Ion-Exchange Membrane Permselectivity: Experimental Evaluation of Concentration Dependence, Ionic Species Selectivity, and Temperature Response

Abstract

Ion-exchange membranes (IEMs) are widely used in reverse-electrodialysis (RED) technology, which can collect the salinity gradient energy between concentrated and diluted solutions and convert it into electromotive force (EMF) to drive power generation and hydrogen production. Recent studies have indicated that the permselectivity of IEMs is vital to determining the performance of an RED stack. In this study, the influences of solution concentration, ion species, and solution temperature on the permselectivity of IEMs were experimentally investigated. The results demonstrate that the permselectivity of IEMs decreases with increasing concentrations of KAc, LiCl, and LiBr solutions for both concentrated solutions (3–5 M) and dilute solutions (0.02–0.2 M). Further, through comparing the LiBr and KBr solutions as well as the LiCl, KCl, and NH₄Cl solutions, respectively, K⁺ demonstrates a higher permselectivity than Li⁺, and both of which are smaller than NH₄⁺ under the same cation and concentration conditions. Moreover, another test was conducted using three potassium salt solutions with different anions, and the experimental permselectivity order is Ac⁻ > Br⁻ > Cl⁻. A slight increase in solution temperature enhances the permselectivity of IEMs due to the increase in ionic mobility. However, an excessive temperature is detrimental to membrane stability and thus reduces permselectivity. It can be seen that ions with low hydration energy, a small hydration radius, and high mobility show a higher permselectivity.

1. Introduction

Ion-exchange membranes (IEMs) are charged polymer membranes with ion-exchange properties, and their core functionality arises from fixed charged groups that enable the selective adsorption and transport of ions in solution [1]. Owing to their efficient separation capabilities, IEMs have been extensively utilized across industrial and energy applications, including water treatment [2,3,4], fuel cells [5], and energy conversion systems [6,7]. Their roles in advancing environmental protection and clean energy technologies have become increasingly prominent. In electrodialysis (ED) for seawater desalination [3], alternating anion-exchange membranes (AEMs) and cation-exchange membranes (CEMs) are subjected to an electric field. This setup creates distinct compartments for cations and anions in brackish water, driving anions toward the negative electrode and cations toward the positive electrode, ultimately yielding fresh water. Similarly, the reverse-electrodialysis (RED) cells [8] leverage a salinity gradient energy (SGE) between concentrated and dilute solutions separated by the AEMs and CEMs, and then the directional ion permeation across the membranes converts the chemical energy into electrical power or hydrogen energy [7,8]. The efficiencies of both ED and RED processes are critically dependent on the permselectivity of the membranes.

Permselectivity is defined as the extent to which a membrane facilitates counter-ion permeation while restricting co-ions, reflecting its ability to discriminate between oppositely charged species [9]. Ideally, the IEMs exhibit a permselectivity of one (100%), enabling complete counter-ion transport and total co-ion exclusion. For instance, in AEMs, positively charged functional groups selectively permit anion passage while effectively blocking the cations. However, permselectivity is not an intrinsic material property; it is dynamically influenced by several factors such as the solution concentration, solute type, and temperature [10]. Due to the importance of permselectivity in practical applications and modeling, many researchers have extensively investigated this property through experimental studies, simulations, and physical/chemical modifications.

Regarding performance tests, for instance, Długolecki et al. [11] and Dong et al. [12] systematically compared commercially available and lab-synthesized IEMs, respectively, proposing characterization methods for electrochemical properties (including permselectivity) to evaluate their efficacies in the RED system. Daniilidis et al. [13] demonstrated that membrane resistance and permselectivity limited the power output of RED cells at high salinities. The permselectivity was high at a high fixed charge density and decreased when the IEMs were exposed to high-concentration solutions due to the suppression of Donnan exclusion [14]. Tedesco et al. [15] optimized RED performance by identifying 0.1 M NaCl as the ideal concentration of the dilute feeding solution for maximizing the power output and energy efficiency. Beyond ionic transport dynamics, Zlotorowicz et al. [16] emphasized the importance of water transport in IEMs and suggested that to enhance the performance of RED, it was beneficial to had water transference coefficients as close to zero as possible. Additionally, Geise et al. [10] studied the specific ion effects on membrane potential and the permselectivity of four commercially available IEMs, and their results showed that counter-ion binding affinity with polymer fixed charge groups and co-ion charge density played an important role in determining the permselectivity of IEMs.

Regarding numerical modeling [17], Fan et al. [1] presented an IEM transport model that employs counterion condensation theory to analytically determine the experimentally inaccessible parameters of ion activity and diffusivity and elucidated the trade-off relationship between the permselectivity and conductivity as a function of the membranes’ properties. Zimmermann et al. [18] investigated the mechanism of partial ion permselectivity and stated the fixed charge density (FCD, defined as the milli-equivalents of charged groups per gram of water in the membrane) was a vital parameter for optimizing membrane permselectivity. Tedesco et al. [19,20] advanced ionic transport theories by predicting permselectivity for specific membranes. Many current models rely on simplifying assumptions (e.g., neglecting the convective effect), which compromises their applicability to real-world scenarios. Furthermore, comparative analyses between AEMs and CEMs are frequently overlooked in many experimental and theoretical studies [21,22,23]. Lots of existing frameworks assume identical operating conditions for both the AEMs and CEMs, thereby ignoring their distinct physicochemical interactions with ions—a simplification that risks misrepresenting their performances in hybrid energy conversion systems.

Regarding working solutions, the permselectivity is the figure of merit indicating the suitability of a salt in the RED unit [18]. Sodium chloride (NaCl) remains the benchmark for characterizing IEM transport properties due to its natural abundance and relevance to seawater desalination and RED applications [15,21]; emerging studies highlight the potential of alternative salts to enhance system performance. Recent advancements in RED-based energy conversion and hydrogen production have spurred interest in optimizing working fluids beyond conventional NaCl solutions. For instance, Giacalone et al. [24] demonstrated that potassium acetate (KAc), cesium acetate (CsAc), and lithium chloride (LiCl) significantly improve closed-loop RED efficiency compared to traditional NaCl electrolytes. Wu et al. [25,26] further validated KAc-methanol-water solutions as superior working fluids, leveraging their synergistic thermodynamic stability and electrochemical activity in RED systems. Their integrated RED–air-gap diffusion distillation (RED-AGDD) model successfully converted low-grade heat into hydrogen energy, establishing the optimal operational parameters for scalability. Similarly, Liu et al. [27] developed a lithium bromide (LiBr)-powered RED heat engine (REDHE), though its efficacy is currently constrained by low permselectivity—a critical technical challenge. Zimmermann et al. [18] analyzed the most promising REDHE system based on a comprehensive thermodynamic model, and as they thought, KNO₃ featured excellent ionic characteristics in terms of permeability through IEMs, while LiNO₃, LiBr, and LiCl had the potential to be high performing in the REDHE system in terms of hydrogen production.

Undeniably, the current research findings mentioned above have important academic value and serve as significant references. The thermodynamic properties of the emerging working fluids have been well characterized, while their electrochemical behaviors remain underexplored. This knowledge gap often leads to oversimplified assumptions in modeling and accurate property evaluation [28]. Such approximations risk inaccuracies in predicting the membrane characteristics and performance of an RED system, underscoring the need to prioritize electrochemical studies (especially permselectivity [29]) for next-generation working solutions.

To enhance the predictive accuracy of permselectivity models, the development of novel approaches is particularly crucial. Kingsbury et al. [17] introduced a new method integrating the Donnan–Manning theory and the Mackie–Meares diffusion model, incorporating key membrane properties such as water uptake and charge concentration. This method initially calculates the ion concentrations at each interface and assumes a linear distribution across the membrane. Subsequently, based on the spatial variation in ion concentrations, critical parameters including ion activity and diffusion flux profiles are computed, ultimately helping derive the membrane’s permselectivity. Although prior studies have tested select commercial membranes, the data remain limited both in terms of quantity and fluid diversity. To address this gap, the present study performs extensive experimental measurements, enriching the dataset while offering broader perspectives for evaluating permselectivity models.

In this study, several groups of non-sodium chloride salts that were promising for the RED system were selected as the experimental electrolytes, including (1) LiBr and KBr, as well as LiCl, KCl, and NH₄Cl, with the same anions but different cations; (2) LiCl and LiBr, as well as KCl, KBr, and KAc solutions with the same cations but different anions, respectively. Then, the authors comprehensively tested the permselectivity characteristics of two kinds of commercial membranes under diversified operational parameters. In the experiments, a custom-designed two-compartment electrochemical cell was engineered to quantify membrane permselectivity across multiple variable dimensions. Systematic measurements were conducted under controlled variations in solution concentration, ion species, and temperature gradient. The subsequent comparative analysis of the acquired dataset revealed the distinct trends in permselectivity evolution, particularly highlighting concentration-dependent threshold effects and temperature-mediated transport phenomena. Through further evaluating the AEMs and CEMs, this work elucidated the differential mechanisms governing co-ion exclusion efficiency and counter-ion transport dynamics. The findings establish correlations between the solution chemistry parameters and membrane performance metrics and provide empirical validation for predictive transport modeling in electrochemical separation processes.

2. Experimental System and Materials

2.1. Experimental System

As depicted in Figure 1, the permselectivity characterization experiment used a custom-designed two-compartment electrochemical cell constructed from polymethyl methacrylate (PMMA) plates, featuring symmetrical chambers separated by a membrane housing assembly, with each chamber possessing a volume of 18 mL. In each test, the tested membrane (Fujifilm Type 10 commercial membrane, Fujifilm, Tokyo, Japan) was positioned in the middle of the two compartments and was clamped by silicone sealing gaskets (thickness: 1 mm). The effective AEM (or CEM) area was tailored to 2 cm² and then subsequently installed into the electrochemical cell system for each test. The reference electrode employs an Ag/AgCl electrode filled with saturated KCl solution, interfaced with a digital multimeter for continuous potential monitoring. Prior to its use, it is immersed in 3 M KCl solution for 2 h to ensure a patent liquid junction and minimize the junction potential. Additionally, the distance from both Luggin capillaries to the membrane must be consistent (1 mm in this experiment). The diagram of the two-compartment electrochemical cell was shown in Figure 2. in which the red and blue indicate the positive and negative of the electrode, and the arrow direction is the direction of the solution flow. Prior to testing, the membranes underwent 24 h of equilibration in 0.1 M NaCl solution. During the operation, the peristaltic pumps circulated pre-thermostated electrolyte solutions at 15 mL/min through the flow channels, while a thermostatic bath maintained operational temperature stability (±0.05 °C). System equilibration was confirmed through real-time conductivity monitoring until <±1.5% variation over 15-min consecutive intervals was achieved, and each group of experiments was repeated at least 3 times. Membrane potential was measured after it stabilized, and inter-experimental flushing with deionized water was employed to eliminate ionic carryover effects. The specifications of instruments and related parameters of two-compartment electrochemical cell are shown in

Table 1. The specifications and features of the instruments used in the experiment.

Equipment	Model	Range	Accuracy
HC & LC peristaltic pumps (Longer Precision Pump Co., Ltd., Baoding, China)	BT300-2 J	0–300 rpm	0.1 rpm
Electronic balance (Changshu Shuangjie Testing Instrument Factory, Changshu, China)	JJ-1023BC	0–1020 g	±0.001 g
Magnetic stirrer (ASONE CORPORATION, Osaka, Japan)	RS-6DN	–	–
Conductivity meter (METTLER TOLEDO International Inc., Zurich, Switzerland)	Mettler 5 Easy Plus	0~500 mS·cm−1	±0.5%
Digital multimeter (Keithley Instruments Inc., Cleveland, OH, USA)	Keithley 2110	1 μV–750 V	120 ppm
Thermostatic water tank (JULABO Labortechnik GmbH, Seelbach, Germany)	Julabo Vivo iTherm-B5	20–95 °C	±0.05 °C
Reference electrode (Shanghai Yueci Electronic Technology Co., Ltd., Shanghai, China)	R303, Ag/AgCl	-	±2.5 mV

and

Table 2. Related parameters of two-compartment electrochemical cell.

Parameter	Value
Length	101.50 mm
Width	80.43 mm
Height	102.12 mm
Luggin capillary–membrane distance	1 mm
Flow rate	1.38 cm/s

, respectively

Permselectivity coefficients (α) were ultimately derived through normalized comparison of experimental potentials (ΔE_exp) against theoretical values (ΔE_theo) calculated via extended Nernst–Planck formalism, expressed as follows:

$$\alpha (\%)={\displaystyle \frac{ \frac{E_{exp}}{E_{th}}+1-2t_g}{2t_c}}\times 100\%$$

(1)

where α is the membrane permselectivity (%), ΔE_exp is experimental potential and ΔE_th is the theoretical membrane potential (mV), and t_g and t_c are the transport numbers of co-ions and counter ions, respectively. In this experiment, it is assumed that t_g = t_c = 0.5 [30].

$$\Delta E_{exp}=\Delta E_{mea}+\Delta E_j$$

(2)

where ΔE_mea is the measured membrane potential, and ΔE_j is the junction potential (mV). (The junction potential was ignored and the measured membrane potential was used to approximately replace the experimental potential in this study.)

$$\Delta E_{\mathrm{th}}={\displaystyle \frac{RT}{zF}}ln({\displaystyle \frac{{\gamma }_{HC}{C}_{HC}}{{\gamma }_{LC}{C}_{LC}}})$$

(3)

where R was the ideal gas constant (8.314 J·mol⁻¹·K⁻¹); T was temperature (K); z was the valence of ions; F was the Faraday constant (96,485 C·mol⁻¹); γ was the activity coefficient of ion [31]; and C was the concentration of solution (mol·L⁻¹) [32]. Detailed specifications of the IEMs from the manufacturer are shown in

Table 3. Specifications of membranes in the experiment.

Membrane Type	δ/μm	SD/%	AR/Ω·cm2	IEC/meq·g−1
AEM-Type 10	125	23	1.7	1.5
CEM-Type 10	135	21	2.0	1.67

.

2.2. Solutions

In the concentration-dependent permselectivity studies, the aqueous KAc, LiCl, and LiBr solutions were utilized with the concentration gradients spanning 3–5 M (as the concentrated solutions) and 0.02–0.2 M (as the dilute solutions). Additional ion-specific investigations incorporated the aqueous KBr, KCl, and NH₄Cl solutions under a fixed concentrated solution concentration (3 M) but with variable dilute solution concentrations (0.02–0.2 M). In the temperature dependence experiments, we employed the predefined salinity gradient pairs: the aqueous KAc (5 M/0.02 M) and LiBr (3 M/0.02 M) solutions. The purity of deionized water was verified through calibrated conductivity measurements. All the experimental salts were of analytical reagent (AR) grade or greater purity.

2.3. Pre-Experimental Result

Prior to formal testing, the preliminary validation experiments replicating the full experimental protocol were conducted using NaCl solutions (99.0% purity, Tianjin DaMao Chemical Reagent Ltd., Tianjin, China). The pre-experimental concentration gradient mirrored the manufacturer-specified conditions: 0.5 M (concentrated) and 0.05 M (dilute). As Figure 3 shows, the consistent transmembrane potential deviations of <2% are observed for both AEM and CEM configurations during the validation trials. The minimal systematic error across membrane types confirmed the operational reliability of the experimental setup and validated its suitability for the subsequent permselectivity characterization under the controlled electrochemical conditions.

3. Results and Analysis

3.1. Concentration Dependence of Permselectivity

The effect of the solution concentration on the permselectivity was examined experimentally, and the solution temperature was set to 298 K during the tests. Regardless of the results for the AEMs in Figure 4a and the CEMs in Figure 4b, the permselectivity shows a significant decline with the increase in the concentrations of either the concentrated solution or the concentration of the dilute solution. As for the CEMs in the 3 M concentrated solution, the permselectivity decreased from 78.4% to 65.6% with the increase in the concentration of the dilute solution from 0.02 M to 0.2 M. The minimum permselectivity of the CEM is 53.9%, which is achieved when the salinity gradient is 5 M/0.2 M at 298 K. As for the AEMs, their permselectivity is obviously higher than that of the CEMs and is much higher in the high-concentration solution. As seen in Figure 4a, the permselectivity of the AEMs can range from 95.9% to 92.8% under the condition of the 3 M concentrated solution. Even when the concentration of the concentrated solution is increased to 5 M, the highest obtained permselectivity is still 85%.

The experimental results are attributed to the following reasons. On the one hand, the decrease in the permselectivity with the increase in solution concentration can be explained by the decrease in the Donnan effect. At lower concentrated solution concentrations, the fixed functional groups within IEMs effectively exclude co-ions (K⁺ in AEMs and Ac⁻ in CEMs) while facilitating counter-ion (Ac⁻ in AEMs and K⁺ in CEMs) transport. However, with the increase in the concentration of the concentrated solution, too many ions will shield the fixed charge of the membrane, weakening the Donnan exclusion of the co-ions. Due to the higher charge density of K⁺, it is more susceptible to be shielded. Moreover, the high-concentration solution condition allows co-ions like hydrated K⁺ in AEMs and bulky Ac⁻ in CEMs to accumulate at the membrane interface, thereby physically obstructing counter-ion migration channels. The larger hydrated radius of Ac⁻ exacerbates pore clogging in the CEMs, resulting in a steeper permselectivity decline. On the other hand, the water transport number also plays a significant role in the observed osmotic selectivity trend. With the increase in solution concentration, the water transport number tends to increase due to the higher osmotic pressure driving more water molecules through the membrane. This increase in water flux can dilute the concentration of counterions near the membrane surface, thereby reducing the effective permselectivity. Moreover, the hydrated water molecules are usually carried by the migrating ions, which further increase the inner resistance and also lead to a decrease in the selective transmission coefficient. Furthermore, the alkaline environment generated by Ac⁻ hydrolysis destabilizes the membrane functionality: sulfonate groups in CEMs undergo partial deprotonation, and thus, the charge density is reduced. The AEMs exhibit better stability and higher permselectivity due to the use of the alkali-resistant quaternary ammonium groups.

The unfavorable effect on the permselectivity of IEMs owing to the increased dilute solution concentration is explained as follows: First, when the dilute solution concentration increases from 0.02 M to 0.2 M, the concentration gradient is reduced to one-tenth of its original value. This SGE reduction directly weakens the driving force for ion migration, slowing the migration rate of counterions and decreasing their flux. As previously noted, the charge screening in concentrated solutions can reduce the permselectivity of IEMs. The dilute solution side remains minimally affected by charge screening at lower concentrations. However, at the higher dilute solution concentration, the diminished effectiveness of Donnan repulsion allows for increased co-ion leakage, thereby reducing the overall permselectivity of IEMs [17]. Second, taking the CEM in Figure 4b as an example, when K⁺ migrates from the concentrated solution side to the dilute solution side, the accumulated K⁺ creates a positive potential that attracts Ac⁻ diffusion from the concentrated solution side to the dilute solution side. With the higher-dilute-solution-concentration condition, the elevated ionic strength on the dilute solution side intensifies this reverse diffusion, which could further counteract the Donnan exclusion effect. Additionally, the increased dilute solution concentration hinders rapid ion diffusion into the bulk solution, facilitating the formation of a high-concentration boundary layer on the interface between the IEM and the dilute solution side. This phenomenon exacerbates concentration polarization effects.

Figure 5 presents the concentration-dependent permselectivity of aqueous LiCl and LiBr solutions in different concentration conditions. The revealed trends partially align with but also diverge from the behavior observed in the KAc solution system. As Figure 5d shows, the permselectivity of LiCl for the CEMs decreases moderately from 81.6% to 74.2% in the 3 M concentrated solution, and LiBr exhibits a similar trend, declining from 85.5% to 73.9% in Figure 5b. Notably, the attenuation of permselectivity with increasing concentrated solution concentration is far less pronounced for these lithium salts compared to KAc. This discrepancy arises primarily from the smaller hydrated radii of Cl⁻ and Br⁻ relative to Ac⁻. The compact hydration shells of Cl⁻/Br⁻ reduce steric hindrance within membrane pores and minimize pore clogging even at the elevated concentrations, whereas the bulkier acetate ions form obstructive ion pairs that amplify the selectivity loss.

In stark contrast, under the experimental conditions, the anion-exchange membranes exhibit lower permselectivity in the LiBr solution (Figure 5a) and LiCl solution (Figure 5c). At high solution concentrations, the permselectivity plummets to 51.7% for LiBr and 51.4% for LiCl—a trend opposite to that of the KAc solution in Figure 4a, which shows a relatively stable performance. This difference can be traced to the unique hydration properties of Li⁺. Unlike K⁺, Li⁺ possesses exceptionally high hydration energy, which stabilizes a dense hydration layer. The expanded hydration shell occupies a significant pore volume in AEMs. It physically obstructs anion migration and weakens Donnan exclusion by altering the dielectric environment around the fixed quaternary ammonium groups. Crucially, the hydrophobic methyl group of acetate enables it to bypass K⁺ hydration barriers through non-electrostatic interactions, whereas the hydrophilic nature of Cl⁻/Br⁻ exacerbates their susceptibility to interference from Li⁺. These findings underscore the critical interplay between ion-specific hydration behavior and membrane structure in governing permselectivity under hypersaline conditions, which provides a mechanistic basis for improving the membranes to some specific ionic environments in RED and ED processes.

3.2. Ion Species Dependence of Permselectivity

3.2.1. The Effect of Cations

Table 4. Hydration characteristics and mobility of cations and anions [33,34,35].

Ions	Hydrated Radius (nm)	Hydration Energy (kJ/mol)	Mobility in Water (10−8 m2/sV)
K+	0.331	−295	7.19
Li+	0.382	−475	4.01
NH4+	0.331	−285	7.63
Cl−	0.332	−340	7.91
Br−	0.330	−315	8.09
Ac−	0.425	−375	4.10

shows the hydration characteristics and mobility of cations and anions [33,34,35]. Figure 6 shows the results of investigating the role of cation species in modulating the performance of IEMs through systematic comparisons of the LiBr and KBr solutions. For the CEM condition, K⁺ demonstrates higher permselectivity (79.1–85.7%) than Li⁺ (73.9–83.5%) under an identical solution temperature and concentration. The enhanced performance of K⁺ stems from its smaller hydration radius and lower hydration energy [18], which reduce steric hindrance and viscous drag during ionic migration. Additionally, the weaker electrostatic interaction between K⁺ and fixed sulfonate groups in CEMs minimizes retardation; thereby, it facilitates the ionic transport process. In contrast, Li⁺, with its compact charge distribution and stronger hydration shell, exhibits lower mobility.

For the AEM conditions shown in Figure 6, the permselectivity of Br⁻ in KBr solutions exceeds that in LiBr by approximately 10%. This divergence arises from cation-dependent interfacial dynamics: the rapid migration of K⁺ mitigates concentration polarization at the membrane surface, preserving the Donnan exclusion efficiency through sustained ionic gradients. The smaller hydration volume of K⁺ also minimizes pore obstruction, ensuring efficient Br⁻ transport. Conversely, the high charge density of Li⁺ strengthens interactions with quaternary ammonium groups in AEMs, partially neutralizing fixed charges and diminishing Donnan exclusion. This charge screening effect, combined with the tendency of Li⁺ to form hydration clusters near the membrane interface, disrupts anion-selective pathways and exacerbates Br⁻ leakage [36].

To further elucidate cation-specific impacts on permselectivity, the performances of three chloride-containing solutions (LiCl, KCl, and NH₄Cl) were systematically evaluated. As shown in Figure 7, the permselectivity trend for CEMs mirrored those observed in bromide systems: NH₄⁺ exhibited the highest permselectivity (up to 91.1%), followed by K⁺ and Li⁺. This phenomenon arises from the synergistic effects of hydration properties, molecular geometry, and kinetic behavior. The tetrahedral structure of NH₄⁺ and its smaller hydration radius jointly enhance its transmission efficiency in the membrane’s three-dimensional pore network. Compared with single-atom ions (such as Li⁺ and K⁺), the symmetrical geometric configuration of NH₄⁺ allows it to adjust its spatial orientation more flexibly, adapting to the irregular shape of membrane pores, thereby reducing steric hindrance during migration [36]. At the same time, its lower hydration energy (compared to Li⁺) weakens the viscous resistance of water molecules to migration, while its higher diffusion coefficient further improves its dynamic response ability [19]. This structural-kinetic synergistic effect allows NH₄⁺ to exhibit a better migration performance in complex pores, consistent with its high permselectivity in CEMs.

In AEMs, the permselectivity of Cl⁻ follows the same cation-dependent order (NH₄⁺ > K⁺ > Li⁺), as Figure 7 presents. High-permselectivity cations like NH₄⁺ migrate rapidly away from the membrane interface, reducing local ionic crowding and preserving the Donnan exclusion efficiency. This dynamic alleviates obstruction to Cl⁻ transport, whereas the slower-migrating Li⁺ exacerbates polarization, weakening Cl⁻ selectivity. The consistency across the aqueous chloride and bromide salts systems underscores a universal mechanism: the cation mobility and hydration characteristics govern the interfacial ion distribution through concentration polarization, indirectly affecting the permselectivity of anions.

3.2.2. The Effect of Anions

As exhibited in Figure 8, the distinct permselectivity behaviors observed in LiBr and LiCl solutions underscore the role of anion characteristics in governing ion transport dynamics. For CEMs, Li⁺ exhibits a marginally higher permselectivity in LiBr (a 1–3% enhancement over LiCl), which is related to a subtle disparity rooted in its anion-dependent and ion-pairing tendencies. The lower charge density of Br⁻ weakens its electrostatic coupling with Li⁺, thus increasing the available number of free Li⁺ ions and improving the migration efficiency. In contrast, the stronger interaction of Cl⁻ with Li⁺ reduces the number of free cations, leading to a slightly diminished Li⁺ permselectivity for CEMs. Analogous trends were observed in the previous sections. In addition to the aforementioned factors, conductivity emerges as another critical determinant. LiCl solutions typically exhibit a lower conductivity compared to LiBr solutions, which renders LiCl more susceptible to pronounced concentration polarization and consequently manifests as reduced permselectivity.

For AEMs in Figure 8, the higher charge density of Cl⁻ amplifies its binding to fixed quaternary ammonium groups, necessitating greater desorption energy and thus leading to lower transport efficiency. Simultaneously, this high charge density facilitates the formation of tight ion pairs between Li⁺ and Cl⁻ due to the strong electrostatic interactions, reducing the proportion of free Cl⁻ ions in the solution. Consequently, the effective migration concentration of Cl⁻ decreases and results in lower permselectivity. Additionally, Br⁻ exhibits lower hydration energy and a smaller hydrated radius compared to Cl⁻, along with greater ionic mobility. Its lower hydration energy enables Br⁻ to shed the surrounding water molecules more effectively during migration. This leads to fewer carried water molecules and reduced transport resistance through the membrane pores.

Then, the permselectivity of IEMs was further examined using three different potassium salt (KCl, KBr, and KAc) solutions. As Figure 9 shows, the observed order of permselectivity of the AEMs is Ac⁻ > Br⁻ > Cl⁻. The trend of Br⁻ and Cl⁻ aligns with the previously noted behavior in lithium salt solutions but exhibits higher values. It is primarily due to the cation effects as the above section discussed. Ac⁻ exhibited the highest permselectivity coefficient among the three anions, reaching up to 95.9%, which can be attributed to the following factors: First, the negative charge of Ac⁻ is distributed across the two oxygen atoms of the carboxylate group, resulting in the lowest charge density. This minimizes the Donnan exclusion with fixed positive charges in the IEMs, thereby reducing the ionic migration resistance. In contrast, both Br⁻ and Cl⁻ require higher dissociation energy for migration. Second, the presence of the methyl group (-CH₃) in Ac⁻ allows it to migrate through hydrophobic pathways if the membrane contains hydrophobic microdomains (e.g., silicone-modified layers [37]). Additionally, the low charge density of Ac⁻ leads to weaker electrostatic interactions with fixed positive charges in the IEMs, resulting in a higher proportion of free Ac⁻ ions in the solution and a stronger driving force for migration [38].

For CEMs, the following permselectivity patterns are observed: the overall order of K⁺ permselectivity across the three solutions is KBr > KCl > KAc. The trends in KBr and KCl solutions are consistent with those observed in the lithium salt systems above; however, K⁺ exhibits markedly reduced permselectivity in the KAc solution, with a particularly significant decline as the dilute solution concentration increases, reaching a minimum value of 58.9%. This anomalous behavior can be attributed to multiple mechanisms. The delocalized negative charge of Ac⁻ across its carboxylate oxygen atoms facilitates the formation of stable K⁺–Ac⁻ ion pairs through multiple binding sites. As the dilute solution concentration increases, the proportion of ion-paired K⁺–Ac⁻ in the solution rises substantially. However, the concentration of free K⁺ ions decreases sharply, reducing both the driving force for K⁺ migration and the mobility of the ion pairs due to their enlarged effective size. Furthermore, Ac⁻, repelled by the fixed negative charges in CEMs, accumulates at the membrane surface due to its bulky structure, forming a dense barrier layer that severely obstructs K⁺ migration [39], with this obstruction intensifying at higher concentrations. Additionally, Ac⁻ is adsorbed onto membrane surfaces or within pores via hydrophobic interactions. The elevated concentration enhances this adsorption and decreases the effective pore size, physically blocking K⁺ transport pathways. The organic nature of Ac⁻ also promotes membrane surface fouling, further compromising the migration efficiency. In contrast, the inorganic anions Br⁻ and Cl⁻ exhibit minimal interfacial interference due to their smaller hydration radii and lack of organic fouling tendencies, allowing K⁺ to migrate smoothly through both the solution–membrane interface and the membrane matrix.

3.3. Temperature Dependence of Permselectivity

To investigate the temperature dependence of permselectivity, systematic measurements were performed for both AEMs and CEMs using aqueous KAc and LiBr solutions across varying temperatures. As shown in Figure 10, the experimental result reveals a characteristic temperature response pattern: the measured membrane potential exhibited a progressive enhancement with an elevation in temperature, while this increasing trend gradually plateaued with the higher thermal condition. Notably, the permselectivity shows relatively minor variations between 298 and 323 K, with both the CEM and AEM conditions achieving the optimal permselectivity at 308 K (CEM: 85.8%; AEM: 60.6%), followed by a decline. However, when the measured membrane potential is high, the relative error in the calculated permselectivity also increases; thus, the maximum value is not pronounced. The result indicates the permselectivity shows a non-monotonic temperature dependence pattern.

The phenomenon in Figure 10 arises from two competing mechanisms influencing membrane behavior. First, the elevation in temperature generates dual beneficial effects: ionic mobility is enhanced through increased diffusion coefficients owing to the reduction in solution viscosity [40], while the charge transport efficiency can be improved, resulting from the diminished concentration polarization at the membrane interfaces. However, excessive thermal energy induction gives rise to the structural alterations in membrane matrices. For example, under the high-temperature conditions, according to the Arrhenius equation, the motion of polymer chains becomes more frequent and unrestricted. This enhanced mobility weakens the interactions between polymer chains, causing the structure of the cross-linked network to become more relaxed. Such relaxation phenomena lead to the enlargement of non-selective micropores, thereby reducing the permselectivity of the membrane. The observed maximum performance at 308 K reflects an optimal equilibrium between enhanced ionic transport kinetics and maintained membrane structural integrity, beyond which detrimental matrix reorganization effects outweigh the mobility benefits.

The data in Figure 11 reveal that the parallel experiments using the aqueous LiBr solution also exhibit temperature-dependent permselectivity patterns fundamentally consistent with the KAc systems reported above. The distinct membrane-specific responses are displayed. Both CEMs and AEMs demonstrate characteristic performance curves with an initial enhancement succeeded by a deterioration upon the elevation in temperature. This non-monotonic behavior parallels the “diffusion kink” phenomenon observed in hydroxide transport studies [41], in which competing thermal activation and structural degradation mechanisms regulate ionic mobility. Notably, for AEMs, it manifests accelerated permselectivity decline (75.1–71.3%) between 318 and 323 K with a distinct transmembrane potential inflection point—a feature that is absent in CEMs. This accelerated degradation primarily stems from the functional group thermolysis, as the AEMs’ charge-anchoring quaternary ammonium groups exhibit markedly higher thermal lability compared to CEMs’ sulfonic acid moieties [42], which is due to their propensity to undergo chemical changes such as the Hofmann elimination reaction and nucleophilic substitution reactions at high temperatures, promoting its degradation. Progressively, the thermal decomposition will reduce the effective charge density, weakening Donnan exclusion effects and Br⁻ transport driving forces.

The phenomenon of permselectivity peaking and then declining with rising temperature is closely related to the competing thermal effects on aqueous/salt transport dynamics within the membrane matrix [43]. The initial temperature elevation enhances salt transport capacity through improved ion mobility, while increased water uptake facilitates ionic migration, collectively boosting permselectivity. However, excessive thermal energy induces membrane swelling, which disrupts this optimization—excessive swelling induces pore overexpansion, generating non-selective micropores that promote parasitic transport pathways, ultimately degrading the permselectivity.

Membrane swelling also plays a critical role in permselectivity degradation. As the temperature increases beyond 318 K, swelling ratios cause pore dilation and charge dilution, facilitating co-ion intrusion and reducing the Donnan exclusion efficiency; the effect of this phenomenon at high concentrations will be amplified [44]. This structural change explains the accelerated permselectivity loss observed experimentally.

The characteristic temperature-dependent permselectivity profile of ion-exchange membranes—an initial enhancement followed by progressive deterioration—arises from competing thermal mechanisms. Below critical temperature thresholds, augmented ionic thermal motion outweighs structural degradation impacts, yielding a progressive permselectivity enhancement. However, thermal excursions beyond these limits reverse this dynamic, and membrane matrix reorganization precipitates a marked permselectivity deterioration. These results establish system-specific optimal temperature windows governed by membrane architecture and material characteristics. The strategic optimization of key membrane parameters (water uptake and fixed charge density) emerges as an effective approach to broaden operational temperature ranges and amplify energy conversion efficiency across thermal gradients.

4. Conclusions

In this study, the influences of solution concentration, ion species, and solution temperature on the permselectivity of membranes were experimentally investigated. The obtained data were evaluated and analyzed, and the following conclusions are drawn:

(1)

The permselectivity of IEMs significantly decreases with the increase in concentration of either the dilute solutions (0.02–0.2 M) or the concentrated solutions (3–5 M). This decline is attributed to the weakening of the Donnan effects and water transport. If ions with large hydration radii and low mobility aggregate in the membrane boundary layer on the concentrated solution side, their corresponding oppositely charged ions are significantly affected by the increase in the concentrated solution;

(2)

The effect of ion species on permselectivity is closely related to the hydration radius and hydration energy. Ions with lower hydration energy can more effectively remove surrounding water molecules, resulting in less transmembrane transport resistance and higher permselectivity. Ions with a higher charge density exhibit poorer permselectivity due to tighter binding with membrane fixed groups during ionic transport. Ions with low mobility or diffusion coefficients are strongly affected by concentration polarization, which impairs their selective permeability and that of their counterions;

(3)

Under moderate solution temperature conditions, the thermal effects on permselectivity are minimal, with ion thermodynamic properties predominantly determining performance. The permselectivity increases slightly with a rise in temperature. However, when the temperature exceeds 318 K, membrane deterioration becomes dominant, significantly reducing the permselectivity of IEMs. Therefore, an optimal operating temperature range exists.