High-Performance Crown Ether-Modified Membranes for Selective Lithium Recovery from High Na⁺ and Mg²⁺ Brines Using Electrodialysis

Abstract

The challenge of efficiently extracting Li⁺ from brines with high Na⁺ or Mg²⁺ concentrations has led to extensive research on developing highly selective separation membranes for electrodialysis. Various studies have demonstrated that nanofiltration membranes or adsorbents modified with crown ethers (CEs) such as 2-OH-12-crown-4-ether (12CE), 2-OH-18-crown-6-ether (18CE), and 2-OH-15-crown-5-ether (15CE) show selectivity for Li⁺ in brines. This study aims to develop high-performance cation exchange membranes (CEMs) using CEs to enhance Li⁺ selectivity and to compare the performance of various CE-modified membranes for selective electrodialysis. The novel CEM (CR671) was modified with 12CE, 18CE, and 15CE to identify the optimal CE for efficient Li⁺ recovery during brine electrodialysis. The modification process included polydopamine (PDA) treatment and the deposition of polyethyleneimine (PEI) complexes with the different CEs via hydrogen bonding. Interfacial polymerization with 1,3,5-benzenetricarbonyl trichloride-crosslinked PEI was used to create specific channels for Li⁺ transport within the modified membranes (12CE/CR671, 15CE/CR671, and 18CE/CR671). The successful application of CE coatings and Li⁺ selectivity of the modified membranes were verified through Fourier-transform infrared spectroscopy, zeta-potential measurements, and electrochemical impedance spectroscopy. Bench-scale electrodialysis tests showed significant improvements in permselectivity and Li⁺ flux for all three modified membranes. In brines with high Na⁺ and Mg²⁺ concentrations, the 15CE/CR671 membrane demonstrated more significant improvements in permselectivity compared to the 12CE/CR671 (3.3-fold and 1.7-fold) and the 18CE/CR671 (2.4-fold and 2.6-fold) membranes at current densities of 2.3 mA/cm² and 2.2 mA/cm², respectively. At higher current densities of 14.7 mA/cm² in Mg²⁺-rich brine and 15.9 mA/cm² in Na⁺-rich brine, the 15CE/CR671 membrane showed greater improvements in Li⁺ flux, approximately 2.1-fold and 2.3-fold, and 3.2-fold and 3.4-fold compared to the 12CE/CR671 and 18CE/CR671 membranes. This study underscores the superior performance of 15CE-modified membranes for efficient Li⁺ recovery with low energy demand and offers valuable insights for advancing electrodialysis processes in challenging brine environments.

1. Introduction

Lithium, an essential component in the evolving clean energy sector, has garnered increasing attention due to its crucial role in powering electric vehicles and driving advancements in battery technologies [1]. Beyond its application in battery production, lithium holds significance in varied industries, including lubricating grease production, primary aluminum, glass, and ceramics, owing to its diverse and unique properties [2,3]. The increased demand for lithium stems from the fast growth of portable electric vehicles and electronic devices, positioning lithium as the pivotal “energy element” of the 21st century [4,5]. In 2023, global lithium consumption was estimated at 180,000 tons, marking a 27% rise from the 142,000 tons consumed in 2022, driven by robust demand from the lithium-ion battery market. The average annual U.S. lithium carbonate price in 2023 reached $46,000 per ton [6], higher than that in 2022 ($37,000 per ton) [7]. The escalating global demand for lithium has underscored the critical need for lithium supply across Asia, Europe, and North America [6]. There has been a concerted effort to establish secure and diversified supply chains to address the increasing demand. In the U.S., identified lithium resources have substantially increased to 14 million tons [6], with brines constituting the major lithium resource (about 70–80% of global lithium deposits) [8]. Therefore, developing effective technologies for selective lithium recovery from brines provides an essential solution.

Electrodialysis [9,10] is a promising technology to recover lithium resourced from brines due to no additional chemicals used and simplicity in operation as compared to other technologies such as chemical precipitation [11], adsorption [12], nanofiltration [13], and solvent extraction [14]. The traditional precipitation–crystallization method is not suitable for brine treatment due to co-precipitation of other elements, affecting lithium purity [15]. The adsorption method has been explored as a cost-effective and environmentally friendly technology for extracting Li⁺ from brines with a high Mg²⁺/Li⁺ ratio. Zhong et al. [16] prepared lithium aluminum double hydroxides (Li/Al-LDHs) to recover lithium from brine with a high Mg²⁺/Li⁺ ratio, showing that the Mg²⁺/Li⁺ ratio was significantly reduced from 301.58 in the original brine to 0.99 in the strip solution. However, ion-sieve adsorption materials typically show low permeability and high dissolution rates in acidic environments [17]. Nanofiltration membranes have also been developed to recover Li⁺ from brines. Li et al. [18] used commercial nanofiltration membranes to separate Li+ from brine with a high Mg/Li ratio, demonstrating selectivity for Li⁺ and 83% Li recovery efficiency. This approach involves high energy consumption, significant capital investment, and the potential for severe membrane fouling and scaling [17]. Solvent extraction generates large volumes of waste that need to be disposed of, making it less environmentally friendly [19].

Among these methods, electrodialysis exhibits distinct advantages in diverse applications, including salt production, desalination, resource recovery from brines, and industrial and municipal wastewater treatment [20]. In addition to its application in water desalination, electrodialysis can be adapted for Li⁺ recovery from brines or geothermal fluids using selective CEM [21]. However, it is challenging to develop high-performing, advanced ion exchange membranes (IEMs) to separate Li⁺ effectively. In brines, Li⁺ typically coexists with other competing ions, such as Mg²⁺ and Na⁺, at significantly higher concentrations than Li⁺ [22,23]. Standard membranes encounter challenges in achieving the selective extraction of Li⁺, particularly due to similar chemical properties of co-existing ions Mg²⁺ and Na⁺ to Li⁺ [24]. Despite the development of membranes with high separation coefficients in Mg²⁺/Li⁺ separation, challenges persist in separating Li⁺ from Na-rich brine, where the separation coefficient diminishes to values below 1 as the concentration of Na⁺ increases [24]. This highlights the importance of the membrane’s selectivity for Li⁺ over other ions like Mg²⁺ and Na⁺ for Li⁺ recovery from brines [25].

Membranes with specific functionalities have gathered increasing attention due to their unique attributes [26,27]. Crown ethers (CEs), which act as host molecules, have been widely integrated into polymers as functional group anchors for various membrane-based separation applications [26,28]. CE demonstrates an outstanding ability to complex with alkali metal ions, creating complexes via ion–dipole interactions between metal cationic ions and the negatively charged oxygen atoms located in the central area of the CE ring [29]. For example, Baudino et al. devised a nanocomposite membrane functionalized with 12-crown 4-ether (12CE) to recover Li⁺ [25]. 18-crown 6-ether (18CE) was also used to prepare materials for efficient Li⁺ recovery [30,31], and 15-crown 5-ether (15CE) was employed to prepare highly selective membranes for recovery of Li⁺ from a solution with high Mg²⁺ content. [32]. Mao et al. enhanced Li⁺ separation efficiency by utilizing 14-crown 4-ether (14CE) [33]. These studies indicate that 12CE (1.2–1.5 Å), 14CE (1.2–1.5 Å), 15CE (1.7–2.2 Å), and 18CE (2.6–3.2 Å) all feature with pore dimensions that can selectively separate Li⁺ [34,35]. In most studies, a singular species of CE has been employed either in modifying or preparing membranes to enhance Li⁺ recovery efficiency [30,31]. Alternatively, membranes prepared with CEs have been used as Li recovery adsorbents [32]. To date, there are no systematic studies on a comparative evaluation of membrane performance in Li⁺ recovery from brines utilizing different CEs via electrodialysis. This research aims to fill the knowledge gap and evaluate the performance and permselectivity of modified IEMs using various CEs.

In this study, selective membranes for Li⁺ recovery were prepared using novel CR671 manufactured by Veolia Water Technologies & Solutions as the pristine membranes. Because 12CE and 14CE have similar cavity sizes, 12CE, 15CE, and 18CE were chosen to modify the CEM CR671 for selective separation of Li⁺ from high Na⁺ and Mg²⁺ brines. Polydopamine (PDA) was used to facilitate the deposition of the complex of polyethyleneimine (PEI) and various CEs (12CE, 15CE, and 18CE) bonded through hydrogen bonds. To reinforce the structure, the cross-linker 1,3,5-benzenetricarbonyl trichloride (BT) was utilized to cross-link PEI, thereby enabling the specific cavities of the CE to selectively transport targeted Li⁺ in brines with high ratio of Mg²⁺/Li⁺ or Na⁺/Li⁺ ions. The primary objectives of this study include (i) the modification of CR671 membranes with various CEs to enhance the membrane permselectivity for Li⁺ recovery from Na⁺-and Mg²⁺-rich brine; (ii) the characterization of the modified membranes using methods like water uptake, zeta-potential measurement, ion exchange capacity (IEC), attenuated total reflectance Fourier-transform infrared (ATR-FTIR) spectroscopy, thermogravimetric analysis (TGA), and electrochemical impedance spectroscopy (EIS) to optimize modification conditions; and (iii) the assessment of the performance and permselectivity of the three modified membranes through the electrodialysis of brines.

2. Methods and Materials

2.1. Materials

The CEM CR671 comprises a cross-linked copolymer with sulfonic acid functional groups derived from vinyl compounds with a 520–540 μm thickness. The chemical modifiers, polyethyleneimine (PEI), sodium dodecyl sulfate (SDS), dopamine hydrochloride (DA), 1,3,5-benzenetricarbonyl trichloride (BT), trimethylol aminomethane (Tris), Na₂CO₃, KCl, and NaCl, were of analytical grade and procured from Sigma-Aldrich (Burlington, MA, USA). Various CEs, including 2-OH-15-crown 5-ether (15CE), 2-OH-12-crown 4-ether (12CE), 2-OH-18-crown 6-ether (18CE), as well as LiCl and MgCl₂ were bought from Fisher Scientific (Waltham, MA, USA). Other analytical reagents were utilized in their pristine state without undergoing additional treatment.

2.2. Membrane Modifications

Figure 1 illustrates the process of modifying the CR671 with 12CE, 15CE, and 18 CE, as well as the optimized amount of chemicals used in the membrane modification following our previous study [36]. The first step was to prepare a 2 g/L DA solution with a pH of approximately 8.5–8.8. The solution turned brown, indicating the formation of PDA [37]. At room temperature, CR671 specimens were submerged upright in the PDA solution for 4 h and named PDA/CR671. Then, PEI, Na₂CO₃, SDS, and various CEs (12CE, 15CE, 18CE) were individually added to deionized (DI) water, with magnetic stirring (350 rpm, 25 °C) for 30 min to facilitate the binding of CEs and PEI through hydrogen bonds. At room temperature, the prepared solution was then applied to the surface of PDA/CR671 and reacted for 10 min. Before the crosslinking, the remaining PEI/CEs solutions were removed. At 40 °C, BT was employed to crosslink PEI in a water bath for 10 min; the prepared membranes were labeled as 12CE/CR671, 15CE/CR671, and 18CE/CR671. The control membranes did not contain CE and were designated PEI/PDA/CR671.

2.3. Membrane Characterization

2.3.1. Functionality and Thermogravimetric Analysis

Before characterization, the membranes were cut into sections and dried overnight at 40 °C for 24 h. The membrane functionality was analyzed using ATR-FTIR (Nicolet IS10, Thermo Scientific, Waltham, MA, USA), and their thermal stability was assessed with a TGA Q500 thermogravimetric analyzer (TA, New Castle, DE, USA).

2.3.2. Surface Charge

Membrane surface charge plays a key role in ion transport during electrodialysis. A SurPass Electrokinetic Analyzer was used to analyze zeta potential (Anton Paar Trading Co., Ashland, VA, USA). 0.01 M KCl solution was used as the electrolyte, and its pH was adjusted using 0.1 M HCl or NaOH solutions. The measurements were taken at a pressure of 400 mbar.

2.3.3. Ion Exchange Capacity

IEC reflects the total amount of functional groups on the membranes. The IEC of the prepared membranes was measured by soaking the membranes in 1 M NaCl solution (200 mL) to saturate them with Na⁺ for 24 h. Subsequently, the adsorbed Na⁺ was replaced by K⁺ using 0.1 M and 0.01 M KCl solutions, and this process was repeated three times. An ion chromatograph was used to quantify the Na⁺ concentrations for determining IEC (IC, ICS 1100 & ICS 2100, Dionex, Thermo Scientific, Waltham, MA, USA).

2.3.4. Water Adsorption Capacities

The ion transport property was influenced by the water adsorption capacities (water uptake, WU) of the CEM membranes during the electrodialysis process. The water adsorption of membranes may vary across different mediums. To assess the water adsorption capacities, the membranes were saturated with DI water and synthetic brines, and the difference in weights before and after exposure to the medium was measured. At 40 °C, the membranes were cut into 4 cm² and dried overnight. The dried membrane was weighed and then immersed in either a Mg²⁺/Li⁺ mixture (2.2 molar ratio) or a Na⁺/Li⁺ solution (3.6 molar ratio). After 24 h, the membranes were re-weighed, and the excess surface water was removed using filter papers. The weight differences between the wet membranes (soaked for 24 h) and the dried membranes, divided by the wet membranes’ weight, were used to calculate the WU.

2.3.5. EIS Analysis

A Gamry electrochemical workstation was employed to test EIS for analyzing the electrical resistance of membranes (Interface 1000, Warminster, PA, USA) [38]. The frequency (f) range applied spanned from 100,000 Hz to 0.1 Hz, operating under alternating current. The membranes (0.8 cm²) were positioned between two stalls, each containing 12 mL of electrolyte solutions. These solutions included two different concentrations: 0.01 M alkali solutions of LiCl, MgCl₂, and NaCl, as well as 0.1 M solutions of LiCl, MgCl₂, and NaCl. The membrane samples were conditioned for 24 h by immersing them in a KCl solution (0.1 M) prior to experimentation. The ion transport time (τ) across the membrane interface and coating layer was calculated using the equation $\tau =1/\left(2\pi fmax\right)$ [39].

2.4. Bench-Scale Electrodialysis and Calculations

An electrodialysis system was used to evaluate the performance of three modified Li⁺-selective CEMs under specific conditions (PCCell GmbH, Heusweiler, Germany). The feed solution comprised a volume of 4 L containing mixtures of Na⁺/Li⁺ solution with a molar ratio of 3.6 (25 g/L NaCl and 5 g/L of LiCl) or Mg²⁺/Li⁺ solution with a molar ratio of 2.2 (25 g/L MgCl₂ and 5 g/L of LiCl) prepared with tap water.

The electrodialysis system was equipped with ten pairs of IEMs, which included eleven modified CEMs and ten anion exchange membranes (AEMs). The electrodialysis experiments were conducted in a continuous-flow, recirculation mode with the feed/diluate and concentrate streams being recycled to the feed solution tank. The flow rates for both streams were set to 180 mL/min, equivalent to a linear flow velocity of 33.5 cm/s. The applied current density was consistently lower than the limiting current density (LCD) measured according to our previous study [10]. Water samples were collected every hour to measure Li⁺, Mg²⁺, and Na⁺ concentrations using an inductively coupled plasma optical emission spectroscopy (ICP-OES, Avio 550 Max, PerkinElmer, Waltham, MA, USA). Figure 2 illustrates the separation mechanism of electrodialysis.

The Li⁺ flux can be calculated as follows (in meq/s-m²):

$$\mathrm{L}\mathrm{i} \mathrm{f}\mathrm{l}\mathrm{u}\mathrm{x}={\displaystyle \frac{\mathrm{E}\mathrm{q}\mathrm{u}\mathrm{i}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{e}\mathrm{n}\mathrm{t} \mathrm{o}\mathrm{f} \mathrm{L}\mathrm{i} \mathrm{r}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{d}\times \mathrm{f}\mathrm{l}\mathrm{o}\mathrm{w} \mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}{\mathrm{E}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e} \mathrm{m}\mathrm{e}\mathrm{m}\mathrm{b}\mathrm{r}\mathrm{a}\mathrm{n}\mathrm{e} \mathrm{s}\mathrm{u}\mathrm{r}\mathrm{f}\mathrm{a}\mathrm{c}\mathrm{e} \mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a}}}$$

(1)

The Li⁺ recovery efficiency can be calculated as:

$$Li recovery efficiency (\%)={(C}_{out}-C_{in})/C_{in}\times 100$$

(2)

where C_in is the Li⁺ concentration of electrodialysis concentrate-in (mg/L); C_out is the Li⁺ concentration of electrodialysis concentrate-out (mg/L).

The transport number was used to determine the transport of cations A (i.e., Li⁺, Mg²⁺, and Na⁺) during electrodialysis.

$${\mathrm{t}}_{\mathrm{A}}={\displaystyle \frac{\mathrm{E}\mathrm{q}\mathrm{u}\mathrm{i}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{e}\mathrm{n}\mathrm{t} \mathrm{o}\mathrm{f} \mathrm{A} \mathrm{t}\mathrm{r}\mathrm{a}\mathrm{n}\mathrm{s}\mathrm{p}\mathrm{o}\mathrm{r}\mathrm{t}\mathrm{e}\mathrm{d}}{\mathrm{E}\mathrm{q}\mathrm{u}\mathrm{i}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{e}\mathrm{n}\mathrm{t} \mathrm{o}\mathrm{f} \mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{c}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{s} \mathrm{t}\mathrm{r}\mathrm{a}\mathrm{n}\mathrm{s}\mathrm{p}\mathrm{o}\mathrm{r}\mathrm{t}\mathrm{e}\mathrm{d}}}$$

(3)

Modified CEMs with lower relative transport number t_A/Li (the calculation is shown in the Supplementary Information) exhibit higher Li⁺ permselectivity, and the improvement of Li⁺ permselectivity can be calculated using the formula:

$$\mathrm{P}\mathrm{e}\mathrm{r}\mathrm{m}\mathrm{s}\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{i}\mathrm{t}\mathrm{y} \mathrm{i}\mathrm{m}\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t}={\displaystyle \frac{{\mathrm{t}}_{\mathrm{A}/\mathrm{L}\mathrm{i}} \mathrm{o}\mathrm{f} {\mathrm{M}}_1}{{\mathrm{t}}_{\mathrm{A}/\mathrm{L}\mathrm{i}} \mathrm{o}\mathrm{f} {\mathrm{M}}_2}}$$

(4)

where M₁ refers to modified membranes with 12CE or 18CE; M₂ refers to the three modified membranes (12CE, 15CE, and 18CE).

The improvement in Li⁺ flux is computed as:

$$L\mathrm{i} \mathrm{f}\mathrm{l}\mathrm{u}\mathrm{x} \mathrm{i}\mathrm{m}\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t}={\displaystyle \frac{\mathrm{L}\mathrm{i} \mathrm{f}\mathrm{l}\mathrm{u}\mathrm{x} \mathrm{o}\mathrm{f} {\mathrm{M}}_1}{\mathrm{L}\mathrm{i} \mathrm{f}\mathrm{l}\mathrm{u}\mathrm{x} \mathrm{o}\mathrm{f} {\mathrm{M}}_2}}$$

(5)

3. Results and Discussion

3.1. Characterization of the Membranes

3.1.1. The Analysis of EIS

EIS analysis presents numerous advantages owing to its inherent characteristics as a steady-state technique. It uses the analysis of small signals and the ability to examine signal reductions over a broad frequency ranging from 0.1 Hz to 100,000 Hz. This capability is facilitated through the utilization of readily available electrochemical working stations.

In Lower-Concentration Solutions

The EIS results for the modified membranes in low-concentration solutions are depicted in Figure 3. In 0.01 M solutions of LiCl, MgCl₂, and NaCl, the EIS of the membranes is presented in Figure 3a–c. Figure 3a reveals that, in contrast to the pristine membranes CR671, which had peaks at a low frequency (0.39 Hz), the modified membranes exhibited conspicuous peaks indicative of Li⁺ transport at higher frequencies. Specifically, 12CE/CR671 operated at 5.01 Hz, 18CE/CR671 at 3.99 Hz, and 15CE/CR671 displayed peaks at 6.35 Hz. Concerning the membranes in the 0.01 M MgCl₂ solution, all membranes displayed significant peaks at higher frequencies. CR671 exhibited peaks at 31.67 Hz, while 12CE/CR671, 18CE/CR671, and 15CE/CR671 were recorded at 24.97 Hz, 24.97 Hz, and 20.03 Hz, respectively (Figure 3b). In the case of 0.01 M NaCl, the membranes demonstrated diminished peaks indicative of Na⁺ transport (Figure 3c). All the modified membranes 12CE/CR671, 15CE/CR671, and 18CE/CR671 exhibited lower transport speed for Li⁺ compared to Mg²⁺ in 0.01 M alkali solutions, as demonstrated by the calculated τ at the maximum frequency (f_max) of 0.032 s, 0.039 s, and 0.025 s for Li⁺ transport, respectively. The rapid Li⁺ transport of 15CE/CR671 is attributed to the size-sieving effect of the CE cavities, which is enhanced by the modification of 15CE.

In Higher-Concentration Solutions

In 0.1 M solutions of LiCl, MgCl₂, and NaCl, the Bode plots of the membranes are presented in Figure 3d–f. In the 0.1 M LiCl solution, at high frequencies, 12CE/CR671, 15CE/CR671, and 18CE/CR671 exhibit prominent peaks at 20 Hz (τ = 0.008 s), 24.9 Hz (τ = 0.006 s), and 15.8 Hz (τ = 0.010 s), respectively (Figure 3d). In the MgCl₂ solution, slight peaks for Mg²⁺ ion transport are observed at low frequencies (Figure 3e). In the NaCl solution, the CR671 membranes recorded a peak at 63.3 Hz, while the modified membranes exhibited peaks at 31.7 Hz (18CE/CR671), 24.9 Hz (12CE/CR671), and 12.5 Hz (15CE/CR671). This indicates that the modified membranes effectively attenuate Na⁺ transport through the membranes (Figure 3f). A comparative examination reveals that monovalent ions (e.g., Li⁺ and Na⁺) exhibited faster transport rates in higher-concentration solutions than multivalent ions (e.g., Mg²⁺), signifying the effective transport of monovalent ions. Meanwhile, the modified membranes consistently demonstrated stable and accelerated transport for Li⁺ regardless of concentration, indicating their good performance for Li⁺ selectivity. In higher concentration LiCl solution, the acceleration of Li⁺ transport through the modified membranes might be due to the combined effects of the concentration gradient, collision frequency, conductivity, and electric field strength.

3.1.2. Physiochemical Properties

ATR-FTIR analysis was utilized to characterize the chemical functionality of the membranes before and after modifications, as illustrated in Figure 4a. When comparing the modified membranes to the pristine CR671 membranes, no substantial changes were observed in the structure of polymer chains. The peak observed at 3200–3400 cm^–1 is attributed to the stretching vibrations of O-H or C-H. For the modified membranes, 12CE/CR671, 15CE/CR671, and 18CE/CR671, discernible peaks are evident, stemming from the integration of -CH₂OH from CEs (12CE, 15CE, and 18CE) onto the surface of CR671. The peaks between 2850–2960 cm^–1 are attributed to the stretching vibrations of C-H. The membranes 12CE/CR671, 15CE/CR671, and 18CE/CR671 exhibit slightly extensive peaks due to the introduction of CEs. These significant peaks provide substantial evidence for the successful coating of CEs onto CR671.

The surface charge can affect the selectivity of the membranes, both at the water–membrane interface and within the membrane pores [40,41]. These charge effects arise from the protonation and deprotonation of the functional groups (e.g., OH, -NH₂, -SO₃H) on the membranes in response to pH changes and the presence of electrolytes [42]. Zeta-potential is a useful technique for assessing charge density on a membrane surface [43]. Figure 4b shows a detailed analysis of the surface charge of the studied membranes. From Figure 4b, the modified membranes,12CE/CR671, 15CE/CR671, and 18CE/CR671, manifested analogous zeta potentials, consistently maintaining values within the range of 4.3 to 479 mV across the pH of 6–9. Despite utilizing distinct CEs, the observed constancy in membrane surface charge implies a uniform surface charge density among the three modified membranes. This observation, in turn, prompts the inference that the surface charge of the modified membranes may not exert a discernible influence on membrane selectivity. Furthermore, an examination of the functional groups of the modified membranes reveals comparable levels, suggesting a homogeneous distribution of modifications across the three membranes. This coherence in modification supports the robustness and consistency of the experimental procedures employed.

The thermal degradation of the membranes was analyzed using TGA within 35–700 °C, revealing four stages as shown in Figure 4c–f. In the first stage (35–250 °C), the weight loss below 100 °C was due to the evaporation of intrinsic or adsorbed water, while the following stage (250–321 °C) involved the decomposition of polymer chains. In the third stage (321–488 °C), a precipitous reduction in membrane weight ensued, attributable to the destruction of polymer chains and structures, culminating in a stabilized state in the final phase (488–700 °C). A comparative examination between the 12CE/CR671, 15CE/CR671, and 18CE/CR671 membranes demonstrated less weight loss in the case of modified membranes than CR671 (3.5%). In Figure S1, after the electrodialysis of Mg²⁺/Li⁺ brines for 12CE/CR671 and 15CE/CR671 and Na⁺/Li⁺ for 18CE/CR671, no significant disparities in thermal stabilities were observed. This implies that the electrodialysis process had a negligible impact on the structural integrity of the membranes.

Following the modification, the thickness of the membranes displayed minimal variations, remaining within the range of approximately 520–550 μm. The three modified membranes showed lower IEC values than the CR671 of 1.89 ± 0.03 meq/g dry membranes. Specifically, the IEC values for the modified membranes were measured as 1.59 ± 0.17, 1.57 ± 0.21, and 1.60 ± 0.24 meq/g dry membranes for 15CE/CR671, 12CE/CR671, and 18CE/CR671, respectively.

3.1.3. Water Adsorption Capacity Analysis in Different Solutions

The permselectivity of membranes is associated with both their WU and the characteristics of the surrounding water medium [44,45]. Figure 5 illustrates the examination of diverse mediums on membrane WU. In DI water, the modified membranes exhibit reduced WU compared to the CR671 membrane. This decline is ascribed to the cross-linking of PEI, resulting in an increased crosslinking density in the modified membranes [46]. Among the modified membranes, the WU of 15CE/CR671 (55.8%) is the lowest in DI water, followed by the 12CE/CR671 (56.5%) and the 18CE/CR671 (58.0%). Contrastingly, in alkali solutions, the WU of the modified membranes surpasses that in DI water, due to the binding of hydrated alkali ions (e.g., Li⁺, Na⁺, Mg²⁺). The modified membranes exhibited comparable WU in the alkali mixture solutions, with values of 72.2% and 62.0% for 18CE/CR671, 71.6% and 59.7% for 12CE/CR671, and 71.8% and 59.6% for 15CE/CR671 in Na⁺/Li⁺ and Mg²⁺/Li⁺ solutions, respectively. The WU in the Na⁺/Li⁺ mixture solution surpasses that observed in the Mg²⁺/Li⁺ mixture solution. This observation suggests that a higher proportion of hydrated Na⁺ is bonded with the CEs in modified membranes than hydrated Mg²⁺ due to the larger hydrated ion radius of Mg²⁺ than Na⁺ [47].

3.2. Permselectivity and Performance of the Membranes in Electrodialysis

The potential impact of various cationic ions competing with Li⁺ is a key consideration for efficient Li⁺ recovery in electrodialysis. Among these ions, Mg²⁺ in brines poses a primary challenge due to the similarity in the ionic size of Mg²⁺ (radius ≈ 0.072 nm) and Li⁺ (radius ≈ 0.076 nm) [48], leading to complications in their effective separation during electrodialysis. Additionally, considering factors such as approximate ion radius and valence state, Na⁺ was identified as another competitor for Li⁺ in brine treatment. As a result, different brine solutions containing Mg²⁺ or Na⁺ were tested to assess the permselectivity of the modified membrane affected by competing ions during the electrodialysis process.

Before experimentation, limiting current density (LCD) was determined following the methodology outlined in our previous study [10]. The LCD is provided in the Supplementary Information. Figure S2 illustrates that for all three membranes, a linear relationship is observed during electrodialysis of the brines between voltage/cell-pair and current density from 2.3 to 29.5 mA/cm². The electrodialysis process was operated under the LCD (15.9 mA/cm² for high Na⁺ brines and 14.7 mA/cm² for high Mg²⁺ brines).

The samples collected from feed solutions before the electrodialysis, diluate-out, and concentrate-out were analyzed at various current densities (2.3, 9.4, and 15.9 mA/cm² for high Na⁺ brine or 2.2, 9.3, 14.7 mA/cm² for high Mg²⁺ brines) at 1 h intervals. After 3 h of electrodialysis, an additional 4 h was employed to evaluate the stability of the four membranes in brines at 15.9 mA/cm² for high Na⁺ brines and 14.7 mA/cm² for high Mg²⁺ brines, which was noted as 15.9–4 and 14.7–4, with “4” indicating after 4 h of additional operation.

3.2.1. Li Separation from High Na⁺ Brines

In brines, separating Li⁺ from a Na-rich solution proves to be more challenging compared to a high Mg²⁺ brine. This difficulty stems from the similar univalent nature of Na⁺ and Li⁺ ions and their lower separation coefficients associated with Na⁺ [49]. To systematically assess Li⁺ recovery efficiency, modified membrane performance, and permselectivity, a Na-rich solution was prepared at a molar ratio of 3.6. It was treated using the bench-scale electrodialysis stack. The evaluation criteria included Li⁺ flux, Li recovery efficiency, relative transport number (t_Na/Li), and Na⁺ leaking rate, as detailed in

Table 1. Membrane permselectivity and performance in high Na⁺ brines.

Membranes	Current Density (mA/cm2)	Current Efficiency (%)	tLi	tNa	tNa/tLi	Li Flux (meq/s-m2)	Li Recovery Efficiency (%)	Na Leaking Rate (%)	Energy Consumption (kWh/kg-Li)
12CE/CR671	2.3	20.3 ± 3.9	0.12	0.88	2.00	4.7 ± 1.2	0.84	1.7	2.81
	9.4	14.8 ± 2.8	0.26	0.74	0.81	34.3 ± 5.7	11.1	4.9	7.05
	15.9	17.3 ± 1.6	0.25	0.75	0.85	62.3 ± 9.8	17.4	9.4	14.84
	15.9–4 *	35.8 ± 8.5	0.25	0.75	0.88	126.9 ± 11.4	22.7	19.6	16.56
15CE/CR671	2.3	38.4 ± 3.9	0.32	0.68	0.61	23.1 ± 3.7	5.1	3.1	0.48
	9.4	17.0 ± 2.9	0.28	0.72	0.76	37.0 ± 7.3	8.2	6.2	5.07
	15.9	39.4 ± 7.4	0.24	0.76	0.93	129.7 ± 14.7	28.8	26.6	5.74
	15.9–4 *	127.4 ± 6.3	0.23	0.77	0.97	407.5 ± 21.5	90.5	86.9	4.26
18CE/ CR671	2.3	80.3 ± 8.5	0.21	0.79	1.06	32.3 ± 5.7	6.6	7.0	0.42
	9.4	19.3 ± 3.9	0.23	0.77	0.96	42.6 ± 7.8	8.6	8.3	5.22
	15.9	17.8 ± 4.0	0.19	0.81	1.15	79.1 ± 9.4	16.0	18.7	10.71
	15.9–4 *	33.4 ± 5.0	0.17	0.83	1.30	117.4 ± 11.4	23.8	32.2	16.85

Note: * electrodialysis experimental results at a current density of 15.9 mA/cm² after an additional 4 h of operation.

.

Table 1 presents t_Li and t_Na for the three modified membranes. At 15.9 mA/cm², t_Li and t_Na were determined to be 0.25 and 0.75 for 12CE/CR671, and 0.19 and 0.81 for 18CE/CR671, respectively; t_Na/Li for 12CE/CR671 and 18CE/CR671 were calculated as 0.85 and 1.15, indicating that 12CE/CR671 and 15CE/CR67 selectively separated and concentrated Li⁺ from Na⁺/Li⁺ solutions. As both time and current density increased, there was an observed rise in Li⁺ recovery efficiency for the three modified membranes. At 15.9 mA/cm², the Li⁺ recovery efficiencies of 12CE/CR671 and 18CE/CR671 were 17.4% and 16.0%, respectively, significantly lower than that of 15CE/CR671 (90.5%). Although Na⁺ leaking rates were observed at 9.4% and 18.7%, respectively, 18CE/CR671 showed that the Na⁺ leaking rate surpassed the Li⁺ recovery efficiency, implying that despite effectively concentrating Li⁺, there is a concurrent concentration of Na⁺. This could be due to the increase in moisture content within the membranes, which increases conductivity and simultaneously lowers the selectivity, coupled with the higher hydration capacity of Li⁺ compared to Na⁺ [44,50].

Among the three modified membranes, 12CE/CR67 and 18CE/CR671 exhibited lower Li⁺ recovery efficiency and Li⁺ flux (62.3 and 79.1 meq/s-m²) at 15.9 mA/cm², along with current efficiencies of 17.3% and 17.8%, respectively, compared to those of 15CE/CR671 (39.4%). At 2.3 mA/cm², the normalized Li⁺ recovery efficiencies for 12CE/CR671 and 18CE/CR671 were 3.8 and 23.6 g/kWh-cm², respectively (Table S1). The observed reductions in conductivity for 12CE/CR671 and 18CE/CR671 were 15.3% and 14.8%, respectively, which were lower than the reduction for 15CE/CR671 (Table S1).

After another 4 h of experiment, the modified membranes demonstrated increased Li⁺ transport permeability in Na-rich brines. The membranes 12CE/CR671 and 18CE/CR671 showed Li recovery efficiency of 22.7% and 23.8%, lower than 15CE/CR671 (90.5%). The Li⁺ flux of 12CE/CR671 and 18CE/CR671 was 126.9 and 117.4 meq/s-m² with the current efficiency of 35.8% and 33.4%, respectively, lower than that of 15CE/CR671 (407.5 meq/s-m²) with higher current efficiency of 127.4%, indicating that 15CE/CR671 possesses a higher permselectivity for Li⁺ than that of 12CE/CR671 and 18CE/CR671. The persistence of modified membranes contrasts with our earlier investigation, where the transfer capacity of the unmodified membranes (CR671) diminished. These suggest that our modified membranes with CEs enhance the viability in the electrodialysis of Na-rich brines.

In altering the membrane by incorporating diverse CEs, the resultant modified membranes demonstrated varying permselectivity for Li⁺ and Na⁺ ions. Numerous factors contribute to these distinctions: (i) The distinct cavity sizes and hydrated radii of the three different CEs and alkali ions (Mg²⁺ > Li⁺ > Na⁺) contribute to the divergent permselectivity measured in the modified membranes [51]^. The 12CE/CR671 and 15CE/CR67 configurations exhibited selectivity for Li⁺ over Na⁺ due to the congruence between the hydrated Li⁺ radius and the cavity diameter of 12CE (1.2–1.5 Å) and 15CE (1.7–2.2 Å). Conversely, 18CE/CR67 demonstrated selectivity for Na⁺ over Li⁺ due to the incongruity between the hydrated Li⁺ radius and the cavity diameter of 18CE (2.6–3.2 Å) [50]; (ii) The observed thermodynamic contrast between the 15CE/Li complex (log k_a = 0) and the 15CE/Na complex (log k_a = 0.44) underscores distinct energetic characteristics in the interactions with Li⁺ and Na⁺, respectively. This disparity in thermodynamic values implies varying affinities and stability for the formed complexes. The inherent instability of the Li-15CE complex suggests a less favorable binding environment, resulting in a selective advantage of 15CE/CR671 for Li⁺ [52,53]; (iii) The higher distribution coefficient of Li⁺ compared to Na⁺ in the solution results in a greater potential of Li⁺ to contact with or reaching the membranes (12CE/CR671), leading to a higher potential permselectivity for Li⁺ over Na⁺ [54]. (iv) The elevated WU, coupled with the establishment of a stable complex between 18CE and H₃O⁺, collectively diminishes the selectivity of the 18CE/CR671 system for Li⁺, thereby facilitating the permeation of Na⁺ through the membranes more readily than Li⁺ [50].

Figure 6 illustrates the improvement of the modified membrane permselectivity and Li flux in high Na+ brines. Membranes 12CE/CR671 and 18CE/CR671 demonstrated permselectivity improvements, approximately 11.9-fold and 10.4-fold, respectively, when compared to PEI/PDA/CR671 at 15.9 mA/cm² (Figure 6a,c). In comparison, 15CE/CR671 exhibited higher permselectivity improvement, approximately 3.3-fold and 1.7-fold, compared to 12CE/CR671 and 18CE/CR671 at 2.3 mA/cm² (Figure 6e). Li⁺ flux increased significantly following an additional 4 h of operation, reaching approximately 8.7-fold and 8.1-fold when comparing 12CE/CR671 and 18CE/CR671 to PEI/PDA/CR671 (Figure 6b,d). 15CE/CR671 exhibited higher Li⁺ flux improvement, approximately 3.2-fold and 3.4-fold, at 15.9 mA/cm² when comparing 12CE/CR671 and 18CE/CR671 (Figure 6f).

3.2.2. Li Separation from High Mg²⁺ Brines

Challenges can arise from Mg²⁺ in brines competing with Li⁺ during the separation process. In our study, a Mg^2+/Li⁺ brine with a molar ratio of 2.2 was utilized to assess the permselectivity and effectiveness of the modified membranes.

Table 2. Membrane permselectivity and performance in high Mg²⁺ brines.

Membranes	Current Density (mA/cm2)	Current Efficiency	tLi	tMg	tMg/tLi	Li Flux (meq/s-m2)	Li Recovery Efficiency (%)	Mg Leaking Rate (%)	Energy Consumption (kWh/kg-Li)
12CE/CR671	2.2	19.6 ± 4.3	0.51	0.49	0.20	16.5 ± 1.3	4.7	0.9	0.56
	9.3	15.3 ± 1.6	0.77	0.23	0.13	71.3 ± 6.7	20.0	2.4	2.21
	14.7	20.0 ± 5.3	0.73	0.27	0.18	121.5 ± 12.7	34.1	5.3	4.56
	14.7–4 *	21.2 ± 4.2	0.71	0.29	0.20	155.0 ± 16.3	43.6	7.2	8.34
15CE/CR671	2.2	38.4 ± 9.9	0.84	0.16	0.08	26.7 ± 5.9	6.8	0.5	0.36
	9.3	17.0 ± 2.9	0.81	0.19	0.11	88.4 ± 15.7	22.3	2.2	1.81
	14.7	39.4 ± 7.4	0.85	0.15	0.10	248.6 ± 47.3	62.9	4.8	2.21
	14.7–4 *	49.7 ± 9.3	0.74	0.26	0.10	318.7 ± 55.6	80.5	5.7	4.02
18CE/CR671	2.2	30.9 ± 4.3	0.48	0.52	0.21	24.2 ± 5.7	6.5	1.3	0.40
	9.3	11.7 ± 3.4	0.67	0.33	0.20	54.8 ± 6.8	14.7	2.8	2.99
	14.7	12.7 ± 2.8	0.74	0.26	0.15	109.8 ± 11.8	29.5	3.9	5.35
	14.7–4 *	19.3 ± 4.9	0.75	0.25	0.14	140.0 ± 18.6	37.6	4.7	9.80

Note: * electrodialysis experimental results at a current density of 14.7 mA/cm² after an additional 4 h of operation.

shows t_Li and t_Mg for the modified membranes. At 14.7 mA/cm², the values for t_Li and t_Mg were 0.73 and 0.21 for 12CE/CR671 and 0.74 and 0.26 for 18CE/CR671, respectively. t_Mg/Li at the same current density was calculated as 0.18 for 12CE/CR671 and 0.15 for 18CE/CR671, indicating selective separation of Li⁺ from Mg-rich solutions, like 15CE/CR67. With increasing time and current density, there was an observed increase in Li⁺ recovery efficiencies for all three modified membranes. The Li⁺ recovery efficiencies of 12CE/CR671 and 18CE/CR671 were 34.1% and 29.5% at 14.7 mA/cm², respectively. The Mg²⁺ leaking rates were 5.3% and 3.9%, indicating higher permselectivity for Li⁺ than Mg²⁺. At 14.7 mA/cm², 12CE/CR67 and 18CE/CR67 showed lower Li⁺ recovery and flux (121.5 and 109.8 meq/s-m²) compared to 15CE/CR67 (248.6 meq/s-m²) with a 39.4% current efficiency. As shown in Table S2, at 2.3 mA/cm² the normalized Li⁺ recovery efficiencies of 12CE/CR671 and 18CE/CR671 were 15.2 and 19.5 g/kWh-cm², respectively, lower than that of 15CE/CR671 (24.7 g/kWh-cm²). The conductivity reduction observed for 12CE/CR671 and 18CE/CR671 was 13.4 and 13.1%, respectively, lower than that of 15CE/CR671 (16.5%).

After another 4 h, the modified membranes exhibited increased Li⁺ transport permeability in Mg-rich brines. The membranes 12CE/CR671 and 18CE/CR671 showed Li recovery efficiency of 43.6% and 37.6%, respectively, lower than 15CE/CR671 (80.5%). The Li⁺ flux of 12CE/CR671 and 18CE/CR671 was 155 and 140 meq/s-m² with current efficiency of 21.2% and 19.3%, respectively, lower than that of 15CE/CR671 (318.7 meq/s-m²), indicating that 15CE/CR671 possessed a higher permselectivity for Li⁺ than that of 12CE/CR671 and 18CE/CR671. Like the Na-rich brines, modifying membranes with CE persisted more than CR671 in Mg-rich brines.

The modified membranes exhibited distinct Li⁺ transport characteristics, and several factors influenced the permselectivity of the modified membranes containing CEs for Li⁺ and Mg²⁺. (i) The lower total energy of the CEs-Mg²⁺ complex (–1082.97 kJ/mol) in comparison to the CEs-Li⁺ complex (–419.19 kJ/mol) indicated the spontaneous occurrence of binding between CEs and alkali ions (e.g., Mg²⁺ and Li⁺). The stronger and more easily formed binding of CEs with Mg²⁺ instead of Li⁺ resulted in the high permselectivity of the three modified membranes for Li⁺ [55]. Our previous FTIR analysis verified this observation. (ii) The increased hydration effect and stronger affinity of Mg²⁺ for water molecules compared to Li⁺ led to greater ease of Li⁺ transport through the membranes than Mg²⁺ [55]. (iii) The positively charged surface of the modified membrane from PEI induced electronic repulsion, particularly toward Mg²⁺, further contributing to the observed selectivity for Li+ over Mg²⁺.

Figure 7 illustrates the modified membranes’ improved Li⁺ flux and permselectivity in high Mg²⁺ brines. The membranes 12CE/CR671 and 18CE/CR671 demonstrated permselectivity improvements, approximately 17.2-fold and 20.4-fold, respectively, when compared to PEI/PDA/CR671 at 14.6 mA/cm² (Figure 7a,c). In comparison, 15CE/CR671 exhibited a greater permselectivity improvement, approximately 2.4-fold and 2.6-fold, in comparison to 12CE/CR671 and 18CE/CR671 at 2.2 mA/cm² (Figure 7e).

After an additional 4 h, significant enhancements in Li flux were observed, reaching approximately 34-fold and 49.8-fold when comparing 12CE/CR671 and 18CE/CR671 to CR671 at 14.7 mA/cm² (Figure 7b,d). At 2.2 mA/cm², 15CE/CR671 achieved approximately 2.4-fold and 2.7-fold greater permselectivity than 12CE/CR671 and 18CE/CR671, respectively, and at 14.7 mA/cm², it showed 2.1-fold and 2.3-fold improved Li⁺ flux (Figure 7f).

The membranes modified with CEs demonstrate higher Li⁺ recovery efficiencies in high Na⁺ brines than in high Mg²⁺ brines while displaying lower permselectivity in high Na⁺ brines than in high Mg²⁺ brines. Several factors contribute to the preference for Na⁺ ion transport over Mg²⁺ ion through the modified membranes: (i) Na⁺ exhibits a higher self-diffusion coefficient than Mg²⁺ [55], implying that Na⁺ has a higher potential to reach or contact the membranes; (ii) Na⁺ binds with 5.6 water molecules, while Mg²⁺ binds with 6.0 water molecules [56], impeding a slower movement of Mg²⁺ from the liquid solution toward the membranes than Na⁺; (iii) The distance between the ion and water molecule for Na⁺–water (0.227 nm) is greater than that for Mg²⁺–water (0.197 nm) [56], indicating that Mg²⁺ has a stronger affinity for water compared to Na⁺; (iv) With a hydrated ion radius of 0.428 nm, Mg²⁺ is larger than Na⁺, which has a radius of 0.358 nm [24], showing that the cavity sizes of CEs favor the transport of Na⁺ over Mg²⁺; (v) After introducing PEI, the surface charges of the modified membranes become more positive, resulting in greater electrostatic repulsion towards Mg²⁺ than Na⁺. These factors collectively reduce the permselectivity of the membranes for Li⁺ in Na⁺/Li⁺ brines.

3.3. Comparison of the Performance and Energy Costs of Different Methods for Li⁺ Recovery

When comparing the 15CE-modified membranes in this study with other membranes used in electrodialysis for lithium recovery, the 15CE-modified membranes exhibit superior performance (Figure 8). In electrodialysis, membranes such as CR67-MK111 and CIMS cationic exchange membranes have achieved Li⁺ recovery efficiencies of 27.5% and 76.5%, respectively [57,58]. However, the 15CE-modified membranes of our study demonstrate a significantly higher Li⁺ recovery efficiency, reaching up to 90.5% in Na⁺/Li⁺ solutions and 80.5% in Mg²⁺/Li⁺ solutions. This study shows improved Li⁺ recovery efficiency compared to other reported membranes used in electrodialysis, which typically achieve recovery rates ranging from 27.5% to 76.5% in brines. These results suggest that the 15CE-modified membranes offer a more effective solution for Li⁺ recovery in electrodialysis applications, making them a promising option for enhancing Li⁺ recovery from brine solutions.

The energy consumption and costs of various methods for Li recovery were compared, as shown in Table S4. Electrocoagulation consumes around 64 kWh/kg-Li for recovering Li⁺ from a solution with 1,000 mg/L of Li⁺ and 7,871.3 mg/L of Na⁺ [62]. Lee et al. [63] reported using hybrid capacitive deionization (H-CDI) to recover Li⁺ from brines containing 1457.4 mg/L of Li⁺, 7,590,000 mg/L of Na⁺, and 9600 mg/L of Mg²⁺, with an energy consumption of 23.3 kWh/kg-Li, costing 1.4 US$/kg-Li. Membrane capacitive deionization (MCDI) and electrodialysis, using CIMS membranes with 500 mg/L TDS and a Li/Mg molar ratio of 1:1 (24.9 mg/L Li⁺ and 86.4 mg/L Mg²⁺), had energy consumptions of approximately 0.3 and 0.33 kWh/kg-Li [64]. Guo et al. [58] used monovalent selective membranes in electrodialysis to separate Li⁺ from brines with concentrations of 140 mg/L of Li⁺, 20,810 mg/L of Na⁺, and 2250 mg/L of Mg²⁺, with an energy consumption of about 95.1 kWh/kg-Li, costing 5.7 US$/kg-Li. In comparison, the 15CE-modified membranes (15/CE-CR671) achieved a high Li⁺ recovery efficiency of 90.5% from Na⁺/Li⁺ brine (molar ratio = 3.6) and 80.5% from Mg^2+/Li⁺ brine (molar ratio = 2.2) through electrodialysis. This process demonstrated lower energy consumption, requiring 4.26 and 4.02 kWh/kg-Li, and reduced energy costs of $0.26 and $0.24 /kg-Li, respectively.

4. Conclusions

This study strategically employed alkali ions ligand CEs to enhance the selectivity of CEM, focusing on optimizing Li⁺ recovery efficiency for the CEM (CR671) through electrodialysis. The membrane modification process involving PDA treatment and the formation of PEI-CE complexes, particularly 12CE/CR671, 15CE/CR671, and 18CE/CR671, successfully created specific channels for Li⁺ ion transport. Characterization techniques, such as ATR-FTIR and EIS, confirmed the effective coating of CEs and the selective capacity of the membranes for Li⁺ transport in all three modified membranes. Bench-scale electrodialysis experiments revealed significant Li⁺ flux and permselectivity improvements in high Na⁺ solutions during electrodialysis. At 15.9 mA/cm², 12CE/CR671 and 18CE/CR671 showed 11.9-fold and 10.4-fold better permselectivity and 8.7-fold and 8.1-fold better Li⁺ flux than PEI/PDA/CR671. At 2.3 mA/cm², 15CE/CR671 exhibited higher permselectivity improvement, approximately 3.3-fold and 1.7-fold. Moreover, 15CE/CR671 demonstrated superior Li⁺ flux improvement, approximately 3.2-fold and 3.4-fold, at 15.9 mA/cm², surpassing the improvements seen in 12CE/CR671 and 18CE/CR671. In Mg²⁺/Li⁺ brines, significant enhancements in Li⁺ flux were observed, reaching approximately 34-fold and 49.8-fold when comparing 12CE/CR671 and 18CE/CR671 to CR671 at 14.7 mA/cm². At 2.2 mA/cm², 15CE/CR671 showed 2.4-fold and 2.7-fold greater permselectivity and 2.1-fold and 2.3-fold better Li⁺ flux than 12CE/CR671 and 18CE/CR671 at 14.7 mA/cm². The 15CE/CR671 membrane achieved a significantly higher Li⁺ recovery efficiency of 90.5% in Na⁺/Li⁺ brines and 80.5% in Mg²⁺/Li⁺ brines with lower energy consumption (4.26 kWh/kg-Li and 4.02 kWh/kg-Li) than other methods. This research is at a technology readiness level of a feasibility and proof-of-concept study. After we have optimized the membrane modification and demonstrated the Li selectivity, our next step is to treat real brine and potentially pilot testing to increase the technology readiness level.