Fabrication of Hollow Fiber Loose Nanofiltration Membrane via Metal-Organic Bonding and PA-PEI Dual Coating for Superior Dye/Salt Separation

Abstract

The efficient removal of dyes and separation from dissolved salts are crucial for the recovery of valuable resources from saline textile wastewater. In this study, hollow fiber membranes were fabricated using the non-solvent-induced phase separation (NIPS) method and then improved with a dual-coating process to create effective nanofiltration (NF) membranes. First, hollow fiber substrates with Fe³⁺ were fabricated using NIPS. Subsequently, the inner surface of the membrane was coated with phytic acid (PA) and polyethyleneimine (PEI), which increased the thickness of the separation layer and reduced the size of the surface pores, thereby improving the separation efficiency. The loose NF membrane exhibited superior water permeance (pure water permeability of 280 L·m⁻²·h⁻¹·bar⁻¹) and, with dye rejection rates consistently exceeding 95%, also remarkable dye/salt selectivity (with separation factors of CR/NaCl: 64.08, CR/Na₂SO₄: 21.21, CBB/NaCl: 14.75, and CBB/Na₂SO₄: 10.74). The flux recovery of the membrane was over 80% for humic acid, and the membrane exhibited favorable stability under acidic and alkaline conditions, confirming its excellent antifouling and stability performance. In conclusion, this study presents a straightforward and effective approach for fabricating hollow fiber loose NF membranes, underscoring their potential for treating hypersaline wastewater and resource recovery.

1. Introduction

With the rapid advancement of the textile industry, it has become the second most polluting sector in China, following the petroleum industry. The processes of textile printing and dyeing have emerged as critical areas of focus for environmental protection and management due to their substantial water consumption and significant wastewater discharge [1,2,3,4]. The complex composition of dyes, characterized by their stable chemical structures and persistent coloration, adversely affects photosynthesis in aquatic ecosystems and bioaccumulates in organisms, resulting in significant impacts on natural water bodies [5,6,7]. Furthermore, wastewater from textile printing and dyeing processes, particularly that from cotton dyeing, contains significant quantities of inorganic salts. The presence of these salts in aquatic environments adversely impacts the efficiency of wastewater treatment systems and escalates operational expenses. Additionally, it results in the depletion of valuable inorganic salt resources [8,9,10]. From the perspective of sustainable development, the efficient screening of dyes and salts, along with their recycling, has significant theoretical and industrial value for the green and sustainable development of related industries.

Membrane separation technology, particularly nanofiltration (NF), is highly regarded for its role in seawater desalination and wastewater reclamation because of its straightforward operation, low energy consumption, and superior performance [11,12,13,14]. However, conventional dense NF membranes face challenges in treating high-salinity wastewater due to their high salt rejection, which exacerbates osmotic pressure and hinders the efficient separation of organic compounds from inorganic salts. These issues compromise the separation efficiency in dye-containing effluents and obstruct effective resource recovery [15,16]. In particular, they fail to enable the selective separation of organic dyes and inorganic salts, making the regeneration of saline resources impractical. To address these drawbacks, loose nanofiltration (LNF) membranes have emerged as promising alternatives [17,18,19,20]. With their relatively porous architecture, LNF membranes deliver substantially improved permeate flux alongside precise dye-salt selectivity, efficiently capturing dye molecules while allowing low molecular weight salts to permeate [21,22,23]. Extensive research in this area has yielded impressive dye-salt separation outcomes. Most of these LNF membranes are flat sheets, and their fabrication processes are relatively complex and energy-intensive [24]. In addition, commercially available flat-sheet NF membranes, such as those in spiral-wound modules, continue to suffer from acute fouling problems, driven mainly by the elevated levels of dyes, organic compounds, and salts in wastewater [25]. This leads to swift reductions in membrane flux, the need for repeated cleaning, and challenges in maintaining prolonged operational stability, ultimately escalating cost and curtailing widespread industrial adoption.

Hollow fiber nanofiltration (HF NF) membranes offer substantial benefits owing to their distinctive geometric and material attributes, such as an elevated surface area-to-volume ratio, reduced susceptibility to fouling, capacity to handle high suspended solid concentrations, and suitability for hydraulic backwashing and air sparging techniques [26,27,28,29]. These features collectively minimize the need for pretreatment and prolong membrane durability. The primary fabrication techniques for HF NF membranes include phase inversion [30,31], interfacial polymerization [32,33], dip-coating [34,35], grafting [36,37], and polyelectrolyte layer-by-layer (LbL) assembly [38,39,40]. Among these, the LbL assembly approach enables precise modulation of the thickness, composition, and architecture of the separation layer, thereby permitting the development of NF membranes with tailored separation properties, and has gained considerable traction in the field [39]. Initial investigations primarily relied on electrostatic interactions among oppositely charged polyelectrolyte moieties to form the separation layer, highlighting the promising merits of HF NF membrane synthesis. However, polyelectrolyte LbL assembly typically entails iterative alternating depositions, which prolongs the assembly duration and complicates procedural workflows. Moreover, the enduring stability of polyelectrolyte-based separation layers poses a barrier to their implementation. Considering the inherent morphology of hollow fiber membranes, numerous researchers have initially engineered reactive sites on the substrate membrane, followed by LbL assembly via coordination, crosslinking, or analogous strategies to construct the NF separation layer, thus broadening the repertoire of HF NF membrane preparation methodologies. For instance, incorporating sulfonated PES into a polyamide dope solution and employing dry-wet spinning yielded a hollow fiber support membrane with a negatively charged surface [41]. Subsequent crosslinking with hyperbranched polyethyleneimine (PEI) refined the pore dimensions and surface charge characteristics of the membrane. The resultant HF NF membrane displayed a pure water flux ranging from 7.0 to 71.2 L·m⁻²·h⁻¹·bar⁻¹, dye retention exceeding 95.5%, and Na₂SO₄ rejection below 10%, highlighting its superior efficacy in dye-salt fractionation.

Metal–organic coordination offers a non-traditional interaction mechanism beyond conventional covalent bonding and is widely applied in membrane fabrication [42]. In contrast to widely used metal–phenol coordination systems [43,44,45], where phenol ligands often possess limited electron-donating ability, leading to compromised membrane integrity and acid–base stability [46], organophosphorus ligands demonstrate markedly stronger electron-donating capacity. This property facilitates the formation of metal–organic coordination bonds with high hydrolytic stabilities [47,48]. Phytic acid (PA), a naturally occurring organic molecule featuring six phosphate groups, exhibits strong electronegativity, which promotes effective chelation with metal ions, yielding complexes with remarkable stability and resistance to hydrolysis [49,50,51,52]. Abundant in oxygen-containing ligands, PA also offers excellent hydrophilicity and environmental compatibility, making it a promising candidate for membrane fabrication and modification applications [53,54,55]. In the context of membrane separation, several notable studies have investigated PA–metal complexes. For example, You et al. [56] utilized hydrogen bonding to self-assemble a molecular layer from hydrolyzed polyacrylonitrile substrates and PA solutions, followed by the introduction of transition metal ions to construct metal–phytic acid coordination networks. Despite these advances, the construction of a dense nanofiltration separation layer on a highly curved hollow fiber membrane presents considerable challenges. The innovative development of a metal-organic coordinated molecular layer on a hollow fiber substrate, followed by the formation of a hollow fiber nanofiltration membrane, offers a promising solution to this issue. However, few studies have focused on metal-organic coordinated molecular layers to construct HF NF membranes for water treatment.

In this study, a novel method for constructing hollow fiber loose nanofiltration membranes is proposed, which involves blending organometallic compounds with coupling and layer-by-layer assembly of coordination monomers. Organometallic compounds are incorporated into the polyethersulfone (PES) membrane matrix via phase inversion to provide a stable host for metal coordination, which subsequently assembles with guest molecules in the coating solution to form a nanofiltration separation layer. Specifically, iron acetylacetonate was incorporated into the casting solution as an iron source and uniformly distributed on the surface of hollow fiber substrate membranes via non-solvent-induced phase separation (NIPS). The membranes were subsequently modified using a dual-coating process. The strongly negatively charged phytic acid (PA) complexed with Fe³⁺ to form a stable metal–organic phosphate coordination, which not only resulted in a super hydrophilic metal–PA complex but also provided additional crosslinking sites for subsequent binding with the cationic polymer polyethyleneimine (PEI). The introduction of PEI contributed additional positive charges to the membrane surface, enhancing its surface electrostatic properties and thus strengthening the electrostatic repulsion of solute ions. The impact of coating the membrane with PA and PEI on its structure and characteristics was investigated. In addition, the use of the obtained membrane for the treatment of simulated textile wastewater was explored. By combining the NIPS method with the dual-coating process, an HF LNF membrane with high permeability and excellent dye/salt separation performance was successfully fabricated. This approach is straightforward, cost-effective, and readily scalable for industrial production, demonstrating its considerable potential for practical applications.

2. Experimental

2.1. Materials and Reagents

Polyether sulfone (PES, BASF, E6020P); polyethylene glycol PEG (Mw = 20,000 g/mol); N,N-dimethylformamide (DMF, >99.9% (GC) Aladdin); Iron acetylacetonate (>98%, Aladdin); phytic acid (PA, 70% aqueous solution, Mw = 70,000); Polyethyleneimine (PEI, 50% aqueous solution, Mw = 70,000); Acid Orange 74 (AO74, M.W. 493.35), Congo red (CR, M.W. 696.66), Eriochrome Black T (EBT, M.W. 461.38), Direct yellow 44 (DY44, M.W. 634.53), Methyl blue (MB, M.W. 799.80), Coomassie brilliant blue (CBB, M.W. 854.02), Acid Fuchsin (AF, M.W. 585.54); NaCl (purity ≥ 99%), Na₂SO₄ (purity ≥ 99%), MgCl₂ (purity ≥ 99%), MgSO₄ (purity ≥ 99%); Humic Acid (HA); deionized water used in the experiments was taken from a laboratory ultrapure water machine.

2.2. Preparation of PES/Fe Based HF Membranes

PES/Fe HF membranes were fabricated using a custom-designed wet and dry spinning apparatus (Figure 1). The spinning solution comprising 20 wt% PES polymer and 10 wt% PEG was dissolved in DMF, along with 3% iron acetylacetonate. This mixture was mechanically stirred at 60 °C for 6 h. Subsequently, the spinning solution was cooled and maintained at a constant temperature of 28 °C for 12 h to eliminate air bubbles. The spinning solution and core liquid (ultrapure water) were co-extruded through a dual-orifice nozzle and solidified at the outlet to form initial fibers. The flow rates for the casting solution and core liquid were set at 10 and 20 mL/h, respectively, with a distance of 21 cm before entering the water-solidification bath. Finally, the prepared PES/Fe HF membrane was stored in ultrapure water for preservation.

2.3. Preparation of PES/Fe-PA-PEI LNF Membrane

The LNF membranes underwent post-treatment using a dual-coating process. The PES/Fe HF membranes were subjected to a vertical drainage process to remove moisture from both their inner and outer surfaces. Subsequently, the PES/Fe HF membranes were immersed in an aqueous PA solution (0–10 g/L) for 30 min. The modified membranes were then rinsed with pure water for 2 min and placed vertically to drain the excess solution from the cavities. After removing the residual liquid, the PA-treated HF membrane was immersed in a PEI aqueous solution (2.5–10 g/L) for 30 min. The pH of the PA solution was 2.4, and that of the PEI solution was 10. Finally, the treated membranes were rinsed again with pure water for 2 min and immersed in a 50% glycerol aqueous solution for 24 h. Subsequently, the membrane sheet was secured inside a PVC tube using an epoxy adhesive and connected to a diaphragm pump to evaluate its filtration performance. All experiments were conducted at 25 °C.

2.4. Characterizations of the Membrane

The functional groups and chemical structure of the film were investigated using attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR, Spectrum Two, PerkinElmer, Waltham, MA, USA) and X-ray photoelectron spectroscopy (XPS, AXIS Ultra DLD, Shimadzu, Kyoto, Japan). The surface morphology and roughness of the film were examined using scanning electron microscopy (SEM, S-4500, Hitachi, Tokyo, Japan) and atomic force microscopy (AFM, Nano Wizard II, Bruker AXS, Baden-Württemberg, Germany), respectively. Prior to testing, the wet films were vacuum-dried in a freeze dryer for 8 h to preserve structural integrity. To observe the cross-sectional structure of the films, brittle fracture was induced using liquid nitrogen to ensure structural integrity. The samples were gold-sputtered prior to testing to enhance their conductivity [57]. The surface hydrophilicity of the films was determined using a contact angle measuring instrument (WCA, SL 200KS, Kino, Shanghai Solen Information Technology Co., Ltd., Shanghai, China) via the hanging drop method [58]. The zeta potential of the film surface was measured using a Surpass™ 3 (Anton Paar, Styria, Austria) and a series of KCl solutions (1.0 mmol/L), with electrolytes in the pH range of 3–10 selected as background electrolytes [59]. For each sample, at least five different locations were measured using the above tests.

2.5. Separation Performance Evaluation

The separation performance of the LNF membrane was measured using a cross-flow filtration apparatus. The flow rate of the cross-flow filtration apparatus was 444.88 mL/min, and the calculated effective membrane area was approximately 4.28 cm². To maintain stable permeate flow during the test, all membrane samples were pre-pressurized at 2 bar for 30 min prior to collecting the permeate, after which the formal test began at a pressure of 0.1 MPa. All the tests were conducted at 25 °C.

The permeate flux of the nanofiltration membrane is calculated by the following formula [15]:

$$J={\displaystyle \frac{V}{A∆t}}$$

(1)

where J represents the pure water flux of the membrane at 0.1 MPa, L/(m²-h); V represents the pure water permeate volume of the membrane at a certain time, L; A represents the effective area of the tested membrane, m²; ∆t represents the time used to test the pure water flux, h.

The filtration and permeation performance of the nanofiltration membrane on dyes and salt was evaluated with dye solution (0.1 g/L AO74, EBT, AF, DY44, CR, MB, CBB) and salt solution (1 g/L NaCl, Na₂SO₄, MgCl₂, MgSO₄) as the feed solution, respectively, and the operating pressure was 0.1 MPa. The retention rates of the dye solution and the salt solution, R, can be calculated by the following equation [60]:

$$R=\left(1-{\displaystyle \frac{C_p}{Cf}}\right)\times 100\%$$

(2)

where R represents the retention rate of dye, %; C_p represents the concentration of leachate, g/L; Cf represents the concentration of feed solution, g/L. The concentration of the dye can be expressed by absorbance, which was measured in this study using a UV-visible photometer (UV-2600, Shimadzu, Kyoto, Japan), and salt concentrations were determined by a conductivity meter (FE38, METTLER, Canton of Zurich, Switzerland). The wavelength range for UV testing is 200–700 nm. Before conducting UV testing, a baseline correction was performed using pure water, followed by absorbance measurements. During the process of the separation test of the fabricated nanofiltration membrane, the pure water was allowed to permeate for about 30 min until its performance was stable, and then the dye/salt solution was poured into the cell, and the rejection of dye/salt was measured after 1 h. The data were recorded every 10 min to calculate the average value after stabilization.

The separation performance of the membranes for dye/salt mixtures was evaluated using a cross-flow filtration device under the same environmental conditions as described above. Different mixtures of solutions were used during the tests, including CR/NaCl, CR/Na₂SO₄, CR/MgCl₂, CR/MgSO₄, and the cationic dye CBB to NaCl, Na₂SO₄, MgCl₂, and Mg₂SO₄ mixtures. The concentrations of the dyes and salts were 0.1 g/L and 1 g/L. The membranes were similarly pre-pressurized with DI prior to testing. The dye/salt selectivity (S) was calculated by the following equation [61].

$$S={\displaystyle \frac{R_{dye}}{R_{salt}}}$$

(3)

Here, S is the dye/salt selectivity; $R_{dye}$ is the dye rejection; $R_{salt}$ is the salt rejection.

2.6. Operational Stability and Anti-Pollution Test of Membrane

The operational stability of the prepared membranes for the separation of organic compounds was evaluated through continuous operation over a 24 h period, utilizing CR (0.1 g/L) and Na₂SO₄ (1 g/L) as the feed solution. The permeate flux and solute retention rate were measured at hourly intervals. Concurrently, tests were conducted under varying operating pressures to observe changes in the pure water flux of the composite membrane, thereby determining the membrane’s maximum pressure tolerance. The membrane filtration was conducted for six cyclic experiments, with data points collected every 30 min, and three groups constituting one cycle. The anti-fouling properties of the composite membrane were assessed using the flux recovery ratio (FRR) as the parameter index [62].

$$FRR={\displaystyle \frac{J_2}{J0}}\times 100\%$$

(4)

$$R_t=\left(1-{\displaystyle \frac{J_1}{J_0}}\right)\times 100\%$$

(5)

$$R_{ir}=\left({\displaystyle \frac{J_0-J_2}{J_0}}\right)\times 100\%=1-FRR$$

(6)

$$R_r=\left({\displaystyle \frac{J_2-J_1}{J_0}}\right)\times 100\%=R_t-R_{ir}$$

(7)

To assess the anti-contamination efficacy of the membrane, an organic contaminant, HA solution, was employed. Initially, deionized water was passed through at a pressure of 2 bar for 1 h until the permeate flux reached stability, after which the initial membrane water permeate flux (J₀) was measured. Subsequently, contamination experiments were conducted for 90 min using a 1000 ppm aqueous HA solution as the feed solution, and the stable permeate flux value (J₁) was recorded. The contaminated membrane was then rinsed with deionized water for 30 min, and its pure water flux (J₂) was measured. This procedure was repeated for a total of three cycles. The flux recovery ratio (FRR), the total flux decline ratio (R_t), irreversible fouling ratio (R_ir), and reversible fouling ratio (R_r) were calculated using established formulae to evaluate the cleaning efficiency and contamination resistance of the PES/Fe-PA-PEI composite membranes. To determine the acid and alkali resistance of the PES/Fe-PA-PEI composite membranes, the modified membranes were immersed in HCl and NaOH solutions with pH values of 3, 5, 7, 9, and 11, respectively, for 24 h. The membranes were then subjected to three successive washing cycles in deionized water for 30 min to remove any residual HCl and NaOH solutions. Subsequently, membrane performance was assessed using a cross-flow filtration device with a mixture of 0.1 g/L CR and 1 g/L Na₂SO₄ under an operating pressure of 1 bar.

3. Results and Discussion

3.1. Characterization of Membrane

The surface functional layer plays a critical role in determining the selectivity and permeance of NF membranes. The morphology and microstructure of the functional layer of the membranes were examined via SEM, as shown in Figure 2. The PES/Fe membrane exhibited relatively large surface pores and lacked a distinct selective layer. In contrast, both the PES/Fe-PA and PES/Fe-PA-PEI membranes featured clearly formed functional layers, distinctly different from those of the unmodified pristine membrane. Cross-sectional SEM analysis revealed that the functional layer of the PES/Fe-PA membrane had a thickness of ~160 nm. After PEI impregnation and coating, the thickness of the functional layer of the PES/Fe-PA-PEI film increased to ~270 nm. Furthermore, surface SEM imaging demonstrated that the PES/Fe membrane exhibited distinct pores. The incorporation of PA and PEI resulted in significant microstructural reorganization, leading to gradual densification of the surface and reduction in pore size. Increasing the thickness of the functional layer and reducing the pore size on the membrane surface will enhance the effective rejection of dyes [57].

To further assess the membrane composition and verify the successful formation of the selective layer on the membrane substrate, the functional groups and chemical structures of both the PES/Fe membrane and modified membrane were examined using XPS and FTIR, as shown in Figure 3. The XPS spectrum (Figure 3a) of the PES/Fe-PA membrane revealed the presence of phosphorus, which was attributed to phosphate groups, indicating that PA had been successfully incorporated [63]. As shown in Figure 3c, after modification with PEI, the composite membrane displayed an N peak. In the high-resolution XPS O spectra of the PES/Fe-PA-PEI membrane (Figure 3b), characteristic peaks of P-O, P=O, and O-H bonds were observed, and the ammonium peak was fitted to the N spectrum (Figure 3c), further corroborating the successful bonding between PA and PEI. Further analysis from the perspective of elemental proportions on the membrane surface revealed that the PES/Fe base membrane contained 4.77% sulfur, whereas no S element was detected in either the PES/Fe-PA or PES/Fe-PA-PEI membranes (Table S1 in SI). This may be attributed to the introduction of PA and PEI coatings. In the FTIR spectra (Figure 3d), the peaks at 1095 cm⁻¹ and 997 cm⁻¹, compared with the PES base membrane, were primarily attributed to the stretching vibrations of the P=O and P-O-C groups in PA, and the peak at 2903 cm⁻¹ confirmed the introduction of PEI [56].

In general, a hydrophilic surface enhances the permeation flux and antifouling capability of a membrane. Consequently, we evaluated the wettability of the membrane surface by measuring the water contact angle (WCA), as depicted in Figure 4a. The PES/Fe based membrane exhibited hydrophobic characteristics, with a WCA of 86°. A substantial number of hydrophilic phosphate groups were anchored on the membrane surface due to the robust metal-organophosphate coordination bonding; as a result, the phosphate hydroxyl groups were exposed on the membrane surface, reducing the WCA to 64°, indicative of favorable hydrophilicity [64]. The contact angle increased following the introduction of PEI. The hydrophilicity of the membrane surface is primarily attributed to the capillary force between the hydrophilic groups on the surface and the membrane pores.

The zeta potentials of the PES/Fe based, PES/Fe-PA, and PES/Fe-PA-PEI membranes were assessed to evaluate their separation performance for solutes with varying charges (Figure 4b). The surface of the base membrane exhibited electrical neutrality across the pH range of 3–10, with an isoelectric point at pH = 5.6. The zeta potential of the PES/Fe-PA membrane was negative, primarily due to the abundance of negatively charged phosphate groups within the membrane [55]. Upon the introduction of PEI, the unprotonated amino group became positively charged over a broad pH range [65], imparting the surface of the PES/Fe-PA-PEI membranes with resistance to contamination by cationic solutes. The membrane surface is nearly neutral or carries a slight positive charge, which may have affected the filtration performance of dyes and salts.

The transmembrane energy barrier refers to the minimum energy required for a substance to traverse a membrane. The higher the energy barrier, the greater the resistance to transport, thereby hindering the passage of molecules and ions through the membrane. In membrane separation, the transmembrane energy barrier is closely related to membrane flux. Consequently, this study examined the impact of modification on membrane flux from a physical standpoint by evaluating its performance. The activation energy for water transmembrane transport, both pre- and post-modification, was calculated using the Arrhenius equation, as depicted in Figure 4c. The introduction of PA reduced the activation energy required for water transport from 3.47 kJ/mol to 2.15 kJ/mol. This phenomenon may be attributed to the introduction of hydrophilic phosphate groups in phytic acid, which increases the membrane’s hydrophilicity, reduces the resistance to water mass transfer, and thereby lowers the activation energy for water transport. The reduction in activation energy may lead to an increase in membrane flux. The secondary assembly of PEI increased the activation energy required for water transport from 2.15 kJ/mol to 4.41 kJ/mol. This increase is attributed to the modification by PEI: long-chain PEI molecules bind tightly to the membrane surface, thereby reducing pore size and increasing the thickness of the separation layer. Additionally, the introduction of PEI reduced the hydrophilicity of the membrane surface. These changes collectively increased the resistance to mass transfer of water through the membrane, thereby raising the activation energy.

3.2. Membrane Filtration Performance

3.2.1. Impact of Fabrication Parameters on the Membrane Filtration Performance

During the crosslinking process, the concentrations of the PA and PEI solutions significantly influenced the structural and functional properties of the membrane. Initially, the optimal ratio of PA to PA-Fe was determined by examining the PA concentration, as shown in Figure 5a. With an increase in PA concentration, the pure water flux of the membrane initially increased and subsequently decreased. At a PA concentration of 5 g/L, the membrane flux reached its maximum at 525 L·m⁻²·h⁻¹·bar⁻¹, while the CR rejection rate remained above 95%. This behavior is attributed to the chelation between the six phosphate groups in the PA molecules and Fe³⁺, which not only enhances the hydrophilicity of the membrane but also reduces the activation energy and preserves effective mass transfer channels. At elevated concentrations, excessive crosslinking between the phosphate groups and Fe³⁺ resulted in the formation of a dense network structure, significantly increasing the transmembrane resistance and causing the flux to decrease to 260 L·m⁻²·h⁻¹·bar⁻¹.

The influence of PEI concentration on the membrane filtration performance was examined under optimal PA concentration conditions, as depicted in Figure 5b. Consistent with prior observations, an increase in the PEI concentration initially resulted in a slight enhancement of the membrane flux, followed by a decline. This phenomenon can be attributed to two primary factors: (1) PEI forms hydrogen bonds with water molecules, thereby enhancing hydrophilicity and facilitating water transport, and (2) the incorporation of PEI increases the thickness of the separation layer and reduces the membrane pore size, consequently diminishing the water flux. The interactions between these factors account for the observed trends. At a PEI concentration of 5 g/L, the membrane flux reached 280 L·m⁻²·h⁻¹·bar⁻¹, with a CR rejection rate of 99% and a sodium sulfate rejection rate of 12%. Excessive PEI concentration resulted in over-crosslinking of PA-PEI, which, while enhancing dye rejection, reduced the flux to 160 L·m⁻²·h⁻¹·bar⁻¹. Therefore, the optimal conditions for constructing PES/Fe-PA-PEI membranes to investigate the dye/salt separation performance were determined to be 5 g/L PA and 5 g/L PEI.

3.2.2. The Dye/Salt Separation Performance of PES/Fe-PA-PEI Membrane

The efficacy of dye/salt separation was evaluated using a simulated feed solution containing CR (100 ppm) and inorganic salts (1000 ppm) (Figure 6a). In all the tested mixtures, the membrane consistently achieved a CR rejection rate exceeding 95%, whereas the salt rejection rate remained below 10%. To assess the influence of the dye molecules, CR and CBB were combined in equal volumes, and the separation efficiency of the PES/Fe-PA-PEI membrane for mixed dyes was evaluated, as shown in Figure 6a. The membrane achieved a 99% retention rate for the mixed dye solution, indicating that the prepared PES/Fe-PA-PEI membrane can efficiently treat mixed dye wastewater. The membrane surface contains positively charged amino functional groups that may adsorb anionic dyes; thus, the cationic dye CBB was selected for a control experiment to investigate the effect of the membrane on CR adsorption (Figure 6b). After continuous operation for one hour, the dye-salt separation efficiency was measured. The CBB retention reached 99%, whereas the retention rates for the different salt components were maintained below 10%, demonstrating excellent dye-salt separation performance. In this experiment, although the positively charged membrane surface may adsorb anionic dyes, the cationic dye CBB was also highly retained (99%), and the salt retention rate was extremely low (<10%), indicating that retention is primarily due to size exclusion rather than electrostatic adsorption. These results suggest that the membrane surface charge has little influence on the selective separation of dyes and inorganic salts; rather, the sieving effect remains the dominant factor.

To further evaluate the separation selectivity with respect to the concentration of inorganic salts, the PES/Fe-PA-PEI membrane was examined using a solution containing CR and Na₂SO₄ at varying concentrations (0.5, 1, 1.5, and 2 g/L). The results are shown in Figure 6c. The rejection rate for CR consistently exceeded 95%, whereas that for salts remained below 20% across all tested concentrations. The increase in the salt retention rate may be attributed to the accumulation of the salt component on the membrane surface during the filtration process, leading to the concentration polarization phenomenon and a subsequent decrease in the permeate concentration.

The dye/salt separation performance of the PES/Fe-PA-PEI membrane developed in this study was assessed in comparison to the latest LNF membranes reported in the literature (Figure 6d). Table S2 shows the details [55,66,67,68,69,70,71,72,73,74,75,76]. The PES/Fe-PA-PEI membrane demonstrated superior performance, evidenced by its high pure water permeability (280 L·m⁻²·h⁻¹·bar⁻¹) and significant dye retention (CR retention of 95%). Its dye/salt selectivity of 21.21 ranks among the highest levels documented in the literature. These findings suggest that the fabricated PES/Fe-PA-PEI membrane exhibits exceptional dye/salt separation efficiency and has considerable potential for practical applications.

3.3. Membrane Antifouling and Stability Performance

The stability of the membrane is a critical factor influencing the overall application cost in practical scenarios. As illustrated in Figure 7a, a mixed solution of CR/Na₂SO₄ was employed as the feed solution, with concentrations of 100 and 1000 ppm for CR and Na₂SO₄, respectively. Continuous filtration was conducted over a 24 h period to assess the long-term stability of the membrane. The normalized flux changes indicate that, during the 24 h continuous operation, the normalized flux of the PES/Fe-PA-PEI membrane remained above 0.9. In contrast, the PES/Fe membrane experienced significant contamination, resulting in a flux reduction to 0.65 after 24 h of operation. This demonstrates the superior anti-pollution properties of the PES/Fe-PA-PEI composite membrane. Furthermore, a pressure test (Figure 7a) revealed that the maximum withstand pressure of the membrane was 0.6 MPa, and it maintained good operational conditions without any rupture, indicating high-pressure resistance performance.

HA was employed as a simulated organic pollutant to further assess the anti-fouling properties of the membrane. As illustrated in Figure 7b, the introduction of organic pollutants resulted in a significant decrease in the permeability of the LNF membrane. This decline was attributed to the formation of a fouling layer on the membrane surface, caused by the repulsion of pollutants and the adsorption of pollutants on the pore walls, which led to a reduction in pore size. Following cleaning with pure water, the pollutants were redispersed, partially restoring the membrane permeability. HA macromolecules possess abundant carboxyl and phenolic functional groups, which carry negative charges in solution. Owing to the weak positive charge on the surface of the LNF membrane, HA molecules may adsorb onto the membrane through electrostatic interactions. However, as evidenced by the test results, after three cycles, the FRR of HA on the PES/Fe-PA-PEI membrane still reached 93.2% (Table S3 in SI), and the HA retention rate remained above 99%. Repeated filtration experiments demonstrated that the prepared PES/Fe-PA-PEI membrane exhibits favorable anti-fouling performance during dye/salt separation.

A mixed solution comprising 0.1 ppm CR and 1 g/L Na₂SO₄ dye salt was employed to assess the stability of membrane separation performance (Figure 7c). The normalized dye flux remained above 80%, and the CR retention rate exceeded 95% across three experimental cycles. Additionally, the retention rate of sodium sulfate increased to 15%, attributed to the accumulation of salt and dye on the membrane surface, a consequence of concentration polarization. Furthermore, nanofiltration membranes are frequently exposed to complex chemical environments, such as strong acids and bases, which can compromise their structural integrity and filtration efficacy. Consequently, it is imperative that the prepared membranes exhibit exceptional chemical stability. The stability and separation performance of the PES/Fe-PA-PEI membranes were evaluated and compared following immersion in solutions with pH values ranging from 3 to 11. As illustrated in Figure 7d, the membrane demonstrated high CR retention and excellent permeability after 24 h of immersion, indicating its superior resistance to acidic and alkaline conditions.

4. Conclusions

In this study, PES/Fe-PA-PEI HF LNF membranes were successfully fabricated by blending metallic ion compounds and dual-coating with PA and PEI. The optimized membrane exhibited high pure water permeability (280 L·m⁻²·h⁻¹·bar⁻¹) and demonstrated exceptional dye/salt separation selectivity, with separation factors of CR/NaCl: 64.08, CR/Na₂SO₄: 21.21, CBB/NaCl: 14.75, and CBB/Na₂SO₄: 10.74. The proposed preparation method is straightforward, highly flexible, and adjustable, facilitating the efficient fabrication and modification of HF membranes with good scalability and significant economic potential. The modified membrane displayed strong resistance to acids and alkalis, maintaining a CR rejection rate above 99% across a pH range of 3–11, with no observable damage to the membrane structure. In antifouling tests, the flux recovery ratio for HA exceeded 80%. Furthermore, the membrane demonstrated excellent operational stability during long-term filtration, sustaining high flux levels and consistent separation performance in dye/salt mixtures. These results suggest that the developed membrane holds considerable promise for the treatment of dye/salt-mixed wastewater and the recovery of valuable resources.