultrafiltration, advanced oxidation processes (AOPs), coagulation‐flocculation, adsorption, and membrane separation are conventional techniques used to eliminate heavy metals from wastewater. These techniques have drawbacks, including a lack of selectivity in eliminating heavy metals, which are present at trace levels compared to typical salts; the production of secondary pollutants, enormous amounts of toxic solid sludge; and high‐energy consumption [23, 24]. Due to high chemical and energy consumption, advanced methods such as AOPs incur high operational costs [25].

Capacitive deionization is an emerging and promising electrochemical technique, considered a robust, economical, and cost‐effective alternative for applications such as desalination, heavy‐metal removal, and radioactive‐metal removal [26]. It has attracted attention over other traditional desalination and heavy metal removal techniques for its advantages of easy operation, low cost, and no secondary pollution, as electrodes in CDI are recyclable by desorbing ions from the electrode surface [27, 28]. CDI is considered an environmentally friendly technique as it relies on electrosorption (electrostatic adsorption). Extreme conditions, such as high pressure and temperature, or high‐energy power, are not required, nor are any chemicals. It is an energy‐efficient process as ions are separated from water rather than water being separated from salts, and the energy consumed during charging can be recovered during discharging. A conventional CDI setup typically features two electrodes made of the same material, with feed water flowing through them. Ions in the feed water move to the oppositely charged electrode, where they are adsorbed into the electrode's pores upon application of a voltage (0.8–2.0 V), generating an electric field, as illustrated in Figure 1a. CDI uses electrosorption, a phenomenon in which ionic species are adsorbed onto electrically charged electrodes. Upon application of a voltage across the two electrodes, oppositely charged ions in the feed migrate to form an electrical double layer (EDL), thereby separating ions from the feed water. Upon removal of external power or reversal of polarity, adsorbed ions are flushed out, allowing for ion discharge to regenerate the electrode for a new cycle. The adsorbed ions are discharged into the bulk solution, forming a brine stream [10, 29]. One CDI operational cycle is a noncontinuous process with an alternative adsorption–desorption cycle, producing two streams in each cycle: a dilute stream and a concentrated stream. The dilute stream is the product water, and the focused stream is the concentrated brine [2, 8]. Electrosorption works on the same principle as the EDL capacitor‐EDLC, where storage of energy occurs via electrostatic adsorption of counter ions, and counter‐ions from electrolyte balance the charges on the surface of electrode, which is similar to the principle of CDI, suggesting that CDI has the characteristics of energy storage and recovery where adsorption–desorption cycle can be regarded as charging and discharging cycle of an energy storage device. EDLC forms at the electrode–electrolyte interface, and in desalination, an EDLC layer forms at the electrode/water interface to produce a stream of fresh water [8, 10]. EDLC comprises a diffuse layer of electrolyte ions in the bulk solution and a Stern layer, which is firmly attached to the surface of the electrode by strong electrostatic forces. The diffuse layer is not firmly bound to the surface of the electrode; rather, it forms in its vicinity due to coulombic interactions between ions and the Stern layer. Electrolyte concentration, electrode material, and pore size are among the factors that can influence the thickness of the EDLC (Figure 1b) [29]. While choosing an electrode material for CDI, it should possess the features such as (i) quick response to adsorption and desorption of ions, (ii) must possess high surface area and capacitance required for electrosorption of ions, (iii) the electrode should be chemically and electrochemically stable and robust, (iv) should show resistance to organic and bio‐fouling, (v) electrode should be easy to assemble and operate, and (vi) should show high electronic conductivity [30]. In this review, we focus on the CDI, a promising technology for deionization, and on the fabrication of promising metal–organic frameworks (MOF)‐based electrode materials in recent years. This section provides a general overview of the fundamentals of a basic CDI system, necessary to understand the enhanced CDI performance of MOF materials.

FIGURE 1.

(a) A conventional CDI setup and (b) a representation of EDL.

1.1. CDI Cell Design and Setup

The architecture of a CDI cell is characterized by symmetry, specifically whether the electrodes are made of the same or different materials. Symmetric architecture electrodes possess the same material and encompass flow‐by CDI, flow‐through CDI, flow‐electrode CDI, and membrane CDI. An asymmetric cell design involves electrodes made of different materials and encompasses hybrid CDI, redox flow desalination, inverted CDI, battery desalination, and rocking‐chair desalination.

1.1.1. Symmetric Cell Design

Symmetric architecture electrodes have the same electrode material on both electrodes and include FB‐CDI, FT‐CDI, FCDI, and MCDI. FB‐CDI (Flow‐by) and FT‐CDI (flow‐through) setups differ in their architecture by the direction of the feed water flow. In FB‐CDI, feed water flows parallel to the electrodes. Therefore, flow‐by CDI, also known as flow‐between CDI, is a type of CDI in which feed water flows between the CDI electrodes. In FT‐CDI, water flows perpendicular to the electrodes, and a porous separator separates the electrodes (Figure 2a,b). A decrease in separator thickness can effectively reduce ionic resistance and increase the desalination rate compared to FB‐CDI. Desalination is diffusion‐controlled in both FB‐CDI and FT‐CDI. FT‐CDI exhibits a higher ASAR value than FB‐CDI for a short half‐cycle; however, FT‐CDI takes longer to desalinate due to the higher‐feed pressure [31, 32, 33]. Both cell designs yield lower electrosorption due to the co‐ion expulsion effect, in which the applied voltage repels co‐ions adsorbed in the electrode pores. The co‐ion expulsion effect observed in FB‐CDI and FT‐CDI is suppressed by ion exchange membranes (IEMs), placed in front of the cathode and anode, respectively; cation exchange membranes (CEM) and anion‐exchange membranes (AEM). IEMs selectively block co‐ions, preventing their expulsion [34]. A cell design that places an IEM in front of electrodes is called membrane capacitive deionization (MCDI) (Figure 2c). The placement of an IEM can substantially increase the SAC value, as ions are completely flushed out during the desorption cycle, thereby enhancing electrosorption in subsequent cycles with limited energy consumption [35]. Unlike FB‐CDI, FT‐CDI, and MCDI, which use fixed electrodes, FCDI uses slurry‐based electrodes and features three flow channels, separated by IEMs. The side chambers are designated for electrode slurry, and the middle chamber is reserved for feed saltwater separated by an ion‐exchange membrane (Figure 2d). Unlike a static CDI cell setup, which operates for a finite time and requires electrode regeneration in the next cycle, the three‐channel cell setup eliminates the disadvantage of electrode saturation, thereby demonstrating the ability to desalinate higher‐salinity feed [31].

FIGURE 2.

Symmetric cell architecture. (a) Flow‐by CDI, (b) flow‐through CDI, (c) membrane CDI, and (d) flow electrode CDI.

1.1.2. Asymmetric Cell Design

Asymmetric architecture features electrodes of different materials and encompasses HCDI, i‐CDI, RCD, BD, and RFD. HCDI architecture utilizes carbon‐based materials and faradaic materials as asymmetric electrodes, coupled with an AEM, to enhance desalination capacity (Figure 3a). A highly concentrated feed can result in faster adsorption in an HCDI setup, where the carbon‐based electrode and the faradaic material‐based electrode remove ions via an EDL and reversible redox reactions or ion intercalation, respectively. The intercalation of ions into the faradaic electrode via chemical bonds, along with the adsorption of feed ions on the surface of the carbon‐based electrodes, reduces energy consumption during desalination and improves the SAC value. Due to the frequent adsorption and desorption cycles during capacitive deionization, the electrodes, which are in direct contact with the feed water, undergo oxidation, and functional groups involving oxygen become attached to the electrode surface, forming oxides, thereby affecting the CDI process [31]. i‐CDI is composed of a cathode electrode with a net charge of positive and an anode electrode with a net charge of negative (Figure 3b). The cathode and anode electrodes are polarized with permanent net positive and negative charges by methods such as acid treatment, sulfonation, and amine treatment. i‐CDI operates oppositely to conventional CDI, in which electrosorption occurs on the polarized electrode surface without an external voltage, and desorption occurs upon the application of an external voltage [36, 37]. RCD is a further modification of HCDI, which operates through simultaneous intercalation and deintercalation using faradaic materials as electrodes. The basic setup of an RCD is a two‐channel CDI cell architecture that includes an AEM or CEM and either a pair of cation‐selective electrodes or a pair of anion‐selective electrodes, which selectively eliminate cations and anions, respectively (Figure 3d). Cations are intercalated into the negative electrode, which is deintercalated at the positive electrode. At the same time, anions pass through the AEM to the positive electrode and vice versa if CEM is used. Consequently, one effluent stream is enriched in ions, and the other stream is diluted. In BD, faradaic materials are used as electrodes to desalinate high‐salinity feed water due to their high electrosorption performance, as cations and anions are intercalated and deintercalated on the electrode surface through faradaic reactions, rather than the formation of EDL (Figure 3c). RFD combines the benefits of FCDI and MCDI, in which the flow of redox‐active electrolyte is through the side channels, and adsorption/desorption occurs due to reversible redox reactions of the ions with the redox‐active electrolyte (Figure 3e). The deionization rate is primarily dependent on the concentration of redox‐active ions. The major limitation of RFD is the tendency of redox‐active ions to move out of the IEM to the effluent stream [10, 29, 31, 38]. Table 1 summarizes different CDI cell designs and their characteristics.

FIGURE 3.

Asymmetric cell architecture (a) HCDI, (b) i‐CDI, (c) BD, (d) RCD, and (e) RFD.

TABLE 1.

Summary of CDI cell design and its characteristics.

FB‐CDI

• The flow of feed water is parallel to the electrodes.

• Adsorption of salt is controlled by diffusion; therefore, desalination takes more time.

FT‐CDI

• The flow of the feed water is perpendicular to the electrodes.

• Desalination is diffusion‐controlled but shows a higher ASAR than FB‐CDI.

MCDI

• IEMs are placed in front of electrodes to prevent co‐ion expulsion.

• Placement of IEMs increases SAC value.

FCDI

• Has slurry‐based electrodes separated by IEMs.

• Continuous electrode regeneration and higher‐feed desalination.

HCDI

• Carbon‐based electrode and a faradaic electrode as asymmetric electrodes coupled with AEM.

• Intercalation of ions occurs at the faradaic electrode, and adsorption occurs at the carbon‐based electrode.

i‐CDI

• Electrodes are polarized with a permanent net positive charge and a net negative charge.

• Operates opposite to conventional CDI.

RCD

• Simultaneous intercalation and de‐intercalation occur at faradaic electrodes.

• Modification of HCDI.

BD

• High‐salinity feeds are desalinated using Faradaic electrodes.

RFD

• Combines the benefits of FCDI and MCDI.

• Redox‐active electrolyte used as electrodes.

1.2. Adsorption and Desorption Cycles in CDI

The adsorption cycle of CDI can operate on constant current adsorption (CCA) and constant voltage adsorption (CVA). In contrast, the desorption cycle of CDI can operate in zero‐voltage desorption (ZVD), reverse‐current desorption (RCD), and reverse‐voltage desorption (RVD) [38, 39]. CVA‐ZVD is the conventional cycle followed in CDI. In the CVA‐ZVD cycle, during adsorption, a constant voltage is applied. Initially, the salt removal process is the fastest, resulting in the lowest effluent salt concentration. Later, the concentration of salt in the effluent gradually returns to the feed value as the salt removal rate decreases with increasing counter EDL voltage [31, 38].

Desorption of counterions in ZVD occurs due to mutual repulsion. It continues until the salt concentration in the macropores of the electrode reaches that of the bulk solution, and desorption decreases as the EDL approaches charge neutrality. The counter‐ion concentration in the EDL is highest at the beginning; therefore, the desorption rate is maximum at the start of the ZVD cycle [31, 38, 40].

In the RVD desorption cycle, all counter ions are terminated, and it is not until charge neutrality is approached that the repulsion is due to mutual repulsion, but also to the reversed electric field. RVD shows a faster and more effective desorption rate than ZVD for high TDS feed because of the influence of the reverse electric field in addition to the mutual repulsion. Because counterion cleanup at the electrode is more pronounced in RVD, both desorption time and cycle time are reduced, thereby increasing salt removal in the subsequent adsorption cycle. Counterion cleanup from the electrode is much more pronounced in MCDI than in classic CDI, as co‐ions in the feed are adsorbed in the electrode's pores. When a reverse voltage is applied during desorption, it expels co‐ions, reducing the efficiency of salt removal. The use of IEMs reduces the co‐ion expulsion in MCDI, which prevents or selectively blocks co‐ion adsorption, thereby enhancing the electrode's adsorption capacity [31, 41].

Tunable and constant effluent concentration can be achieved by employing the CCA or RCD cycle for adsorption and desorption, with a constant current applied, resulting in continuous salt removal and thereby a constant effluent salt concentration [38, 42, 43]. The EDL voltage continues to increase as counterions in the EDL increase and the electrolyte concentration decreases; therefore, the required voltage to maintain a constant current also increases. When the voltage reaches a preset threshold, the desorption cycle initiates. The counterion adsorption flux and the co‐ion expulsion flux generate ionic current. Ionic current, because of co‐ion expulsion, is reduced in MCDI because of involvement of IEMs, and constant current is only because of counter ion adsorption flux; therefore, CCA can only operate under MCDI and not for classic CDI, and similarly, a better clean‐up of electrodes is provided by RCD during the desorption cycle in the presence of IEMs. The advantage of producing a constant, tunable effluent concentration with the CCA and RCD cycle is particularly effective for industrial brines with high TDS [31, 38, 44, 45]. Adsorption in constant‐current mode has a higher energy recovery rate; energy can be recovered by attaching an external load. Lower energy consumption and more energy‐saving features are observed in CC mode. Adsorption in constant‐voltage mode suffers energy losses due to ohmic and non‐ohmic resistance. Although CV mode exhibits higher energy consumption, it is more widely studied in academic and commercial fields than CC mode due to the difficulty of controlling voltage and operational challenges. CV mode consumes more energy and experiences greater energy loss than CC mode, but the most energy‐efficient mode may not necessarily exhibit the highest removal efficiency. CV mode consumes more energy than CC mode, but it cannot achieve the required removal efficiency. Therefore, high‐energy efficiency does not always ensure high removal efficiency [43, 44, 46, 47, 48].

To obtain the required removal efficiency with minimal consumption of energy and flexible operation, a new charging mode, the advanced staged voltage mode (SV) for MCDI, has been studied, leveraging both CV and CC modes. Although CC mode shows higher energy efficiency than any other modes, SV mode can be an efficient operating strategy to obtain the needed water quality. SV mode is typically operated in a two‐staged voltage mode (2SV), where an ascending voltage is applied, followed by a descending voltage under various TDS concentrations. The energy consumption and removal efficiency of 2SV are then estimated systematically through LSV (linear) and ESV (exponentially) staged voltage modes [49] (Figure 4). Table 2 summarizes different operation modes of CDI.

FIGURE 4.

Operation of SV (staged voltage) in the MCDI process. Reproduced from ref [49]. Copyright (2020), with permission from Elsevier.

TABLE 2.

Key characteristics of different operating modes of CDI.

Operation mode	Description/features	References
Constant voltage adsorption (CVA)	A constant voltage is applied during the adsorption step. Adsorption initially reaches a maximum, then slows.	[31, 38]
Constant current adsorption (CCA)	Constant current is applied in the adsorption step until the voltage reaches a preset threshold. Constant effluent concentration can be achieved.	[38, 42, 43]
Zero voltage desorption (ZVA)	Desorption occurs when the external power is turned off. Therefore, spontaneous desorption occurs, with desorption initially reaching a maximum and slowing as the EDL approaches neutrality.	[31, 38, 40]
Reverse voltage desorption (RVD)	A reversed voltage is applied during the desorption step. Electrode cleanup is more pronounced. It is more significant in MCDI than classic CDI.	[31, 41]
Reverse current desorption (RCD)	A reversed current is applied for the desorption step until it reaches a preset threshold. It is more effective for industrial brines with high TDS. IEMs are required for a complete electrode cleanup during desalination.	[38]

1.3. Key Parameters for CDI Performance and Evaluation Criteria

Various parameters are used to evaluate CDI performance, including salt adsorption capacity (SAC), charge efficiency, desalination rate, cycle stability, and water recovery. This section explains these parameters with different terms and symbols.

1.3.1. SAC (Salt Adsorption Capacity)

SAC, also known as desalination capacity or salt‐removal capacity, is a key performance metric for assessing a CDI system. SAC estimates the quantity of salt removed per unit area of the electrode (Equation (1)). When the salt concentration reaches equilibrium, the SAC value is calculated. At equilibrium, charging of the cell is complete, and no more adsorption occurs.

$$\text{SAC}=\frac{(C_{\mathrm{o}}\hspace{0.17em}-C_{\mathrm{e}})\times \hspace{0.17em}V}{m}$$ (1)

C _o = Initial concentration of salt solution in mg L⁻¹ unit.

C _e = Final or the equilibrium concentration of salt solution measured in mg L⁻¹ unit.

m = Mass of the electrode material measured in g units.

The electrode material's salt absorption capacity against the operation time of desalination in min⁻¹ is known as the average salt absorption rate (ASAR) (Equation (2)).

$$\text{ASAR}=\frac{\mathrm{SAC}}{t}$$ (2)

t = Time of electrosorption.

1.3.2. Charge Efficiency (CE)

CE is the adsorbed equilibrium salt‐to‐charge ratio at the electrode (Equation (3)). Higher CE indicates lower energy consumption.

$$\mathrm{CE}\hspace{0.17em}=\frac{\mathrm{SAC}\times \hspace{0.17em}F}{\frac{\int I.\mathrm{d}t}{m}}$$ (3)

F is the Faraday constant (96,485 C g⁻¹).

I is the current in A unit.

1.3.3. Desorption Capacity (DC) and Energy Regeneration Efficiency (ERE)

DC and ERE are calculated by Equations (4) and (5)

$$\mathrm{DC}\hspace{0.17em}=\frac{(C_{\mathrm{o}}\hspace{0.17em}-C_{\mathrm{d}})\times \hspace{0.17em}V}{m}$$ (4)

C _d = Concentration of feed solution after complete desorption, measured in mg L⁻¹

V is the total volume of feed solution measured in L.

$$\mathrm{ERE}\hspace{0.17em}=\frac{{\mathrm{SAC}}_1}{{\mathrm{SAC}}_n}$$ (5)

1.3.4. SRE (Salt Removal Efficiency)

SRE is calculated by Equation (6)

$$\mathrm{SRE}\hspace{0.17em}=\frac{C_{\mathrm{o}}\hspace{0.17em}-C_{\mathrm{e}}}{C_{\mathrm{o}}}\times \hspace{0.17em}100$$ (6)

1.3.5. Coulombic Efficiency (CoE)

CoE in FCDI is calculated by Equation (7)

$$\mathrm{CoE}\hspace{0.17em}=\hspace{0.17em}\raisebox{1ex}{$Q_{\text{discharge}}$}\!\left/ \!\raisebox{-1ex}{$Q_{\text{charge}}$}\right.$$ (7)

Q _discharge = Recovered charge during discharging measured in A.h.

Q _charge = Applied charge measured in A.h.

1.3.6. Energy Consumption

Equation (8) calculates energy consumption:

$$E\hspace{0.17em}=\frac{E_{\mathrm{cell}}\int I.\mathrm{d}t}{m_{\mathrm{rem}}}$$ (8)

E _cell =  Applied voltage in V.

M _rem = Amount of ions eliminated from the salt solution measured in mg.

1.3.7. Water Recovery (Υ)

Water recovery (Υ) is the ratio of the treated water produced to total processed water by the system (Equation (9)).

$$Y=\raisebox{1ex}{$V_{\text{desalinated}}$}\!\left/ \!\raisebox{-1ex}{$V_{\text{processed}}$}\right.$$ (9)

V _desalinated = Volume of dilute water.

V _processed = Volume of total water processed.

Υ depends on factors like flow channel volume, flow rate, and charge–discharge time.

2. Metal–Organic Frameworks (MOFs) as Electrode Materials for CDI

Carbon‐based electrodes were initially considered ideal electrode materials for CDI due to their high surface area and electrical conductivity. Carbon nanotubes (CNTs), carbon aerogels (CAs), graphene, mesoporous carbon (MC), and their composites are among the carbon‐based materials that have been explored since activated carbon (AC) was first reported as an electrode for CDI. Later, MOF‐derived porous carbon (PC) emerged as a promising material for CDI electrodes due to its high specific surface area (SSA)and porosity.

Because the electrode material is one of the most important factors influencing the treatment effect of CDI technology, many researchers have conducted studies on improving electrode materials. Among them, much attention focused on carbon‐based materials, which included AC, CA, carbon nanofiber (CNF), CNT, MC, and graphene, as they have been considered a suitable material for CDI electrodes due to their advantages of having good stability, high economic efficiency, and tunable pore size. However, its application in CDI is limited due to its poor hydrophilicity and low performance, thereby attracting much attention to modifications of carbon materials to optimize surface area, porosity, hydrophilicity/hydrophobicity, and, in turn, increase deionization capacity. Modifications such as doping with metal oxides, conductive polymers, and MOFs to synthesize composite materials that improve CDI performance by leveraging the advantages of each component [50]. Carbon composites have shown great potential to improve CDI performance, but certain challenges, such as composite stability, ion dissociation and agglomeration during the electrosorption process, hinder electrode longevity and performance.

MOFs, a class of organic framework materials, are highly porous and crystalline, composed of inorganic metal clusters or ions that form porous networks via self‐assembly using polydentate organic ligands, with metal nodes interconnected via coordinate bonds that expand in three dimensions. MOFs can incorporate multiple metals, guest molecules, and donor ligands, making them highly flexible. They also combine the properties of organic and inorganic porous materials. These properties of MOFs have attracted many researchers and have also been applied in storage and separation, sensing, electrochemistry, and catalysis [51, 52].

MOFs exhibit significant compositional and structural tunability, as multiple metals, guest molecules, and donor ligands can be incorporated to alter their composition. In contrast, selecting ligands of different lengths and creating defective coordination sites can induce structural alterations. Different synthetic procedures can induce these alterations by controlling particle size, coordination number, ion or charge‐transfer properties, topologies, and dimensions [53]. Pyrolysis of MOFs can incorporate redox‐active components into the synthesized carbonaceous materials by heat treatment, decomposing the organic linkers and transition metals present in the MOFs. The resultant carbonaceous material, with metal sites that form redox sites, is distributed on a carbon matrix that supports electrosorption.

Several reviews have been published on MOF‐derived carbon materials as electrodes for CDI, as well as on carbon‐based materials for CDI, focusing on different electrode materials reported, their preparation conditions, their advantages and disadvantages, and the factors affecting their CDI performance [54, 55, 56]. This review highlights how modifying MOF‐based materials can enhance electrosorption performance and elucidate the mechanisms underlying desalination and ion capture.

2.1. Desalination

For early applications of MOFs as CDI electrode materials, researchers focused on MOF‐derived carbons due to their diverse structures, developed pores, adjustable pore sizes, high SSAs, and functionalized pore surfaces. Based on their structural arrangement, MOFs are mainly of the MOF, zeolitic imidazolate framework (ZIF), and material Institute Lavoisier framework (MIL) series. Most of the directly carbonized MOFs are MOF‐5, ZIF‐8, and MIL‐88 [57, 58]. MOF‐5 has a structure similar to that of zeolites and possesses a high surface area. Chang et al. carbonized MOF‐5 to prepare PC for CDI electrodes. MOF‐5 was synthesized and then carbonized at temperatures from 500°C to 1000°C to prepare PC. The results indicated that MOF‐5 was completely carbonized at 900°C, and ZnO completely disappeared and only carbon was left, showing maximum SSA as well as pore volume, because of which PC‐900 showed the maximum desalination capacity of 9.39 mg g⁻¹ with a specific capacity of 107.74 F g⁻¹ [59]. ZIFs are a subclass of MOFs having zeolite topologies that show good chemical stability and structural diversity, in which imidazolate ligands are bridged to tetrahedrally coordinated transition‐metal ions. Liu et al. prepared PCP (PC polyhedra) via carbonization of ZIF‐8 at 800°C, 1000°C, and 1200°C and used them as CDI electrode materials. The results indicated that PCPs‐1200 had the highest SSA, specific capacity, and pore volume, thereby achieving an adsorption capacity of 13.86 mg g⁻¹, demonstrating the important role of carbonization temperature in the electrochemical performance of PCPs, which influences charge transfer resistance and specific surface area [60]. Later, Xu et al. inserted CNTs into ZIF‐8, and then pyrolyzed it to synthesize a hierarchical porous CNT (hCNT)/PCP hybrid. The fabricated hybrid hCNT/PCP showed an electrosorption capacity of 20.5 mg g⁻¹, attributed to its hierarchical porous structureand a 3D conducting network that enables high electrosorption capacity, a high specific surface area, and good cycle stability [61]. In the same year, Wang et al. for the first time used a bimetallic ZIF composed of ZIF‐67 and ZIF‐8 as a template to fabricate nitrogen‐doped PC, having a large surface area, highly dispersed N species derived from the ZIF‐8, and high graphitic structure from ZIF‐67 and the resulting CDI electrode exhibited a capacity of 16.63 mg g⁻¹ due to the combined effect of high surface area, high graphitic structure and high N content [62]. Bimetallic MOFs (BMOFs) were further studied for CDI applications by Ding and coworkers. The group synthesized a series of BMOFsbased on ZIF‐67 and ZIF‐8 with varying Zn:Co molar ratios and used them as precursors to synthesize PC with varying particle sizes and graphitization degrees. PC derived from Zn:Co = 3:1 (PC‐ZnCo‐3) showed a superior salt removal capacity of 45.62 mg g⁻¹ at 1.4 V. The high performance is due to improved electrical conductivity and a large accessible ion surface area of BMOF‐derived carbon [63]. Lavoisier skeleton series materials are called the MIL series. They have the advantages of high specific surface area, high porosity, and unsaturated matching sites compared to zeolitic imidazolate and graphite materials. Xu et al. prepared sPCRs (shuttle‐like PC rods) by pyrolysis at high temperature and subsequent etching of rod‐shaped MOF MIL‐88(Fe) as a template. sPCRs‐900 showed an electrosorption capacity of 16.2 mg g⁻¹ and good cycle stability. The results indicated that sPCRs had good morphology and a high specific surface area, and that their pore volume and specific surface area increased with increasing carbonization temperature [64]. Synthesizing MOF‐derived carbon with a controllable surface area remains difficult, despite MOF‐derived carbons making significant progress in this field. Thereby, to address the issue of synthesizing PC derived from MOFs with a controllable surface area, wang et al. synthesized ZICarbon, RMCarbon, and ZFCarbon from carbonizing three different MOF structures‐ ZIF‐8, RT‐MOF‐5, and Zn Fumarate, respectively, at about 1000°C. The results showed the order of specific surface area as ZFCarbon > RMCarbon > ZICarbon, due to differences in Zn/C ratios. Increasing the Zn/C ratio in MOF precursors nearly proportionally increases the specific surface area and pore volume of the resulting PC, thereby improving desalination performance (Figure 5). The enhanced desalination performance is attributed to the well‐developed pores and diverse structure of MOF‐derived carbon, which provided greater ion accessibility than the smaller micropores of AC, making it a potential electrode material for CDI applications [65]. To further increase the electrosorption performance of MOF‐derived carbons, MOFs can be modified to increase the specific surface area. For example, wang et al. in 2017 worked on expanding the pores of MOF‐derived carbon via a surfactant‐template strategy. The material after carbonization under nitrogen exhibits expanded pores, including micropores and mesopores, thereby enhancing deionization performance. Before, most MOF‐derived carbon was limited to micropores, which affected ion transfer [66]. Doping of other elements into MOF‐derived carbon can further modify the pore structure of the MOF‐derived carbon, which, along with increasing specific surface area, also increases the hydrophilicity and conductivity of MOF‐derived carbon materials. Shi et al. highlighted that MOF‐derived carbon materials can be combined with conductive structures to improve deionization performance and improve electron transport by integrating MOF‐derived PC rods with a graphene network. Composites of MOF‐derived carbon with other carbon materials integrate the advantages of both materials while mitigating their shortcomings. However, surface modifications and carbonization at high temperature, the use of alkalis and strong acids, and other environmental factors can cause collapse of the MOF's structure [67].

FIGURE 5.

Synthetic process of porous carbon from Zn‐containing MOFs.

In 2019. Wang et al. directly incorporated a MOF, ZIF‐67, into CDI electrodes, along with conductive polypyrrole nanotubes, thereby highlighting the potential of MOFs as promising electrode materials for CDI. Previously, MOFs were mainly pyrolyzed to produce PC materials for CDI electrodes [68]. The work on MOF‐based materials as electrode materials for CDI was expanded by Xu et al. in 2019, in which commercially available CNTs replaced the complex synthetic procedure and the high production cost of polypyrrole. A ZIF‐67/CNT hybrid was synthesized and directly used as a CDI electrode. When compared to the pristine MOF, the combination of both ZIF‐67/polypyrrole and ZIF‐67/CNT enhanced the electrochemical performance and electrical conductivity as well as improved CDI performances. Still, as MOF structures are powdered, without polymeric binders, they cannot be directly used to fabricate electrodes, leading to poorer electrical conductivity, a decrease in the accessible surface area and poorer CDI performance [69]. To address this drawback, Zhang et al. prepared a binder‐free nanopatterned MOF (npMOF) electrode by directly patterning ZIF‐67 nanocrystals on carbon cloth (CC). The synthesized npMOF exhibited improved accessible surface area, electrical conductivity, and a desalination capacity of 21.3 mg g⁻¹ [70]. Subsequent research on MOFs focused on improving their unsatisfactory electrochemical stability and electronic conductivity. Therefore, considering the pseudocapacitive properties of MOFs and the high conductivity of graphene, Feng et al. in 2021 designed a MOF/holey graphene composite as a CDI electrode material. The porous hierarchical structure of the synthesized composite increases the electrolyte/electrode contact area, thereby demonstrating good cyclic stability and strong ion electrosorption capacity [71]. To summarize, initial research on MOFs as electrode materials for CDI mainly focused on MOF‐derived carbons for CDI electrodes, and efforts were made to improve the specific surface area of MOFs by tuning morphologies, structures, and pores through doping, integrating with conductive structures, or templating agents. Efforts to directly incorporate MOFs into CDI electrodes were also made. This section highlights recent modifications to MOFs that impact the final composite properties and desalination performance. Recently fabricated MOF‐based electrodes for desalination, modifications adapted during fabrication, and CDI performance are summarized in Table 3.

TABLE 3.

Summary of recent MOF‐based electrode materials for desalination, modifications adapted, and CDI performance.

Rational design strategy	Fabricated electrode	MOF material	Modification adapted	Initial NaCl concentration, mg L−1	Elestrosorption capacity (mg g−1) @ Voltage (V)	Reference
Morphological	CoTi‐800/PPy	CoTi‐ZIF‐9	Incorporation of Polypyrrole	500	77.3 @ 1.2	[72]
	Ni‐MOF@CNT	2D Ni‐MOF	Integration of CNTs	500	53.57 @ 1.2	[73]
	Ni‐MOF/NiCoFe‐PBA	2D Ni‐MOF	Introduction of Prussian blue analog	250	63.19 @ 1.2	[74]
	ZPC0.5	ZIF‐L (Zn)	Coating of polydopamine (PDA) on ZIF‐L	500	29.60 @ 1.2	[75]
	ZIF‐L@mPDA‐C	Zn.Co‐ZIF‐L	Coating of mesoporous polydopamine on ZIF‐L	500	56.5 @ 1.4	[76]
	NiCoMn‐HNA	NiCo‐ZIF‐L	Integrating MOF with LDH on carbon fiber	500	103.1 @ 1.2	[77]
	NC@GC/CNTs	ZIF‐L and ZIF‐67	ZIF‐67 coated on ZIF‐L template (MOF‐on‐MOF strategy)		77.33 @ 1.4	[78]
	MoC/Mo2C/CoNC	ZIF‐8 and ZIF‐67	MoC/Mo2C nanocrystrals embedded on Co/N‐doped carbon nanoframe via MOF‐on‐MOF strategy	500	73.6 @ 1.2	[79]
	NC(M)/HMCS	ZIF (M)	Synthesized via interface coating	10.0 × 10−3 M	28.9 @ 1.2	[80]
	NC(M)@HMCS	ZIF (M)	Synthesized via space encapsulation	10.0 × 10−3 M	32.1 @ 1.2	[80]
	CoNi‐LDH@HMCS	ZIF‐67	Confinement of ZIF‐67 into HMCS	584	36.41 @ 1.2	[81]
	Ni@NiO@C//AC	Ni‐MOF	Spherical morphology of the composite	1500 µS/cm	2.18 mmol/g @ 1.2	[82]
	MOF‐Fhy‐C	Zn‐MOF	Fungal hyphae integration to MOF	585	40.8 @ 1.2	[83]
Metal node	Ag‐‐MOF	Ag‐MOF	Ag‐MOF synthesized with Ag nodes and organic linkers	500	115.16 @ 1.2	[84]
	HCNF/Mg	Mg‐MOF	Mg‐MOF synthesized with Mg nodes followed by carbonization	500	22.10 @ 1.2	[85]
	Fe‐MOF/CNT	Fe‐MOF	Fe‐MOF synthesized with Fe nodes and integrated with CNT	1000	80.575 @ 1.6	[86]
Doping	N‐PC	Al‐MOF	N‐doping	500	19.9 @ 1.2 24.7 @ 1.4	[87]
	NPC‐800	MOF‐5	N‐doping	500	24.17 @ 1.2	[88]
	NGZC	ZIF‐8	N‐doping	500	16.29 @ 1.2	[89]
	N‐C	Zn/Fe bimetallic MOF	N‐doping	500	CDI‐19.82 @ 1.2 MCDI‐30.19 @ 1.2	[90]
	NPFe‐N‐C	Zn/Fe bimetallic MOF	N‐doping with embedded Fe nanoparticles	500	CDI‐29.13 @ 1.2 MCDI‐44.63 @ 1.2	[90]
	SAFe‐N‐C	Zn/Fe bimetallic MOF	N‐doping with isolated single Fe atom integrated	500	CDI‐36.75 @ 1.2 MCDI‐57.80 @ 1.2	[90]
	PBA@Co/N‐HC	ZIF‐67 (Co)	Co and Ni doping	500	33.37 @ 1.2	[91]
	3D Ag@NC	ZIF‐67 (Co)	Ag nanoparticles doping	1000	105.29 @ 1.4 121.11 @ 1.6	[92]
	P‐doped ZIF‐8	ZIF‐8 (Zn)	P doping	500	35.95 @ 1.2	[93]

2.1.1. Morphological Modifications

The morphology of MOFs‐derived carbon plays an essential part in charge and ion transport in CDI. For the development of high‐performance materials, it is necessary to focus on the rational design of the pore structure, morphology, and chemical composition of the MOF materials. Polypyrrole and CNTs are ideal electrode additives known for their high conductivity, which can enhance charge storage in composite materials by increasing the charge‐transfer efficiency between the electrolyte and the electrode materials, thereby improving removal efficiency. As discussed above, in 2019, Wang et al. first incorporated conductive polypyrrole nanotubes into ZIF‐67, in which the 3D PPy nanotube network interconnects each MOF particle by intertwining PPy nanotubes with ZIF‐67 polyhedra, serving as a bridge for electron transfer between MOF particles. The ZIF‐67/PPy hybrid exhibited a desalination capacity of 11.34 mg g⁻¹ [68]. The same group replaced polypyrrole nanotubes with commercially available CNTs, and FESEM and TEM images of the hybrid ZIF‐67/CNT composite show a 3D network‐like structure in which CNTs are inserted into ZIF‐67 polyhedra, with CNTs bridging neighboring ZIF‐67 polyhedra [69]. Therefore, the group first highlighted the potential of integrating high‐electrical‐conductivity PPy nanotubes and CNTs with high‐porosity, regular‐micropore MOF‐based materials as a promising material for CDI electrodes beyond traditional carbon electrodes with PPy and CNTs forming ZIF‐PPy‐ZIF and ZIF‐CNT‐ZIF electron‐conductive pathways. The integration of PPy and CNTs into MOFs for desalination was further studied. Recently, Cao et al. synthesized two forms of PPy: nanoparticles and nanotubes, and incorporated them into a bimetallic ZIF (CoTi‐ZIF‐9), which has a microcube‐like structure. The PPy nanoparticles, approximately 100 nm in size, had a rougher surface, were stacked in a disorderly manner, and then assembled into a nanoflower‐like morphology. The PPy nanotubes, with 200–500 nm diameter and dozens of micrometers in length, were stacked and interconnected to form a 3D network‐like morphology, with a correspondingly hollow interior and smooth surface. The group primarily prepared a range of nanocomposites of CoTi‐800/PPy, naming them according to the different forms of PPy and CoTi‐800 feed ratios. Figure 6a–f shows PPy with a nanotube‐like morphology successfully grown on CoTi‐800 cube, and Figure 6g–l shows PPy with a nanoparticle kind of morphology polymerized successfully on the CoTi‐800 carbon surface. The results indicated that the CoTi‐800/PPy composite performance is superior to that of PPy or CoTi‐800 alone, because of the combined effect between CoTi‐800 and PPy. The CoTi‐800/PPy‐2 electrode demonstrated a 77.3 mg g⁻¹ rate of desalination, attributed to the introduction of PPy which is rich in nitrogen, and because of the nanotubular kind of morphology of PPy forming a 3D blockchain like network in the composite for Na⁺ and Cl⁻ to move through the surface of electrode and thereby, the diffusion path gets shorten within the composite structure, and improving transport of ion [72]. Hu et al. interconnected CNTs with hierarchically structured 2D Ni‐MOF nanosheets, where CNTs serve as a conductive network to enhance conductivity by providing an efficient electron‐transfer pathway between Ni‐MOF nanosheets and reducing their restacking. The 2D Ni‐MOF nanosheets, due to their unique morphology distinct from the bulk layered structure, expose more active sites, thereby enhancing electronic conductivity by influencing the bandgap energy and the density of states present near the Fermi energy level. The 2D Ni‐MOF@CNT composite synthesized via a one‐pot hydrothermal method (Figure 7) exhibited fast interfacial charge transfer when evaluated as a CDI electrode, with a desalination capacity of 47.58 mg g⁻¹ at 1.2 V. The composite additionally exhibited a lithium‐ion extraction capacity of 30.29 mg g⁻¹ [73]. The same group further studied ways to mitigate the stacking of 2D Ni‐MOF sheets, thereby exposing their active sites and enabling increased sodium ion embedding and de‐embedding. Therefore, the group synthesized a NiCoFe‐PBA (Ni‐MOF/NiCoFe‐Prussian blue analog) composite via an in situ growth method (Figure 8). The growth of PBA on the surface of Ni‐MOF mitigates the challenges of limited charge transfer and particle agglomeration encountered by Prussian blue, and it also prevents the stacking of Ni‐MOF 2D sheets, thereby exposing more active sites. Therefore, the composite has enhanced structural integrity and thereby an excellent desalination capacity of 63.19 mg g⁻¹ with a NaCl solution of 500 μS cm⁻¹ [74]. Many PCs derived through carbonization of MOFs have been widely used as electrodes for CDI; however, 2D PC is considered a promising electrode material because of its convenient ion transport pathway and exposed active sites, especially 2D‐ZIFs, a subclass of N‐rich MOFs having flattened and thin leaf‐like shape, which, upon carbonization, transforms into nanosheets exhibiting high specific surface area. However, most of these 2D‐ZIFs, particularly the 2D foliated ZIF‐L crystal with Zn metal nodes, are structurally unstable; therefore, carbonization of ZIF‐L leads to structural collapse, shrinkage, and even aggregation, resulting in 2D structural deterioration and functional degradation. As a result, the ZIF‐L‐derived carbon (ZC) exhibits poor mesoporous structure, low electrical conductivity, and self‐aggregation, leading to poor CDI performance. Therefore, to maintain the 2D morphology, coating of the MOF surface with a polymer has been considered an effective strategy, as a polymer‐coated shell interacts with the MOF surface and exerts an outward force, and preventing shrinkage of the MOF structure during pyrolysis, thereby maintaining the original 2D morphology. Fang et al. coated a polydopamine (PDA) shell coating on the surface of ZIF‐L before pyrolysis. The prepared ZIF‐L‐derived 2D hollow leaf‐shaped nitrogen‐doped MC composite electrode exhibits a desalination capacity of 29.6 mg g⁻¹, which is attributed to the protective effect of PDA on the 2D structure of ZIF‐L, thereby exposing more adsorption sites, enlarging pore size, and ensuring sufficient contact area between the active sites and the electrolyte [75]. An appropriate thickness and pore characteristics of the PDA shell are considered to affect the pore structure, composition, morphology, and electrochemical performance of the derivatives. Therefore, to modulate the porosity of the PDA layer, Li et al. encapsulated a leaf‐like Zn, Co‐ZIF‐L core with a mesoporous PDA layer, resulting in ZIFL@mPDA precursor. The mPDA layer during pyrolysis acts as a rigid interface, preventing the collapse of the ZIF‐L's backbone. The ZIFL@mPDA‐C electrode obtained after pyrolysis showed higher SAC than many reported carbon‐based CDI electrodes, which can be attributed to the easier release of volatile organic compounds from the mesoporous region where the PDA layer is thinner, leading to explosive CNT growth, facilitating electron transfer and electrochemical performance [76]. Jiang et al. integrated leaf‐like NiCo‐ZIF‐L and NiCoMn‐LDH to form a hetero‐nanosheet array of MOF/LDH and used the fabricated NiCoMn‐HNA electrode as an anode for dechlorination. The hollow MOF/LDH heteronanosheet effectively reduces LDH self‐aggregation. NiCoMn‐HNA has a leaf‐like morphology, with a honeycomb‐like surface structure resulting from randomly distributed interwoven nanosheets that form a 3D porous mesh resembling a honeycomb. This morphology of NiCoMn‐HNA facilitates ion diffusion and redox reactions by increasing surface area and preventing electrolyte ions from becoming congested. The electrode exhibited a dechlorination capacity of 103.1 mg g⁻¹ [77]. The MOF‐on‐MOF strategy is another cutting‐edge approach to producing well‐defined hybrid MOFs, integrating the advantages of individual MOF components by constructing a core‐shell MOF@MOF heterostructure. Zhang et al. constructed a core‐shell structured 2D ZIF‐L@ZIF‐67 using the MOF‐on‐MOF strategy. The ZIF‐L@ZIF‐67 hybrid MOF, interestingly, had a morphology different from the original 2D leaf‐like ZIF‐L. Pure ZIF‐67 has a 3D rhombic dodecahedral shape with a distorted, rough surface. When ZIF‐67 is coated on the ZIF‐L template, the edges of the ZIF‐L nanoleaves become rough. Therefore, ZIF‐L@ZIF‐67 exhibits an angled surface and an elongated hexagonal morphology. The ZIF‐L@ZIF‐67, when subjected to pyrolysis, formed the NC@GC/CNTs composite, which possess 2D leaf‐like morphology, attributing that the MOF‐on‐MOF strategy that prevents 2D ZIF‐L from agglomeration or collapsing in the carbonization step and achieves a greater specific surface area and higher porosity than ZIF‐L‐C, demonstrating that the introduction of ZIF‐67 into ZIF‐L can decrease structural collapse, thereby accelerates transport and diffusion of electrolyte ions by forming interconnected pore channels required for fast charge/discharge process [78]. Pang and coworkers synthesized ultrafine MoC/Mo₂C nanocrystals that are uniformly embedded in conductive Co/N‐doped carbon nanoframe (MoC/Mo₂C/CoNC) via a MOF‐on‐MOF strategy. The MOF‐on‐MOF strategy enables the fabricated composite to exhibit structural stability and compositional advantages, facilitating charge transfer and ion adsorption kinetics [79]. Hollow MC nanospheres (HMCS) possess some unique physical and chemical characteristics, including high surface area, large pore volume, and low density. Modifications of HMCS‐based nanomaterials offer the advantages of abundant mesopores and micropores, excellent stability and durability, and a large number of active sites. Tang et al. synthesized two types of ZIF(M)‐derived nitrogen‐doped carbon and HMCS nanocomposites via interface coating (NC(M)/HMCS) and space encapsulation design (NC(M)@HMCS). Modification via the interface‐coating method involves the self‐assembly of nanoparticles, thereby altering the interface's chemical composition and structure. The space‐encapsulating method involves confining guest substances within the host material's internal space. As a result of a better interconnected 3D framework and better utilization of the internal space of hollow carbon NC(M)@HMCS composite, it exhibits higher electrochemical performance than NC(M)/HMCS [80]. The same group confined ZIF‐67 into HMCS through a space‐confined strategy (Figure 9). The synthesized yolk‐shell structured ZIF67@HMCS nanocomposite, in which HMCS acts as the shell and ZIF‐67 as the core, when subjected to etching, resulted in an LDH@HMCS nanocomposite with a 3D hollow nanocage superstructure that prevents self‐stacking of LDH and facilitates ion transportation [81]. Xue and coworkers carbonized Ni‐MOF under an argon atmosphere to prepare spherical Ni@C. The Ni@C has a PC skeleton with embedded Ni nanoparticles. A composite Ni@NiO@C was synthesized by carbonizing Ni@C a second time under an air atmosphere, resulting in surface oxidation of the Ni nanoparticles. The synthesized Ni@NiO@C has a spherical morphology, providing an open, porous structure that promotes greater salt‐ion diffusion. Surface oxidation additionally makes the spherical surface more porous. The spherical morphology of the composite Ni@NiO@C is not easily disrupted during desalination, indicating its structural stability. The composite is rich in mesopores and has low charge‐transfer resistance, with a desalination rate of 2.18 mmol g⁻¹ in a 1500 μS cm⁻¹ solution of NaCl at 1.2 V, attributed to its large specific surface area and rich electroactive sites that accept more sodium ions due to the novel spherical morphology of Ni@NiO@C. The strong synergistic effect of a large surface area and a rich supply of electroactive species promotes ion diffusion and accelerates charge transfer [82]. Chen et al. developed a porous, binder‐free carbon electrode with a fungal hyphae sheet surface, which acts as a supporting substrate for the growth of MOF crystals. Fungal hyphae have a continuous, fibrous morphology with many functional groups on their surfaces, making them an ideal supporting substrate. The developed MOF‐fungal hyphae‐derived carbon (MOF‐Fhy‐C) exhibited a 3D, open, and porous morphology with interconnected fibrous networks and a rougher surface than pristine fungal carbon. Functional groups such as ‐COOH and ‐OH were identified on the fungal surface by FTIR analysis, suggesting a greater number of active sites. The MOF‐Fhy‐C electrode significantly enhanced surface area, with a SAC value of 40.8 mg g⁻¹ and excellent cycling stability over 30 consecutive adsorption‐desorption cycles [83].

FIGURE 6.

Structural characterization and Morphology of composites of CoTi‐800/PPy. The first figure on the left shows the CoTi‐800/PPy composite structure. SEM image of (a,b) CoTi‐800/PPy‐1; (c,d) CoTi‐800/PPy‐2; (e,f) CoTi‐800/PPy‐3; (g,h) CoTi‐800/PPy‐4; (i,j) CoTi‐800/PPy‐5; (k,l) CoTi‐800/PPy‐6; (m,n) XRD image of Composite CoTi‐800/PPy; XPS image of (o) Co 2p spectrum and (p) Ti 2p spectrum in CoTi‐800/PPy Composites. Reproduced from ref [72]. Copyright (2024), with permission from Elsevier.

FIGURE 7.

Representation of the synthesis of Ni‐MOF@CNT composite by the one‐pot hydrothermal synthesis method. Reproduced from ref [73]. Copyright (2024), with permission from American Chemical Society.

FIGURE 8.

Illustration of the synthesis method of the composite Ni‐MOF/NiCoFe‐PBA. Reproduced from ref [74]. Copyright (2026), with permission from Elsevier.

FIGURE 9.

Yolk‐shell‐structured layered double hydroxide (LDH)@hollow mesoporous carbon spheres (HMCSs), prepared via ion etching of template ZIF‐67@HMCSs. Reproduced from ref [81]. Copyright (2025), with permission from Tsinghua University Press.

2.1.2. Metal Node Modifications

MOFs are composed of organic ligands and metal ions. A carbonaceous substrate with controllable composition, porosity, and morphology can be achieved by carefully selecting the MOFs constituent components. Most widely studied MOFs are transition‐metal‐based, such as Ni‐MOF, Co‐MOF, Mn‐MOF, and Zn‐MOF. Among them, MOFs assembled with Zn nodes have been most widely studied. Park and coworkers first reported Zn‐MOF‐derived carbon for CDI performance in 2014, which showed a high specific surface area of 3000 m² g⁻¹ and a SAC value of 11 mg g⁻¹ in 10 mM NaCl at 1.20 V [94]. Recently, Wei et al. prepared Ag‐based MOFs featuring Ag nodes and organic ligands via an organic linker‐confinement strategy. The coordination interactions allow a homogeneous dispersion of Ag metal ions on the organic substrate at the atomic level, thereby preventing particle aggregation. Ag metal nodes facilitate faradaic conversions, and their desalination mechanism involves the direct reaction of active Ag⁺ with Cl⁻ to form AgCl in an aqueous medium. Simultaneously, Ag⁰ in Ag‐MOF reacts with Cl⁻ to convert to AgCl under an electric field. Therefore, the release and capture of Cl⁻ under an electric field in Ag‐MOF is due to the reversible conversion between AgCl and Ag in the Ag‐MOF structure. The advantages of Ag‐MOF in Cl⁻ capture are attributed to its ordered porous structure, which facilitates ion transport, and to the organic linker's coordination structure with Ag⁺, which prevents Ag agglomeration and enhances stability (Figure 10). Ag‐MOF, when subjected to selective capture of Cl⁻ in an HCDI cell setup, exhibited an improved Cl⁻ removal capacity of 121.52 mg g⁻¹ at 20 mA g⁻¹ with a NaCl concentration of 500 mg L⁻¹, and capacity retention of 96.93% at 50 mA g⁻¹ [84]. One critical concern with transition‐metal‐based MOFs is the risk of these heavy metals leaching into the water stream. To study and address this, Yao et al. prepared hierarchical PC nanofibers derived from electrospun Mg‐MOF‐having CNFs (HCNF/Mg). The use of Mg‐based MOFs provides safety during the desalination process, as Mg is a trace element essential to human health and therefore poses no potential threat to human health. Additionally, using only Mg(CH₃COO)₂ and DMF, the Mg‐MOF synthesis is environmentally friendly and straightforward, mitigating the potential risk of contamination from heavy metals. The developed composite exhibits an electrosorption capacity of 22.10 mg g⁻¹ and a rapid salt adsorption rate of 2.2 mg g⁻¹ min⁻¹ for a NaCl concentration of 0.5 g L⁻¹ at 1.2 V [85]. Akulwar et al. sought to achieve large‐scale electrode material manufacturing with controlled electrode structures to avoid contamination or deactivation, and adopted a one‐step hydrothermal synthesis approach to fabricate the Fe‐MOF/CNT electrode material. The salt‐removal process in Fe‐MOF/CNT involves a faradaic reaction in which Fe(II) ions participate in sodium‐ion intercalation and deintercalation, thereby driving the redox process that underpins the composite's high pseudocapacitance. The Fe‐MOF/CNT electrode material showed an improved salt absorption capacity of 80.575 mg g⁻¹ at 1.6 V. An increase in desalination capacity is attributed to the Fe‐MOFs redox potential, which enhances electrosorption in aqueous solution [86].

FIGURE 10.

(a) Schematic illustration of Ag‐MOF synthesis method. The evolution of (b) Ag‐NPs and (c) Ag‐MOF electrodes during Cl⁻ capture and recycling process. Reproduced from ref [84]. Copyright (2024), with permission from Wiley‐VCH.

2.1.3. Doping

To modify the pore structure of MOFs or MOF‐derived carbon, other elements can be doped to tune the material's surface properties. Element doping with other elements has been shown to enhance the material's hydrophilicity, conductivity, and specific surface area, thereby improving the electrode material's CDI performance. Nitrogen doping is known to strengthen deionization performance by improving electrical conductivity and wettability, and by increasing salt adsorption via faradaic reactions. It also increases pseudocapacitance through metal‐ligand bond interactions between Na⁺ ions and nitrogen, thereby promoting electrosorption. Considering these advantages of nitrogen doping, Yu et al. synthesized nitrogen‐doped PC (N‐PC) by carbonization of a mixture of waste PET‐derived MOFs and ammonium. The developed N‐PC exhibited a large surface area and an optimal meso‐to‐micropore ratio, with a maximum SAC value of 19.9 mg g⁻¹, driven by rapid ion diffusion from mesopores and micropores and by faradaic reactions [87]. Wang et al. successfully synthesized another nitrogen‐doped hierarchical PC by carbonizing a mixture of MOF‐5 and melamine, in which melamine serves as both a nitrogen source and a porogenic agent [88]. Hou and coworkers attempted to synthesize N‐doped carbon using a low‐cost method, as most N‐doped electrodes rely on complex synthetic protocols and costly raw materials. Therefore, a novel low‐cost N‐doped carbon (NGZC) was successfully synthesized through a one‐step in situ growth process using N‐doped graphene quantum dots (N‐GQDs) and ZIF‐8 as starting materials. The N‐doped GQDs increased the mesopore volume and the pyrrolic‐N ratio in NGZC, thereby facilitating charge transfer and pseudocapacitive behavior. NGZC projected a SAC value of 17.86 mg g⁻¹ [89]. Jafari and coworkers studied the impact of implanting a single Fe atom into an N‐doped carbonaceous framework derived from ZIFs, as Fe‐based substrates are considered to have potential due to their abundance, redox properties, adjustable valence states, and cost‐effectiveness. Hence, to study the impact, they synthesized three carbonized products: N—C (nitrogen‐doped carbon), N‐PC with embedded NPFe–N–C (Fe nanoparticles), and an isolated single Fe atom within SAFe–N–C, N‐doped PC structure (Figure 11). The SAFe–N–C electrode exhibited enhanced performance with a salt‐removal capacity of 57.80 mg g⁻¹ in an MCDI setup using a 500 mg L⁻¹ feed NaCl solution at 1.2 V. The improved performance is due to the effectively exposed adsorption sites formed by highly dispersed N and Fe atoms within the carbonaceous matrix, which facilitate electron transport and prevent Fe aggregation, thereby maintaining structural integrity [90]. Prussian blue analogs (PBA) are prone to aggregation and exhibit inherently poor electrical conductivity. The PBA electrode deforms the crystal lattice during continuous intercalation and deintercalation of Na⁺. To address this issue, Li et al. constructed CoNi‐PBA@Co/N‐HC heterostructure by growing cobalt‐doped Ni‐PBA on ZIF‐67‐derived Co/N‐doped hollow PC (Co/N‐HC). The synergistic effect between these two materials yields a composite electrode, CoNi‐PBA@Co/N‐HC, with a convenient Na⁺ diffusion pathway, enhanced redox activity, and improved charge transport in an HCDI cell setup [91]. Wei et al. demonstrated the effect of anchoring ultrafine nanoparticles of Ag in a 3D PC skeleton. 3D Ag@NC composite was fabricated via a MOF‐assisted carbonization replacement strategy with 3D ordered ZIF‐67 as an initial substrate, combined with galvanic replacement reaction, which is a spontaneous redox process governed by the difference in standard electrode potential, i.e., metal with lower reduction potential when it encounters metal ions in solution with higher reduction potential, undergoes spontaneous oxidation and dissolution. At the same time, metal ions are reduced and deposited from the solution. The spontaneous displacement reaction prevents the agglomeration of Ag nanoparticles, which is usually induced by high temperature reduction (Figure 12). The 3D‐ordered Ag@NC showed a dechlorination performance of 105.29 mg g⁻¹, and good cyclic stability of 84.12% after 100 cycles [92]. Liu et al. modified the local electronic structure of ZIF‐8, P atoms hetero‐coordinated, leading to the P—C bond formation via substituting the sp²‐hybridized nitrogen atoms, resulting in enhanced charge transfer efficiency, increased electron density, and enhanced pseudocapacitive behavior. P‐doped ZIF‐8, due to alterations in bond nature, leads to core‐electron entrapment and polarization of nonbonding lone‐pair electrons, thereby facilitating charge transfer and electrosorption and improving electrical conductivity. The P‐doped ZIF shows a desalination capacity of 35.95 mg g⁻¹ in a 500 mg L⁻¹ NaCl solution at 1.2 V [93].

FIGURE 11.

Top and side views of sodium ion adsorbed on (a) single‐layer graphene, (b) nitrogen‐doped single‐layer graphene, and (c) Fe–Nitrogen–doped single‐layer graphene. Top (d) and side (e) views of the differences in charge density of sodium ion adsorbed on different samples (blue areas are electron‐deficient areas, whereas yellow areas are concentrated negative charge areas). Reproduced from ref [90]. Copyright (2025), with permission from Elsevier.

FIGURE 12.

Illustration of synthesis procedure of Ag@NC and 3D Ag@NC. Metal–organic frameworks (MOFs) for heavy‐metal removal using CDI. Reproduced from ref [92]. Copyright (2025), with permission from Elsevier.

2.2. Ion Capture

Design and Fabrication of advanced electrode materials are among the most influential factors affecting deionization performance. As discussed above, for decades, PC materials have been used as electrodes for CDI due to their electrical conductivity and high porosity. But these materials face challenges, including limited desalination capacity and low charge efficiency. Recently, PC derived from MOFs has attracted attention for CDI applications because of its controllable pore sizes, diverse metal‐ligand interactions, and uniform structures, which effectively contribute to the pseudocapacitive behavior originating from the metal nodes and organic linkers present in MOFs, and its porous structure enhances adsorption. However, MOF‐derived carbon faces practical challenges, including a complex synthesis procedure, an uneven pore‐size distribution, poor conductivity, and low structural stability. The earlier section discussed modifications to MOFs, their impact on the final composite, and how these modifications can alleviate challenges. This section discusses recent MOF‐based electrodes designed and fabricated for enhanced electrosorption of specific ions, as the electric field applied in a CDI system enhances ion‐adsorption; however, for selective ion adsorption, it is essential to design and fabricate electrode materials that exhibit ion‐specific adsorption. Therefore, the specificity of electrode materials for enhanced adsorption selectivity, as well as the deionization mechanism underlying this selectivity, is discussed.

2.2.1. Cations

2.2.1.1. Uranium Capture

The rapid growth of the nuclear industry to replace fossil fuels and provide significant electricity without emitting CO₂ creates a need for the extraction and recovery of nuclear fuel, uranium, using suitable technology. Ion exchange, adsorption, chemical precipitation, biological treatment, and membrane separation are among the technologies used to extract U(VI). The above methods, however, have drawbacks, including low adsorption efficiency, poor selectivity, and difficulty in handling residual solvents. CDI is a promising technique for its easy electrode regeneration and low power consumption. To achieve a higher desalination capacity, the fabricated electrode material should possess properties such as high electron conductivity, a larger accessible surface area for ions, and greater electrochemical stability in water. AC, clay minerals, and zeolites have been widely used and have demonstrated their advantages for uranium removal. MOFs with exceptional surface area, porosity, and high thermal and chemical stability have emerged as promising for selective uranium adsorption. MOF‐derived PC is chemically stable during pyrolysis; however, the absence of controlled morphological structure and graded porosity limits its performance. Integration of MOFs with conductive nanostructures, such as graphene, reduced graphene oxide, and CNTs, or modification of MOFs through polymer addition or functionalization with heteroatoms or functional groups can enhance the functionality of the fabricated material. Recently, Hu et al. synthesized a 3D PC carbonized from the composite MOF‐199@PVP/MWCNTs (CMPC) for improved electrosorption of U(VI) from water. In the synthesized hybrid nanocomposite, Cu‐MOF serves as a template for PC, PVP introduces N‐doping and provides additional active sites, and prevents aggregation of MOF‐199, and the integration of MWCNTs into MOF‐199 results in an electron conducting network‐ “MOF‐MWCNTs‐MOF”, connecting adjacent MOF particles for enhanced electrical conductivity for uranium adsorption and stability. The composite CMPC exhibited an outstanding U(VI) removal rate of 95.2%, which stayed over 85.7% even after 5 CDI cycles. To elucidate the possible electrosorption mechanism, XPS spectra were examined. According to the analyzed data, initially, electrostatic attraction between the CMPC electrode surface and UO₂ ²⁺ ion occurs. Later, the abundant active sites present in the CMPC electrode drive UO₂ ²⁺ ions within the material. The addition of PVP and MWCNTs provides additional pseudo‐capacitance, enhancing electrosorption. During electrosorption, applying a voltage reduces the concentration of UO₂ ²⁺ at the electrode and deposits UO₂, thereby enhancing the uranium removal rate [95]. Hu et al. fabricated MOFC‐N (hollow Cu‐MOF‐derived nanoporous carbon) via pyrolysis followed by chemical etching with HNO₃ and applied it as FCDI electrodes. The prepared MOFC‐N has a porous octahedral structure with a concave, rough surface, formed by a simple HNO₃ etching process that removes Cu sites, gradually developing a characteristic open structure. Upon application of an external voltage of 1.2 V, UO₂ ²⁺ cations are attracted to the MOFC‐N electrode surface and electrodeposited readily due to the high reduction potential of UO₂ ²⁺, which is partially reduced at the electrode surface to form complexes; as a result, the removal efficiency of UO₂ ²⁺ is enhanced. The MOFC‐N maintained a 98% removal efficiency of U(VI) after 40^th operational cycles [96]. Liu et al. adapted an environmentally friendly and facile solvothermal method to prepare HKUST‐1 and its composites by incorporating varying concentrations of GO. The composite HKUST‐1/GO showed excellent U(VI) deionization performance of 288.68 mg g⁻¹ at 1.2 V. The composite's electrosorption capacity remained pH‐independent, demonstrating high aqueous stability and acid‐base tolerance. U(VI) adsorption is stronger in the composite, attributed to an increase in oxygen‐containing functional groups that facilitate U(VI) binding to their oxygen atoms [97]. Recently, MXenes have shown potential to enhance the performance of MOF‐based composites for U(VI) adsorption due to their excellent electrical conductivity and mechanical strength, which improve electron transfer and the structural integrity of MOF composites, making them suitable for radioactive metal adsorption. MXenes interact readily with MOFs via functional groups, such as ‐O, ‐F, and ‐OH, which provide active sites for U(VI) adsorption. U(VI) adsorption is further increased by the high surface area provided by the layered structure of MXene. The synergistic effects of MXenes and MOF enhance adsorption capacity and selectivity by enhancing structural stability and exposing more active sites. Li et al. combined UiO‐66‐NH₂ with MXene, which has aerogel‐like properties, via a hydrothermal method followed by freeze‐drying, where TEM images showed that UiO‐66‐NH₂ was uniformly distributed on the MXene surface. The synthesized aerogel‐like MOF has a mesoporous structure, while its gelatinous nature resists degradation and maintains capacity retention of >95% after five adsorption–desorption cycles. The adsorption process of the electrode material is pH‐dependent, which in turn alters uranium speciation. The FTIR spectra of the fabricated electrode after adsorption showed a new U(VI) peak, indicating successful capture of U(VI). Analysis of XPS spectra reveals that the Zr—O bond participates in U(VI) binding, with U (VI) distributed uniformly across the surface [98]. Wan et al. integrated amidoxime‐functionalized CoZn‐MOF with MXene to form an MMA composite, which exhibited an adsorption capacity of 2322.4 mg g⁻¹ and removed 95.6% of ions from a solution containing 500 mg L⁻¹ at 1.2 V within 2 h in a FCDI cell setup. The MMA electrode adsorption efficiency for uranium remained above 90% over a pH range of 3–6, indicating the high stability of the amidoxime‐uranium complex in acidic medium. A 0D–2D heterojunction forms between MXene and MOF, which facilitates electron transfer and enhances the migration of uranyl ions. Further, the strong chelate formed between amidoxime and uranyl ions enhances the adsorption efficiency [99].

2.2.1.2. Copper Ion Capture

Industrial emissions are increasingly contaminated with Cu²⁺ ions, posing a serious threat to human health and the environment. CDI technology has great potential for removing and recovering Cu²⁺ due to its advantages of easy regeneration, no secondary pollution, and low energy consumption. AC, modified AC, reduced graphene oxide, and doped reduced graphene oxide are some of the electrode materials used in CDI for Cu²⁺ removal. However, these materials focused solely on metal‐ion removal and lacked selectivity for Cu²⁺ ions. Therefore, the mechanism and specific interactions between the electrode materials and Cu²⁺ were difficult to understand due to competing ions, which also disrupted adsorption and Cu²⁺ selectivity. Further, MOF‐derived carbon materials showed promising selectivity for Cu²⁺. Recently, Wang et al. fabricated a MOF‐derived carbon loaded with Co(PO₃)₂ (cobalt metaphosphate) ‐NC‐Co(PO₃)₂ by low‐temperature gas/solid phosphating. Co(PO₃)₂ is a cobalt‐based metaphosphate, which, when loaded on MOF‐derived carbon, can increase redox active sites and improve electrochemical performance. The encapsulation of Co(PO₃)₂ inhibits structural collapse during pyrolysis and imparts structural stability, thereby preventing agglomeration and leaching into the electrolyte solution. NC‐Co(PO₃)₂ demonstrated an enhanced electrosorption capacity of 95.41 mg g⁻¹ at 1 V in a 50 mL Cu²⁺ solution because of the binding tendency of metaphosphate groups to Cu²⁺ and interfacial redox reactions induced by the labile valence state of cobalt. A small portion of Co²⁺ is converted to Co³⁺, contributing to the redox‐active nature of the material, providing reactive active sites that facilitate electrosorption by electron transfer and the weak reduction of Cu²⁺ to Cu⁺, and during desorption, Co³⁺ is reduced back to Co²⁺. The P—O bonds and metaphosphate groups easily bind to metals by changing their position, thereby favoring adsorption. The FTIR data show weakened P—O and PO3‐ peaks, which may be attributed to the binding with Cu²⁺, and the reappearance of the P—O bond after desorption was observed, thereby improving CDI performance [100]. For the selective removal of Cu²⁺, Wu et al., by mechanical grinding and pyrolysis method, synthesized S‐Co/MOF, which showed improved electrochemical performance and stability because of the formation of an interconnected conductive network inside the hollow nanofibers. The fabricated electrode showed Cu²⁺ ion selectivity, even at low concentrations. The enhanced selectivity is due to the S element in S‐Co/MOF, which selectively binds Cu²⁺ and precipitates it as CuS on the electrode surface. The MOF enables rapid diffusion and adsorption of Cu²⁺ ions into the material. Density functional theory (DFT) calculations further confirm the high binding energy of Cu²⁺ with S‐Co/MOF. The fabricated electrode exhibited an electrosorption capacity of 57.21 mg g⁻¹ at a 20:10:1 concentration ratio of Na⁺:Cu²⁺:Zn²⁺ [101]. To address the self‐aggregation of electrode materials and the active‐ substance degradation during charge‐discharge cycles due to the Jahn‐Teller effect, Nie et al. proposed an in situ polymerization technique to coat a polypyrrole (PPy) layer on an iron‐based MOF, forming the MIL@PPy composite. The conductive PPy layer reduces self‐aggregation of the composite by mitigating interfacial interactions at the interface. The MIL‐88B(Fe) reduces divalent copper to monovalent copper via charge transfer through iron atoms, thereby MIL‐88B(Fe) adsorbs Cu(I). At the same time, the MIL@PPy electrode adsorbs most of the copper ions by forming a copper chelate with nitrogen atoms on its surface. Therefore, a stable blue phthalocyanine‐like copper chelate forms, enhancing the copper ion adsorption capacity. DFT calculations were employed to study the mechanism of electrosorption, showing that the MIL@PPy composite enhances copper ion adsorption by stably chelating them to conductive PPy molecular chains, forming DNA‐like molecular strips for ion storage across the electrode surface. MIL@polypyrrole‐1 composite demonstrated an adsorption capacity of 228 mg L⁻¹ at a voltage of 1.2 V for a 100 mg L⁻¹ Cu²⁺ concentration [102].

2.2.1.3. Lithium Capture

Lithium recovery from wastewater is pivotal because of its applications in electric vehicles (EVs) and portable electronics such as smartphones, laptops, and tablets. A crucial application of lithium is in the production of lithium‐ion batteries (LIBs), which require high‐purity lithium. The global energy transition has driven up lithium prices in just two years. The low selectivity, complex regeneration process, and high‐energy consumption of conventional methods have driven the search for alternative methods. LFP (LiFePO₄) and LMO (LiMn₂O₄) are the two most used materials for the selective extraction of lithium. LFP exhibits a size‐screening effect that prevents the insertion of larger Na⁺ and Mg²⁺ ions, thereby facilitating Li⁺ transport. However, the Low conductivity and specific capacity of LFP limit its application, as they lead to slower adsorption. LMO exhibits a spinel‐type crystal structure, which promotes rapid Li⁺ ionic transport but also leads to poor cycle stability and lattice distortion due to the Jahn‐Teller effect, limiting its application. MOFs and the integration of MOFs with other materials have shown promise for efficient and selective extraction of Li⁺. For highly selective and efficient extraction of lithium from brine, Xu et al. embedded LiMn₂O₄ nanoparticles in carbon networks derived from MOFs (LMO‐C) by incomplete calcination. The assembled LMO‐C exhibited a clear hollow cavity structure, with LMO nanoparticles encapsulated within the carbon network. LMO‐C, when used as a positive electrode in a hybrid CDI cell, showed an electrosorption capacity of 3.3 mmol g⁻¹, attributed to its hollow‐cavity structure, which elevates the diffusion coefficient of lithium ions and enhances carbon retention in nanoparticles, increasing electronic conductivity and structural stability. The results of the DFT calculation indicate that the insertion of Li⁺ ions into the LMO lattice is more thermodynamically favorable than Mg²⁺. Therefore, LMO‐C has a selective preference for Li⁺. The Raman results indicate the intercalation and deintercalation of Li⁺ ions via a reversible structural conversion in LMO‐C. The assembled LMO‐C//AC system reached separation 24.5 in a solution with a high magnesium‐to‐lithium ratio [103]. Hong et al. conducted molecular dynamics (MD) simulations on three ZIFs, namely ZIF‐8, ZIF‐8‐Cl, and ZIF‐8‐Br, that have a common Zn metal node but differ in the functional groups on the imidazolate linkers to evaluate Li⁺/Na⁺ selectivity. The MD stimulation results indicate the potential to alter functional groups within the ZIF‐8 layer with Br, as ZIF‐8‐Br has the smallest aperture and cavity among the three ZIFs, thereby exhibiting the highest Li⁺/Na⁺ selectivity [104]. He et al., to increase the electrical conductivity, to facilitate ion diffusion, and to provide enhanced active sites for electrosorption, synthesized M‐HITP/LDH/CC, a conductive heterostructured electrode, wherein NiCoMn‐LDH, mixed‐metal layered double hydroxide (LDH), was first grown on CC, subsequently coordinated with HITP, a N‐rich ligand, to fabricate a 2D conductive MOF. The fabricated M‐HITP/LDH/CC composite electrodes exhibit an electrosorption capacity of 59.08 mg g⁻¹, attributed to the hierarchical architecture that facilitates lithium‐ion diffusion, and to the integration of Mn, Ni, and Co into the LDH framework, which enhances pseudocapacitive characteristics. At the same time, the N‐rich ligand, HITP, provides additional redox sites and lithium‐specific sites. A superior selectivity for lithium ions, with a separation coefficient of 15.1, was observed in the presence of competing ions. DFT calculations show that electrons are depleted from the LDH substrate and accumulate at C–N–M ring structures within HITP, demonstrating charge transfer from LDH to M‐HITP. The migration energy barrier for Li⁺ in M‐HITP/LDH is lower than that in NiCoMn‐LDH, facilitating rapid intercalation of Li⁺ ions. The N group in HITP increases the nucleophilic character of the M‐HITP/LDH electrode, thereby enhancing electrostatic attraction. The nucleophilic regions enhance the material's ion affinity. The excellent coordination affinity of the N group for Li⁺ and compatibility of the Li⁺ ion and M‐N₄ further enhance the Li⁺ ion selectivity of the electrode [105]. Li et al. developed a ternary composite by combining Fe‐tetrahydroxybenzoquinone (Fe‐THBQ), a conductive MOF, along with the MoSe₂@MXene framework. The c‐MOFs provide enhanced active sites for lithium‐ion capture; simultaneously, MoSe₂@MXene enhances ion transport and provides structural stability. Thereby demonstrating a 4.05 mmol g⁻¹ of lithium capture capacity, attributed to the layered morphology of MoSe₂, providing channels for rapid diffusion of lithium ions, while Fe‐THBQ enhances lithium adsorption through reversible redox reaction, and Fe‐THBQ and MoSe₂@MXene heterointerface form coordination bonds of Fe—Se, promoting charge transport. DFT studies show the mechanism of selective Li⁺ removal. The electron density is concentrated at the component of Fe‐THBQ@MoSe₂@MXene electrode interface, suggesting strong electron coupling at this interface. The electronegativity of MXene creates a unidirectional electron transfer pathway (Fe‐THBQ $\to$ MoSe₂ $\to$ MXene), where Fe‐THBQ loses an electron, and MXene gains an electron. This directional mechanism of charge transfer enhances electron transfer and conductivity, thereby significantly improving Li⁺ adsorption. Li⁺ shows the lowest migration energy barrier in calculations of migration energy for different cations, suggesting Li⁺ has a faster transport rate in the presence of other competitive cations [106]. To address the issue of low electrosorption capacity, poor regeneration, and low ion transport of conventional LiAl‐LDH, He et al. prepared a Zn‐doped LDH to fabricate LiZnAl‐LDH on CC, followed by HHTP (2,3,6,7,10,11‐hexahydroxytriphenyl) coordination to develop Zn‐HHTP/LDH heterostructured electrode, which demonstrated π‐d conjugation at the interface, which improves charge transport and increases diffusion of lithium ion with an electrosorption capacity of 50.25 mg g⁻¹. The incorporation of HHTP via Zn²⁺ coordination forms a 2D honeycomb conductive MOF, enabling LDH to provide a conductive pathway for lithium‐ion transport. To understand the selective adsorption performance of the fabricated electrode, DFT calculations were performed. The optimized structure model of the fabricated electrode indicates LDH‐to‐Zn‐HHTP charge transfer, with electron depletion in LDH and electron accumulation in the C–O–Zn ring of HHTP, suggesting enhanced electron transport, increased electrode surface reactivity, and reduced charge transfer resistance, thereby accelerating reversible adsorption/desorption of lithium ions. The calculated adsorption energy of lithium ions in Zn‐HHTP/LDH is significantly higher than that of competing cations, suggesting a stronger interaction of lithium with the metal‐oxygen sites of the electrode and thereby enhancing lithium selectivity. Therefore, the synergistic effects of a reduced migration barrier for lithium ions, higher adsorption energy, and stronger coordination with the Zn‐O₄ site contribute to superior lithium selectivity [107].

2.2.2. Anions

2.2.2.1. Fluoride Capture

Fluoride is a highly harmful pollutant and extremely difficult to eliminate. Fertilizer production, metal processing, semiconductor manufacturing, and glassmaking industries are the major sources of fluoride contamination in wastewater from these industries. Traditional techniques used for fluoride removal from wastewater are chemical precipitation, coagulation, ion exchange, membrane separation, electrodialysis, and adsorption. Common adsorbents for defluorination include zeolites, metal oxides, AC, and hydroxyapatite. These conventional techniques exhibit poor selectivity and limited functionality for specific groups. In contrast, an electric field enhances the adsorption efficiency of fluoride ions in a CDI cell setup compared to conventional adsorption. Therefore, CDI technology for defluorination applications holds promise. Commonly utilized carbon‐based CDI electrode materials, such as carbon fibers, CAs, and CNTs, lack specific sites for fluoride binding, thereby decreasing adsorption capacity and selectivity. The special layered structure of LDHs and MXenes also promises high fluoride removal efficiency with additional morphological modifications to enhance their stability in aqueous solution. Recently, Li and coworkers developed an amino‐functionalized lanthanum‐based organic framework‐derived carbon electrode via a solvothermal method, followed by carbonization, and studied the amino functionalization on the LaBDC/C‐NH₂ electrode surface, which becomes protonated due to the abundant H⁺ ions in water, leading to electrostatic attraction between F⁻ ions and protonated NH₃ ⁺ on the electrode surface, thereby facilitating selective F⁻ adsorption. F⁻ ions replace OH⁻ from the hydrated lanthanum oxide (La(OH)₃.nH₂O), forming a new La–F ligand. Therefore, the LaBDC/C‐NH₂ electrode has a stronger affinity for F‐ because of the complexation occurring intracellularly between fluoride and the electrode. The developed LaBDC/C‐NH₂ electrodes exhibited adsorption of 73.8 mg g⁻¹ at a constant current of 2 A m⁻², with removal performance for F⁻ ranging from 8 to 0.17 mg L⁻¹ [108]. Qiu et al. fabricated a Ti/N MOF‐derived PC for defluorination via a solvothermal method followed by carbonization. The synthesized TiNC after the carbonization step has abundant micropores and mesopores along with abundant active sites. Nitrogen‐doping enhances hydrophilicity and thereby yields an adsorption capacity of 33.99 mg g⁻¹ at 1.2 V with an initial fluoride concentration of 50 mg L⁻¹ [109]. The XPS spectra of TiNC before and after adsorption help elucidate the possible defluorination mechanism involved. TiNC forms hydroxy titanium compounds TiO(OH)₂ as a consequence of hydrolysis and hydroxylation. F⁻ is readily adsorbed because it shows a greater affinity for TiO^2+. Upon adsorption, the ion‐exchange reaction between F⁻ and OH⁻ occurs, forming a Ti—F bond. Therefore, the high adsorption capacity of the TiNC electrode is attributed to the enhanced F⁻ migration to the fabricated electrode surface under applied voltage. Kang et al. synthesized a spherical, BMOF, Ce/Zn‐BTC, via a solvothermal method followed by pyrolysis, and transformed it into a CeO2@C composite that preserves the spherical structure of the original MOF, leading to enhanced pore volume and specific surface area and thereby improved electrochemical performance. The composite was fabricated in an HCDI cell setup, with CeO₂@C as anode material and birnessite‐type MnO₂ as cathode material. The F⁻ removal occurs via electrostatic attraction followed by ion exchange, which replaces surface OH⁻ groups with F⁻. The defluorination process first involves activation of Ce to Ce—OH in solution, followed by fluoride capture via hydrogen bonding. Fluoride ions then, through electron transfer or sharing, undergo ion exchange with ‐OH to form Ce—F, thereby facilitating defluorination. The fabricated HCDI cell setup exhibited a fluoride removal efficiency of 22.03 mg g⁻¹ for a concentration of 50 mg L⁻¹ fluoride solution at 1.2 V [110]. Xiao et al. synthesized bead‐chain‐like nanotube MOF‐801@PPy via the solvothermal method, in which nanotubes of polypyrrole act as chains to connect particles of MOF‐801 and thereby improve the hydrophilicity and conductivity of the composite. The presence of ‐COOH, ‐OH, and other functional groups in MOF‐801 facilitates rapid adsorption of water molecules via a synergistic effect of hydrogen bonds and van der Waals forces. XPS data further validate the replacement of the ‐OH group on Zr—OH by F⁻, as evidenced by a relative decrease in Zr—OH content postadsorption. DFT calculations were studied to elucidate the insertion and extraction process of fluoride ions in the MOF‐801@PPyNT electrode. DFT studies demonstrated higher electron‐accepting density in MOF‐801 and higher electron‐donating density in polypyrrole at the interface than in other regions, suggesting a more efficient charge‐transfer pathway. The binding energy calculations for the metal cluster of MOF‐801 and different ions showed an exceptionally high binding energy with F⁻, demonstrating high electrode selectivity and thereby efficient fluoride removal [111].

2.2.2.2. Phosphorus Capture

High levels of phosphorus are the primary drivers of eutrophication and algal blooms, thereby threatening the sustainability of aquatic ecosystems. Phosphorus is primarily present as an anionic form in solution (H₂PO₄ ⁻, HPO₄ ²⁻, and PO₄ ³⁻). Difficulties with the disposal of massive sludge, adsorbent regeneration, and stringent environmental conditions are among the challenges faced in conventional phosphorus removal technology; therefore, there is a need to develop more effective and eco‐friendly phosphorus elimination technology. CDI, owing to its low cost, nonchemical reagent requirements, and environmental friendliness, has attracted considerable attention as a promising electrochemical technology for phosphorus removal. Conventional PC, known for its excellent conductivity and high specific surface area, has limited active sites and disordered pores, thereby limiting its applicability for phosphorus adsorption. However, metal/carbon composites have shown potential for phosphorus uptake due to their abundant metal active sites and ordered structure, making MOF‐derived carbon and MOF composites efficient electrode materials for phosphorus adsorption. To introduce N doping and form a stable carbon skeleton, Chen et al. fabricated N@Zr/C, a MOF/biomass‐derived zirconium‐nitrogen codoped carbon via carbonization of Zr‐MOFs and bagasse. Nitrogen‐groups supported on bagasse and Zr‐MOFs provide phosphorus‐affinity sites, thereby enhancing phosphorus removal. N@Zr/C composite, when applied for phosphorus removal, as an anode under 1.2 V, demonstrated an electrosorption capacity of 32.30 mg g⁻¹. The FTIR and XPS results suggest substitution of surface hydroxyl groups bonded to Zr by HPO₄ ²⁻ via ligand exchange reaction and forming monodentate or bidentate Zr—O—P inner sphere complex. Under the influence of an electric field, surface hydroxyl groups (‐OH) are converted to ‐OH₂ ⁺, which can be readily substituted from the active metal site. N‐doping facilitates phosphorus trapping by modulating the metal‐oxide reaction center. Therefore, the formation of ‐OH₂ ⁺ and N groups facilitates the ligand‐exchange process [112]. A novel composite, LDH@MOF/PANI, was successfully fabricated by Song et al. using polyaniline (PANI) as a conductive substrate and MOF as a self‐sacrificial precursor for LDH, thereby improving the pore structure and enriching phosphorus‐binding sites. The LDH@MOF/PANI, when applied as an electrode for CDI, showed superior adsorption capacity. The superior adsorption capacity can be attributed to ligand exchange, anion exchange, hydrogen bonding, and electrostatic attraction. The XPS and FTIR results confirm that phosphorus adsorption occurs through chemisorption. Under the influence of the electric field, the ‐OH groups on the electrode surface are protonated to ‐OH₂ ⁺. The formation of ‐OH₂ ⁺ facilitates ligand exchange at phosphorus, forming an M—O—P inner‐sphere complex. The phosphorus further undergoes anion exchange with the LDH component, leading to its insertion into the LDH interlayer in LDH@MOF/PANI. The ‐N=/‐NH‐ group of PANI also becomes protonated under the influence of the electric field, which further contributes to the electrostatic attraction of phosphorus by binding to the protonated amino group via hydrogen bonding [113]. Wan et al. studied the effect of gadolinium doping on MIL‐101‐NH₂ and its efficiency in the electrosorption of phosphorus and arsenic(V). Therefore, GF‐MILN (Gd‐doped Fe‐MOF) compounds were synthesized by a simple solvothermal method, in which Gd doping optimizes Fe‐MOF electronic structure and provides oxygen vacancies to enhance electrochemical activity. 0.75GF‐MILN demonstrated highest performance for As(V) and P removal. The XPS and FTIR results of the fabricated electrode before and after adsorption of As(V) and P indicate As/P—O—Fe/Gd bond formation, implying Gd—OH and Fe—OH groups are involved in the bond formation, and the hydroxyl groups are replaced by As(V) or P via ligand exchange during the adsorption process. EPR analysis showed that Gd doping increased the number of adsorption sites by increasing the number of unsaturated coordination bonds [114]. Bi et al. incorporated PEDOT (poly(3,4‐ethylenedioxythiophene)) into an yttrium‐based MOF consisting of amino linkers (YMOF‐NH₂). The theoretical calculation showed that incorporating PEDOT reduced oxygen vacancies and increased the density of states near the Fermi level, thereby improving structural stability and electronic conductivity. The theoretical calculations show strong covalent bonding between the yttrium metal sites and HPO₄ ²⁻. Initially, through weak coordination bonds, HPO₄ ²⁻ ions get bonded to the surface of PEDOT, forming PEDOT‐‐‐‐‐PO₄H, and later bind with the amino group of YMOF‐NH2 through hydrogen bonds. These weak interactions help in the initial adsorption of phosphorus. Further, through intermolecular interactions and thermodynamic forces, the HPO₄ ²⁻ ion binds to the yttrium metal site, forming a strong coordination bond and a stable Y—PO₄H bond. The adsorption mechanism suggests that PEDOT provides additional active sites and a stable electronic structure, thereby optimizing phosphate adsorption [115]. Ding et al. synthesized a layered nanoscale MOF via a simple hydrothermal method and integrated it with CNTs to form electron channels that facilitate N/P removal. Zr‐BTB@CNTs (ZB@C) was developed and applied for capacitive removal as an anode for phosphate and nitrate. It showed NO₃ ⁻ electrosorption capacity of 222.1 and 159.9 mg g⁻¹ for PO₄ ³⁻ in 120 min at a voltage of 1.2 V. To understand the mechanism of electrosorption of nitrate and phosphate, a comprehensive analysis of XPS, FTIR, and DFT simulations was conducted. The DFT analysis showed coordination of P to the Zr atom and the formation of the Zr—O—P bond, as evidenced by a depletion of electron density at the Zr center, indicating the electron‐withdrawing effect of phosphate. XPS analysis revealed the formation of a stable Zr—O—P inner sphere complex via ligand exchange between surface hydroxyl groups and phosphate ions. FTIR spectra of the electrode before and after nitrate electrosorption revealed unique structural changes, including the exposure of abundant uncoordinated Zr⁴⁺ sites, which facilitate inner‐sphere complex formation via ligand exchange, leading to the formation of a new coordination bond between the oxygen atoms of NO₃ ⁻ and Zr⁴⁺ [116].

To conclude, this section highlights the unique advantages of MOFs for the design and fabrication of ion‐specific electrode materials, the impact of modifying MOFs, and their potential to enhance adsorption selectivity. Also, the mechanism underlying the enhanced adsorption selectivity is discussed, and the roles of FTIR, XPS, and DFT studies in elucidating it are highlighted. Table 4 summarizes the recent MOF‐based electrodes fabricated for ion capture, the mechanisms underlying them, and their CDI performance.

TABLE 4.

Summary of recent MOF‐based electrode materials for ion capture, modifications, underlying mechanism, and their performance.

	Fabricated Electrode	MOF material	Modification	Mechanism of adsorption	Initial concentration, mg L−1	Electrosorption capacity (mg g−1) @ Voltage (V)	Reference
Cations
Uranium	MOF‐199@PVP/MWCNTs (CMPC)	MOF‐199 (Cu)	Introduction of PVP and MWCNTs	Enhanced reduction of UO2 2+ with PVP and MWCNTs providing additional pseudocapacitance	60	410.3 @ 1.2	[95]
	MOFC‐N	MOF‐199 (Cu)	MOF derived carbon subjected to chemical etching with HNO3	Chemical etching develops open structure leading to enhanced UO2 2+ reduction	240	ASRR = 17.45 µg cm−2 min−1 @ 1.2	[96]
	HKUST‐1/GO	HKUST‐1 (Cu‐MOF)	GO incorporation into HKUST‐1	GO incorporation increases oxygen containing functional group thereby facilities UO2 2+ reduction	100	164.99 @1.2	[97]
	UiO‐66‐NH2@MXene (U‐M)	Zr‐MOF	UiO‐66‐NH2 uniformly distributed on MXene surface	The Zr—O bond of UiO‐66‐NH2 enhance binding of uranium.		229.1	[98]
	MMA	CoZn‐MOF	Amidoxime‐functionalized CoZn‐MOF was integrated to MXene	Enhanced electron transfer and strong chelate formation between amidoxime and uranyl ions	500	2322.4 @ 1.2	[99]
Copper	NC‐Co(PO3)2	ZIF‐8 (Zn)	Co(PO3)2 loaded to MOF derived carbon	Strong binding of Cu2+ to metaphosphate group thereby enhanced electrosorption	50	95.41 @ 1	[100]
	S‐Co/MOF	ZIF‐67 (Co)	Co‐MOF combined with thiourea	S‐selectively bind to Cu2+ and precipitates as CuS on electrode surface	25	55.39 @ 1	[101]
	MIL@PPy	MIL‐88B (Fe)	Polypyrrole coated to MIL‐88B	Cu2+ binds to N‐atom of PPy chain forming a stable chelate	100	215 @ 1.2	[102]
Lithium	LMO‐C//AC	Mn‐MOF	LiMn2O4 nanoparticles embedded in MOF‐derived carbon	Intercalation and deintercalation via structural conversion of LMO‐C	28.57 mmol/L	4.71 mol/g @ 1.2	[103]
	M‐HITP/LDH/CC	Conductive MOF(C‐MOFs)	Coordination of N‐rich ligand‐HITP to mixed metal LDH	Enhanced charge transfer from LDH to HITP and increased nucleophilic character due to N‐rich ligand, thereby enhancing electrostatic attraction	423	59.08 @ 1.4	[105]
	Fe‐THBQ@MoSe2@MXene	Fe‐THBQ (C‐MOF)	Integration of C‐MOF with MoSe2@MXene framework	Enhanced electron transfer pathway (FeTHBQ $\to$ MoSe2 $\to$ MXene)	10 mmol/L	4.05 mmol/g	[106]
	Zn‐HHTP/LDH/CC	Zn‐HHTP (C‐MOF)	Integration of Zn‐HHTP with LDH	Enhanced charge transfer from LDH to Zn‐HHTP	10 mmol/L	50.25 @ 1.2	[107]
Anions
Phosphorus	N@Zr/C	Zr‐MOF	N‐doping	Adsorption via ligand exchange and formation of Zr‐O‐P inner sphere complex with N‐doping facilitates P trapping	5	5.35 @ 1.2	[112]
	LDH@MOF/PANI	Al‐MOF	Introduction of LDH and PANI	Adsorption via ligand exchange, anion exchange, hydrogen bonding and electrostatic attraction	1	0.126 @ 1.2	[113]
	GF‐MILN	MIL‐101 (Fe‐MOF	Gd‐doping	Hydroxy groups of Gd and Fe replaced by P and As via ligand exchange.		As: 220.7 P: 112.80	[114]
	PEDOT/Y‐MOF‐NH2	Y‐MOF	PEDOT incorporation into yttrium MOF with amino linkers	Adsorption via formation of stable Y‐PO4H coordinate bond		14.12 @ 1.2	[115]
	ZB@C	Zr‐MOF	Integration of CNTs	Formation of Zr‐O‐P inner sphere complex via ligand exchange	500	P: 159.9 NO3 −: 222.1	[116]
Fluoride	LaBDC/C‐NH2	Lanthanum‐based MOF	Lanthanum‐based MOF functionalized with amino group and carbonized	Amino functionalization increases electrostatic attraction of F− ions, facilitating selective adsorption	100	73.2 @ 2.2 Am−2	[108]
	TiNC	MIL‐125 (Ti)	MIL‐125 functionalized with NH2 and carbonized.	Amino functionalization leads to enhanced adsorption.	50	33.99 @ 1.2	[109]
	CeO2@C/MnO2	Ce/Zn bimetallic MOF	CeO2@C used as anode and birnessite‐type MnO2 as cathode in HCDI setup	F− ions are captured via electrostatic attraction followed by ion exchange	50	22.03 @ 1.2	[110]
	MOF‐801@PPy	MOF‐801	Polypyrrole nanotubes introduced to MOF‐801	Zr—F bond formed via ligand exchange and enhanced charge transfer due to incorporation of PPy	50	22.92 @ 1.4	[111]

3. Challenges and Future Perspective

MOFs, MOF‐derived carbon, and MOF‐based composites have been shown to be promising electrode materials for CDI in water treatment, owing to their tunable pore structures, high surface areas, and functionalizability. However, they face several technical and practical challenges, which include

• 1. Water stability of MOFs and chemical durability: Hydrolysis of metal‐ligand bonds can degrade MOFs in water. Saline environments with high ionic strength and extreme pH conditions can accelerate structural collapse of MOF structures. Chemical durability of MOF structures is also affected by deionization through repeated redox reactions, as these reactions can generate hydroxyl radicals or form oxides during the oxygen‐water redox process, leading to MOF degradation. Therefore, more comprehensive research is needed to improve selective deionization and remove target contaminants, as competing ions can reduce the structural stability of MOFs and their adsorption efficiency. Selective deionization will be achieved by modifying the MOF surface properties and pore sizes. One of the major challenges hindering the practical application of MOFs is their stability, as continuous operation requires long‐term cycling stability, which many MOFs cannot yet provide. The choice of hydrophobic ligands and high‐valence metals during MOF synthesis prevents water penetration. Hydrophobic ligands protect MOFs from hydrolysis, and high‐valence metals can form stronger bonds with ions, thereby imparting water stability [117]. Some reported MOFs that can be used in harsh environments include MIL‐101(Cr), which remains stable in strong acid and base solutions. MIL‐101 (Fe) and MIL‐101‐Fe‐NH₂ show water, acid, and base stability. In the UiO series, UiO‐66 remains stable in water, acid‐base, and also stable in various organic solvents. UiO‐67 shows excellent acid‐base stability along with adjustable hydrophilicity and hydrophobicity. In the BUT series, BUT‐17 shows strong stability in strong acids and bases, as well as in boiling water. BUT‐12‐15 are among the most acid‐stable MOFs. BUT‐8(Cr) is one of the most acid‐stable MOFs reported [118, 119, 120, 121, 122].
• 2. Limited electrical conductivity in MOFs: Most MOFs exhibit poor electrical conductivity, leading to slow ion transport and low charge efficiency. Much research has focused on increasing the electrical conductivity of MOFs by fabricating composites such as MOF/CNT, MOF‐derived carbons, and other hybrids. However, much of this work increases cost and complexity. Conductive MOFs exhibit graphene‐like properties, with strong π–π stacking and in‐plane conductivity. However, they are not as conductive as graphene. c‐MOFs have not been explored much for CDI applications. Research on modifying c‐MOFs by doping or combining them with other 2D materials, such as MXenes, holds promise [123]. Some ways to increase the electrical conductivity of MOFs include choosing metal ions or ligands with multiple valencies and metal‐ligand orbital overlap. Multiple valencies can facilitate interconvertible oxidation states to form a redox‐active framework, and good overlap can increase charge delocalization. 2,5‐dihydroxybenzoquinone is one of the highly explored redox‐active ligands, and can show a valence tautomerism up to three different valence states. The frontier orbitals of the ligand have energies similar to those of many transition metals. Additionally, the use of planar, highly conjugated organic ligands such as triphenylene, phthalocyanine, naphthalocyanine, etc. can enhance interlayer π‐π interactions, thereby favoring charge transport. Further, encapsulation of electroactive molecules such as TCNQ (7,7,8,8‐tetracyanoquinodimethane), iodine, and methyl viologen within the MOF framework can create a pathway for electron transport and modulate the electrical conductivity of the parent MOF [124].
• 3. Electrode regeneration and reusability: Repeated regeneration cycles may degrade the MOF structure and reduce adsorption capacity, as the strong interfacial interactions and Jahn‐Teller effect lead to lattice distortions, causing MOF nanoparticles to aggregate, leading to uneven exposure of active sites, leaching of metal centers, and reduced charge transfer, thereby significantly affecting the longevity of the electrode and increasing the chance of dissolution. Therefore, modifying MOFs with protective polymer layers or functionalizing them to mitigate volume changes caused by frequent adsorption and desorption cycles needs further research. Therefore, Future studies should focus on understanding electrode regeneration, as very little attention has been given to it when the electric field is removed or reversed [125].
• 4. Scalability issues: Most MOFs and their composites, as well as MOF‐derived carbons, are fabricated under controlled conditions. For practical CDI applications, large‐scale synthesis becomes difficult due to the use of expensive organic linkers and toxic solvents like DMF for solvothermal synthesis, with a long reaction time required, which hinders large‐scale synthesis of MOFs with postprocessing modifications needed for activation and improved performance of MOFs, adds to the cost, and makes it economically challenging.
• 5. Environmental concerns: Continuous operation and long‐term cycling of MOF‐based electrodes can lead to potential metal ion leaching, resulting in secondary contamination, as most MOFs can lose their structural integrity and morphology during repeated cycles of charging and discharging, attributed to their low mechanical strength, leading to structural collapse and thereby disrupting the electron transport pathway. Synthesizing MOFs with high‐valence metal nodes (Al³⁺, Zr⁴⁺, Ti⁴⁺) can improve structural integrity by forming stronger bonds with ions. Choosing metals and ligands with matching softness or hardness can enhance bond strength and water stability [126]. Additionally, the use of toxic solvents during MOF synthesis raises safety concerns for practical applications. Therefore, MOF synthesis using greener synthesis methods with water or ethanol as solvents should be considered for future studies.
• 6. Fouling issues: Natural organic matter and suspended solids in wastewater or seawater can clog MOF pores, thereby reducing adsorption and permeability. Most of the research studies are based on synthetic solutions prepared in the laboratory and do not use real wastewater or seawater. Hence, the effect of fouling on electrode materials is rarely discussed or studied. Therefore, incorporating real contaminated water or seawater to study the performance of fabricated electrodes is necessary. Immobilizing MOFs on a hydrophilic substrate, introducing hydrophilic groups, or properly optimizing MOF pore size to exclude large foulants are a few ways to improve MOF antifouling.
• 7. Challenges in Structural optimization: Fabricating novel hybrid materials with precise control and balance of pore size, surface charge, hydrophilicity, and new chemical and physical functionalities remains difficult, as fabricating thermodynamically stable materials, and achieving atomic precision in the hybridization of MOF components are essential. Therefore, developing novel materials by combining different materials to achieve a balance of stability, conductivity, and adsorption can provide CDI with a promising future.
• 8. Computational studies: Computational methods, such as DFT, MD simulation, and continuum models, can enable the optimization of materials, the understanding of mechanisms, the simulation of ion movement inside nanopores under different concentrations, and the study of competitive adsorption in mixed‐ion systems. Hence, proper utilization of computational models can define the next generation of CDI research.

The field of CDI can evolve from a laboratory technique into a next‐generation, energy‐efficient water‐treatment method by fabricating advanced materials capable of selective ion removal. Materials such as MXenes capable of fast ion transport, covalent organic frameworks (COFs), conductive MOFs, ion‐intercalation materials (battery type), heteroatom‐doped carbons, and materials capable of providing hierarchical porous structures need further research and development. Most CDI technologies are based on classical EDL; future CDI systems need to be developed beyond classical EDL to treat higher‐salinity water, operate continuously, and reduce energy consumption. Efforts to integrate CDI technologies with other water treatment techniques, such as RO, electrodialysis, and adsorption, also hold promise. Therefore, with the combined efforts from academia and industry, further improvements in CDI technology can be achieved, enabling large‐scale water deionization at low cost.

Author Contributions

Bency Jose: writing original draft, software, and editing. Sib Sankar Mal: conceptualization, validation, writing – reviewing, and editing.

Funding

This work was supported by Alexander von Humboldt‐Stiftung.

Conflicts of Interest

The authors declare no conflicts of interest.

Biographies

Bency Jose is currently a Ph.D. candidate under the supervision of Dr. Sib Sankar Mal in the department of chemistry at the National Institute of Technology, Karnataka. She hails from Mumbai, and completed her bachelor's degree from St. Xavier's College, Mumbai, and her master's degree from Sophia College, Mumbai. Her main research interests are in the design and synthesis of redox‐active electrode materials based on polyoxometalates, metal–organic frameworks, and MXenes, and in exploring their applications.

Sib Sankar Mal received his Ph.D. in chemistry from Jacobs University, Bremen, in 2008, under the supervision of Prof. Ulrich Kortz. He then completed postdoctoral work at the University of Ottawa, Canada, in 2011, and subsequently moved to the University of Hamburg, Germany, as an Alexander von Humboldt postdoctoral fellow. After completing his postdoctoral work, he joined the National Institute of Technology, Karnataka, India, as an assistant professor in 2013, where he currently serves as an associate professor in the department of chemistry. His primary research areas are energy storage, energy conversion, renewable energy, electrochemistry, catalysis, and polyoxometalates. He has published over 100 peer‐reviewed papers (H‐index: 32) and edited books. He has been awarded the Early Career Research Award by the Science and Engineering Research Board (SERB), India, and was recently honored with the Alexander von Humboldt Research Group Linkage Grant from Germany.