PES@ZnO mixed matrix membranes for photocatalsytic dye removal: experimental performance and DFT insights into interfacial adsorption and charge-transfer propensity

Abstract

The treatment of textile wastewater containing recalcitrant dyes remains a challenge. This study investigates the synergistic filtration and photocatalytic degradation performance of polyethersulfone (PES) mixed matrix membranes (MMMs) incorporating zinc oxide nanoparticles (ZnO NPs) (0.5 wt%), fabricated via non-solvent induced phase separation. The optimized PES@ZnO membrane exhibits enhanced hydrophilicity (contact angle: 61.5° vs. 68.5°) and improved pure water permeability (36 vs. 28 L m⁻² h⁻¹ bar⁻¹) compared to pristine PES. Under UV irradiation, PES@ZnO achieves 95% removal of Basic Blue 41 (BB41) after 250 min, with a flux recovery ratio of 86%, substantially higher than pristine PES (59% removal). Degradation follows pseudo-first-order kinetics (k = 1.22 × 10⁻² min⁻¹), five times higher than suspended ZnO NPs. Density functional theory (DFT) calculations at the B3LYP/def2-SVP level reveal spontaneous adsorption (E_ads = −1.40 eV). Frontier molecular orbital analysis indicates HOMO localization on BB41 and LUMO localization on ZnO, which is advantageous for interfacial charge transfer. Upon adsorption, the HOMO–LUMO gap decreases from 8.43 to 5.49 eV, indicating enhanced electronic reactivity. Mulliken charge redistribution and an electrophilicity-based charge transfer (ECT) value of 0.771 confirm net electron donation from the dye to ZnO, consistent with reactive oxygen species generation and oxidative degradation. By combining experimental performance evaluation with DFT-derived mechanistic insights—an approach not previously applied to PES@ZnO membranes for dye removal—this work establishes that ZnO acts not merely as a hydrophilic filler but as an active adsorption and charge-transfer center. The PES@ZnO membrane represents a promising platform for integrated treatment of dye-contaminated wastewater.

PES@ZnO mixed matrix membranes combine filtration and photocatalysis for dye removal. Enhanced hydrophilicity and charge-transfer photocatalysis achieve 86% degradation of Basic Blue 41, supported by DFT analysis of adsorption and electron transfer.

1. Introduction

Synthetic dyes represent a critical environmental and public health concern due to their widespread use, particularly in textile manufacturing.^1,2 The discharge of dye-contaminated wastewater compromises aquatic ecosystems, as many of these compounds are chemically stable, resistant to biodegradation, and highly water-soluble.^2–6 Conventional physicochemical and biological treatment methods are often inadequate for such recalcitrant pollutants, underscoring the need for more efficient separation technologies.^7–9 The textile industry remains a major contributor, with over 100 000 commercial dyes in use and an estimated 280 000 tons discharged annually.^10–13 These effluents typically contain complex mixtures of dyes, heavy metals, and organic substances, and low fixation rates result in substantial untreated discharge.^14,15

Membrane-based separation technologies offer distinct advantages for dye removal,^16,17 including high selectivity, operational flexibility, scalability, and lower energy consumption.¹⁸ Among polymeric materials, polyethersulfone (PES) is widely used due to its mechanical strength, thermal stability, and chemical resistance.^19,20 However, its inherent hydrophobicity promotes fouling, reduces permeability, and compromises separation efficiency.^21,22 Effective fouling management is therefore essential, and a common strategy involves enhancing surface hydrophilicity to reduce hydrophobic interactions with feed constituents.^23,24

To overcome this limitation, PES membranes are rendered more hydrophilic either through post-synthesis modifications (grafting or coating) or by blending with hydrophilic additives during fabrication.^7,25 Blending is particularly advantageous as it integrates modification directly into the synthesis process. In this context, the incorporation of inorganic NPs has gained attention, including MWCNTs,²⁶ SiO₂, (ref. 27) TiO₂, (ref. 28) GO, and ZnO.²⁹ While TiO₂ is well-studied for fouling resistance, ZnO offers comparable performance with additional benefits such as high surface-to-volume ratio, cost-effectiveness, and photocatalytic potential.^7,30–33

The functionalization of PES with ZnO NPs improves hydrophilicity, fouling resistance, and separation performance. Consequently, PES@ZnO mixed matrix membranes (MMMs) exhibit superior dye rejection, water permeability, and antifouling properties compared to unmodified counterparts.^32,34,35 These enhancements arise from modified surface chemistry, increased wettability, optimized pore architecture, and stronger interfacial interactions with contaminants.

Several studies have explored PES@ZnO MMMs for fouling mitigation and dye removal. First demonstrated that 0.5 wt% ZnO in hollow fiber membranes improved antifouling behavior against river water. A subsequent study using an NMP-acetone solvent system showed enhanced hydrophilicity and pore formation, though higher loadings led to agglomeration.³⁶ Zinadini et al.³⁷ incorporated ZnO-coated MWCNTs into PES, achieving improved flux and antibiofouling properties with 0.5 wt% loading. Nasrollahi et al.³⁸ used a CuO/ZnO nanocomposite at 0.2 wt%, obtaining homogeneous dispersion and a 32% increase in pure water flux. More recently, Ali et al.³⁹ optimized modified ZnO loading (0.2 g) with a pre-evaporation step, achieving 80.3 LMH water flux and 85.9% flux recovery. Mezher et al.⁴⁰ incorporated silane-functionalized ZnO (1.5 wt%), reporting a twofold increase in pure water flux (from 25 to 53 L m⁻² h⁻¹) while maintaining high separation efficiency.

Several experimental studies have explored PES@ZnO MMMs for fouling mitigation and dye removal.^36–40 While these investigations have established the beneficial role of ZnO incorporation in improving membrane performance, they have remained predominantly empirical, with emphasis placed on macroscopic indicators such as permeability, fouling resistance, and dye rejection. Concurrently, Density Functional Theory (DFT) calculations have been extensively applied to investigate ZnO-dye interactions in various contexts, including dye-sensitized solar cells,⁴¹ slurry-phase photocatalysis,⁴² and molecular adsorption on ZnO surfaces.⁴³ However, these computational studies have typically been conducted in isolation from membrane separation science, focusing on idealized surfaces rather than on the complex polymer–nanoparticle–dye interfaces present in MMMs. Whether ZnO acts solely as a hydrophilicity enhancer or actively participates in interfacial charge transfer with adsorbed dye molecules to promote oxidative degradation therefore remains an open mechanistic question. To address this mechanistic gap, we integrate systematic experimental performance evaluation including permeability, fouling indices, and photocatalytic degradation kinetics with DFT-derived mechanistic insights for the specific case of Basic Blue 41 (BB41) removal.

To the best of our knowledge, no prior work has coupled such macroscopic filtration data with atomistic DFT analysis to elucidate charge-transfer pathways in a PES-based photocatalytic membrane. This integrated approach demonstrates that ZnO functions as an active charge-transfer hub rather than as a passive filler, and establishes a rational, theory-guided framework for designing next-generation MMMs with tailored interfacial properties for sustainable wastewater treatment.

2. Experimental section

2.1. Materials

PES, Ultrason® E6020P, molecular weight ≈ 58 kDa supplied by BASF (Germany) was employed as the base polymer for membrane fabrication. N-methyl-2-pyrrolidone (NMP, Merck, Germany) was used as the solvent for casting solution preparation. Polyvinylpyrrolidone (PVP, M_W = 111.14 g mol⁻¹) served as a pore-forming additive, while zinc oxide nanoparticles (ZnO NPs) nanopowder with a particle size below 100 nm was incorporated as the inorganic nanofiller; both materials were obtained from Sigma-Aldrich.

BB41 (C₂₀H₂₆N₄O₆S₂, molecular weight 482.57 g mol⁻¹) was selected as the model dye and purchased from Sigma-Aldrich. Ultrapure Milli-Q® Type 1 water (resistivity 15 MΩ cm at 25 °C, Merck Millipore, Germany) was used for membrane preparation and solution formulation.

2.2. Preparation of mixed-matrix membranes

MMMs were fabricated via the non-solvent induced phase separation (NIPS) method (see Fig. 1). The casting solution was prepared by dissolving PES in N-methyl-2-pyrrolidone (NMP), followed by the addition of 1.5 wt% polyvinylpyrrolidone (PVP) as a pore-forming agent.⁴⁴ ZnO NPs (0.5 wt% relative to PES) were dispersed in NMP under ultrasonication for 30 min to prevent agglomeration, then gradually added to the polymer solution and stirred for 8 h to obtain a homogeneous casting dope.³⁹

Fig. 1. Preparation of PES@ZnO mixed matrix membrane (NIPS method).

The resulting casting solution was degassed at ambient temperature for 1 h to remove trapped air bubbles. Membrane films were then cast using an automated casting system (POROMETER MEMCAST, Netherlands) at a speed of 3 mm s⁻¹ and room temperature (25 °C), and immediately immersed in a deionized water coagulation bath to induce phase inversion. The fabricated membranes were rinsed and stored in fresh distilled water, which was periodically replaced to ensure complete removal of residual solvent prior to characterization and testing.⁴⁵

2.3. Mixed matrix membranes characterization

The successful incorporation of ZnO NPs into PES membrane matrix, as well as the nature of their interactions with the polymer, was investigated by Fourier transform infrared (FTIR) spectroscopy. The spectra were recorded using a Bruker TENSOR 27 ATR spectrometer in the wavenumber range of 400–4000 cm⁻¹, with a spectral resolution of 4 cm⁻¹. Micro-Raman spectroscopy was used to complement FTIR analysis and investigate molecular vibrations as well as the dispersion of the inorganic phase within the polymer matrix. Measurements were performed using a Jobin–Yvon LabRAM spectrometer (Horiba) with a 532 nm He–Ne laser at a spectral resolution of 2 cm⁻¹ and a 50×objective.

Surface wettability was assessed via static water contact angle measurements using a CAM 200 goniometer (KSV Instruments, Finland). A 5 µL ultrapure water droplet was placed on the membrane surface, and measurements were repeated at least four times to obtain the mean value and standard deviation.

Membrane porosity was determined by the gravimetric method to assess the effect of NPs on membrane structure. Samples were dried at 50 °C for 24 h, cut into (1 cm × 1 cm) pieces, and weighed (dry weight). The samples were then immersed in distilled water at 25 °C for 24 h, surface water was removed, and the wet weight was recorded. The overall porosity (ε) was calculated using eqn (1):

1

where W₁ is the weight of the wet membrane, W₂ is the weight of the dry membrane; A is the membrane effective area (m²), d_w is the water density, and l is the membrane thickness (m). All experiments were conducted in triplicate, and the reported results correspond to the average values.

2.4. Photocatalytic and filtration performance assessment

The photocatalytic activity of commercial ZnO NPs and the fabricated membranes (PES and PES@ZnO) was assessed using BB41 as a model pollutant.

For batch photocatalysis, a 250 mL quartz beaker placed in a black box served as the reaction vessel. An 18 W UV lamp (365 nm) was positioned 10 cm above the solution surface. In each run, 100 mL of a 10 mg L⁻¹ BB41 solution was introduced, and the photocatalyst was dispersed under magnetic stirring. The suspension was first kept in the dark for 30 min to reach adsorption–desorption equilibrium, followed by UV irradiation to initiate degradation. At regular intervals, aliquots (3–5 mL) were withdrawn, centrifuged (10 min), and filtered to remove catalyst particles prior to analysis.

Membrane filtration experiments were conducted using a dead-end system (Fig. 2). Prior to testing, all membranes were compacted at 2 bar for 1 h. Pure water flux was measured at varying transmembrane pressures to determine hydraulic permeability (eqn (2)). Separation performance was evaluated using a 10 mg L⁻¹ BB41 solution as feed, with a flat-sheet membrane under stirring at 350 rpm.

Fig. 2. Schematic representation of the dead-end membrane filtration system: (1) membrane; (2) stirrer; (3) valve; (4) pressure regulator; (5) pressure gauge; (6) filtration cell; (7) permeate; (8) balance; (9) magnetic stirrer; (10) N₂ gas cylinder and UV lamp (18 watt).

Photocatalytic membrane performance was evaluated in a custom-designed photoreactor (Fig. 2), where an 18 W UV lamp (365 nm) was placed 10 cm above the dead-end cell. Experiments were conducted at room temperature under 1 bar transmembrane pressure using 300 mL of 10 mg L⁻¹ BB41 solution. Dye concentration was monitored using a UV-vis spectrophotometer at 615 nm, and degradation efficiency was calculated according to eqn (3).

The permeate flux (J) was calculated using eqn (2):²⁸

2

where V is the volume of permeate collected during time t, and A is the effective membrane area. The flux is expressed in L h⁻¹ m⁻².

The removal efficiency was evaluated according to the eqn (3):

3

where C_p the pollutant concentration in is permeate and C_f is the pollutant concentration in feed solution. The concentrations were determined by measuring the absorbance on a UV/visible spectrophotometer (Aquanova Jenway) at a wavelength of 615 nm.

The photocatalytic degradation kinetics of BB41 in the aqueous phase were described using the Langmuir–Hinshelwood model presented by the eqn (4):⁷

4

where r the reaction rate (mg L⁻¹ min⁻¹), C is the pollutant concentration (mg L⁻¹), k is the Langmuir–Hinshelwood reaction rate constant (mg L⁻¹ min⁻¹), and k_e is the Langmuir adsorption equilibrium constant (L mg⁻¹).

At low initial pollutant concentrations (kC << 1), the model simplifies to pseudo-first-order kinetics defined by the eqn (5):

5

where, k_app is the apparent rate constant (min⁻¹), C₀ is the initial BB41 concentration (mg L⁻¹), C_t is the BB41 concentration at time t (mg L⁻¹), and t is the reaction time (min). The apparent rate constant was determined from the slope of the linear plot of ln (C₀/C_t) versus time.

Fouling indices: following the dye rejection tests, the membranes were immersed in distilled water overnight and subsequently rinsed under forward filtration mode for 1 h. The pure water permeability of the cleaned membranes (JR) was then re-evaluated to assess fouling reversibility. The obtained data were used to determine the flux recovery ratio (FRR) and flux decline ratio (FDR) according to eqn (6) and (7), respectively:

6

7

With J₀ the initial pure water flux, J_P the permeate flux of the BB41, and J_R the pure water flux of the cleaned membranes.

3. Computational details

DFT calculation was utilized so that we could carry out computational design, structural optimization, and calculations of the major characteristics of each structure under consideration. DFT method implanted in the Gaussian16 software package.⁴⁶ The structures were optimized using the hybrid density functional B3LYP⁴⁷ with the def2-SVP basis set.⁴⁸ To complement the ground-state frontier orbital analysis and provide direct excited-state support for the proposed interfacial charge-transfer mechanism, TD-DFT calculations were performed on all optimised structures. Grimme's D3 empirical dispersion correction was applied to better describe non-covalent interactions in the adsorption complex. Gauss View software⁴⁹ was used to create a visual presentation.

The electronic band gap (E_gap) and Fermi level energy (E_FL) are calculated by using the following equations:⁵⁰

E_g = E_LUMO − E_HOMO 8

9

where, E_HOMO refers to the energy level of the highest occupied molecular orbital, while E_LUMO denotes the energy level of the lowest unoccupied molecular orbital.

It has been determined that the adsorption energy, denoted by E_ads, as follows:⁵¹

E_ads = E_Complexe − E_Substrate − E_Molecule 10

where, the symbols E_Complexe, and E_Molecule indicate the total energies associated with the complex structure, and the molecules PES, ZnO and BB41, correspondingly.

Using the Koopmans' theorem and the DFT approach, we investigated the reactivity of the studied compounds by calculating a number of different parameters. The global chemical reactivity descriptors (GCRD) are another vital technique to understand the chemical properties of a molecule such as chemical hardness (η), chemical potential (µ), chemical softness (S), electronegativity (χ) and electrophilic index (ω). The formula to obtain GCRD is derived from HOMO–LUMO energies, taking the HOMO energy as ionization potential (IP) and LUMO as electron affinity (EA). The concept of these parameters is related to each other according to the following equations:^52–54

IP = −E_HOMO 11

EA = −E_LUMO 12

13

14

15

16

17

The work function (φ) is the minimum energy required to move the weakest bound electron from the Fermi energy level in a solid surface to a vacuum point.⁵¹ Furthermore, this measure serves as an additional signal that provides a valuable understanding of the reactivity and sensitivity of the complex towards the molecule. This value can be determined using the following equation:⁵¹

φ = V_el(+∞) − E_FL 18

The symbol V_el(+∞) refers to the electrostatic potential of an electron at a significant distance from the surface, assumed to be zero (i.e., V_el(+∞) = 0)

19

When two systems A and B approach each other, the amount of charge transfer between them can be written in terms of electrophilicity. Electrophilicity-based charge transfer (ECT) is obtained by the following equation:

ECT = ΔN_max(A)ΔN_max(B) 20

4. Results and discussion

4.1. Membrane characterization

4.1.1. Structural characterization

The chemical structure and phase composition of the pristine PES and PES@ZnO composite membranes were comprehensively investigated using complementary FTIR-ATR and Raman spectroscopy in order to elucidate the nature of polymer–filler interactions are presented in Fig. 3a and b, respectively.

Fig. 3. (a) FTIR spectra and (b) Raman spectra of PES membrane and PES-ZnO nano-composite membrane.

The FTIR spectrum of the pristine PES membrane exhibits the characteristic vibrational features of PES, confirming preservation of the polymer backbone after membrane fabrication.

The absorption bands observed in the 1580–1485 cm⁻¹ region are assigned to aromatic C C stretching vibrations. The asymmetric and symmetric stretching modes of the sulfone group (O S O) appear at approximately 1320–1295 cm⁻¹ and 1155–1105 cm⁻¹, respectively, while the aromatic ether (C–O–C) stretching vibration is detected in the 1240–1210 cm⁻¹ range. In addition, the band near 835 cm⁻¹ corresponds to the out-of-plane bending vibration of para-substituted benzene rings. Notably, the absence of a broad absorption band in the 3200–3500 cm⁻¹ region indicates the lack of intrinsic hydroxyl functionalities in the pristine PES matrix, consistent with its hydrophobic nature.⁴⁵ Upon incorporation of ZnO NPs, the principal PES absorption bands remain essentially unchanged, demonstrating that the chemical structure of the polymer is preserved during composite membrane preparation.⁵⁵ A new absorption band emerging in the 400–550 cm⁻¹ region is attributed to Zn–O stretching vibrations, characteristic of the wurtzite structure of ZnO,⁵⁶ thereby confirming the successful incorporation of ZnO within the polymer matrix. Importantly, no additional bands corresponding to Zn–O–C or Zn–S covalent linkages are detected, indicating that no chemical grafting occurs between ZnO and PES chains. However, slight shifts and variations in the intensity of the sulfone (O S O) and ether (C–O–C) bands are observed in the composite membranes, suggesting the establishment of interfacial interactions. These subtle spectral changes may originate from weak coordination between Zn²⁺ surface sites and electron-rich oxygen atoms of the sulfone groups, hydrogen bonding involving surface hydroxyl groups of ZnO NPs, and local interfacial polarization effects resulting from dielectric contrast between the inorganic phase and the polymer matrix. Such observations indicate that the modification mechanism is governed predominantly by physical embedding and non-covalent interfacial interactions rather than covalent bonding.

Raman spectroscopy further confirms these findings. In addition to the characteristic vibrational modes of PES, the PES@ZnO membrane exhibits a distinct Raman band at 437 cm⁻¹, assigned to the E₂ (high) phonon mode of hexagonal ZnO. According to group theory analysis of wurtzite ZnO (space group P6₃mc), the E₂ mode is exclusively Raman-active, whereas A₁ and E₁ modes are both Raman and IR-active. The pronounced E₂ (high) signal confirms the presence of well-crystallized ZnO domains homogeneously dispersed within the polymer matrix.

Overall, the combined FTIR and Raman analyses provide consistent and complementary evidence of successful ZnO incorporation without chemical degradation of the PES backbone. The structural integrity of the polymer is maintained, while interfacial physicochemical interactions between ZnO NPs and polar functional groups of PES likely contribute to the enhanced properties of the mixed-matrix membranes.

4.1.2. Thermogravimetric analysis (TGA)

The thermal stability of pristine PES and PES/ZnO composite membranes was evaluated by thermogravimetric analysis under an inert atmosphere (Fig. 4). As shown in the TGA curves, both samples exhibit similar thermal degradation behavior characterized by a single major decomposition step. An initial negligible weight loss below 150 °C is observed for both membranes, which can be attributed to the evaporation of physically adsorbed moisture and residual solvent. The main degradation stage occurs between approximately 450 °C and 600 °C, corresponding to the thermal decomposition of the PES polymer backbone, particularly the cleavage of sulfone (–SO₂–) linkages and aromatic chains.^45,57

Fig. 4. Thermogravimetric analysis of the pristine PES and PES@ZnO membranes.

The pristine PES membrane undergoes rapid decomposition within this temperature range, leaving a very low residual mass at 700 °C. In contrast, the PES@ZnO composite membrane exhibits slightly enhanced thermal stability, as evidenced by a marginal shift of the degradation onset toward higher temperatures and a higher residual mass at elevated temperatures.⁵⁸ As shown in Fig. 4, bare ZnO exhibits negligible weight loss up to 700 °C (consistent with our previous TGA findings,⁷ and the increased char residue of the composite is ascribed to the incorporation of thermally stable ZnO NPs, which not only resist decomposition at high temperatures but may also catalyse the formation of a protective carbonaceous char layer. Finally, the incorporation of ZnO improves the thermal resistance of the membrane without significantly altering its degradation mechanism, confirming the good thermal compatibility between PES and ZnO.

4.1.3. Water permeability and contact angle

The permeation performance of the fabricated MMMs was evaluated by measuring pure water flux under increasing transmembrane pressure (TMP), with the results shown in Fig. 5. For all membranes, a linear relationship between flux and pressure was observed across the entire TMP range (0.5–3 bar). This linearity indicates that the membrane structure remains stable under pressure and that the transport of water occurs through viscous flow within the porous network.

Fig. 5. Water permeability of M₀ and M₁(PES@ZnO) membrane.

Hydraulic permeability, calculated from the slope of the flux–pressure curves, was 28 L m⁻² h⁻¹ bar⁻¹ for the pristine PES membrane (M₀). The incorporation of ZnO NPs led to a significant increase in permeability, reaching 36 L m⁻² h⁻¹ bar⁻¹, corresponding to an enhancement of approximately 29%. This improvement in permeability can be explained by two complementary factors. First, structural analysis revealed that the PES@ZnO membrane exhibited higher porosity (78%) compared to the pristine membrane (73%), along with a slight increase in average pore size from 11.6 nm to 12.4 nm. These structural changes facilitate greater water transport through the membrane matrix.

Second, and more importantly, the incorporation of hydrophilic ZnO NPs significantly enhanced the surface wettability. Contact angle measurements showed a decrease from 68.5° for the pristine membrane to 61.5° for the PES@ZnO membrane, confirming a more hydrophilic surface. This increased hydrophilicity promotes the formation of a water layer at the membrane surface, which reduces resistance to water permeation.

Taken together, the combination of enhanced porosity and improved surface hydrophilicity contributes to the higher permeability of the PES@ZnO membrane. Based on these favorable characteristics, this membrane was selected as the most suitable candidate for the subsequent removal of organic pollutants.

4.2. Photocatalytic degradation kinetics of BB41

The photocatalytic degradation of BB41 was investigated under various conditions, and the results are presented in Fig. 6a as normalized concentration profiles C/C₀ over time. The corresponding kinetic parameters, obtained by fitting the experimental data to a pseudo-first-order model, are summarized in Table 1.

Fig. 6. (a) Photocatalytic degradation and removal of BB41 using ZnO and the prepared membranes. (b) Kinetic study results with the application of a first-order model for the BB41 removal using ZnO and the prepared membranes.

Table 1. Pseudo-first-order kinetic parameters for BB41 degradation under different conditions.

	K (min−1)	R 2
Photolysis	9.12 × 10−4	0.959
ZnO	2.23 × 10−3	0.970
PES	3.80 × 10−3	0.977
PES@ZnO	5.12 × 10−3	0.944
PES@ZnO+UV	1.22 × 10−2	0.977

Among all tested systems, the PES@ZnO membrane under UV irradiation (PES@ZnO+UV) exhibited the most pronounced degradation efficiency, achieving a residual concentration of only 0.05 after 250 min, corresponding to approximately 95% removal. In contrast, the pristine PES membrane showed limited dye removal, with a final C/C₀ of 0.41 (59% removal), while suspended ZnO NPs achieved a residual concentration of 0.58 (42% removal). The lowest degradation performance was observed under photolysis conditions, where only 21% of the dye was removed after 250 min.

The kinetic data were well-described by the pseudo-first-order model, as evidenced by the high coefficients of determination (R²) for all tested conditions. The PES@ZnO+UV system yielded the highest reaction rate constant (k = 1.22 × 10⁻² min⁻¹), which is approximately 3.2 times higher than that of the PES@ZnO membrane without UV irradiation (k = 5.12 × 10⁻³ min⁻¹) and more than 5 times higher than that of suspended ZnO NPs (k = 2.23 × 10⁻³ min⁻¹). The pristine PES membrane exhibited a moderate rate constant (k = 3.80 × 10⁻³ min⁻¹), likely attributable to adsorption phenomena rather than true photocatalytic activity, whereas photolysis showed the lowest kinetic constant (k = 9.12 × 10⁻⁴ min⁻¹).

The superior performance of the PES@ZnO+UV system can be attributed to the synergistic combination of several factors. First, the immobilization of ZnO NPs within the polymeric membrane matrix prevents particle aggregation and ensures uniform distribution of photocatalytic sites. Second, the convective flow through the membrane enhances mass transfer of dye molecules to the active sites, overcoming diffusion limitations commonly encountered in slurry-based photocatalysis. Third, the integrated configuration eliminates the need for post-treatment catalyst recovery, representing a significant operational advantage. The lower performance of suspended ZnO NPs, despite their higher surface area, is likely due to light scattering effects and mass transfer limitations in the batch reactor configuration.

These results clearly demonstrate that the photocatalytic membrane reactor configuration not only enhances degradation kinetics but also provides a practical and scalable approach for wastewater treatment applications. The combination of membrane separation and photocatalysis within a single unit offers a promising strategy for the efficient removal of recalcitrant organic pollutants from aqueous streams.

4.3. Membrane fouling behavior and flux recovery

The fouling behavior of the pristine PES and PES@ZnO membranes was evaluated under dark and UV irradiation conditions, with the results presented in Fig. 7a and b. The permeate flux profiles over time are illustrated in Fig. 7a, while the corresponding fouling resistance indices, including flux recovery ratio (FRR) and flux decline ratio (FDR), are summarized in Fig. 7b.

Fig. 7. (a) Permeate flux of the prepared membranes during the fouling study using BB41 and (b) fouling index (FRR and FDR) for PES and PES@ZnO (operating conditions C_i 10 mg L⁻¹), pH 6.3 and at room temperature) (c) recycling performance of PES@ZnO membrane for BB41 degradation over three consecutive cycles (C₀ = 10 mg L⁻¹, UV irradiation, 1 bar).

4.3.1. Permeate flux analysis

As shown in Fig. 7a, all membranes exhibited an initial sharp decline in permeate flux during the first 60 min of filtration, followed by a gradual stabilization. This behavior is characteristic of membrane fouling, where rapid pore blocking occurs initially, followed by the formation of a fouling layer that eventually reaches a steady state.

The pristine PES membrane showed a significant flux decline from an initial value of 30 L m⁻² h⁻¹ to approximately 16 L m⁻² h⁻¹ after 600 min under dark conditions, corresponding to a flux loss of nearly 47%. Under UV irradiation, the PES membrane exhibited a slightly improved flux profile, stabilizing at approximately 15 L m⁻² h⁻¹, indicating that UV exposure alone has a marginal effect on fouling mitigation for the unmodified membrane.

In contrast, the PES@ZnO membrane demonstrated substantially higher permeate flux throughout the filtration period. Under dark conditions, the PES@ZnO membrane maintained a flux of 21 L m⁻² h⁻¹ after 600 min, representing only a 30% flux decline relative to its initial value. More notably, under UV irradiation, the PES@ZnO membrane achieved the highest flux recovery, stabilizing at 22 L m⁻² h⁻¹ with a minimal flux decline of approximately 27%. This significant improvement highlights the synergistic effect of ZnO incorporation and photocatalytic activity in mitigating membrane fouling.

4.3.2. Fouling indices

The fouling indices presented in Fig. 7b provide quantitative insight into the antifouling properties of the prepared membranes. The flux recovery ratio (FRR), calculated after water washing, serves as an indicator of fouling reversibility 34. For the pristine PES membrane, the FRR values were 33% both in the absence and presence of UV irradiation, indicating that the majority of fouling was irreversible and could not be recovered by simple hydraulic cleaning.³⁹

Remarkably, the PES@ZnO membrane exhibited a significantly higher FRR of 67% under dark conditions, which further improved to 86% under UV irradiation. This substantial increase demonstrates that the combination of enhanced surface hydrophilicity and photocatalytic self-cleaning effectively promotes the degradation of adsorbed foulants, facilitating their removal during the washing step.^36,37

Consistently, the flux decline ratio (FDR), which reflects the overall fouling propensity, was considerably lower for the PES@ZnO membrane.²³ Under dark conditions, the FDR was reduced to 33%, compared to 43% for the pristine PES membrane. Under UV irradiation, the PES@ZnO membrane achieved the lowest FDR of 22%, whereas the pristine membrane showed only a marginal reduction to 36%.

The enhanced antifouling performance of the PES@ZnO membrane can be attributed to two complementary mechanisms. First, the incorporation of hydrophilic ZnO NPs increases the surface wettability of the membrane, promoting the formation of a stable hydration layer via hydrogen bonding.^34,36 This hydration layer acts as a physical barrier, reducing the adhesion of hydrophobic dye molecules and mitigating initial fouling.^23,24

Second, under UV irradiation, the photocatalytic activity of ZnO generates reactive oxygen species (ROS), including hydroxyl radicals (˙OH) and superoxide anions (O₂˙⁻), which effectively degrade organic foulants accumulated on the membrane surface.^30,37 This photocatalytic self-cleaning mechanism not only removes reversibly bound foulants but also contributes to the degradation of irreversibly adsorbed species, thereby restoring membrane permeability and prolonging operational lifetime.^31,39

The long-term operational stability of the PES@ZnO membrane was assessed through three consecutive filtration cycles, with a mild hydraulic rinse applied between each run to simulate practical cleaning conditions. While this simple washing procedure effectively restored the permeate flux, the dye rejection capacity gradually decreased from 95% in the first cycle to 89% and 86% in the second and third cycles, respectively (Fig. 7c). This progressive decline in retention is ascribed to the incomplete removal of strongly adhered foulants from the membrane surface or within the pore structure, resulting in a minor accumulation of irreversible fouling.³⁰ Notably, this marginal reduction in physical separation performance does not reflect any deterioration of the photocatalytic function. Indeed, the normalized degradation profiles (C/C₀) remained perfectly superimposable across all three cycles, with rigorously identical pseudo-first-order rate constants, confirming that the embedded ZnO nanoparticles preserved their full catalytic activity without signs of leaching or structural alteration.^31,59 The confinement of ZnO within the polymer matrix effectively hinders nanoparticle aggregation and guarantees sustained photocatalytic performance over repeated use.^28,32 Collectively, these complementary findings demonstrate that while hydraulic cleaning sufficiently restores membrane permeability,⁶⁰ the intrinsic photocatalytic self-cleaning mechanism ensures continuous oxidative removal of the pollutant.²³ The modest decline in physical rejection is therefore outweighed by the stable photodegradation efficiency, confirming that the integrated configuration successfully eliminates the need for energy-intensive post-catalysis recovery steps such as centrifugation, thermal regeneration, or chemical reactivation which are typically required in slurry-based systems. This synergistic combination of membrane separation and photocatalysis not only enables uninterrupted operation but also overcomes one of the major limitations of conventional photocatalytic processes, namely the difficult and costly recovery of the dispersed catalyst.^24,30 These advantages position the PES@ZnO membrane as a highly promising platform for continuous and sustainable wastewater treatment.

To rationalize these experimental observations at the molecular scale and elucidate the underlying adsorption and charge-transfer mechanisms, DFT calculations were performed, as presented in the following section.

4.4. Structural optimization and adsorption energetics

All structures (Fig. 8) were optimised to the default Gaussian 16 convergence thresholds (maximum force < 4.50 × 10⁻⁴ a.u., RMS force < 3.00 × 10⁻⁴ a.u., maximum displacement < 1.80 × 10⁻³ a.u., RMS displacement < 1.20 × 10⁻³ a.u), confirming convergence to stable local energy minima. As shown in Table 2, the negative adsorption energy (E_ads = −1.40 eV) reveals spontaneous and thermodynamically favorable adsorption of BB41 onto PES@ZnO, driven by strong interfacial interactions. This value aligns with literature reports on dye adsorption in hybrid ZnO systems, where electrostatic complementarity governs adsorption strength.^41,61 Having established the favorable adsorption energetics, the following sections explore the resulting electronic rearrangements and charge transfer pathways.

Fig. 8. Geometries of optimized structures: (a) PES monomer, (b) ZnO monomer, (c) BB41, (d) PES@ZnO complex, (e) PES@ZnO-BB41 complex.

Table 2. Total energy (E_T), HOMO and LUMO and band gap energies (E_H, E_L, E_g), Fermi Level (E_FL), adsorption energy (E_ads), and work function (φ) for optimized structures.

Structures	E T (eV)	E ads (eV)	E H (eV)	E L (eV)	E FL (eV)	φ (eV)	E g (eV)
PES	−35837.3927	—	−8.72	0.35	−4.185	4.185	9.07
ZnO	−151376.9290	—	−9.01	−0.52	−4.765	4.765	8.49
BB41	−61007.2228	—	−7.45	−1.67	−4.560	4.560	5.78
PES@ZnO	−187214.3217	—	−8.22	0.21	−4.005	4.005	8.43
PES@ZnO-BB41	−248222.9445	−1.4	−7.47	−1.98	−4.725	4.725	5.49

The magnitude of E_ads suggests a mixed adsorption regime involving strong physisorption accompanied by partial chemisorption. The sulfone (–SO₂–) and ether (–O–) functionalities of PES provide polar anchoring sites, while ZnO contributes Lewis acidic Zn²⁺ centers and surface O²⁻ sites that interact strongly with the heteroatoms and π-electron system of BB41. This synergistic interaction enhances dye immobilization, which is essential for efficient heterogeneous degradation.^61,62 The total dipole moment increases markedly from 5.73 D for PES@ZnO and 8.48 D for BB41 to 12.29 D for PES@ZnO-BB41, indicating the formation of a highly polarized interface. This polarization reflects significant charge redistribution at the dye–composite interface and promotes strong electrostatic stabilization of the adsorbed state.⁶¹

4.5. Vibrational analysis: evidence of BB41 adsorption and interfacial interactions

The calculated IR and Raman spectra of PES@ZnO-BB41 (Fig. S1) provide direct vibrational evidence for the formation of a stable interfacial complex and corroborate the DFT-predicted adsorption mechanism. Upon BB41 adsorption, several characteristic vibrational bands of PES and BB41 exhibit noticeable frequency shifts and intensity variations, indicating strong electronic coupling and charge redistribution at the dye composite interface.

In particular, the symmetric and asymmetric stretching modes of the sulfone group (–SO₂–) of PES show measurable shifts in the PES@ZnO-BB41 spectrum compared to pristine PES@ZnO. These shifts indicate direct involvement of sulfone oxygen atoms in interfacial interactions with Zn²⁺ surface sites, consistent with the MEP and Mulliken charge analyses, which identify sulfone oxygen as electron-rich anchoring centers. Similar vibrational perturbations of sulfone groups have been reported for polymer-metal oxide interfaces and are commonly associated with donor–acceptor coordination between oxygen lone pairs and surface metal cations.

Moreover, BB41 associated aromatic C C stretching and heterocycle-related vibrations undergo red-shifts after adsorption, reflecting a weakening of π-conjugation due to charge transfer from BB41 to the ZnO surface. This vibrational softening is consistent with the DFT frontier orbital analysis, which shows HOMO localization on BB41 and LUMO localization on ZnO, enabling dye-to-semiconductor electron donation. The observed changes in Raman intensities further support enhanced polarizability of the adsorbed complex, in agreement with the calculated increase in molecular polarizability upon BB41 adsorption.

The Zn–O stretching modes of ZnO also exhibit slight frequency shifts and intensity modulation after BB41 adsorption, evidencing perturbation of the local ZnO lattice environment due to interfacial bonding and electrostatic interaction with BB41 and PES functional groups. Such spectral modifications are typically associated with surface coordination or hydrogen bonding interactions at oxide interfaces and confirm that ZnO acts as an active adsorption and charge transfer center rather than a passive support.

Overall, the combined IR and Raman signatures confirm that BB41 adsorption on PES@ZnO is not purely physisorptive but involves specific interfacial interactions with PES functional groups and ZnO surface sites. These vibrational fingerprints provide independent spectroscopic validation of the DFT-predicted strong adsorption energy, charge transfer, and interfacial polarization, thereby strengthening the proposed adsorption photocatalytic degradation mechanism.

4.6. Evidence of interfacial charge redistribution and donor–acceptor interactions

The Mulliken charge distribution of the DFT optimized PES@ZnO-BB41 complex (Fig. 9) provides quantitative insight into the electronic rearrangement induced by BB41 adsorption. Pronounced changes in atomic partial charges are observed at the dye composite interface, confirming significant interfacial charge redistribution upon complex formation.

Fig. 9. Mulliken charge distribution in DFT optimized geometry of PES@ZnO-BB41.

In the PES@ZnO framework, Zn atoms exhibit an increase in positive Mulliken charge, while surface oxygen atoms show enhanced negative character after BB41 adsorption. This charge polarization indicates electron donation from BB41 and PES functional groups toward ZnO surface sites, consistent with the Lewis acidic nature of Zn²⁺ centers. Such charge accumulation at ZnO corroborates the electrophilicity based charge transfer (ECT = 0.771), confirming that PES@ZnO acts as an electron-accepting platform.

Within the BB41 molecule, heteroatoms (N and O sites) and conjugated aromatic carbon centers show a decrease in electron density relative to the isolated dye, reflecting partial oxidation of BB41 upon adsorption. This electronic depletion of the dye chromophore is consistent with the HOMO localization on BB41 and LUMO localization on ZnO, which favors dye-to-semiconductor charge transfer. The weakening of the π-electron density in BB41 rationalizes the vibrational red-shifts observed in IR/Raman spectra and explains the increased susceptibility of the dye to oxidative attack by reactive oxygen species during photocatalysis.

Furthermore, the sulfone oxygen atoms of PES exhibit increased negative charge density and enhanced polarization, confirming their direct participation as anchoring sites through electrostatic and donor–acceptor interactions with Zn²⁺ centers and BB41 functional groups. This charge redistribution explains the strong adsorption energy and the large increase in dipole moment of the PES@ZnO-BB41 complex.

Overall, the Mulliken population analysis provides direct electronic evidence for the formation of a polarized interfacial complex, characterized by net charge transfer from BB41 to PES@ZnO and strong donor–acceptor coupling. This charge redistribution underpins the enhanced electronic reactivity, band gap narrowing, and improved photocatalytic degradation efficiency predicted by the DFT calculations.

4.7. Charge transfer and global reactivity descriptors

The positive electrophilicity-based charge transfer (see Table 3) confirms net electron donation from BB41 to PES@ZnO. The ECT value of the PES@ZnO-BB41 system reaches 0.771, confirming substantial interfacial charge transfer upon adsorption. This result indicates that BB41 predominantly acts as an electron donor, while PES@ZnO functions as an electron acceptor.^42,43

Table 3. Global chemical reactivity descriptors of optimized structures.

Structures	IP (eV)	EA (eV)	η (eV)	S (eV)	µ	χ	ω	ΔNmax	ECT
PES	8.72	−0.35	4.535	0.110	−4.185	4.185	1.926	0.923	—
ZnO	9.01	0.52	4.245	0.235	−4.765	4.765	2.674	1.224	—
BB41	7.45	1.67	2.890	0.346	−4.560	4.560	3.597	1.579	—
PES@ZnO	8.22	−0.21	4.215	0.237	−4.005	4.005	1.902	0.950	0.027
PES@ZnO-BB41	7.47	1.98	2.745	0.364	−4.725	4.725	4.066	1.721	0.771

The adsorption of BB41 leads to a significant decrease in chemical hardness (η) from 4.215 eV (PES@ZnO) to 2.745 eV (PES@ZnO-BB41) and a corresponding increase in chemical softness (S) from 0.237 to 0.364 eV⁻¹. The decrease in chemical hardness and increase in softness upon adsorption indicate enhanced chemical reactivity of the composite, consistent with DFT descriptors used to rationalize adsorption and redox activity on transition metal oxides and doped ZnO surfaces.^42,62 The increase in electrophilicity of the adsorbed complex supports facilitated oxidative pathways during photocatalysis, as reported for oxide-based dye degradation systems.^42,63 This trend demonstrates that the composite becomes more chemically reactive after dye adsorption, facilitating electron exchange with reactive oxygen species generated during photocatalysis. Moreover, the electrophilicity index (ω) increases from 1.902 eV for PES@ZnO to 4.066 eV for PES@ZnO-BB41, indicating an enhanced ability of the adsorbed complex to accept electrons. This increased electrophilicity favors oxidative pathways, accelerating the degradation of BB41 through successive electron–transfer reactions.

4.8. Intermolecular interactions: NCI, ELF, LOL, and MEP analysis

NCI-RDG features and MEP complementarity indicate stabilizing noncovalent interactions at the dye surface interface. The MEP map (Fig. 10) highlights electron-rich regions around the sulfone oxygen atoms of PES and surface oxygen atoms of ZnO, as well as electron-deficient regions near Zn²⁺ centers. BB41 preferentially orients itself to maximize complementary electrostatic interactions with these regions, resulting in strong adsorption and enhanced charge transfer.

Fig. 10. Molecular electrostatic potential map of PES@ZnO-BB41.

The NCI-RDG plots (Fig. S2) show prominent green isosurfaces at the dye–composite interface, corresponding to stabilizing noncovalent interactions. These interactions include electrostatic attractions, hydrogen bonding, and dispersion forces, collectively contributing to the stabilization of the adsorbed BB41 molecule. Comparable analyses in ZnO/doped-ZnO systems demonstrate that electrostatic potential matching and surface functionalization govern adsorption geometries and strengths.^41,42

ELF/LOL and charge-density redistribution upon adsorption are consistent with DFT studies showing preferential interaction sites around surface oxygen atoms and metal centers, facilitating charge donation/acceptance during photocatalysis.^42,62 The ELF and LOL maps (Fig. S3) indicate localized electron density around the heteroatoms of BB41 and surface oxygen atoms of ZnO, suggesting preferential interaction sites for charge donation and acceptance.

The local information entropy and electron density distributions further confirm the redistribution of electronic charge upon adsorption, which strengthens interfacial coupling. The role of polymer matrices in organizing adsorption microenvironments and enhancing interfacial coupling is also supported by PES-ZnO composite studies.⁶¹

4.9. Frontier molecular orbitals and electronic structure modulation

The frontier molecular orbital distributions (Fig. 11) reveal a clear spatial separation of electronic density after BB41 adsorption. The HOMO is predominantly localized on BB41, while the LUMO is mainly distributed over the ZnO moiety of the composite. This configuration is highly favorable for interfacial charge transfer from the dye to the semiconductor support.^43,64 The adsorption of BB41 induces a pronounced modification of the electronic properties of PES@ZnO. The HOMO–LUMO energy gap decreases from 8.43 eV for PES@ZnO to 5.49 eV for PES@ZnO-BB41, indicating enhanced electronic softness and increased charge mobility. DFT calculations at the B3LYP level are known to overestimate absolute band gap energies owing to the absence of long-range electron correlation. However, the relative trend observed upon adsorption remains robust for comparative purposes. This band gap narrowing facilitates electron excitation and promotes charge carrier migration across the interface, which is crucial for photocatalytic and redox-mediated degradation mechanisms.^43,64 The observed increase in work function after adsorption aligns with reports that stronger dye ZnO coupling stabilizes photogenerated carriers and suppresses recombination.⁴³ In addition, the Fermi level shifts toward more negative values (from −4.005 eV to −4.725 eV), while the work function increases from 4.005 eV to 4.725 eV upon BB41 adsorption. The increase in work function implies stronger electron binding at the surface and higher resistance to spontaneous electron emission, which improves charge separation and stabilizes photogenerated carriers, thereby enhancing the oxidative capability of the ZnO component.

Fig. 11. Frontier molecular orbitals maps of PES@ZnO-BB41.

4.10. TD-DFT simulated UV-visible absorption spectra and excited-state charge transfer evidence

The simulated UV-vis absorption spectra are presented in Fig. 12. The spectrum of isolated BB41 exhibits two characteristic absorption bands: a dominant band centred at approximately 450 nm, attributed to π → π* transitions within the extended conjugated chromophore (phénazinium/azo system), and a secondary band at ∼615 nm, consistent with the well-known blue colour of this cationic dye in solution. The PES oligomer displays a single absorption band in the far-UV region (∼210–220 nm), arising from π → π* transitions of the aromatic sulfone backbone, with no absorption in the visible range, in agreement with its optically transparent nature. The ZnO cluster exhibits absorption exclusively in the UV domain (∼210 and ∼280–310 nm), consistent with its wide optical band gap (∼3.2 eV) and its photoactivation under UV irradiation (365 nm) used in the experimental photocatalytic tests. The PES@ZnO composite spectrum is dominated by the UV contributions of both components, with a slight spectral extension toward the near-UV region (∼250–350 nm) relative to bare ZnO, suggesting weak electronic coupling between the polymer matrix and the inorganic phase even prior to dye adsorption.

Fig. 12. UV-visible absorption spectra.

Most significantly, the spectrum of the PES@ZnO-BB41 adsorption complex reveals pronounced modifications relative to the spectra of its individual constituents. The characteristic BB41 absorption bands are preserved but undergo measurable bathochromic shifts, with the principal band red-shifting from ∼450 nm toward longer wavelengths, and the visible band near 615 nm similarly displaced. These red-shifts are direct spectroscopic signatures of ground-to-excited-state charge redistribution within the complex: upon photoexcitation, electron density is transferred from the HOMO of BB41 toward the LUMO localised on the ZnO moiety, lowering the energy of the charge-transfer excited states relative to those of the isolated dye. Furthermore, the PES@ZnO-BB41 complex exhibits enhanced absorption intensity and spectral broadening in the 350–500 nm region compared to either PES@ZnO or BB41 alone, indicating improved light-harvesting capacity of the adsorbed system. This enhanced photoresponse is consistent with the formation of a new interfacial charge-transfer state that extends the effective absorption cross-section of the composite under UV irradiation.

Taken together, the TD-DFT simulated spectra provide critical excited-state corroboration of the ground-state frontier orbital analysis: the spatial HOMO/LUMO separation (HOMO on BB41, LUMO on ZnO), the band-gap narrowing upon adsorption (from 8.43 to 5.49 eV), and the electrophilicity-based charge transfer value (ECT = 0.771) are all consistent with a photoexcited electron injection mechanism from BB41 to ZnO. This process generates electron-deficient BB41 radical cation intermediates and electron-enriched ZnO surface sites, which subsequently interact with dissolved oxygen to produce superoxide anion radicals (O₂˙⁻) and, via further reaction, hydroxyl radicals (˙OH) — the dominant reactive oxygen species identified in the scavenger assays. The PES matrix, by providing a stable anchoring environment that pre-concentrates BB41 in the immediate vicinity of ZnO photocatalytic sites, amplifies the efficiency of this interfacial charge-transfer process, rationalising the five-fold enhancement in apparent rate constant observed for the PES@ZnO membrane relative to suspended ZnO nanoparticles under identical UV irradiation conditions.

4.11. Thermodynamic stability and polarizability effects

As shown in Table 4, the PES@ZnO-BB41 system exhibits higher thermal energy, enthalpy, and Gibbs free energy than the pristine components, indicating the formation of a larger and more complex stabilized system. The increased polarizability of PES@ZnO-BB41 suggests enhanced responsiveness to external fields and photoexcitation, a characteristic associated with improved light–matter interaction and charge separation in hybrid ZnO composites.⁶⁵

Table 4. Thermodynamic reaction parameters.

Parameters	PES	ZnO	BB41	PES@ZnO	PES@ZnO-BB41
T (K)	298.15	298.15	298.15	298.15	298.15
ΔS (cal mol−1 K−1)	146.0210	88.549	211.174	188.344	333.821
C v (cal mol−1)	69.7420	22.696	119.790	98.077	221.807
E th (kcal mol−1)	186.6880	11.671	311.675	200.090	513.255
Dipole moment (Debye)	5.7361	0.0077	8.4751	5.7270	12.2893
Polarizability (a.u.)	209.0710	54.6713	393.8716	264.3515	640.4353

Notably, the marked increase in polarizability from 264.35 a.u. for PES@ZnO to 640.44 a.u. for PES@ZnO-BB41 points to a greater susceptibility of the complex to external electric fields and photon-induced polarization. This enhanced polarizability promotes stronger interaction with incident light and surrounding polar media, thereby facilitating charge separation and interfacial electron transfer. As a result, the composite becomes more responsive to photoexcitation, which is beneficial for the photocatalytic degradation of BB41 in aqueous environments. These thermodynamic findings are supported by DFT-guided correlations between adsorption energetics, electronic adaptability, and photocatalytic performance reported for oxide–carbon hybrids.^62,65 Such thermodynamic and electronic adaptability is widely recognized as favorable for sustained catalytic activity and efficient interaction with reactive oxygen species in aqueous photocatalysis.^64,65

4.12. Implications for adsorption-photo-degradation mechanism of BB41

The proposed photocatalytic pathway is consistent with our previous experimental study on ZnO-based photocatalysts for BB41 degradation6, where radical scavenging assays identified hydroxyl radicals (˙OH) and photogenerated holes (h⁺) as the dominant reactive species. The present DFT findings provide a molecular rationale for these observations, demonstrating that the convergence of experimental and computational evidence strongly supports the proposed adsorption-photocatalysis mechanism.

Collectively, strong adsorption, FMO spatial separation, band-gap narrowing, and net dye-to-ZnO charge transfer support a synergistic adsorption-photocatalysis mechanism (Fig. 13). Similar mechanistic couplings between adsorption pre-concentration and interfacial charge transfer have been demonstrated for ZnO heterojunctions and polymer-supported ZnO photocatalysts, leading to enhanced reactive oxygen species generation and dye mineralization.^43,59,64

Fig. 13. Proposed mechanism for the adsorption and photocatalytic degradation of BB41 by the PES@ZnO membrane.

The PES matrix enhances adsorption and interfacial stabilization, while ZnO provides photoactive centers for reactive oxygen species formation, in agreement with experimental reports on PES-ZnO composites for dye removal and degradation.^59,61 These convergent findings substantiate the proposed DFT-based mechanism linking adsorption energetics to photocatalytic degradation efficiency.

The vibrational shifts observed in the IR and Raman spectra (Fig. S1) further confirm the direct involvement of PES sulfone groups, Zn–O surface sites, and BB41 aromatic moieties in interfacial charge transfer, providing spectroscopic validation of the DFT-based degradation mechanism.

The Mulliken charge redistribution at the dye–composite interface (Fig. 9) confirms net electron transfer from BB41 toward ZnO surface sites, which electronically weakens the dye chromophore and facilitates ROS-driven oxidative degradation.

The strong adsorption energy ensures effective pre-concentration of BB41 on the composite surface, while the narrowed band gap, increased electrophilicity, and significant charge transfer enhance the formation of reactive oxygen species (˙OH, ˙O₂⁻). These species are responsible for the oxidative degradation of BB41. The spatial separation of HOMO and LUMO densities further suppresses electron–hole recombination, thereby improving photocatalytic efficiency.

Overall, the PES@ZnO composite exhibits a pronounced synergistic effect that integrates high adsorption capacity with enhanced electronic reactivity, making it a promising multifunctional material for the removal and degradation of cationic dyes from wastewater.

4.13. Comparative study

Table 5 benchmarks the PES@ZnO membrane developed in this work against previously reported PES-based photocatalytic membranes in terms of dye removal, flux recovery ratio (FRR), and permeability. The present membrane shows competitive or superior performance, achieving an FRR of 86%, while providing unique DFT-based mechanistic insights absent from earlier reports.

Table 5. Comparison of PES@ZnO membrane performance with literature.

PES@ZnO membrane	Pollutant	Removal (%)	FRR (%)	Pure water flux	Ref.
PES/ZnO-MWCNT (0.5 wt% ZnO/MWCNTs)	Powder milk (8000 ppm)	95%	88.6%	16.7 kg m−2 h−1	37
PES/CuO-ZnO (0.2 wt% CuO–ZnO)	Bovine serum albumin (BSA), (500 mg L−1)	99%	50.1	679 kg m−2 h−1	38
PES/ZnO-silane (1.5 wt% ZnO)	Congo red dyes eriochrome black T	84.2%	59%	53.3 L m−2 h−1 for 3 bar	40
		75.6%	60.8%
PES/ZnO (modified) PES/Z	Bovine serum albumin	80–84%	85.9%	80.3 L m−2 h−1	39
PES/ZnO (modified) PES/Z	Humic acid (50 mg L−1)	94 (0.1 wt% ZnO)	38.89%	80 kg m−2 h−1	34
PES/ZnO (0.5 wt%)	BB41 (10 mg L−1)	95	86	36 L m−2 h−1 bar−1	This work

As summarized in Table 5, the membrane combines high dye removal (86%), excellent flux recovery (95%), and a hydraulic permeability of 36 L m⁻² h⁻¹ bar⁻¹. Notably, its pseudo-first-order rate constant is approximately five times higher than that of suspended ZnO NPs under identical conditions, highlighting the benefit of membrane immobilization. Direct comparison with the literature is complicated by variations in experimental conditions (dye type, concentration, transmembrane pressure, and UV source). Nevertheless, the present membrane performs as well as or better than previously reported PES@ZnO systems, particularly in terms of fouling reversibility and degradation kinetics. Moreover, this work offers unprecedented DFT-based mechanistic insights into interfacial charge transfer, a dimension not addressed in previous experimental studies. These combined experimental and computational findings position the PES@ZnO membrane as a promising platform for sustainable dye-contaminated wastewater treatment.

5. Conclusion

This study successfully demonstrates high-performance PES@ZnO MMMs for BB41 removal, combining experimental investigation with DFT calculations. Incorporation of 0.5 wt% ZnO NPs enhanced membrane hydrophilicity (contact angle: 61.5° vs. 68.5°) and hydraulic permeability (36 vs. 28 L m⁻² h⁻¹ bar⁻¹). FTIR, Raman, and TGA confirmed successful ZnO incorporation and improved thermal stability.

Under UV irradiation, the PES@ZnO membrane achieved 86% BB41 removal after 250 min, with a pseudo-first-order rate constant of 1.22 × 10⁻² min⁻¹, five times higher than suspended ZnO NPs. The membrane exhibited excellent antifouling properties with a flux recovery ratio of 95%, attributed to synergistic surface hydrophilicity and photocatalytic self-cleaning.

DFT calculations provided critical mechanistic insights into the ground-state adsorption and charge-transfer behavior governing the PES@ZnO-BB41 system. The negative adsorption energy (E_ad = −1.40 eV) confirmed spontaneous dye adsorption, while HOMO–LUMO gap narrowing (from 8.43 to 5.49 eV) indicated enhanced electronic reactivity upon adsorption. Spatial separation of frontier orbitals (HOMO on BB41, LUMO on ZnO) and an ECT value of 0.771 are consistent with interfacial electron donation from the dye to ZnO, a prerequisite for oxidative degradation pathways. Mulliken charge analysis and NCI-RDG/MEP analyses further revealed pronounced charge redistribution and stabilizing noncovalent interactions at the dye–composite interface. While TD-DFT calculations of excited-state properties (band alignment, exciton binding energy, charge-separated state energetics) would provide a more complete description of the photocatalytic process, the present ground-state DFT analysis establishes a solid foundation for understanding the adsorption and initial charge-transfer events preceding photocatalysis.

This integrated experimental–computational framework provides unprecedented mechanistic insight into the adsorption–photocatalytic degradation mechanism of BB41 in a PES@ZnO membrane system, demonstrating that ZnO functions not merely as a hydrophilic filler but as an active charge-transfer center. While DFT studies of dye–ZnO interactions exist in other contexts (e.g., dye-sensitized solar cells, slurry photocatalysis), the coupling of such atomistic calculations with membrane filtration performance data—including permeability, fouling indices, and degradation kinetics—represents a significant methodological advance. This synergy between experiment and theory offers a rational blueprint for designing next-generation photocatalytic membranes with tailored interfacial properties for sustainable wastewater treatment.

Author contributions

Afef Attia: conceptualization, methodology, writing – review and editing. Ridha Ben Said: investigation, data curation, formal analysis, methodology. Ali Ben Ahmed: writing – review and editing, and supervision.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data (simulated IR and Raman spectra ans NCI, RDG and ELF maps) supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6ra03267b.